For current production lasers, the manufacturers' Web sites often provide basic specifications. For older lasers, it's often difficult to obtain detailed specs so estimates based on physical size, and then testing may be the only option.
The sections in this chapter are arranged approximately in alphabetical order by manufacturer.
Unless otherwise noted (below), these data were obtained from brochures, spec sheets, or user manuals for each laser. Contributions and corrections welcome.
All values are in parts per billion (ppb).
<--- Frequency Stability Time Scale ---> Model Type/AP Sec Min Hour Day Year Life ------------------------------------------------------------------------------- *Aerotech S100 (2) SM S +/-1 +/-2 +/-3 (8 hours) *Axsys 150 (8) DM M +/-2 +/-6 +/-20 (24 hours) *Coherent 200 DM S +/-2 (5 min) +/-10 long term *Excel 1001A/B/F AZ M 20 (unspecified time) Frazier 100 I2 S *HP-5500A/B/C, 5501A/B AZ M +/-2 +/-20 *HP/Agilent 5517 (all) AZ M +/-2 +/-20 *HP/Agilent 5518A, 5519A/B AZ M +/-2 +/-20 Jenaer ZL 600 DM M 2.0 *Lab for Science 200 DM S 0.03 0.05 0.2 0.5 *Lab for Science 210 DM S 0.03 0.05 0.2 0.5 *Lab for Science 220 TZ S 0.01 0.02 0.05 0.2 *Lab for Science 260 TM S 0.02 0.02 0.1 0.4 *Laseangle RB-1 (3) DM S 0.01 0.1 Laser Metric Systems SFL-1 DM M 2 (unspecified time) Limtek LS 10.3 GP M 20 (unspecified time) *Mark-Tech 7900 DM M +/-2 (const. temp.) Mark-Tech 7910 (6) DM M +/-2 (const. temp.) *Melles Griot 05-STP-901 (4) DM S +/-1 +/-4 +/-4 (8 hours) *Melles Griot 05-STP-91X (2) DM S +/-1 +/-2 +/-3 (8 hours) Micro-G ML-1 DM S 0.08-0.15 (unspecified time, const. temp.) NEOARK NEO-262 TZ M 1 (unspecified time) NEOARK NEO-92SI-NF I2 S 0.025 Newport NL-1 DM S Nikon DM M NPL Hexagon ?? S 0.01 +/-2 NPL I2 543 nm I2 S +/-0.25 NPL I2 633 nm I2 S +/-0.2 *Optodyne L-109 DM M *Optra Optralite AZ M Renishaw ML10 ?? M Renishaw XL80 ?? M REO SHL DM S +/-2 +/-4 +/-6 SIOS SL 02 DM S +/-2 +/-10 +/-20 SIOS SL 03 DM S +/-1 +/-2 +/-10 *Spectra-Physics 117 (5) DM S *Spectra-Physics 117A (4) DM S +/-1 +/-4 +/-4 (8 hours) *Spectra-Physics 117C (5) DM S *Spectra-Physics 119 (7) LD S +/-2 *Teletrac 150 (Short) (8) DM M +/-2 +/-6 +/-20 (24 hours) *Teletrac 150 (Long) (8) DM M +/-2 +/-6 +/-20 (24 hours) Winters 200 I2 S 0.025 *Zygo 7701/7702 (9) DM M +/-2 +/-10 +/-100 *Zygo 7705 (9) AZ M +/-10 +/-20 +/-200 *Zygo 7712/7714 (9) DM M +/-0.5 +/-1 *Zygo 7722/7724 (9) DM M +/-0.5 +/-10
The * denotes lasers that are covered in detail elsewhere in this chapter.
AP (Application) Legend:
A metrology laser can generally also be used for scientific/research applications since all have very tightly controlled optical frequency. And while the converse is often (but not always) true in principle, it's not usually practical or worthwhile except for experimental purposes since metrology systems may require laser characteristics (like two-frequency) that aren't present in laboratory stabilized lasers. In addition, the optics and cabling/electronics requirements would likely make their adaptation potentially complex, if possible at all.
With only a single photodiode sampling the beam, only the amplitude of one of the two polarized modes can be stabilized. However, the frequency stability will still be quite good once the laser tube has reached thermal equilibrium and its power has leveled off. However, why frequency stabilization was not also implemented is a mystery as it would have been a very straightforward enhancement - a second photodiode and difference amp!
Lasers based on the Aerotech technology are now sold as the Melles Griot 05-STP-909 and 05-STP-911 (0.5 and 1.0 mW minimum output power, respectively). The Melles Griot lasers are physically and functionally very similar to the Syncrolase 100 but it is not known whether the electronics is as well. However, all indications are that very little has been changed since acquiring the technology from Aerotech when they took over the company. Melles Griot calls them frequency stabilized lasers though their description seems to indicate that the same amplitude stabilization technique as the Syncrolase is used. And, if you'd like to order a few, the Melles Griot price (in 2009) is about $3,600 each! :) Searching for "Melles Griot 05-STP-909" should return a spec sheet. Here are specifications for the Melles Griot versions:
In addition to output power, the Syncrolase came in two versions based on whether a pair of DC wall adapters were used to power an internal HeNe laser power supply and the locking controller, or whether the laser head had a standard Alden connector to attach to an external lab-style HeNe laser power supply. (Melles Griot has since discontinued the dual DC adapter versions.)
One of the unique features of this system is that rather than using a resistance heater over a substantial part of the HeNe laser tube as is done in most commercial stabilized HeNe lasers, it uses a coil surrounding the OC mirror mount stem to directly heat the metal mount via RF induction. A very simple MOSFET driver can provide over 10 W directly to the mount resulting in a very rapid response. Based on tests I've done, I estimate that at maximum RF power, it will increase the temperature of the mirror mount stem itself by greater than 1 °C per second. This is more than an order of magnitude faster than traditional resistance heaters surrounding the glass portion of the tube. A thermocouple in close proximity to the mirror mount stem senses its temperature and is used both to switch the feedback loop on when hot enough, as well as to shut the heater off if the temperature goes too high. Warmup to fully stable operation still takes 30 minutes or so because the rest of the laser has to come into thermal equilibrium as well as the mirror mount stem. But, initial locking is very quick - a couple of minutes - and once locked, it should use less power and be more immune to ambient temperature variations, and the faster response also improves frequency stability. In addition, the use of this technique allows for the possibility at least in principle of converting almost any HeNe laser tube with a suitable mode structure into a stabilized laser by simply attaching the very compact controller to its output end. However, in practice, minor details like the mirror mount stem dimensions and the exhaust tip-off usually being in they way make this rather difficult.
The DC adapter (either version, 1 or 2 required depending on the model of the laser) is rated 13 VDC, 1.3 A. Measurements show it to have an open circuit output of 16.5 V. The plug is 2.5 mm center positive. I do not know the official specifications for the the external HeNe laser power supply (where required), but based on the length of the tube and other typical Aerotech tubes, it probably around 1,500 V ad 5 mA. Since there is a 7812 +12 V regulator in the controller (see the schematic below), the output of that DC adapter must be greater than about 14.5 V to assure proper regulation. So, at least once the feedback loop is closed, the input voltage should never dip below 14.5 V.
The HeNe laser tube in the Syncrolase 100 is about 7 inches long. A common 6 to 9 inch tube with cathode-end output (high voltage far away from the electronics!) would probably work except that the mirror mount stem needs to be a about an inch long with the exhaust tip-off cut off close to the end-cap so as not to interfere with the coupling coil assembly. Very few tubes have these characteristics, though some are close enough to be usable in a pinch. In addition, using too long a tube might result in a second longitudinal mode being present if the Output Adjustment is set so the main lasing line is too close to the neon gain center.
The gate of a power MOSFET is driven by a simple oscillator, running at between 500 kHz and 1 MHz (I measured about 700 kHz on one unit). The feedback signal is summed into the gate junction from the error amp and serves to modulate the output of the induction heater to maintain lock once the operating temperature has been reached. The coil is just short of 9 turns of #24 AWG wire close wound on a 1.35 cm form. Due to the way the leads enter through the back of the form, the final turn is short changed! :) This is probably not terribly critical though.
For details on theory and implementation see U.S. Patent #4,819,246: Single Frequency Adapter.
A schematic diagram of the electronics for the Syncrolase 100 can be found at Schematic of Aerotech Syncrolase 100 Controller. This may not yet be quite complete and numerous errors are possible since the PCB is tight, it is a 4 layer board, and the soldermask is almost totally opaque. It was not much fun to trace the circuit. Part numbers are not available for a half dozen components because (1) they might have been obscured and (2) there were several added parts that appear to be in the "oops" category. :-) But I bet this schematic provides infinitely more information than what's available anywhere else! :)
The RF driver consists of a HEX Schmitt trigger (MC14584BCP similar to a CMOS 40106) with one section used as the oscillator and the remaining sections paralleled to buffer its output. An RC network converts the squarewave of the oscillator to a roughly triangle waveshape at the MOSFET gate. The output of the Error Integrator feeds into the gate as well with the effect of modifying its DC offset. Since the MOSFET gate threshold is fixed, this produces a modulation effect which is a combination of amplitude and pulse width, with the net result of controlling the amount of RF power transferred to the HeNe laser tube mirror mount stem. A significant part of the capacitance in the waveshaping network is the internal input capacitance of the MOSFET gate itself, and this may exceed 1 nF. Thus, it's possible that if the MOSFET needs to be replaced, the value of the capacitor between the gate and ground (C13) may need to be adjusted as well to maintain approximately the same net capacitance and waveshape. The MOSFET gate capacitance can vary by a factor of over 2:1 between MOSFETs with the same part number, or by even more if a MOSFET with otherwise acceptable specifications is substituted. On the unit I have, it was about 1.3 nF.
Newer versions include a ULN2003 Darlington array for something. But I haven't dug deep enough to be sure if it's part of the MOSFET driving circuitry or replaces the MOSFET entirely. They may also use a semiconductor sensor in place of the thermocouple - it looks like a 1N4148 with no markings. That's probably much cheaper!
The control functions are implemented in the four sections of a TLC27L4CN quad op-amp as follows:
The output of A1A also feeds the Over-Temperature protection circuit that is supposed to turn the heater off if the temperature goes too high. However, on early units, this almost looks like an afterthough with its adjustment pot hanging in mid-air!
When powered up, the temperature sensor is initially cool so the RF driver comes on at full power. When the mirror mount stem reaches the operating temperature (something like 80 °C in 30 seconds or so), the feedback loop becomes active and the Sync LED comes on. However, since the remainder of the laser tube is still increasing in temperature due to the normal heating of the discharge and hasn't reached thermal equilibrium, lock may be lost several times as the overall tube expands and the controller then needs to keep *reducing* the temperature of the mirror mount stem to maintain the distance between the mirrors constant. When the mirror mount temperature gets to be too low, the system will go back to continuous heating based on the hysteresis of the Sync Enable Comparator. After a half hour or so, the laser tube will reach thermal equilibrium and the system will then remain locked forever. (Unfortunately, many people take this literally and leave the laser on until it dies, which is considerably sooner than forever!)
Here are some photos:
The coupling coil assembly on the first Syncrolase 100 of mine had disintegrated due to excessive temperature. (Actually the magnet wire and its insulation is in fine shape but the plastic form on which the coil was wound or embedded is no more and it's not even possible to determine much about it.) I've tested the induction heating winding a test coil on a tube made from insulating plastic sheet. The effect is impressive considering the simplicity of the circuitry (see the schematic below) raising the temperature of a dummy mirror mount stem by more than 1 °C per second even with a coil that is probably far from optimal.
I do not know for sure if the cause of the destroyed coil form was due to a part failure rather than simply a result of the laser being been left on for 7 years continuously! :) The HeNe laser power supply was indeed dead, probably due to the tube being very hard to start and impossible to run for more than a few seconds regardless of power supply or ballast resistance. So it's possible that when the tube decided it was tired of doing its thing and the power supply shorted out, the controller ended up cycling on the over-temperature condition. The Melles Griot manual does warn against running without the laser on. And, electrical tests seem to indicate that the controller is working properly.
So, the Over-Temperature (OT) adjustment might have been incorrect and too high all along. Since it's not something that affects normal behavior, it would be all too easy to neglect setting it properly! I've also been told by the former owner that this laser always ran very hot. If the tube fails - even if someone forgets to plug in its wall adapter! - the heater tends to be on and bad things can then happen if the OT setting is too high. Ask me how I found out. :( :) OK, I'll tell you. I acquired another Syncrolase with a good tube but that would not stabilize. I traced the problem to what I believe may have been a short in the temperature sensor and then adjusted it to operate at a reasonable temperature set-point. But I accidentally left the controller powered after turning off the laser and went away. When I returned (after lunch!), the entire assembly was too hot to touch and the platic coil form at is cover had melted!!! Apparently, either the OT setting was way too high (it's possible someone before me messed with it) or it isn't effective or was broken.
Interestingly, on one of those rare occasions where I was able to get the tube to remain on long enough with a lab power supply to watch a few mode sweep cycles, it is a classic FLIPPER! I suppose that the flipperitis could have happened in its old age (it is also weak - about 0.7 mW - and with brown crud in the bore), but normally the flipper or non-flipper status of a tube doesn't change over the course of its life. I do have another Aerotech laser head that would screw right on to the controller but it too is a flipper! :( :) In fact, its behavior shown in Plot of "Flipper" Aerotech OEM1R HeNe Laser Head During First Part of Warmup and the merged version in Plot of "Flipper" Aerotech OEM1R HeNe Laser Head During First Part of Warmup (Combined) looks virtually identical to that of the Syncrolase tube (over the few mode sweep cycles I could see before the tube went out). But, even more interstingly, the flipping of the tube in the plots ceases entirely and it becomes perfectly well behaved once nearly warmed up as shown in Plot of "Flipper" Aerotech OEM1R HeNe Laser Head at Transition to Normal Behavior (Combined). Perhaps that tube was intended for a Syncrolase as it in unusual in having the required long mirror mount stem and short cutoff exhaust pipe. Perhaps it was a reject due to the flipping. Or perhaps for unknown reasons, all these tubes flip when cold. Since the Syncrolase 100 would be operating well beyond this point, there's a chance that the flipping is irrelevant and it would work just fine. In fact, that one working genuine Syncrolase tube is also a flipper until it warmed up! More on this below.
I built a replacement coil using the wire from the first dead Syncrolase on a roll of plastic. It works, though the temperature response is faster probably because the thermocouple is not in the same location as the original. So, it locks more quickly, but also loses lock more frequently during warmup but is otherwise functional. Perhaps changing the temperature set-point would correct that. It's amazing how much variability can be tolerated with this design.
Adjusting the temperature set-point is an interesting exercise. Ideally, it should be slightly above the equilibrium temperature of the laser head with only the laser tube powered. Set too high and the laser will run excessively hot, but there will be a fewer number of lost lock events during warmup. Set too low and it may lose lock eventually when the tube equilibrium temperature exceeds the set-point temperature.
One way to do the adjustment might be to initially set the Temp. Gain pot (R1) fully CW (for a very low temperature) and power *only* the laser head (not the controller) for at least an hour so it reaches thermal equilibrium. Then, power up the controller and slowly turn R1 CCW to slightly beyond the point where the SYNC LED goes out. Monitoring the Temperature Amp output (A1 pin 1) will indicate how effective this is. The voltage on A1 pin 1 should remain between approximately 1.5 V and 2.75 V when the laser is locked. If it goes below about 1.5 V, the feedback loop is disabled and the heater turns on full (SYNC LED OFF). This state continues until the temperature increases to the point where A1 pin 1 exceeds about 2.75 V and the feedback loop is enabled (SYNC LED ON). Better to start out with the temperature set-point adjusted too low should the over-temperature protection fail. :( :)
I built another temporary coil for the first laser to check it out. This coil is wound on a plastic cylinder found in a junk pile that was glued to the remains of the original coil form. The Epoxy seems to stick rather well, which is a bit surprising. I didn't have any #24 AWG magnet wire, so I used #20, which just fits 9 turns in the available space. The laser works quite well now except that the speed of heating is not quite as fast, possibly due to the coil being slightly longer and larger in diameter. However, this is probably of little consequence in the grand scheme of the Universe. :) Lowering the RF frequency improved the response, though there was no resonance.
Finally, I built a new coil form for the third laser. It has approximately the same dimensions as the original so it behaves very well. But the plastic is too think and there is very little clearance between the form and mirror mount stem. So the genuine Syncrolase laser head won't fit because its tube is too off-center. (This must have been a result of the way it was manufactured since its beam is well centered.) But my "flipper" head fits just fine and works just fine. :)
If anyone has an Aerotech Syncrolase or Melles Griot 05-STP-909/910/911/912 laser, dead or alive, that they no longer need, or one that they'd like evaluated), please contact me via the Sci.Electronics.Repair FAQ Email Links Page.
I would suggest adding a separate temperature sensor used only for protection. The circuit could be as simple as a 10K NTC thermistor and fixed resistor or rheostat in a voltage divider, a zener diode and a 2N3904 or similar transistor in parallel with the one in the existing OT circuit. When the transistor turns on due to the resistance of the thermistor decreasing, it would shut down the heater drive. These parts would easily fit in the available space. There's even a spare hole in the coil form for an additional temperature sensor (at least in the ones I've seen). Since it's only for OT, the sensor can be further from the coil.
Interchanging the Output Adjust and Photodiode inputs to the Error Integrator (A1D) would cause the heater to turn off with no or low optical power. The only difference in functionality is that the laser would lock on the opposite side of the neon gain curve, equivalent to selecting the orthogonal polarization to the photodiode (by rotating the laser head 90 degrees).
Adding a second photodiode for dual polarization stabilization would also be beneficial since the relative intensity of the two modes would be the relevant variable, not the absolute intensity of a single mode. This ratio would still be valid at very low total output power.
Another modification (or complete redesign depending on your point of view!) that would enable the Syncrolase (or any thermally-stabilized laser) to run at the minimum temperature to assure reliable operation would to have a temperature set-point that is based on the ambient temperature of the environment, not a fixed setting. In principle, this can easily be accomplished by counting mode cycles from a cold start. Since each mode cycle represents a precise change in temperature, this would enable the laser to operate at a temperature of ambient plus a constant known to be greater than the heating from the laser tube current. A microcontroller could be used for the implementation, left as an exercise for the student. :)
This of course assumes that the ambient temperature remains relatively constant, but this is often the case with real lab environments. The Zygo metrology lasers with digital controllers compute the number of mode cycles (they call them "mode slews) needed to reach operating temperature based on the actual tube temperature when the laser is switched on, though they may still operate at the temperature required for worst case conditions.
While the Melles Griot 05-STP-9XX is not supposed to melt down due to a fault condition like not powering the laser, it is not known if the design has actually been improved or rather that they are simply depending careful adjustment of the over temperature pot.
So, while two data points may not be conclusive, it would seem that that almost any tube that can be stabilized using the conventional heating blanket technique can also be stabilized using the Syncrolase controller if its mirror mount stem will fit inside and extend far enough into the induction heater coil. Where the tip-off is not too long but interferes with the coil assembly, simply removing the plastic cover may gain enough clearance. Of course, if you happen to be friendly with the tip-off person at a HeNe laser tube manufacturer, simply ask them to pinch-off and trim the tip-off closer to the tube! :) For longer higher power tubes, the internal preamp gain would need to be reduced to allow the Output Adjust pot to lock at higher power. Of course, for such tubes, the position on the gain curve over which the output is pure single mode would be reduced.
And flippers will work just fine, thank you. :-) And as noted above, 3 of 3 Aerotech tubes from Syncrolase lasers were flippers, at least when cold!
The real challenge would be to find space for the modified photodiode assembly. One option would be to replace the polarizing beam-splitter cube with a 45 degree 5 percent plate beam-splitter. Then, mount a 5 mm diameter PBS above this with the photodiodes attached to it. As far as the additional electronics. the newer versions of the Syncrolase, and presumably the 05-STP-909/910/911/912 as well have less space available, but there's probably still enough space to float an IC above the other components. :) Or, if everything were converted to surface mount, there would be plenty of space!
The HeNe laser head is powered from a standard Laser Drive 6.5 mA, 2,100 V power supply brick via a HV BNC connector. There is no special control or regulation of this supply - it's turned on by the main power switch. But some thoughtful engineer included a high resistance bleeder to discharge the HV caps in the power supply brick after power is removed. :)
The HeNe laser tube itself is a Melles Griot (not made by Coherent!) model, labeled 05-LHR-219-158. It has similar dimemsions to an 05-LHR-120, a common 2 mW (rated) random polarized laser. But, the -158 may mean it has been specially selected to have a well behaved mode sweep cycle (not a flipper!) for this application. It may also be filled with isotopically pure gases and an AR-coated HR (to minimize back-reflections from the HR's outer surface). The tube itself puts out more than 2 mW when new - possibly up to 4 mW or even more - but the polarizing and beam sampling optics sucks up some of it. In addition, depending on the particular version, there is either a dielectric filter or polarizing filter in the end-cap. The dielectric filter cuts the output by about half but the this can be varied by 10 percent or so (though I'm not sure if this is intentional or just a byproduct of it being angled). The polarizing filter allows continuous adjustment of output power. (In both cases, the adjustment is done by loosening a set-screw and rotating the end-cap). According to the CDRH sticker, the output beam is supposed to be less than 1 mW. Given the wide swings in output power during warmup (see below), even with 50 percent attenuation, the peak output power may approach 1 mW. But regardless of the type of end-cap, only a single polarization ever exits the laser since the internal beam sampler blocks the other one.
There is a thin film heater attached to a thick rubber jacket between the tube and laser head cylinder. A beam sampler assembly consists of a pair of Beam-Splitter Cubes (BSCs) in series and two photodiodes, each associated with one of the BSCs. The first BSC is a polarizing beam-splitter and reflects the full power of one polarized mode to its photodiode. Thus, the beam that passes through it is linearly polarized with the orthogonal orientation. The second BSC reflects 10 or 20 percent of this mode to its photodiode. So, the output beam from the laser is pure linearly polarized and has slightly less output power than one of the polarized modes of the tube. The controller monitors the lasing modes and maintain cavity length using the heater so that a pair of orthogonally polarized longitudinal modes straddle the gain curve. The beam sensor assembly can be rotated to align the photosensors with the 2 orthogonal lasing modes as this is arbitrary from tube to tube, and orientation within the cylinder, but should remain fixed for the life of the tube.
The controller can be set up to run on various input voltages from 100 VAC to 240 VAC by changing the position of a small PCB that plugs into the AC entrance assembly, and plugging in the appropriate fuse. However, it seems that the HeNe laser power supply always runs on 115 VAC from a tap on the main power transformer so it doesn't need to be capable of 230 VAC operation, even though the one that's in there has that option - the wire for 230 VAC is not used! The output of the HeNe laser power supply is rated 2,100 V at 6.5 mA with no start delay.
The user controls consist of one (1) power switch. There are indicators for AC power and Status. After a warmup period of 20 minutes or so for the laser head to reach operating temperature, the Status indicator will change from Wait (red) to Ready (green). Doing anything that causes lock to be lost will result in a shorter delay of a couple minutes to re-establish it.
The internal circuitry of the controller box is relatively simple and includes a pair of LM3403 quad op-amps, a 741 op-amp, and LM311 voltage comparator, along with a TO5 power transistor on a heatsink to drive the heater.
Here is the pinout of the circular control connector as determined by my measurements. There may be errors.
Pins Wire Color Function Comments -------------------------------------------------------------------------- 1,2 Blk/Wht Heater Power ~22 ohms 3,4 Blk/Red Temp Senseor ~880 ohms at 25 °C, ~1.2K when locked 5,6 Blk/Blu Photodiode 1 Anode is pin 5; Approximately 250 uA max 7,8 Blk/Grn Photodiode 2 Anode is pin 8; Approximately 50 uA max
It would appear that the difference in sensitivities is the way it's supposed to be since this was similar on 3 heads. (However, the readings on an analog VOM for the photodiodes did differ on 2 heads I tested - I'm not sure what, if any significance, that has.) This makes sense given that the sampling is done from the main beam. One polarization orientation is blocked entirely and thus the associated photodiode gets its full intensity. The other mode would then seem to be sampled at about 20 percent intensity. The controller and laser head are normally a matched pair and there is an adjustment inside the controller to equalize the responses.
The heater consists of a serpentine thin file metal pattern on a rubbery backing material that wraps completely once around the tube.
The temperature sensor extends the length of the tube and is buried within the heater backing, technology unknown.
I picked up a controller and 3 laser heads in two separate eBay auctions for a grand total of $22.50 + shipping. The serial number on one of the heads matched that of the controller and while this head was initially hard to start, after running it for awhile on my HeNe laser test supply, it now starts normally.
The controller originally had a dead HeNe laser power supply brick (Laser Drive 314S-2100-6.5-2, 2,100 V at 6.5 mA) which is likely the reason it was taken out of service. I replaced that with an Aerotech LSS-5(6.5) which seems to be happy enough. Using a laser power meter, one of the two modes of the laser (the one present in the output beam) could be seen cycling up and down between about 0.60 and 1.40 mW with the orientation of the beam sensor assembly adjusted for maximum peak power. Each cycle took longer and longer as the tube warmed up to operating temperature, helped along by the heater. After about 15 minutes, it would appear to try to "catch" at certain power levels but couldn't quite remain there. (This behavior may have had nothing to do with the feedback control though.) Then suddenly, after about 20 minutes, the Ready light came on and a few seconds later, it locked rock stable at 0.95 mW. :) A second laser head behaved in a similar manner but with a slightly higher final output power of 1.02 mW. No adjustments were needed inside the controller despite the fact that the second head's serial number didn't match the controller's serial number. Possibly, even better stability or slightly higher stabilized output power could be achieved with some fine tuning. (The 1.02 mW head actually had higher peak power than the 0.95 mW head. The difference is probably in part due to the photodiode sensitivities.) With the fixed filter end-caps installed, the output power dropped to around 0.50 mW. I rather suspect that these are normal power levels for this system. (This was later confirmed when a manual with detailed specifications turned up.) The third head had its cables cut but I finally scrounged a replacement control connector from a box of junk in the garage and jerry-rigged the HV BNC for testing. That laser head now works as well. It also came with an adjustable polarizer in its end-cap. With that installed on either of the other heads, the output power could be varied continuously from near 0 mW to about 1 mW.
Note that the Ready light comes on and then the laser locks in at the proper phase of the next mode cycle. So, basically the pea brain in the controller (no actual CPU of any kind!) decides that conditions are suitable and enables the feedback loop. The final "decision" is based the cycle duration being longer than some magic number (around 1 minute). :) I've also seen the ready light come on even if the laser doesn't start and when one of the previously locked heads was plugged back in after a few minutes of cooling. In the latter case, the laser was indeed locked though it might not have been able to maintain it continuously since the tube was probably no longer really warm enough.
There are actually two feedback loops in the controller. During warmup, the heater is driven to a fixed temperature based on the resistance between pins 3 and 4 of the Control connector. Once the period of the mode cycle exceeds a fixed time (guessing somewhere around 60 seconds), the control loop based on the difference of the photodiode outputs is enabled. The same signal that switches from the temperature feedback to mode feedback turns the Wait indicator goes off and the Ready indicator on. More on this in the next section.
Plot of Coherent Model 200 Stabilized HeNe Laser Head During Warmup and Plot of Coherent Model 200 Stabilized HeNe Laser Head Near End of Warmup show the output power variation due to mode cycling. Note how it seems to "snap" into regulation once the time is right. :) There are roughly 90 mode cycles during warmup prior to lock. The internal optics account for the large variation in output power. The HeNe laser tube itself has a normal mode sweep of only a few percent.
Another Coherent 200 system I have has a fully functional controller but a fully dead laser head. It is very hard start, impossible to run, and way beyond end-of-life. So, that gave me an excuse to go inside.
The Coherent 200 laser head can be disassembled in a reversible manner with fewer individual parts than the Spectra-Physics 117/A or the essentially identical Melles Griot 05-STP-901. However, it doesn't come apart as easily, using a press-fit for the tube/heater sandwich.
As noted above, the tube was found to be way beyond end-of-life. If it could be convinced to start (on a lab power supply), it would not run at any reasonable current and produced no output at all. There was sputtered aluminum coating on the holes near the cathode end-cap and even through holes in the cathode can near the center of the tube. This system had obviously been left on continuously for a large number of years. It was probably not even in use for a good portion of that time, forgotten and lonely in a corner of a lab, wasting its life producing coherent stabilized photons no one was using until there were no more! :) That seems to be the destiny of so many stabilized HeNe lasers. I'll be searching for a suitable replacement tube. The original tube, a 05-LHR-219 (with or without a -158), doesn't show up in any list I've seen) but an 05-LHR-120 has nearly the same dimensions and will run on the same power supply. So, as long as one can be found that is well behaved (non-flipper, wedged HR), it will almost certainly work fine. Other random polarized laser tubes of similar length can also be adapted but may require replacing the HeNe laser power supply and coming up with a creative mounting scheme if diameter is smaller.
An operation manual and application notes for the Coherent 200 can be found at Ajax Electronics Laser/Optics Manuals under "Coherent".
Everything is in Schematic of Coherent Model 200 Stabilized HeNe Laser. Note that most of the part numbering is totally arbitrary as there were *no* part numbers on the PCB except for the PCB connectors (and I only have J2 in the drawing). This is a late revision with PCB artwork dated 1997, though that probably only means that there was a PCB fab run in 1997, since the artwork itself was obviously hand taped. :) I guess some important customer just had to have more of these lasers made well after they would have been considered very obsolete by Coherent. :)
The controller has two feedback loops. The Preheat Loop, which is active while the tube is warming up, drives the heater in the laser head to a fixed temperature (set by a pot). The temperature sensor in the laser head is not a common NTC thermistor, but something that increases in value with increasing temperature. It has a resistance of around 800 to 900 ohms at room temperature, but well over 1K ohms at operating temperature. The preheat loop prevents the mode feedback loop from going active until the temperature is sufficiently high. Only after this occurs, does a timer begin to look at mode changes, and switches from the preheat loop to the mode feedback loop once their period exceeds around 60 seconds. The mode feedback loop uses the difference between the orthogonally polarized A and B modes in a simple PI control loop to drive the heater. Should the laser not stabilize as evidenced by mode changes still occurring, the preheat loop will be switched back on to try again. At least, that seems to be how it's supposed to work. However, a system with a laser tube that doesn't start (or a bad HeNe laser power supply) will likely turn on READY shortly after being powered up even though it is obviously not working correctly. Well, I guess it IS quite stable - dead with a frequency of exactly 0.0000000000 Hz and an output power of exactly 0.0000000000 mW! :)
Here is the adjustment procedure. A multimeter (preferably an analog VOM, with a needle!) or oscilloscope is required. A 14 pin "DIP Clip" will come in handy, and a laser power meter and temperature probe are desirable but not essential. A hex wrench to set the output polarizer orientation and small flat blade screwdriver to adjust the pots will also be needed.
This should be done from a cold start at an ambient temperature close to where the laser will typically be used. If the laser had been on, it should be turned off and allowed to cool down for a half hour minimum before proceeding.
A printout of the Schematic of Coherent Model 200 Stabilized HeNe Laser will come in handy.
Mode A and B adjustment
The balance between the two polarized modes will affect the location of the lasing line on the neon gain curve. The following sets the two mode amplitudes to be equal, which places the modes equidistant on either side of the gain curve. However, it should be possible to offset the modes if desired, if a different location or slightly more output power in Mode A (the output beam) is desired. However, it's not possible to place either mode precisely at the top of the gain curve.
The HeNe laser tube and ballast resistors dissipate almost 12 W (1.8 kV at 6.5 mA). The temperature set-point must be selected such that it is slightly above what would result from the tube and ballast power alone. At an ambient temperature of 18 °C, the required temperature set-point ends up being around 40 °C, a difference of 22 °C. I do not know exactly how this is affected by a change in ambient temperature. If the difference remains constant, the head must run at 62 °C for the maximum allowable operating temperature of 40 °C (from the specifications in the Coherent manual). Such a high operating temperature seems unrealistic.
One way to estimate the value for the temperature set-point is power only the laser HeNe laser tube (not the heater) by disconnecting the Control cable and allow it to reach thermal equilibrium (at least 1/2 hour). Measure its temperature and then reconnect the Control cable and adjust the Temperature set-point to be about 5 °C higher, or so that the mode sweep goes through an additional 15 full cycles.
The following assumes an ambient temperature of 18 °C:
Note that the range of 5 to 10 V is my estimate. The Coherent manual shows a graph with the voltage at 12 V at the time of lock (which would then likely drop down to under 10 V after thermal equilibrium). But there is no description or indication of what ambient temperature was used. Perhaps some key piece of information is missing. While there's no problem adjusting the temperature so the laser locks and is stable at any given ambient temperature or a reasonable range around it like +/-5 °C, I don't see any practical way the laser could be set up to operate over the entire 0 to 40 °C range spec'd in the manual without running excessively hot, especially under typical conditions (below 25 °C). It would make more sense if R2 was a sensor for ambient temperature so that the temperature set-point was an offset from ambient rather than actual temperature, but R2 looks like an ordinary resistor.
If the laser will be used in an environment where the ambient temperature is much different than where it was tested, readjustment may be needed. The official Coherent Adjustment Procedure (CAP) probably sets the temperature so high that this would not be required over the full spec'd temperature range of 0 to 40 °C, but that shouldn't be necessary unless the laser is to be used near in a sauna. :)
Mode feedback gain adjustment
Finally, power off for 1/2 hour and confirm that the laser will then stabilize properly (after the warmup period) when powered back on.
One other thing that's recommended while the case is opened is to check R39 and R40, the third and forth resistors from the right in the first row at the front of the PCB. These are the current limiting resistors for the Wait and Ready indicators, respectively, and were originally 510 and 1K ohms, both apparently 1/4 W (by size and appearance). There are other current limiting resistors inside the indicator packages, but the voltage across R39 and R40 may still be high enough to greatly exceed the 1/4 W ratings of the original resistors. If so, the PCB will probably be darkened beneath them as well. Measure the voltage across R39 and R40 when their respective indicator is lit. If either is more than 20 V and the resistor is only 1/4 W, replacement is highly desirable, especially for R40 which will be stressed possibly for years on end. :) Suitable values are 1K, at least 1/2 W for both. Yes, Ready won't be quite as bright but it will be much happier! Proper replacement will require removing the PCB but this is just five screws and several connectors. Space the new resistors off the PCB a bit to further aid in cooling. The PCB is easily damaged, so use a proper desoldering tool to remove the old resistors and clean up the holes. Or just cut the leads off at the bodies of the old resistors and solder to those.
The first step in tube replacement is to find a suitable tube. Melles Griot probably won't even sell you a tube, and if they did, it would cost $300 to $400! Although a common type, this seems to be harder to find surplus than it would appear. Most of those that turn up on eBay seem to be the 05-LHP-120 - the polarized version - which is useless for this purpose. Once a suitable candidate tube has been found, it needs to be tested for non-flipper behavior. A tube that is a flipper may still be useful if the flipping is consistent, or if it disappears when the tube warms up, but a totally well behaved non-flipper is most desirable.
CO-200 laser head disassembly:
CO-200 laser head reassembly:
I found an old 05-LHR-121 laser head with a good tube, extracted the tube, and spent way too much time installing it in a CO-200 laser head that had a nearly dead tube. This included a liberal application of duct tape and bailing wire. :) But it works. I knew that this particular tube was a flipper and expected to simply pick the proper mode polarity such that it would lock on the opposite side of the gain curve from the one that flipped. However, it turned out to only flip until it warms up for about 4 minutes or 113 half-mode cycles, then abruptly it stops flipping and becomes well behaved. I have an Aerotech tube with similar behavior, cause unknown.
I don't think this is in what might be called original condition, but it does start right up without problems (no hard-start tube!) and has decent power (3.2 mW or more total from the tube). It locks normally with 1.2+ mW in a single mode.
All in all though, much more effort is required to do a tube replacement on the CO-200 than the SP-117/A.
The controller that went along with this laser head also had minor problems. I had to replace the usual toated dropping resistor for the READY LED but also had to totally rebuild the READY LED assembly itself - both LEDs and their current limiting resistors were fried to a crisp. :( :)
Excel has only a single type of laser, the 1001, a Zeeman-split HeNe laser with a split/REF frequency between 1.5 and 3.0 MHz. But there are at least 2 different case styles. The 1001A and 1001F seem to be similar or identical to the smaller HP/Agilent lasers. I don't know what the difference is between them. The 1001B may be more along the lines of an HP-5518A with a built-in optical receiver. But it's not clear if the 1100B 6DOF Calibration System uses the 1001B laser. :)
Here are the specifications for the Excel 1001A/F lasers (mostly from the file linked above):
Several photos of the Excel 1001F laser can be found in the Laser Equipment Gallery (Version 3.00 or higher) under "Excel Precision HeNe Lasers".
The two most interesting ones are:
Here are some observations/comments:
ID Function --------------------------------------- J1 Polarization mode photodiodes J2 Tube heater J3 Not installed, ???? J4 HeNe laser power supply J5 Input power J6 Backpanel LEDs
So, I need to find a dead Excel laser to go inside. :)
Wire Pin Color Function ----------------------------------------------------------------- A-D NC E ~REF (Zeeman beat signal from internal optical F REF receiver's differential line driver.) G Black Ground H Green Ground J Orange +15 VDC K Red +15 VDC L White NC (-15 VDC on HP/Agilent cable) M +15 VDC N,P NC (Cable Shield on HP/Agilent cable) R Signal Return (REF) S Ground (Optical Receiver) T +15 VDC (OPtical Receiver) U NC (Cable Shield on HP/Agilent cable)
The general approach is shown in Interferometer Using Two Frequency HeNe Laser. The capabilities of this approach are quite impressive. A typical example is the HP-5501B laser head from the HP-5501A Laser Interferometry Measurement System, which enables a position/distance resolution down to better than 10 nm (that's nanometer as in 0.000000001 meter!). More information on inteferometers based on two frequency lasers including descriptions of the optical components can be found in the section: Interferometers Using Two Frequency Lasers. What follows relates mainly to the laser technology.
Here is a comparison of most of the HP two frequency metrology laser models. Of these, as of Winter 2010, only the 5517A/B/C/D and 5519A/B are still in production:
(6) (4,5) Reference Maximum Beam Model Case Tuning Frequency Velocity Diam. Comments ------------------------------------------------------------------------------ 5500A Huge :) PZT 1.5-2.0 MHz 0.4 m/s 6 mm (1) 5500B Huge :) " " " " " 6 mm (1) 5500C Huge :) " " " " " 6,9 mm (2) 5501A Small " " " " " " " (3) 5501B Small Thermal " " " " " " (3) 5517A Large " " " " " 6 mm 5517B Small " 1.9-2.4 MHz 0.5 m/s " 5517BL Small " " " " " " 5517C Small " 2.4-3.0 Mhz 0.7 m/s 6,3,9 mm 5517D Small " 3.4-4.0 MHz 1.0 m/s 6 mm 5517DL Small " >4.4 MHz >1.1 m/s 6,9 mm 5517E Small " 5.5-6.5 MHz 1.6 m/s 6 mm (8) 5517EL Small " ???-??? MHz ??? m/s 6 mm (8) 5517F Small " >7.0 MHz >1.7 m/s 6,3,9 mm (8) 5517FL Small " ???? MHz ??? m/s 6,3,9 mm (8) 5517G Small? " ???-??? MHz ??? m/s ??? (8) 5517GL Small? " ???-??? MHz ??? m/s ??? (8) 5518A Large " 1.5-2.1 MHz 0.4 m/s 6 mm SN below 2532A02139 (2) " " " " " 1.7-2.4 MHz 0.453 m/s " SN 2532A02139, above (2) 5519A Large " 2.4-3.0 MHz 0.7 m/s " (2) 5519B Large " 3.4-4.0 MHz 1.0 m/s " (2)
The 5518A and 5519A/B have a single optical receiver built-in. And of all the HP metrology lasers, the 5519s are unique is having a built-in DC power supply so they simply plug into the wall and feed their REF and MEAS signals to the associated measurement processor/display.
Like the 5500C, the 5518A or 5519A/B can be used in the normal way (e.g., in a 5528A Laser Measurement System), but are generally intended to be set up stand-alone without any additional optical receivers in a 5529A Dynamic Calibrator. For example, the 5519A laser head can be mounted on a cart and aimed through interferometer optics at a cube-corner (retro-reflector) or plane mirror on a tool whose motion needs to be measured precisely.
The polarizing beam-splitter that detects the modes is deliberately oriented so that the separation isn't perfect and a small amount both F1 and F2 are present in each. This results in a beat frequency being generated which is used to produce the reference signal (REF) and to confirm that there is enough beam power to be usable.
Although locking typically occurs in 4 minutes (READY comes on solid), some lasers (perhaps the 5517F and 5517G) may require 20 minutes. My 5517E takes about 9 or 10 minutes. But a fully stable frequency output requires 90 minutes for lasers with a non-vented cover or a vented cover but no fan. Those with a vented cover and a fan require only 45 minutes. From my observations, the frequency oscillates slightly immediately after locking with a period of order of minutes. The amplitude of these oscillations gradually decreases with time and eventually becomes very small. However, the laser likely still meets accuracy specifications during this time.
An internal optical receiver samples part of the output beam and is used to generate the references signal and to confirm that there is enough beam power to be usable.
The 5501B is the only laser to use Pulse Width Modulation (PWM) rather than pure analog to drive the heater inside the laser tube. This was probably done to reduce power dissipation in the electronics, but does result in modulation of the optical frequency by the PWM.
There's no way to tell the version (e.g., 5517C) or reference frequency (e.g., 2.3 MHz) of the tube itself by inspection of the assembly or from its label. They don't have that information explicitly, only a part number:
The difference in tube part numbers for same model lasers isn't entirely clear. It may be a combination of the size of the beam optics and other special features like a particularly high REF frequency or high output power option.
As noted, the older 5501A and 5500C use physically identical tubes with PZT tuning. The tubes in the 5500A and 5500B are the same and functionally similar to those in the 5501A and 5500C, but the construction differs enough to make it impractical to substitute for those. None of these tubes are compatible with any of the other lasers. The chart below shows the Physical (P) and Reference frequency (R) compatibility of the various thermally-tuned HP/Agilent lasers:
(a) (b) 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 0 1 1 1 1 1 1 1 1 1 7 7 7 7 8 8 9 9 (c,d) B A B C D A A A B ----------------------------------------------- (e) 5501B PR R P P P R 5517A R PR PR P P P 5517B P PR P P R 5517C P P PR P R 5517D P P P PR R (a) 5518A R PR PR P P P (b) 5518A P R P PR P P 5519A P R P P PR P 5519B P R P P P PR
If anyone has additional info defining what these other options mean, please contact me via the Sci.Electronics.Repair FAQ Email Links Page.
There are photos of various HP/Agilent metrology lasers in the Laser Equipment Gallery (Version 2.32 or higher) under "Hewlett Packard HeNe Lasers". These include the 5501A laser head and tube, the 5501B laser head, (which is physically similar to the 5517B/C/D except for the connectors), and the 5517A, 5519A, and 5519A/B laser heads. I think the older 5501A tube looks much cooler than the newer ones. :)
The most significant difference between the various lasers is in the Zeeman split reference frequency. A higher frequency enables a faster slew rate for position and velocity measurements. As of Winter, 2010, all the 5517s and 5519s are current Agilent products. General information, descriptions, and specifications may be found by going to Agilent and searching for "laser positioning laser heads" or a specific model number like "5517C". Some of the specifications from the datasheet:
These sound quite incredible but 1 ppm is a frequency of about 474 MHz (1/1,000,000 of 474 THz, the optical frequency corresponding to a wavelength of 633 nm). Thus 0.1 ppm is 47.4 MHz, 0.02 ppm is 9.5 MHz and 0.002 ppm is 0.95 MHz. So, still impressive, but quite reasonable for a well designed stabilized HeNe laser. However, what is somewhat unique about the 5517 and some of the other HP/Agilent lasers is that this absolute accuracy is achieved without the need for any periodic testing or adjustments, by virtue of the design of the mode sampling and locking electronics.
With respect to selecting among the various laser models, if your application has no need for the higher REF frequency (often called the split frequency), there is no advantage to getting a laser like a 5517D as opposed to a 5517B. In fact, the lasers with a lower REF frequency tend to have higher output power and thus may be easier to set up and align especially in multiple-axis configurations. They also tend to be less expensive on the surplus market, though the Agilent price isn't that much different. The only disadvantage of a laser with higher output power is that there can be enough of a detected MEAS signal due to slight angular misalignment of interferometer optics like the 10706A to result in a reading even if the beam to the tool or whatever whose position or velocity is to be measured is blocked or misaligned. The interferometer cube contains a polarizing beamsplitter and if the F1 and F2 orientation are not precisely aligned with the polarizer, there will be a small amount of F1 mixed with F2 and vice-versa even without the reflection from the mirror on tool. With a 400 µW laser and single axis, the required angular accuracy to avoid a false MEAS signal is well under 1 degree with the optical receiver threshold at its default most sensitive setting. And even if the alignment is perfect, polarizing beam splitter and AR coatings are not perfect so there can still be residual mixing. None of this matters once the return beam is aligned since the MEAS signal will be much stronger than the bogus one, but it can be confusing. Increasing the threshold may be desirable to avoid the issue.
And a note about that impressive spec'd lifetime of 50,000 hours - about 6-1/4 years of continuous use. HP lasers used to last a long long time and it wasn't unusual to find an HP laser running fine after 8 years. But I rather suspect this is no longer the case. I've seen many late model (2004 to 2006) Agilent 5517s that were going down hill well before 6 years including at least one that was essentially dead after less than 3 years. These were standard 5517Bs or 5517Cs pulled from semiconductor fabs, either because they failed in normal use, or because they were rejected during preventive maintenance due to low power or the REF frequency going out of spec (which is usually related to the power decline). Thus, even a late manufacturing date is no longer assurance of a healthy laser. Nor would even a close inspection of the HeNe laser tube, as the they appear identical except for the Agilent label - perhaps that's enough! So if you are buying these things new, it probably pays to go for the extended warranty. :)
Waveplates at the output of the HeNe laser tube convert the left and right-hand circularly polarized Zeeman split modes to linearly polarized modes that are orthogonal and aligned with the horixontal and vertical axes of the laser. These two modes usually differ in optical frequency by between 1.5 and 4 Mhz (depending on the specific laser). (Some recent versions of the 5517 may actually go to 6 MHz or more.) The X and Y polarizations are sent down different paths in the metrology application. One is generally a reference length and the other is the distance to be measured or tracked. (It's the change in path length difference that matters so they could both move if desired.) The two beams rather than creating an interference pattern are beat together in a detector that outputs a difference (or heterodyne) signal. If the relative distance between the two beam paths changes by one half wavelength of the laser (about 632.8 nm but accurate to many significant digits!), the phase of the difference signal will change by 360 degrees. The laser also generates an electrical signal from beating the signals together internally. This constant reference is compared to the detector signal and an electronics package counts off the phase shifts and uses it to determine the distance traveled.
A moderately powerful cylindrical permanent magnet does the Zeeman splitting resulting in a pair of circularly polarized outputs at two very slightly different frequencies. F1 is designated the lower frequency and F2 is designated the higher frequency. For the 5501A/B, F1 is vertical (perpenticular to laser base) while F2 is horizontal (parallel to laser base). For the 5517A/B/C/D, 5518A, and 5519A/B, F1 is horizontal (parallel to laser base) while F2 is vertical (perpendicular to laser base). (Exactly why HP switched orientations between the two model series is not clear as there is no benefit to one over the other and it just causes confusion.) The difference F1 and F2 is between 1.5 and 4 MHz depending on the model (as listed above) and also the specific sample of the laser. The distance between the mirrors in the 5501B and all later lasers like the various 5517s is feedback controlled by a heating coil wrapped around the bore inside the tube to force the cavity to maintain the position of the lasing lines symmetric on the Zeeman split neon gain curves as shown in Axial Zeeman Split HeNe Laser Mode Behavior. A 1/4 waveplate converts the circular polarized output to orthogonal horizontal and vertical polarized components which are used externally. F1 is reflected from the thing being measured or tested (e.g., disk drive servo writer) and F2 is reflected from a fixed reference. The difference frequencies (F1-F2) and (F1-F2)+dF1 are then analyzed to determine precise position, velocity, or whatever. This approach has lower noise, greater stability, and is therefore more accurate than the common single frequency interferometer. By using cavity length control to lock the difference frequency to a known reference, the actual optical wavelength/frequency can be set very accurately. Using the MHz range beat signals makes signal processing straightforward and is more immune to noise than the baseband optical signals.
Interestingly, the actual beat or reference frequency does NOT need to be super stable over the long term. Rather, it is the difference between the reference and the return signals that matters and that only depends on the motion of the target reflector, the optical frequency of the meausrement beam, and the speed of light. Thus, although the optical frequency needs to be known to high precision (+/-0.1 ppm for the standard lasers; +/-0.02 ppm for those calibrated to MIL STD-45662), the exact beat frequency of each laser is not precisely controlled or even precisely measured and recorded or used anywhere in the calculations. This is one reason why the listings above include only a range of values. Any given sample will operate somewhere within that range but the exact value is somewhat random depending on the specific characteristics of the tube/magnet assembly, and to a lesser extent, the specific place on the neon gain curve that the lasing line is parked. However, the beat frequency will be relatively constant over the life of the laser. While one might think that locking the difference frequency to a crystal reference would be even better - and the technique is patented - it's not clear that this would be better and might actually be worse. The difference frequency relative to the mode position can change for any number of reasons. But forcing the amplitude of the two modes to be equal as is done in the HP lasers centers the modes on the split neon gain curves, which should be very stable in terms of wavelength.
All of the HP lasers use conventional dual polarization mode stabilization to lock the lasing lines to the split neon gain curve. However, the two signals are not from adjacent longitudinal modes as with most common laboratory stabilized HeNe lasers, but are the two Zeeman split sub-modes differing in frequency by a few MHz instead of many 100s of MHz. In fact, both are the same cavity mode but shifted slightly higher and lower than would be predicted by c/2*L. One twist on the implementation is that the 5501B and those below it on the chart use a Liquid Crystal Device (LCD) polarization rotator to alternately sample the horizontal and vertical polarized modes, and subtracting sample-and-hold to compare them in the error amp driving the heater, rather than the polarizing beamsplitter and dual photodiodes used in many other dual polarization mode stabilized lasers including the 5500C and 5501A. The LCD approach does have a sort of elegance as well as practical benefits. Since the same optical path and photodiode is used for both polarization modes, the sensitivity is identical so the mode balance should be perfect assuming the LCD polarization rotation is 90 degrees. Since the intent is to park the modes symmetrically on the split neon gain curve, this is perfect and thus requires no offset adjustment over the life of the laser as the output power of the tube declines. And, the LCD and associated electronics may in fact be cheaper than a high quality polarizing beam splitter. However, it also creates some artifacts as a result of the digital switching, resulting in small cyclical variations in optical frequency over a period of 2 or 3 seconds. These are of no consequence for most metrology applications, but do detract from the elegance of these lasers.
In fact, the thermally tuned lasers have only one adjustment associated with stabilization, and that is for the temperature setpoint at which the controller switches from pre-heating to optical locking. The resistance change of the actual heater coil is used to sense temperature and there may apparently be variation from one tube to the next. But this is an extremely non-critical setting and won't affect accuracy, only possibly the temperature range over which the laser will remain locked. (5517s with the Newest Digital Control PCB may have no adjustments, or at least none that are obvious!)
One oddity with respect to the thermally tuned laser tubes is the patent reference that appears on the label of all newer ones at least: "Licensed by Patlex Corporation Under Patent No, 4,704,583". The title of this patent is: "Light Amplifiers Employing Collisions to Produce a Population Inversion", filed in 1977 but not granted until November of 1987. The most curious thing is that there appears to be very little of relevance in the patent other than its association with laser action! Nothing in the patent diagrams or text has any obvious connection to the tube assembly design. In fact, the exact same text exists on other more mundane things like a Carl Zeiss-badged Siemens LGK 7634, a bog standard 2 mW random polarized HeNe laser head. I've heard that Patlex is actually a bunch of lawyers and I bet they made out or are making out quite well. :)
There are two possible "simple" causes of the lasing frequency shifts resulting from Zeeman splitting: Mode pulling (which tends to attract each lasing mode towards its respective gain center) and magnetically-induced birefringence of the plasma (which results in the effective cavity length differing for each lasing mode's polarization). Normal mode pulling seems unlikely as its effect in a HeNe laser is typically in the 10s of kHz range at most, 2 orders of magnitude lower than the MHz difference frequency present with these lasers. So, that would leave birefringence, but birefringence effects are even worse at predicting some of the observed behavior. For example, why does cavity loss affect split frequency so dramatically?
The following explanation may be totally bogus but it has the attractive property that using some not-so-hairy math, it is able to predict the approximate behavior of real Zeeman lasers. So here goes:
The mechanism for the shift of the Zeeman modes away from the cavity modes is a type of mode pulling. Normal mode pulling is a form of mode competition where two or more lasing modes are competing for the same population of excited particles. The lasing modes will then tend to be attracted slightly toward the center of the gain bandwidth curve. This will be true in a solid state laser where the gain is homogeneously broadened - all atoms are in the same environment since they are fixed in a solid, usually a crystal lattice. The only time it will be true with a HeNe laser with a Fabry-Perot cavity is where modes are equally positioned on opposite sides of the Doppler-broadened neon gain curve. Since the Doppler broadening is due to motion, photons traveling one way will stimulate excited atoms on one side of the gain curve and photons of the same frequency traveling in the opposite direction will stimulate excited atoms on the opposite side of the gain curve. Even though the Zeeman-split modes are shifted in frequency with respect to each-other, they satisfy this condition when they are near the center of the split gain curves - which is the normal locked position.
The basic mode pulling equation (with gobs of assumptions!) is:
GB FS = FSR * (1 - ---------) GB + CB
CB will be equal to FSR/Finesse where for high-R mirrors with equal reflectivity, Finesse = pi*sqrt(R)/(1-R). For the types of mirrors in HeNe lasers with R close to 1, a good enough approximation will be 3/(1-R). Where one of the mirrors is HR as is the case with HeNe lasers, this will then be near 6/(1-R).
There should probably be a factor of 2 thrown in since what we're interested in is not the shift of a single mode, but the change in distance between the two modes due to both of them shifting by the same amount in opposite directions. However, this can arguably be offset by the fact that the modes are positioned about half way down the gain curves, not far down where the mode pulling effect would be greatest. So, accept that this hand waving cancels out. :)
The HP-5517 laser tubes have a 127 mm (5 inch) cavity corresponding to an FSR of about 1.2 GHz. We assume the neon (Doppler-broadened) GB to be 1.6 GHz. Plugging in some numbers:
1.6 GHz FS = 1.2 GHz * (1 - --------------------------- 1.6 GHz + 1.2 GHz/Finesse
Here are 2 lasers with OC mirrors with reflectivity that has been measured (dissected barcode scanner tube and 5517B tube) and several others where the reflectivities can be estimated based on difference frequency specifications (5517A/C/D):
OC Mirror Cavity F2-F1 F2-F1 Reflectance Finesse Predicted Range Tube type -------------------------------------------------------------------------- 99.0% 600 1.5 MHz 1.2-1.6 MHz Barcode Scanner 98.75% 500 1.8 MHz 1.5-2.0 MHz HP-5517A 98.5% 400 2.24 MHz 1.9-2.4 MHz HP-5517B 98.0% 300 3.0 MHz 3.0-3.4 MHz HP-5517C 97.5% 200 3.7 MHz 3.4-4.0 MHz HP-5517D
OK, so I kind of picked the reflectivities for the 5517A/C/D mirrors to make the results reasonable. With the increasing cavity loss, the output power of lasers with higher REF frequencies will tend to be lower. However 97.5% may simply be too low to lase at all or with useful power on a tube of this length. Eventually, I will measure 5517A, 5517C, and 5517D OCs. But I don't have any tubes that I'm willing to take to bits at the present time, partially due to (1) the physical and emotional trauma that would result and (2) the fact that I haven't located the special chants and incantations required for metrology laser tube sacrifice. :) If anyone has done this, has certifiable 5517A/C/D tube bits, or has a 5517 tube that's already cracked or broken they'd be willing to contribute to the cause, please contact me via the Sci.Electronics.Repair FAQ Email Links Page.
This approach also predicts that the shift and thus beat frequency will be a maximum with the modes centered as they are when locked. Of course, all these nice results based on numerous assumptions may be wishful coincidence, but they are close to what is observed and don't require delving into esoteric plasma physics. Whew! :)
Now back to your regularly scheduled programming. :)
But HP lasers from day 1 (the original 5500A around 1969 even before it had an official model number) and Agilent lasers to the present have all had both 1/4 and 1/2 waveplates with the basic design unchanged over more than 40 years. Further, both waveplates are in mounts that allow the tilt of each one to be adjusted around one of its principle axes. Why? Some possibilities as to the reasons for this more complex setup are as follows:
Real HeNe laser tubes exhibit some small random amount of birefringence both from the fine structure of the mirror coatings as well as from unavoidable geometric asymmetry in their construction. Without a magnetic field or explicit polarization control measures such as a Brewster window or plate, these tend to lock the polarization of the longitudinal modes to a fixed orientation about the tubes optical axis, and 90 degrees from it. Adjacent modes will almost always be orthogonally polarized. In a HeNe laser with an axial magnetic field such as one from HP/Agilent, this will result in the Zeeman modes being slightly (or not so slightly) elliptically polarized rather than pure circularly polarized. So, the orientation of the 1/4 waveplate will matter and only certain orientations (2 or 4) will convert these to orthogonal linearly polarized modes. But the resulting linealry polarized modes in general won't be aligned with the system's X and Y axes, so the 1/2 waveplate is then required to rotate them to match. (The magnetic field will also never be perfectly symmetric or uniform, though I don't know whether this is ever a significant factor in affecting the mode polarization.)
The adjustable tilt allows the exact retardation of each waveplate to be altered slightly. I find it somewhat hard to believe that the reason is simply to be able to use cheapo waveplates that might not always be exactly 1/4 or 1/2 wave! However, this explanation can't be entirely discounted since the accuracy of the retardaion is critical to producing F1 and F2 modes that are purely linear and precisely orthogonally as required for the metrology applications. And, the waveplates are made from what looks like optical-grade mica whose discrete layers preclude the ability to select the exact retardation by controlling thickness. And whether mica waveplates were originally selected based on low cost or zero order or temperature stability or being very thin to avoid significantly shifting the beam when tilted or being what the designers had laying around is not known either. But there might be another reason for this "feature" - namely to further compensate for some deficiency in the modes coming out of the tube, again related to deviation from being purely circular. Or something. ;)
Thus, there are in fact 4 degrees of freedom, though clearly the tilt has a much smaller effect than the rotation. And without a full understanding (possibly including hairy math!), it's difficult to really come up with an adjustment procedure that will work in general. That's the bad news.
The good news is that from experience, swapping the entire waveplate assembly between HP/Agilent tubes is likely to result in acceptable performance without any adjustments other than making sure that the overall orientation is the same. Therefore, it would appear that these errors are usually small. But it does seem that it is sometimes necessary to touch up the tilt of the 1/4 waveplate in order to produce the best mode purity in addition to optimizing the waveplate orientations. Adjusting a waveplate assembly from scratch (random orientation and tilt) may be much more challenging though since it would first be necessary to match the orientations of each of the waveplates to a waveplate assembly to at least allow the laser to lock. Without a stable output, going further would be virtually impossible.
So much for hand waving. :-) One might think that a good source for information relating to Hewlett Packard's metrology laser technology would be patents. But so far, patent searches have turned up almost nothing of relevance. If anyone has knowledge or references related to the waveplate issue or anything else of relevance, please contact me via the Sci.Electronics.Repair FAQ Email Links Page.
The only actual waveplate fine tuning procedure to maximize the F1/F2 separation and achieve optimal orientation I've ever seen for one of the early HP-5500 lasers with built-in interferometer optics. The goal then is to minimize any return of the component that is supposed to be transmitted. It simply says something like:
But there is no explanation as to why this procedure is the way it is. And they don't mention the tilt adjustment on the 1/2 waveplate for some reason.
I did tests of WavePlate (WP) sets from three HP-5517 lasers using a linearly polarized HeNe laser. The results are as follows:
<--------- Orientation ----------> ID# Laser Input 1/4 WP 1/2 WP Output --------------------------------------------------------- 1 5517C 9mm +20° +20° +32.5° +45° 2 5517B 6mm -20° -20° +12.5° +45° 3 5517B 6mm +20° +20° +32.5° +45°
(I doubt that the specific type of 5517 laser or its beam diameter makes any difference. The accuracy of my measurements on orientation is within +/-2° for input and +/-1° for output, though the latter at +45° is probably quite precise based on theory.)
The Input is the orientation required for the polarized HeNe laser to produce a pure linearly polarized beam at the output of the WPs, and thus also of the orientation of the optical axes of the 1/4 WP. (Only if aligned with the slow or fast axis of a 1/4 WP will the polarization remain linear, a requirement for these tests.) As expected, the output orientation is the same in all cases since the desired output will be rotated by 45° to align with the X and Y axes. This is a result of the conversion from circular to linear polarization by the 1/4 WP, at 45° with respect to its optical axes. The orientation of the 1/2 WP was inferred from the transfer function from input to output. Now, it's quite possible that the orientation of the 1/4 WP was chosen at random and not actually determined for the specific tube unless it was found that the adjusting the 1/2 WP alone would not meet specifications. However, I have found that it may be necessary to iteratively tweak the 1/4 WP and 1/2 WP to achieve best purety of the F1/F2 modes - the same result could not be obtained by adjusting only the 1/2 WP. Indeed, with a genuine original HP/Agilent tube that has its WPs optimally adjusted, the purety of the F1/F2 modes is nearly perfect. So, perhaps they start at 0° (or +/-20°!) for the 1/4 WP orientation and go from there.
With the Zeeman tube producing a pure two-frequency left and right circularly polarized beam, all of these WP assemblies would result in pure orthogonal linearly polarized outputs oriented along the X and Y axes. This was confirmed by placing another 1/4 waveplate in the linearly polarized HeNe laser's beam at 45 degrees to produce pure left and right circularly polarized inputs to the WPs. The results were linearly polarized outputs oriented along X and Y. In neither WP assembly, was there any indication that the tilt of either WP was adjusted for other than pure 1/4 or 1/2 wave as the extinction when set for optimal linearly polarized outputs was nearly perfect.
Experiments with common barcode scanner HeNe laser tubes inside strong magnetic fields have also shown similar characteristics. With these as well, the beat frequency would come and go as the tube heated and expanded with this effect becoming more pronounced when the magnetic field encompassed the entire tube as it does with the HP lasers.
I also did some tests with and without the presence of the waveplates from an HP laser. Without waveplates, there was absolutely no indication of a polarization preference in the output beam at any time indicating pure circular polarization. When the optics were installed and aligned to the original blue paint, the symmetry of the beat waveform, if nothing else, was polarization dependent. In addition, just after the output beat appeared as well as just before it disappeared, the polarizer would suppress the beat entirely when oriented so that its axis was parallel or at 90 degrees to the axis defined by the blue paint. These perhaps weren't quite as dramatic as the effects I was hoping for but confirmed some of the speculation at least.
Laser |<-------- Wavelength -------->| Optical Type Vacuum Air Frequency ------------------------------------------------------- 5501A 632.99???? nm 632.81???? nm 473.6122?? THz 5501B 632.991372 nm 632.816759 nm 473.612234 THz 5517A 632.991372 nm 632.816759 nm 473.612234 THz 5517B 632.991372 nm 632.816759 nm 473.612234 THz 5517BL 632.991372 nm 632.816759 nm 473.612234 THz 5517C 632.991354 nm 632.816741 nm 473.612248 THz 5517D 632.991354 nm 632.816741 nm 473.612248 THz 5517DL 632.991354 nm 632.816741 nm 473.612248 THz 5517E 632.99???? nm 632.81???? nm 473.6122?? THz 5517FL 632.991354 nm 632.816741 nm 473.612248 THz 5517G 632.99???? nm 632.81???? nm 473.6122?? THz 5518A 632.991372 nm 632.816759 nm 473.612234 THz 5519A 632.991354 nm 632.816741 nm 473.612248 THz 5519B 632.991354 nm 632.816741 nm 473.612248 THz
I'm not sure what accounts for the two different wavelengths (and thus optical frequencies) among these lasers. There are no obvious physical differences to account for it. The tubes, beam samplers, and relevant portions of the control electronics are all identical. So, it's possible there was a change in isotopic gas-fill or pressure or something else between 5517A/5517B/5518A/5501B lasers and those that came after them. The difference of approximately 12 MHz is still way lower than the commercial-grade error spec of +/-0.1 ppm (roughly +/-47 MHz), so it really doesn't matter. For the Military-grade lasers, the exact optical frequency is measured and included in the calibration report. But a report for one laser I saw had the optical frequency over 10 MHz away from the spec'd value anyhow. My contact at NIST doesn't even know whether it's an actual change in optical frequency or simply an upgrade to the calibration in the measurement electronics!
I've compared the optical frequency of multiple 5517 and 5501B lasers and have found no evidence of any real difference, let alone one averaging 12 MHz as shown in the specifications above. Here are some data. These lasers are listed in more or less the order in which they were tested:
Locked REF/ Balanced Laser Laser Output Split Frequency ID Type Power Freq. Difference Notes/Comments ------------------------------------------------------------------------------- 1 5517B 660 µW 2.3 MHz -2.30 MHz Faulty beam sampler was replaced 2-0 5517B 480 µW 2.4 MHz -1.44 MHz Laser 2 with its (new) beam sampler 2-1 " " " " " " -1.35 MHz Laser 2 with beam sampler 1 2-2 " " " " " " -6.75 MHz Laser 2 with beam sampler 2 3 5501B 450 µW 1.9 MHz 0.00 Mhz Other lasers referenced to Laser 3 4 5517E 120 µW 6.3 MHz -2.10 MHz Only laser with digital Control PCB 5 5517D 120 µW 3.6 MHz -8.25 MHz 6 5517C 260 µW 2.7 MHz -15.60 MHz Tube run at 4.0 mA (not 3.5 mA) 7 5517C 240 µW 2.9 MHz -9.00 MHz " " 8 5517C 210 µW 2.7 MHz -8.58 MHz " " 9-0 5517D 80 µW 3.7 MHz -9.10 MHz Laser 9 with its beam sampler 9-1 " " " " " " -7.10 MHz Laser 9 with beam sampler 1 9-2 " " " " " " -11.10 MHz Laser 9 with beam sampler 2 10-0 5517A 550 µW 1.7 MHz -7.63 MHz Laser 10 with its beam sampler 10-1 " " " " " " -2.13 MHz Laser 10 with beam sampler 1 10-2 " " " " " " -9.33 MHz Laser 10 with beam sampler 2 10-3 " " " " " " -3.44 MHz Laser 10 with beam sampler 3 11 5501B 220 µW 2.1 MHz -10.70 Mhz 12 5501B 150 µW 1.8 MHz -8.65 Mhz Tube run at 4.0 mA (not 3.5 mA) 13 5501A 100 µW 2.0 MHz -23.48 MHz Really high mileage! 14 5501A 50 µW 2.1 MHz -25.42 MHz " " 15 5501A 35 µW 2.0 MHz -34.47 MHz " " 16 5517A 410 µW 1.6 MHz -0.34 MHz Tube run at 4.0 mA (not 3.5 mA) 17 5517D 405 µW 3.7 MHz +75 MHz Rebuilt with non-HP/Agilent tube
Diagram of Test Setup for HP/Agilent Laser Optical Frequency Comparison shows the way these measurements were made. Photo of Test Setup for HP/Agilent Laser Optical Frequency Comparison shows how ugly it really is, but the scope on the right is displaying the actual beat signal of a pair of 5517Bs. For later measurements, two external HeNe laser power supplies (the white boxes and the Variac) were added to eliminate the FM introduced by the switching noise/ripple of the internal HeNe laser power supplies, but they weren't used for these tests and wouldn't have affected the results since the high frequency FM would be averaged out.
The 5517 lasers were enclosed in standard cases (non-vented for all except the 5517E) and allowed to reach equilibrium (2 hours minimum). (Removing the cover may significantly change the optical frequency once equilibrium is reached.) The lasers in the photo don't have any cloths on, but, well, that's another matter! :) The 5501As were run naked so that the "Photodiode Offset" adjustment could be performed. (More below.) It might be best to do this via a hole in the cover as they do drift significantly with the cover in place. But I wasn't *that* enthusiastic!
Laser 3, the first and healthiest 5501B to be tested, was chosen arbitrarily to be the reference for optical frequency. The Balanced Frequency Difference is the frequency of the mid-point between F1 and F2 for the subject (ID) laser minus the frequency of the mid-point between F1 and F2 for Laser 3. There's still a +/-1 MHz or more uncertainty due to variations in the specific lock point of the two lasers being compared during any given run.
The three 5501As are very well used and weak, but it was easy to obtain a beat between them and the 5517A, Laser 10 (which happened to be the 5517 laser tested and thus conveniently left in place!). 5501As have a "Photodiode Offset" adjustment, which moves the lasing line on the split gain curve. It's the square pot (R4) on the Lock Reference PCB, clockwise rotation decreases optical frequency. I could have set them all to have a 0 MHz difference frequency, but this would have resulted in grossly unbalanced mode amplitudes for these high mileage lasers. So, they were adjusted according to the HP procedure - maximizing the F1-F2 REF frequency, which centers the lasing line between the split neon gain curve. Before doing this, it wasn't even possible to see the difference frequency with Laser 13 likely because it was too high for my instrumentation. This was probably because parts of Laser 13 had been swapped, including the tube, without making any adjustments. The Photodiode Offset adjustments on the other 2 5501As were quite close to optimal. However, this is a single turn pot which adjusts the mode ratio from 1:2 to 3:2, and thus the optical frequency varies significantly with very small changes in its position - possibly 50 MHz or more end-to-end. Going only by the REF frequency - which isn't perfectly stable - it's quite likely that there will be an uncertainty of 5 or 10 MHz. So, best would be to adjust this pot (make it a 10 turn pot!) through a hole in the cover after the laser has reached thermal equilibrium. And if what you want is a precise optical frequency and don't mind some possible mode imbalance, adjust it with respect to a reference laser like an iodine stabilized HeNe laser instead of for maximum REF frequency! :) And, although the optical frequency changes with the cover installed, the Photodiode Offset adjustment could still be optimal if the change is due to the tube temperature, and thus the gas pressure increasing. That would still maintain the same mode balance.
Laser #17 was rebuilt by a company other than HP/Agilent. (I'm not at liberty to reveal the company name.) It is believed to have had its original laser tube removed from the magnet/optics assembly and replaced with a laser tube that is not from Agilent. This explains the large offset in optical frequency, which could result from any number of factors. (See below.) The offset is probably of little practical consequence as long as it remains relatively constant.
Aside from laser #17, based on these data, there is really no consistent difference in average optical frequency based on laser type and if anything, it goes the wrong way! And note the change resulting from the swap of the beam sampler. Beam Sampler 1 was originally on Laser 2 and was resulting in the optical frequency dancing around, then swapped with Beam Sampler 2 which resulted in a large frequency offset, then with a third Beam Sampler which was finally well behaved and now remains in that laser. I have no reason to suspect anything is wrong with either Beam Sampler 1 or 2 and did test them for basic functionality with a voltage source and polarizer. All beam sampler assemblies I've checked regardless of what laser they came from have exactly the same part number though it's possible that the optics inside differ in some subtle way depending on laser type. There are at least two versions of the housing - one with a small aperture for 6 mm optics and another with a large aperture for 9 mm optics, but beyond that I don't know of any differences. Lasers 3 through 7 definitely have their original beam samplers. Though I don't have minimum specs for the 5517E, Laser 4 is probably relatively high mileage. And lasers 5 and 9 are high mileage as evidenced by their low (below spec) output power. Even though the output power of Lasers 6, 7, and 8 is well within spec, they are also definitely high mileage lasers being extremely slow start and unable to run on the normal 3.5 mA discharge current. This in itself shouldn't have a large effect on optical frequency unless the tube actually runs hotter (in which case the optical frequency should increase, more below). But the lock point temperature adjustment has not been changed on these lasers, so the equilibrium bore temperature should be similar to that of the others though the equilibrium laser tube envelope and laser temperature will be slightly higher.
So, the actual optical frequency may be dominated by the amount of use (number of hours on the tube) which also tends to correlate with a decline in output power. This may overwhelm any real or fictitious optical frequency offset found in the specifications. While I don't know what the original output power was for most of these lasers, those with 400 µW or more start very quickly or instantly and are likely relatively young (usage-wise). Laser 1 is known to have been taken out of service due to a bad LCD in the beam sampler, so it could have seen relatively little use.
There has been research showing that the neon gain center frequency tends to decline with use due to a drop in tube pressure and other factors. Helium has an effect on lasing center frequency of about +22 MHz/Torr, so a loss of He due to gas entrapment on the tube walls or cathode, leading to a drop in its partial pressure, can easily account for these large frequency differences. (Loss of He due to diffusion through the tube walls would also result in a decline in its partial pressure, but this loss mechanism should be minimal.) Major factors include:
Cause Sensitivity Comments ------------------------------------------------------------------------------ Helium Pressure +22 MHz/Torr Pressure of He decreases with use Neon Pressure -25 MHz/Torr Pressure of Ne decreases with use Neon Isotopic Ratio +10 MHz/% of 22Ne Ratio of 22Ne:20Ne Decreases with use Temperature +280 kHz/°C Affected by specific lock point
Both He and Ne partial pressures descrease over the life of the tube but because the fill ratio is between 5:1 to 9:1 of He:Ne, the decrease in He pressure dominates and a frequency drift downward of several MHz/year is quite reasonable. If not filled with a pure Ne isotope, the Ne isotope ratio also will change slightly as the 22Ne will be trapped at a slightly a higher rate than 20Ne. Note the strong dependence on the Ne isotope ratio, a 1 GHz range! So, just over a 1 percent change in the ratio at the time of manufacture could account for the 12 MHz difference in nominal frequency specifications. And, for any given measurement, there is uncertainty in the actual lock point as the laser warms up but that's probably only a maximum of +/-1 MHz or so. The 280 kHz/°C is for a tube about 8-1/2 inches long - similar in length to most of the HP/Agilent tubes. However, note that for the HP/Agilent lasers, the temperature of the tube envelope is not controlled, only the mirror spacing rod for the 5501B and 5517s, and not at all for the 5501A. So, the temperature of the interior of the laser may have a significant impact on the optical frequency. This differs from many other stabilized HeNe lasers where a large portion of the tube is wrapped in a heater.
The above has been distilled from the paper: "Frequency stability measurements on polarization-stabilized He-Ne lasers", T. M. Niebauer, James E. Faller, H. M. Godwin, John L. Hall, and R. L. Barger, Applied Optics, vol. 27, no. 7, 1 April 1988, pp. 1285-1289.
So, it's quite possible that any differences in the optical frequency of these lasers when they were new is totally swamped by changes due to use. For example, if a laser has been run 24/7 for 3 years (middle age for these lasers!), its optical frequency could have gone down by 10 to 15 MHz due to the decline in gas pressure and isotope ratio changes. But the ultimate conclusions may be that (1) it's not worthwhile to assume anything about the nominal optical frequency on used HP/Agilent lasers, but if the optical frequency can be measured (or compared to that of a new laser), (2) the frequency shift may be a means of estimating how many hours or years they've been on! :-)
Despite all these potential source of variability, for an application requiring an accurate stable optical frequency reference like calibration of a wavementer, a healthy 5517 laser (any version) is probably a better choice than a laboratory stabilized HeNe laser like a Spectra-Physics 117A. The reason is that the design of the 5517 inherently locks to a balanced mode state, with no adjustments and little in the electronics to drift with age to change this. Lasers like the SP-117A have separate photodiodes and pre-amps for the two mode signals as well internal adjustments that can affect the lock point. Furthermore, the optical frequency specifications of all HP/Agilent lasers are known (even if there is an unexplained discrepancy of 12 MHz going from the 5517B to 5517C). This is not the case for many other stabilized HeNe lasers. And, if needed, the laser can easily be packed up and sent to NIST or elsewhere to have its optical frequency measured precisely without fear of it changing either from a few bumps during shipment, or over time if turned on periodically rather being run 24/7. I wouldn't recommend other HP lasers like the 5501B simply because healthy ones are becoming harder and harder to find. And the 5501A uses a different locking design which is similar to that of the other (non-HP/Agilent) lasers. However, a healthy 5518A or 5519A/B would also be suitable, using the same design as the 5517.
Also see the section: Comparing the Optical Frequencies.
So, if you're salivating for an HP/Agilent laser and can't live without one, they cost somewhere between $8,000 and $12,000 new depending on options! Used (or "previously owned" - which would be classier!) HP/Agilent lasers can be had much cheaper but caveat emptor. The only ones most people can afford for personal use would be found on eBay. But most of these lasers that end up being resold are taken out of service because they have an end-of-life tube. Interferometry lasers used for metrology are often run 24/7 from the day they are installed until they die. Even though the lifetime of the special HeNe laser tubes used in these lasers may be 50,000 hours, that's still only about 6-1/4 years. And guess where they then end up? :) If the seller hasn't powered the laser head (or doesn't admit to it) and lists the laser "as-is" with no returns, chances are excellent that it will serve as a nice door stop but not much else, at least not without some effort. Unfortunately, except for the 5519A which plugs into a standard wall socket, these lasers require +/-15 VDC for power with a Military-style connector that you won't find at Radio Shack. It's easy to "hot wire" power from inside, but unless the seller is familiar with this sort of thing or has the mating power supply and cable, it may be better to just get a DOA warranty in writing and accept that you may have to pay shipping both ways if the laser is only good as a door stop.
These lasers also show up at surplus dealers but they tend to ask higher prices than would be considered acceptable for basic tinkering and many seem content to simply have the laser gathering dust than to let it go for a realistic price if untested. But I've also heard of at least one instance where such a laser was found at a garage sale. That price was almost certainly right!
Also, note that when looking for lasers like this on eBay or elsewhere, the clothes these lasers wear are of little importance. Newer Agilent OEM 5517 lasers (which are mostly what show up as late model surplus in 2010) tend to have a thin cheaply made gold-ish (alodined) aluminum shroud rather than the beige or gray two piece case of most older HP lasers. It has a feeble attempt at a rubber gasket all around to seal it but this really doesn't work well and only makes reassembly a royal pain. What's inside is the same, though lasers built after the early 2000s will likely have the newer Digital Control PCB rather than the analog Control, PCB but they are otherwise identical and functionally equivalent. And in the trivial triviality department, Agilent's only concession to style seems to be in the color of the front and back plates: Beige for 5517Bs, silver for 5517Cs, and gold for 5517Ds! :) But this is only true of some samples and there doesn't appear to be any way to predict which ones.
Regardless of who is selling the laser, if they are able to power it, the three most important things to ask of them would be:
"Yes" and "a few seconds or less" means the HeNe laser tube and power supply are probably good and happy working together. A laser that takes awhile to start may still be fully functional, but it can be annoying to wait 10 minutes for a beam, and associated equipment may expect the laser to be ready within a fixed amount of time.
Note that the 5501B is the one exception where the laser tube isn't turned on until near the end of the locking process. Thus a beam that doesn't appear immediately on the 5501B is a feature, not a bug. :) This was probably done in the design to minimize the DC current consumption during warmup to be backward compatible with the 5501A. My preference, where this isn't an issue, is to bypass the transistor switch and enable the tube immediately as with all the other lasers.
A "yes" answer to this question alone is usually sufficient to confirm proper operation with usable power for many purposes. However, cold start to READY on solid may be over 10 minutes and up to 20 minutes for a few lasers like the 5517E, 5517FL, or 5517G, and some other 5517s using the Newest Digital Control PCB, but these lasers are almost non-existent surplus. Any 5517 laser from before the year 2000 - which will be most of those found on eBay even in 2010 - should come ready in the typical 4 minutes. A slightly longer time like 5 minutes is of no consequence, but if a pre-2000 laser takes several minutes longer, it's probably very low power and requires the extra time for the power to increase enough as the tube warms up for the laser to be convinced there is enough power available. However, such a laser could still be useful.
However, some types of data processing systems like the HP-5508A Measurement Display will produce a hard error if the laser takes more than 10 minutes to become ready. (Power cycling the 5508A once the laser is ready will generally get around this even if it powers the laser, as it's likely to become ready much quicker once warmed up.)
For many applications, much less than spec'd minimum power is quite sufficient. Even if the seller is unable to measure the output power, as long as READY comes on solid, it is probably at least 80 µW for all the lasers except the 5501A, which will lock at much lower power - down to 40 µW or less. Even this is sufficient for a single axis system.
Where the laser passes these tests, it will probably be more than adequate for an experimental, demo, test, educational, or research system which doesn't have many measurement axes and isn't run continuously for years. However, before considering such a laser for installation in a semiconductor wafer stepper producing next generation multi-core processors, many additional tests would need to be performed to determine its present health and life expectancy. In some installations, the laser is swapped out after a fixed number of hours, like fluorescent lamps! :) While in others, they are replaced at the point where they are at the hairy edge of meeting HP/Agilent specs. Either of these approaches makes sense where where the cost of down time is extremely high. So, even though they may start instantly, run reliably, and have decent output power much greater than the HP/Agilent minimum, if their REF frequency is found to be at or above the range for that model laser, they may be flagged for replacement during preventive maintenance. (REF frequency tends to increase with use and is related to the decline in output power.) An example would be a 5517B outputting 500 µW with a REF frequency of 2.6 MHz. The spec'd REF frequency range of a 5517B is 1.9 to 2.4 MHz. Details are beyond the scope of this presentation, but there may be a writeup in the future. Stay tuned.
(If anyone has an HP/Agilent laser and has a record of the output power and REF frequency when new either from measurements, the label, or original paperwork, and what they are now, and if possible, an estimate of how much it has been run, please contact me via the Sci.Electronics.Repair FAQ Email Links Page. This will aid in my attempt to more accurately estimate previous use and life expectancy for these lasers.)
Even if the laser plays dead, it could just be a blown fuse or something else that's easily repaired. Or, it could be *really* slow start.
For details, see the section: Common Problems with HP/Agilent 5517 Lasers.
And even if the HP/Agilent laser tube is certifiably dead, it is possible to install an inexpensive barcode scanner tube in its place that results in a usable system, at least for experimentation or demos. This isn't for the casual user, but if you're up to a modest challenge and have some basic mechanical and electronic skills, see the section: Installing a Common HeNe Laser Tube in an HP-5517 or 5501B.
For more information on alternatives to purchasing new HP/Agilent lasers and critical issues in their selection and testing, see the companion document: Considerations in Evaluating Used or Rebuilt Hewlett Packard/Agilent Metrology Lasers.
And, if you do come across one of these lasers at a garage sale, just splurge, pay the $2 they're asking, and take the risk. :-)
So, in short, the laser itself won't function any better if run continuously compared to being turned on at most 90 minutes before needed as long as it doesn't affect the environment in such a way as to change the calibration. (90 minutes is HP/Agilent's spec for warmup to full accuracy on an unvented laser, only 45 minutes on one with forced air cooling). And for less critical applications, simply waiting until READY comes on solid may be adequate. It should be possible to test for the overall effect by making a measurements of a known length in each axis when the laser comes READY, after 90 (or 45) minutes, and after 24 hours. If any differences found are acceptable, there is nothing to be gained by continuous operation.
Where the laser might be used for a few hours a week, as in a diamond turning machine at a custom optics house, this should effectively extend the life of the laser to infinity.
(The following deals with retrofitting systems using 5501A or 5501B lasers. For really old systems using 5500A/B/C lasers, a few more issues are present since the 5505A Measurement Display is more tightly coupled to the laser and somewhat more is involved to keep it happy.)
It is often possible to install a more modern 5517 laser in place of a 5501A or 5501B. 5517 lasers are still in production and used working units are also readily available at very reasonable cost.
Only a few relatively minor differences need to be accommodated. With only a bit of resourcefulness, the total cost (excluding the laser and labor) for this type of conversion will be under $100 in most cases:
However, even a tube deemed to be dead by Agilent due to low power or an inability to stay lit, may often be made usable for many applications (especially where only 1 or 2 measurement axes are required), for a test or educational system, or as an emergency spare, with at most some relatively minor low cost modifications to the laser, or possibly even simply an adjustment. But if the output power is so low that the beam actually disappears periodically while warming up, there won't even be a beat signal and such a tube is only good as a high tech paperweight with built-in magnetic paper clip holder. :)
Assuming the tube is usable, except for late model Agilent 5517 lasers based on digital Control PCBs, all of these lasers are very serviceable as far as the electronics are concerned. Pre-2000 lasers - most of what's found surplus even in 2010 - will certainly have the older Analog Control PCBs. For these, even most of the HP house-numbered ICs have standard equivalents available from major electronics distributors, and none of the other electronic parts are special. Operation and service manuals are available which include detailed adjustment and troubleshooting information and complete schematics. And parts units can be obtained on eBay at low cost. Except for a blown fuse of my own doing, dried up electrolytic capacitors on really old lasers, a blown line driver chip, and bad REF photodiode, I've yet to see an Analog Control PCB with any serious problems including defective proprietary HP ICs. However, on a 5501B, the heater driver transistors and main fuses were blown as a result of dried up electronic capacitors on the Connector PCB. So, for 5501B lasers, it's probably good preventive maintenence to replace all 4 large electrolytic capacitors on the connector PCB on a laser more than 10 or 15 years old as a precaution. This is the only situation I know of where a high ESR/low uF capacitor will result in actual damge to other components. For more on the 5517 laser in general and the digital Control PCBs in particular, see the section: HP/Agilent 5517 Laser Construction.
Now, if you're independently wealthy and would like to have Agilent repair your laser, I've heard that an evaluation is about $500. Essentially, they confirm that it's an HP or Agilent laser and then tell you how much it will cost to repair, if they are willing to repair it at all. For a single failure, the cost is a flat rate between $1,500 and $2,000, but the evaluation fee will be applied toward that, thank goodness. :) A "single failure" probably includes a blown fuse, broken resistor, dried up capacitor, or bad IC. I don't know whether something like a degraded LCD in the beam sampler or blown HeNe laser power supply would qualify as a single failure, or if two dried up capacitors would be charged (no pun....) for separately. And, it's almost certain that if you read the fine print, the flat rate would exclude a weak or dead HeNe laser tube that required replacing the tube assembly even though it is technically a "single failure". In that case Agilent would simply return the laser after collecting their evaluation fee.
For amusement, go to Find-A-Part: Agilent's Test and Measurement Parts Catalog and enter a laser model like "5517D". If you're not independently wealthy, you better be sitting down when viewing the prices. For example, (in 2009) the cover is $344, the Control PCB is $1075, the HeNe laser power supply is $496, and a small screw is a bargain at $1.24 each. However, prices for the operation and service manuals are not totally ridiculous - $28.44 for the 5517A and $42.67 for the 5517B/C. But the exact parts available for each model laser seem to be somewhat random and forget about even being able to order a new tube assembly (or parts). They are listed as: "Not orderable, contact Agilent for repair service". Right. :-)
And before doing something silly, getting inside HP/Agilent lasers is trivial. On the large lasers (5517A, 5518A, 5518A/B) it's just a matter of removing the 4 tiny screws on top and gently levering up the cover using a knife blade. On the small lasers (5501A/B, 5517B/C/D), rotate the front turret so the large hole is at the bottom. That will expose a slotted head screw - a 1/4 turn fastener. Push in and rotate 1/4 turn counter-clockwise and the front plate will pop off. The covers or shroud can then be removed. The only reason I've gone to this level of detail is that I had an academic type ask me if that screw was for tuning the laser frequency! :)
Also see the section: Common Problems with HP/Agilent 5517 Lasers (which applies to other lasers like the 5501B as well). For operation and service manuals, see the section: Additional HP/Agilent Resources.
I also have backup copies of the same PDFs at Sam's Bakcup of Agilent Laser and Optics System Design Manual.
Note that the file for Chapter 7Y does not exist on the Agilent Web site and I haven't been able to find it elsewhere. I've left the link in place should it magically appear.
The HeNe laser tube in the 5500A/B is generally similar to the one in the 5500C and 5501A, but isn't quite identical and thus is not interchangeable, at least not without some work. The original patent for the 5500A/B laser tube is: U.S. Patent #3,771,066: Gas Laser. The most notable obvious differences between the 5500A/B tube and the one in 5500C and 5501A are in the PZT connector at the rear which is a ring (rather than a center terminal) that allows the waste beam from the HR mirror to escape, and the optics assembly at the front of the tube assembly which only has the beam expander - the waveplates are mounted externally (though strictly speaking these aren't part of the tube itself). And the glass tube is simply clamped to the mounting feet, which are not part of the tube assembly.
Rather than using a portion of the main beam for feedback, there's a shielded can with a photodiode behind a 1/4 waveplate and motor driven rotating polarizer that samples the waste beam from the back of the tube. The photodiode signal is used in a feedback loop to lock the laser so the modes are of equal amplitude. (See: U.S. Patent #3,701,042: D.C. Motor Circuit for Rotating a Polarizer and Providing a Detector Synchronizer Signal for a Laser Stabilizing System.) Ironically, this is actually closer in function to the LCD optical switch of the 5501B and later lasers, than the polarizing beam samplers of the 5500C and 5501A that followed the 5500A. Since the 5500C/5501A tube has no waste beam exiting the laser tube, duplicating this function would be a bit of a challenge.
The 5500A is in the same size case as that of the 5500C. The main difference between the 5500A and 5500C is what's at the front of the laser. The 5500A has interferometer optics and detectors for both REF and MEAS within the case. The 5500C has two channels of optical receivers but no interferometer optics. However, it was possible to install linear interferometer optics inside the 5500C to give it 5500A functionality.
The 5500A is also unique among HP lasers since it is the only one with a run-time (hour) meter!
There are photos of a 5500A in the Laser Equipment Gallery (Version 2.49 or higher) under "Hewlett Packard HeNe Lasers".
For several original articles introducing HP's interferometer-based measurement system using the 5500A, see the Hewlett Packard Journal, August 1970.
Also see Dave Meier's HP Laser Interferometer Evolution Page which includes a links to the early HP catalog pages.
I have a 5500A laser (see gallery pages, above) which appears to be from around 1970 based on the date code found on a 74H10 TTL IC in the optical receiver. Except for the shape of the beam expander mount and color of the ballast resistor cover, my 5500A appears identical to the laser shown on the last page of the August 1970 HP Journal. An external HeNe laser power supply was used to perform initial tested before being connected to a 5505A Measurement Display. The laser tube starts and runs flawlessly with a raw output (after the beam expander but before the waveplates) of at least 370 µW and possibly as high as 450 µW. (The power varies with temperature as the tube warms up if not feedback stabilized and I didn't run it long enough by itself to determine the actual maximum power.) Even the low end of 370 µW would be considered excellent power for a much newer 5501A tube. The output power of the laser is between 106 and 150 µW (again depending on the temperature as it's not locked). If the locked output is anywhere near the higher end of this range, then it's basically like it was when it was last serviced. There is a note inside the laser saying: "120 µW August 1978". Perhaps the tube was also replaced at that time. The reason for the large difference between tube output power and laser output power is that the waveplates cut the power by 15 to 20 percent, and the internal interferometer optics suck up approximately half of the remainder since most of the F1 frequency component doesn't exit the laser.
When first attached to a 5505A, the laser powered up and locked instantly, and within a couple minutes, I was able to make sub-micron measurements! But, then at some point while my back was turned, the original HeNe laser power supply inside the laser head failed. Hard to believe! Not like the thing has probably been turned on for the first time in 20+ years! :) I don't know if the failure was in the two transistor driver, or inside the potted HV module, which is beautifully made in clear semi-flexible plastic with no obvious damage. But there could be a shorted turn in the inverter transformer or a capacitor breaking down. The driver transistors passed ohmmeter tests and were getting equally warm, but the output was only going to around 1 kV and then dropping to 0 V, never lighting the tube. So, I replaced it with a small brick power supply from a barcode scanner, installed inside the original aluminum can to preserve authenticity. Unless one knew exactly where to look, there would be no way to tell that it wasn't totally original.
One thing that's probably only of curiosity value is that both the HeNe laser HV power supply and the PZT HV power supply are driven from a common oscillator which must be running for the PZT tuning to work. Without tuning, the 5505A readout may still function, but the RESET button will keep flashing. Newer versions of the 5500C, as well as the 5501A use independent self-oscillating inverters in ugly bricks made of hard tan potting compound for these two power supplies. The earliest 5500Cs are probably similar to the 5500A.
It's extremely easy to align the interferometer with my home-built authentic replica of the retroreflector mount shown in the 1970 HP Journal article. As long as it adjusted so the return beam enters the optical receiver aperture or even the tiny alignment holes in the laser head turret, the system is happy.
And here is the genuine imitation authentic setup hot off my time machine:
More information and photos from early HP manuals and brochures, and elsewhere can be found at Dave Meier's HP Laser Interferometer Evolution Page.
The cable wiring is given in the next section since it is the same for the 5500A and 5500C.
More to come.
There are photos of a 5500C in the Laser Equipment Gallery (Version 2.48 or higher) under "Hewlett Packard HeNe Lasers".
Also see Dave Meier's HP Laser Interferometer Evolution Page which includes a link to the early HP catalog pages.
The 5500C uses a HeNe laser tube with PZT tuning that appears identical to the one in the 5501A, though the part number differs. (The 5500A has a very similar, though not identical tube. See the description and patent reference in the previous section.) The beam sampler for the feedback stabilization is of the common modern polarizing beam-splitter variety with the control loop driving the PZT of the laser tube to adjust cavity length. But, unlike the 5501A which only requires DC power supplies, the 5500 requires the mating 5505A Measurement Display to even turn on and stabilize since its HeNe laser power supply and PZT power supply are controlled by the 5505A. Although the HeNe laser power supply could be run open loop with a variable DC voltage, this would not provide current regulation. However, the PZT power supply of later 5500Cs which appears to be a potted module inside more potting, can be used as a stand-alone PZT, PMT, or other variable HV low current power supply since its output is fairly linear with respect to input from 0 to 15 V, which is multiplied approximately by somewhere between 100 and 200 to produce the output voltage. (Although I have not seen it specifically stated, the PZT power supply appears to be capable of more than 2 kV based on the 5501A schematics.) Both HV Control and PZT Control are really just the power input to a self oscillating inverter. (Very early versions of the 5500C and the 5500A have the inverter transformers and other high voltage components potted inside metal cans with the driver circuitry on separate PCBs fed from a common oscillator.)
The pinout for the self contained PZT power supply module is:
5500A/C and 5505A connector pinout
Pin Function ------------------------ A Gnd B DOPPLER (A) C +5V D LOCK (A) E HV CON F REF TRIP G -15V H BEAM AL J PZT MON K REF (A) L GND M REF (B) N DOPPLER (B) P NC R LOCK (B) S LASER I T +15V U PZT CON
If constructing your own cable, the wires to pins B and N should be shielded twisted pair, shield to pin A, and the wires to pins K and M should be shielded twisted pair, shield to pin L. The shield probably isn't critial for relatively short cables, but use the twisted pair. Size the voltage (+5 and +15) and Gnd wires to handle a couple amps. HV Control may also need to supply some current.
The 5501B is a functional replacement for the 5501A. Locking of the 5501B typically takes 5 to 9 minutes compared to 10 seconds or so for the 5501A, but this is of no consequence for machines that are run for hours or years. In terms of optical characteristics, and power requirements and reference signals (including connector pinouts), they are equivalent. However, the 5501B lacks the Diagnostic (J3) connector of the 5501A, so other system components may not be happy and some substitutes may need to be provided. Going the other way doesn't have this issue, but if a 5501A is installed in place of a 5501B, it may be necessary to press the Retune button from time-to-time whereas there is no such button or need on the 5501B! This may be anywhere from a few hours to never, but it would be a good idea to do this periodically at convenient times between measurement runs, at least until the system has reached thermal equilibrium. Performing a Retune cycle does not compromise the accuracy in any way. Once the Retune LED goes out, it's ready to go again. Even from a cold start, a laser may go 12 hours or more without requiring a Retune. After that, once a day may be more than sufficient.
Operation and service manuals for the HP-5501A and HP-5501B may be found on the Hewlett Packard/Agilent Metrology Laser/Interferomter Page.
Compared to the 5500C, the 5501A is in a much smaller lighter case with simplified optics and totally different electronics. See Interior of the HP-5501A Laser Head - Left Side and Interior of the HP-5501A Laser Head - Right Side. The HeNe laser tube dominates the interior space in both views. The high voltage piezo driver power supply brick is visible under the magnets at the center of the tube. The HeNe laser power supply brick is underneath the output end of the tube. The piezo driver electronics circuit board at the far right end of the right side view. The optical sensor circuit board is at the far left of the left side view.
The naked tube is shown in HP-5501A Laser Tube Removed From Magnet and Output Optics Assembly. The normally enclosed part is really just a very think walled fine-ground bore inside an outer glass envelope. A spring (visible through the glass at the left) at the rear holds the PZT, HR mirror, bore, and OC mirror in place. No adjustment is possible. There are distinct multiple spots on the card because the output window is at a slight angle and not AR coated.
Both the HeNe laser power supply and piezo power supply run off the -15 VDC power supply. An interlock switch (easily defeated) prevents operation with the cover removed. In the 5500A and 5500C, these power supplies are regulated by the 5505A Measurement Display. In the 5501A, the potted power supply bricks have no inputs other than power. Rather, current and voltage regulation are accomplished by controlling the input current. For the HeNe laser power supply in the 5501A, as well as later versions of the 5500C, while the passive HV components are buried in potting compound, the two 2N5192 driver transistors are mounted on the outside of the brick and are replaceable. However, from my experience, when the transistors blow, there is probably a fault in the potted section so replacing them doesn't help, I've successfully replaced the 5501A HeNe laser power supply with a common barcode scanner brick, the Laser Drive model 103-23. This has an input rance of 21 to 31 VDC at less than 0.5 A, and an output of 1.1 to 1.5 kV at 3.5 mA (fixed). The 3.5 mA is a bit higher than the labeled current on most 5501A tubes, but seems to be acceptable and actually beneficial for some high mileage tubes that like to run at a slightly higher current. But, adjustable versions of these supplies are readily available. I connected the supply between the HV Control (white/green wire) and -15 VDC (purple wire) with the pot set fully CCW (max current). This assures that the 5501A current regulator will not attempt to compete with the brick's internal regulator. However, with some HeNe laser power supplies, it may be possible to use the 5501A's regulator to *reduce* the current in a stable manner. This is left as an exercise for the student as it may not work in general.
The output of the laser tube is passed through a quarter wave plate to convert the circular polarization to orthogonal linear polarization components, and then through a half wave plate to rotate the linear polarization by an arbitrary, but fixed angle to line the two linearly polarized components up with subsequent optics. These waveplates are adjustable with respect to orientation around the optical axis of the laser as expected. But the angle of each waveplate along one of its principle axes with respect to the optical axis of the laser is also adjustable - presumably to optimize the 1/4 or 1/2 wave performance. They are both very thin and may be zero order waveplates, possibly made of optical grade mica. The beam is then expanded and collimated and passed through an angled partially reflecting plate located just beyond the collimating lens on the laser tube assembly. This deflects about 20 percent of the beam to a polarizing beamsplitter which sends each component to its own photosensor to provide the frequency control feedback. A control loop uses these signals to adjust the PZT, and thus resonator length, so that the two signals are of equal amplitude. The difference of the two signals is the frequency/phase reference.
The laser stabilization control algorithm is actually dirt simple: The voltages from the photodiodes corresponding to the two polarization components are compared in an integrator which maintains the PZT voltage at a level so they are equal. (There is an adjustment to compensate for slight differences in amplitude resulting from beamsplitter ratio and photodiode sensitivity.) While crude and simple to implement, this approach is adequate to achieve the needed stability. The electronic reference signal is derived from the slight residual difference frequency present in one of the polarization components.
While the spacer rod has a very low coefficient of thermal expansion, it isn't exactly zero, so as the system heats up (over hours), the cavity length will still change slightly. Eventually, the PZT voltage may be unable to compensate. The PZT voltage is compared with fixed upper and lower limits which are well within the range over which locking is assured. When either limit is passed, the "Tune Fault" flag is set turning on the "Retune" LED and asserting the "Retune_Status" signal. The laser may be retuned via a pushbutton or external TTL signal). This clamps the PZT control voltage at its lowest value for a short time and then releases it to ramp up to the lock point. Requiring external intervention (whether manually or by computer) assures that a measurement will never be made when the laser isn't stable, nor will one in progress be interrupted due to the laser relocking unexpectedly.
The 5501A laser head requires +15 VDC and -15 VDC for power. (There is also a +5 VDC pin but it is an output according to the manual.) The two voltages (and common) are all that is needed to operate the laser head but an interlock switch (on the right side at the rear of the case) must be depressed to turn on the laser tube. I haven't yet looked at the output with a photodiode or scanning Fabry-Perot interferometer but after a few seconds, the "Retune" LED goes off, similar to if the "Retune" button is pressed. And then there is a stable reference signal. I have since acquired an operation and service manual for the HP-5501A laser head which confirms the information above.
HP-5501A reference connector J1
Pin Function Socket View --------------------------------------------- A A Accessory +15 VDC fused o B +15 VDC return D o o B C Reference (difference) frequency o D Complement of J1-C C
HP-5501A power connector J2
Pin Function Socket View --------------------------------------------- A +15 VDC input D o o A B -15 VDC input C +5 VDC output (test-point) C o o B D Power ground
HP-5501A diagnostic connector J3
Pin Function I/O Comments ------------------------------------------------------------------------------ A +15 VDC TEST O Sample for diagnostics B -15 VDC TEST O Sample for diagnostics C +5 VDC TEST O Sample for diagnostics D SYS COM - Ground/return E Retune_CMD- I Active low to initiate PZT tune/check cycle. F Retune_Failure O Active high output indicates failure of PZT tune/check cycle. J Retune_Status O Active high when tune/check cycle is in progress. K Laser_Cur_Err O Active high indicates laser tube current is outside acceptable limits. L Error O Logical OR of J3-J, J3-K, and PZT voltage outside of specifications. M L I Mon Test O Laser current sample for diagnostics. N PZT Mon Test O PZT voltage sample for diagnostics. P Ref OK Status O Active low diagnostic signal indicates laser is properly tuned.
Like the 5501A, the 5501B also requires only +/-15 VDC to power up. There is no case interlock on the laser I have, though one is shown in the manual so I assume this is either an addition or deletion depending on version. When power is applied, at first, only the +/-15 VDC power LEDs come on. After 3 to 6 minutes, the "READY" LED begins to flash at about a 1 second rate. After another 1.5 minutes or so, the "Laser On" LED comes on and the beam appears. Finally, a minute or so after that, the READY LED comes on solid and remains that way. Note that this method of turning on the laser only after the temperature set-point has been reached is unique to the 5501B. All the other HP/Agilent lasers turn on the laser with power-on. Specific times for one test beginning from a cold start at an ambient temperature of about 65 °F were: (min:sec) 3:15, 1:35, and 0:48. The first of the times is called "preheat" and is determined by how long it takes for what HP calls the "laser rod" to reach operating temperature. The laser rod is the large glass bore of the laser tube to which the mirrors are clamped at either end. It thus controls cavity length. The temperature is sensed by disabling the heater drive and measuring the resistance of the heater coil every 25.6 seconds. The warmup is much shorter if the laser is restarted after having been running: 1:00, 1:20, and 0:50. Only after the READY LED is on solid, do the reference signals appear. The 5501B adjusts the cavity length so that the two polarized components of the beam (the Zeeman split longitudinal modes) have equal power. Interestingly, there is only one photodiode sensor which is alternatively switched between beams using a liquid crystal polarization rotator. A sample-and-hold then outputs to the error amplifier of the optical mode control feedback loop.
There are two outputs of about 5 to 6 V p-p (centered about 0 V), 180 degrees out of phase. For this laser, the reference frequency is about 1.80 MHz. There is no need for a "Retune" button as with the PZT based system of the 5501A. Also unlike the 5501A, there are no other signals to or from the 5501B (no large connector), only the +5 VDC output on the power connector, and a fused +15 VDC output on the reference connector.
Although the control board inside the 5501B looks similar to that of the "small" 5517 lasers, it is not interchangeable with them as some functions like the heater drive have are on the small "connector PCB" at the back-end of the case.
There were two problems with the first 5501B I acquired that I had to deal with. The first was that the tube wouldn't stay on stably at the 3.5 mA setting (fixed) of the power supply but works fine at 4 mA. Such a condition is usually due to the tube having been run for a long time, which wouldn't be surprising with a surplus 5501B laser head. Since the existing power supply has no current adjustment, I needed to find a similar size HeNe laser power supply brick (1"x1.5"x4" or smaller) that will run on 15 VDC to replace it that can be set for 4 to 4.5 mA. The tube seemed healthy enough otherwise. I installed one that runs the tube at 4 mA but draws more DC input current than the original, and possibly for that reason, the controller aborts and resets after about 1 second when it turns the laser on. For now, to get around this, I have connected the HeNe laser power supply directly to the raw -15 VDC and added a transistor to drive its enable input when the original laser power turns on. That appeared to work fine. But after replacing the cover, the laser tube wouldn't come on. :( I discovered that it needed the room light to start! I had thought this to be a relatively rare malady for HeNe laser tubes, but more common for neon lamps and glow-tube fluorescent lamp starters. However, it turns out that a decent percentage of HP/Agilent HeNe lasers start more quickly when illuminated. So, there is now a decorative red LED shining on the back of the tube which is lit when the laser is powered. An HeNe laser power supply with a higher starting voltage would probably make this kludge, oops, feature, unnecessary. But no one will ever know about it. :) While many of these higher mileage HP/Agilent lasers can benefit from this addition, since the 5501B turns the laser on and expects it to come on quickly, it is more critical than with the other lasers like the 5517s that really don't care whether the laser is outputting a beam or not, until they actually try to lock. However, in either case, if the laser takes too long to lock, associated equipment like the 5508A Measurement Display may flag it as a failure.
HP-5501B reference connector J1
Pin Function --------------------------------------------- A Accessory +15 VDC fused B +15 VDC return C Reference (difference) frequency D Complement of J1-C
HP-5501B power connector J2
Pin Function --------------------------------------------- A +15 VDC input B -15 VDC input C +5 VDC output (test-point) D Power ground
The HP-5525A was used in the original HP interferometer introduced around 1970 and includes the HP-5505A Measurement Display and the HP-5500A two-frequency HeNe laser head. The 5500A laser has the interferometer optics built-in and thus only requires an external retroreflector (cube-corner) on the moving part to be mesaured. The HP-5525B upgraded to the 5500C laser head which requires external interformeter optics but allows for two axis measurements (with a pair of 5505As!). The 5526A seems to have added a variety of options and but it's not clear how it really differs from the 5525B.
The 5525A/B and 5526A can be set up in the field with relative ease with a minimum number of individual components and no need for a control computer as its basic functions are built-in to the HP-5505A. It provides for the stand-alone precise measurement of position and velocity. But straightness and angle are not directly supported.
The 5505A implementation of the display function is all done in MSI TTL logic with a pair of 36 bit counter/registers for REF and DOPPLER (same as MEAS for other HP lasers), with a decimal adder/subtractor to generate the result. This is all on multiple PCBs and while there is one labeled "Program", there is no actual microprocessor controlling the system.
The 5525A, 5525B, and 5526 all require the 5505A display but differ in the laser and options. (There may be some minor changes required to convert an older 5505A to be used in a 5526A system.) The following is from the N4MW HP 5526A Documentation Page which also has links to the actual HP catalog pages for each system.
5500Cs have also been showing up with internal linear interferometers like the 5500A. I haven't seen any reference to this as a standard product though. I wonder if they were retrofits for customers who found their original 5525A configuration adequate or whined when their 5500As went bad and wanted an exact replacement.
The HP-5525A/B and HP-5526A are very obsolete, but many are still in use. 5505As show up on eBay, often for next to nothing. To non-interferometer geeks, the set of Nixie tubes is probably more valuable than a working unit! However, being so old, they often have problems, and at least some of the ICs like the Nixie tube drivers are proprietary parts and no longer available.
For info (or lack thereof):
The laser connector on the back of the 5505A is the same type and has the same pinout as that on the 5500A and 5500C heads. The 5508A supplies +/-15 VDC power for the laser head. It also controls both HeNe laser power supply current regulation and PZT laser tuning.
To use the 5505A with a 5500A, all that's required is a 05500-60025 cable and a retroreflector (cube-corner) as shown in Original HP-5525A with HP-5500A, HP-5505A, and Retroreflector - View 1. (Additonal photos can be found in the section above on the 5500A laser.) It's straightforward to make a cable. The connectors are standard and everything is wired 1:1 at both ends. To use the 5505A with the HP-5500C laser also requires external interferometers optics. All of the standard configurations that have separate outgoing and return beams should work.
To use the 5508A with other HP laser heads will require a custom cable and possibly a separate optical receiver which can be any version of the 10780 (A, B, C, F, U). However, some circuitry may need to be added to the 5505A to keep it happy by making it think it still has control of PZT tuning.
FWIW in the "well that's interesting department", here is the board set from another 5505A. This is a rather vintage sample, Serial Number: 2016A01966, which puts its manufacturing date around 1970:
Slot Name Part # Additional Markings ----------------------------------------------------------------- A1 Analog Board 05505-60001 Series 1920 03L A2 Clock Board 05505-60002 B3 Series 952-2 03F A3 Accumulator Board 05505-60034 Series 1920 01403F A4 Accumulator Board 05505-60034 Series 1920 01403F A5 Adder Board 05505-60005 Series 952 03F A6 Algorithm Board 05505-60006 Series 952 00203F A7 Program Board 05505-60007 Series 2240 23103F A8 Function Board 05505-60058 Series 1920 23903F A9 Multiplier Board 05505-60049 Series 1948 23103F A10 D-Register 05505-60010 03F A11 Display Board 05505-60011 Series 1324 03F A12 Power Suppy Board 05505-60012 Series 1940 01503F
The HP-5528A includes the HP-5508A Measurement Display, an HP-5518A two-frequency HeNe laser head on heavy duty tripod, a variety of interferometer optics, and optional environmental sensors, and other stuff. :) So, this system can be set up in the field with relative ease with a minimum number of individual components and no need for a control computer as its basic functions are built-in to the HP-5508A. It provides for the stand-alone precise measurement of position, velocity, angle, and straightness by using the appropriate interferometer assemblies.
The 5508A implementation of the display function consists of X16 frequency multipliers for REF and MEAS, which are then applied to separate 16 bit up counters. These initiate a non-maskable interrupt to the microcontroller when either exceeds the half way point (the MSB gets set). They are then stopped while separate small "swallow counters" absorb pulses occurring while the interrupt is processed and the position is updated. The microcontroller is kept rather busy, but since it doesn't have all that much else do do, should be quite happy. :)
Although the HP-5528A is considered obsolete by Agilent, it's still very useful and surplus systems or components are now much cheaper. The Agilent 5529A Dynamic Calibrator is the replacement for the 5528A. Rather than a dedicated display, it requires a PC (not included). But aside from the slightly higher REF frequency of the 5529A laser head (generally irrelevant in these types of typically slow speed applications), the precision is no better than that of the 5528A.
(Both of these are also accessible directly from Agilent. Search for "5528A".)
The laser connector on the back of the 5508A is the same type and has the same pinout as that on the 5517 and 5518A laser heads. The one "No Connect" pin on the 5517 connector (pin A) is used to drive the MEAS beam level indicator on the front of the 5508A. The meter reading seems to be proportional to the current flowing out of this pin, from an internal +5 VDC source, with approximately -2 mA being full scale. Pins B and C that are also unused on the 5517 lasers now get ~MEAS and MEAS. (They are connected to line drivers on the 5517 lasers but only used for testing.) The 5508A supplies +/-15 VDC power for the 5518A laser head.
There are several other connectors on the rear of the 5508A for various environmental sensors (temperature, pressure, etc.) and even a remote control. (I'd like to see that!) There is also a IEEE-488/GPIB/HP-IB interface for control and data acquisition.
To use the 5508A with a 5518A, all that's required is a 10793A/B/C cable, which is wired 1:1 at both ends. To use the 5508A with other HP laser heads will require a custom cable, and possibly a separate optical receiver which can be any version of the 10780 (A, B, C, F, U). Except for ~MEAS and MEAS, all signals are wired 1:1 for 5517s. However, I'm not sure whether all versions of the 5508A will work over all velocities with any of these except the 5517A. The 5517B/C/D have REF frequencies, and result in possible MEAS frequencies, that may be too high, at least under some conditions. It is also possible to use the 5508A with the 5501A/B lasers, but the connectors are totally different. For the 5519A/B laser, these same issues apply. The connector of the 5519A/B provides only the REF and MEAS signals as this laser head has an internal power supply that plugs into the AC line. To use the 5508A with 5500A/C lasers would require additional circuitry to provide the HeNe laser current control and PZT tuning.
One way to get around the REF frequency issue is to build a divide-by-two circuit for REF and MEAS that goes between the laser and 5508A. This is simply a dual differential line receiver, a pair of D flip-flops, and a dual differential line driver. Add a switch to select straight through or divide-by-two if desired. Of course, measurements values will now be halved unless a plane mirror interferometer is used, in which its doubling will be exactly offset by the halving in the divider!
The only thing that won't work when using laser heads other than the 5518A without additional effort is the beam level meter on the 5508A front panel, fed from pin A of the laser connector. This seems to require a current of 0 to 2 mA to Ground from an internal +5 VDC source. The test-point on the outside of all 10780 receivers generates a voltage related to signal level, but simple voltage to current converter circuit (1 transistor and a few resistors) is then needed to interface to the meter input. If you're not a purist, this can just be ignored as it is not used anywhere and its only purpose is to aid in optical alignment and confirmation of adequate signal. But the 10780 test-point services the same function.
To use a 5501A/B laser with the 5508A, wire the REF signals from the laser into E,R and F,R of the 5508A connector with twisted pairs, and the MEAS signal from the 10780 into pins B,D and C,D of the 5508A connector with twisted pairs. Power (or at least -15 VDC) for the laser may have to come from a separate supply as I do not know if the 5508A has enough capacity on the -15 VDC output for the 5501A/B lasers (more than 1/2 A) since the 5517 lasers require only 20 mA.
I've attached my 5508A to my measurement test setup and initially have been using a 5517D laser head with it. I'm a bit surprised that this even works with the 5508A as it has a REF frequency almost double the maximum of even the later versions of the 5518A laser. I don't know if it will run at full velocity, but for modest speeds, the readings seem to be fine. But I intend to add the divide-by-two circuit as insurance.
When warming up, the difference frequency only appears for perhaps 10 percent of the time during mode sweep - only when the Zeeman modes are near equal amplitude on the split neon gain curves. The difference frequency is maximum and the output power is minimum at the center of this region, which is also where it will eventually lock. This is normal behavior for these lasers based on what is shown in Axial Zeeman Split HeNe Laser Mode Behavior. Note that while there may be some other longitudinal modes present, at least one of the other pairs is too weak to lase and thus there will be no beat except from the main pair, and then only when very close to being positioned symetrically on the Zeeman-split neon gain curves. In fact, only the main F1/F2 mode is present when locked. While other rogue mods would not produce any beat signal, they would result in problems in the interferometer and possible transient errors.
Unlike most other internal mirror HeNe laser tubes, there are no mirror adjustments. The mirrors are held in place against the thick glass bore (or mirror spacing rod as HP calls it) by spring pressure alone. So, the ends of the bore and mirrors must be ground to a precision sufficient for alignment to be near perfect. The distance between the mirrors is 127 mm in the 5517B corresponding to an FSR or longitudinal mode spacing of about 1.18 GHz. The purpose of having multiple glass backing disks behind the OC mirror is not really known. But along with the appropriate length spacing rod, these would provide a means of selecting cavity length by using a variable number of them without having to manufacture multiple (glass) tube designs. So, everything could be identical except for the spacing rod and number of backing disks installed behind the HR mirror and in front of the OC mirror, and perhaps the length of the associated spring(s). Having at least one backing disk may be desirable to reduce stress on the mirrors from the springs, but some tubes don't have any behind the HR mirror. So backing disks may simply be present to accomodate different spring sizes so the bore is properly positioned. I do not yet know if other model lasers have different cavity lengths. Visual examination of 5501B, 5517A, 5517B, 5517D, and 5519B tube assemblies from the front after removing the beam expanders and waveplates has been inconclusive. And the glass tubes from 5517B and 5517D lasers appear physically identical in all dimensions that matter from the outside - I have yet to smash a 5517D tube to go deeper! I was expecting the low REF frequency 5517A to have some obvious difference compared to the high REF frequency 5517D, like only a single backing disk behind the OC and maybe a shorter spring, but this does not appear to be the case. So, if they are all the same, what is the purpose of the extra space in front with 5 backing disks and the extra length spring? If anyone has taken more of these tubes to bits or X-rayed them, or can measure the non-Zeeman difference frequency between longitudinal modes (and thus the FSR yielding the cavity length), and can thus shed some coherent light on this topic :), please contact me via the Sci.Electronics.Repair FAQ Email Links Page.
In terms of cavity length, the 5517 tubes are very similar to typical 1 mW barcode scanner HeNe laser tubes. Their relatively low power - always under 1 mW and often much less - may be accounted for by the combination of the magnetic field, precisely specified OC mirror reflectance, and possibly isotopically pure helium and neon and fill pressure, to achieve the desired Zeeman beat frequency range for each model of the 5517. This is consistent with the generally lower output power for higher REF frequency lasers - 5517Ds almost always have much lower power than 5517As when new. The heater connections with red and purple wire stubs sticking out can be seen at the left of the tube. The purple one also attaches to the cathode via a piece of springy sheet metal, no welds. The anode connection goes through the outer glass envelope but there does not seem be a glass seal into the bore, but simply a hole drilled in it to coincide with the anode location. Even after totally disassembling a tube, it's not clear what prevents the discharge from bypassing the bore unless there was some other sealant present that got lost in the process. But there are no traces of any, nor anything visible in two other tubes. The fit between the bore and surrounding glass cylinder is quite close but even this wouldn't normally guarantee that the discharge goes through the bore, especially on hard-start tubes. The magnet is a single piece cylinder of what looks like AlNiCo with an inner diameter a few mm larger than the tube. It extends at the left and right ends of the tube to exactly where the discharge begins and ends.
I have measured the tube+ballast voltage on one 5517B laser to be about 1,675 V at 3.5 mA. This may be a bit greater than typical since the tube was high mileage, though with normal start and run behavior. The ballast was 100K so the actual voltage across the tube was 1,325 V, This is still somewhat greater than expected for a laser with a power output of less than 1 mW. Part of the reason may be that the magnet probably slightly increases the tube voltage.
Here are the 5517B HeNe laser tube parameters as best I have been able to determine them so far, compared with typical short barcode scanner HeNe laser tubes. Except for tube voltage (which I haven't measured and am only estimating) and tube current (which is always spec'd to be 3.5 mA), the parameters for the 5517B are for one very healthy sample. The total length, planar HR mirror, divergence (without beam expander), and beam diameter are similar or identical for other models. The parameters for the barcode scanner tube are for a range of typical models. The divergence for a particular model barcode scanner tube is usually achieved by either the specific curvature of the outer surface of the OC mirror glass or with an external lens glued to it, but the cavity design including the OC RoC (radius of curvature) is the same for all.
5517B Barcode Parameter Tube Scanner Tube -------------------------------------------------------------------- Output Power 1.0 mW 0.5 to 1.0 mW Total Length 194 mm 125 to 155 mm Cavity Length 127 mm 115 to 145 mm HR Mirror Radius of Curvature Planar Planar OC Mirror Radius of Curvature 140 mm 250 to 300 mm OC Mirror Reflectance 98.5 % 99.0 to 99.5 % Beam diameter at output surface ~1 mm 0.4 to 0.6 mm Beam divergence ~10 mR 1.7, 2.7, 8 mR Discharge Length 100 mm 60 to 75 mm Operating Current 3.5 mA 3 to 4 mA Operating Voltage (tube only) 1.3 kV 0.7 to 1.1 kV
Of note here is the relatively low OC reflectance of 98.5 percent. This is one of the primary parameters that determins the REF frequency. All other factors being equal, reducing the reflectance increases cavity loss which increases the REF frequency.
Plot of Hewlett Packard Model 5517C Stabilized Laser During Warmup shows how a typical 5517 laser behaves. This is a 5517C rather than a 5517A but the locking algorithm is the same. Note that the entire warmup period from laser on to locked is only around 3.5 minutes because of the internal location of the heater for the active mode as noted above. A laser with the more common external heater could take 20 minutes or more to lock. The control algorithm is a bit more sophisticated than used on some other stabilized lasers, checking periodically for the status, and switching from "Warmup Mode" to "Optical Mode" about half way through the warmup period, at which point the READY LED starts flashing. A short while after it locks is when the READY LED comes on solid.
Plot of Hewlett Packard Model 5517C Stabilized Laser Near End of Warmup shows the 5 mode cycles just before locking and the final transition to the locked state. The peculiar shape of these Zeeman-split modes is clearly evident in this expanded view.
The beat frequency is shown for the last 5 cycles and after locking in both these plots. This is the actual measured frequency captured along with the vertically and horizontally polarized modes and total output power. (Showing the frequency plot earlier would be a mess.) The beat only appears for a small percentage of the mode cycles with some variation during the time it is present, peaking when the F1 and F2 amplitudes are equal, and only when F1 is falling with increasing temperature. There is no beam when F1 and F2 are equal but F2 is rising with increasing temperature. The reason for this becomes evident from the simplified diagram in Axial Zeeman Split HeNe Laser Mode Behavior, or more accurately in HP-5517 Zeeman Split HeNe Laser Mode Behavior. The second diagram has been specifically crafted based on the mode plots, above, and thus more accurately represents the actual mode behavior of the 5517. Both show 5 snapshots of most of a mode sweep cycle starting with the cavity being 1/4 wavelength too short at the top and ending with being 1/8 wavlength too long. (The case of 1/4 wavelength too long would be the same as the first, 1/4 wavelength too short). Only when the longitudinal mode is centered between the Zeeman-split neon gain curves will there be a beat. In addition, the mode amplitudes are changing rapidly as the cavity expands at those high slope locations on the gain curves. When the cavity length changes (longer or shorter) by 1/4 of the lasing wavelength of approximately 633 nm, the amplitudes are again equal, but the two separate longitudinal modes are oscillating far apart and there is no beat. Note that the red and blue plots include the F1 and F2 amplitudes, but also may have contributions from other longitudinal modes derived from the same split gain curve which thus have the same original circular polarization.
Note that as the tube ages with use, the gain declines and the width of each gain curve that is above the lasing threshold decreases. Eventually, with a really high mileage tube, there may be no overlap at all and the beam will probably disappear for a part of the mode sweep cycle. But it is exactly at that point where the Zeeman beat would be generated, so it will also disappear entirely. Lasers are generally taken out of service long before this happens, but I recently found one whose output power was so low that this behavior was present - or absent depending on your point of view!
Here is how the article "An Instant-On Laser for Length Measurement" by Glenn M .Burgwald and William P. Krugein describes the operation of the laser tube in the Hewlett-Parckard Journal, Aug., 1970.
"If an axial magnetic field is applied to a laser which is free from polarization anisotropy in either the mirrors or the plasma tube, the output splits into two frequencies of left and right circular polarization. First-order theory predicts that the frequency splitting is proportional to magnetic field strength and to the ratio of line Q to cavity Q. In the new laser, magnetic field strength is adjusted for a difference frequency of about 2.0 MHz. Line center is virtually midway between the displaced lines, so proper cavity tuning can be assured by adjusting for equal intensities of the lines."
And they show a gain curve diagram even simpler than the one above. See that article for more details.
The reason for the peculiar shape of the mode plot is not clear. It is probably due to a combination of Zeeman modes and normal longitudinal modes. However, the gain profiles need to be asymmetric to account for it. The simpified explanation of Zeeman splitting rarely takes into account the actual profile of the magnetic field which no doubt stretches and distorted the gain profiles. So, drawing a pair of nice bell-shaped gain curves really isn't accurate. Near the ends of the tube, there will be little magnetic field and the gain curve there will not be shifted very much. In the center, the field will be maximum and the gain curves will be shifted by a large amount. Thus, the neon atoms will see a variation in magnetic field along the bore - a summation or integral of gain curves varying with frequency. This is depicted in HP-5517 Zeeman Split HeNe Laser Mode Behavior. Here, the gain curves have been modified so that the results would be roughly similar to what was in the plots, above. And HP-5517 Zeeman Split HeNe Laser Mode Behavior Versus Mode Position on Gain Curve shows one complete mode cycle along with little split gain curves.
Note that the locked mode amplitudes are not equal and that the locked beat frequency varies a bit after locking and does not stay at its maximum value as would be expected if the stabilization was optimal. This is not a quirk of the particular laser I'm using for these experiments as I've tested dozens with similar behavior. So, they aren't as perfect as we might hope! :)
Also note the two smaller modes on the tails of the gain curves in the diagrams. They aren't present in all of these lasers. But if they are and align with the X and Y axes, then the only effect will be to slightly decrease the MEAS or detected REF signal level with respect to laser output power. However, if the are not aligned with the X and Y axes (e.g., at 30 degrees), they will cause level changes in the envelope of the signal from the optical receiver's photodiode due to self-interference in the interferometer. This is similar to what would happen if the primary Zeeman modes were misaligned, or not pure. The consequences could be transient position errors but only during motion. The end-points would be accurate.
Some later versions of the 5517 lasers have a totally redesigned electronics board with more digital circuitry. I don't know exactly when this changeover took place, or whether it's simply a special option, but several 5517Cs and 5517Ds from before 2000 had the old style and a 5517C from 2005 had the new style. So, perhaps Agilent insisted on reinventing the wheel. :)
The glass 5517A HeNe laser tube itself is similar or identical to that of the 5501B. However, the tube's enclosure appears to have been cost-reduced: It is a base metal casting rather than being constructed from precision machined parts. (This is also true of the 5518A and 5519A/B.) All the other 5517s use tube enclosures that are physically similar to that of the 5501B. (Perhaps newer versions of the 5517A have the higher quality construction.)
The output optics consists of a beam expander/collimator (the black and silver object just to the left of the power supply danger label) and an additional optical assembly to the left of this whose front and rear halves contain what appear to be AR coated optical quality mica pelicles oriented at slight, but different angles. The front and rear sections can be rotated independently and they were sealed with blue paint once the perfect orientations were found. The two mica (or whatever) pieces of the optics assembly (just after the beam expander) are adjustable waveplates. The first one is a 1/4 waveplate to convert the circular polarization of the Zeeman split output of the HeNe laser tube to linear polarization and the second one is a 1/2 waveplate to rotate the resulting linearly polarized components to be aligned along the horizontal and vertical axes. These can then be separated out with a polarizing beamsplitter at the detectors.
The HP/Agilent lasers do not employ any sophisticated method of stabilization such as locking the Zeeman beat frequency (which chances slightly depending on where the modes are on the neon gain curve) to a crystal reference. They simply use the amplitudes of two orthogonally polarized signals in an analog feedback circuit as is common with most other stabilized HeNe lasers. However, here, the two polarizations are of the two Zeeman split components of the single oscillating mode rather than two separate longitudinal modes. The error signal is the difference between their amplitudes, which is forced to zero by temperature tuning of the cavity. And, in fact, there is no real need to have the frequency be precisely known or even constant over the long term, as long as it is stable over the short term. More below.
The warmup/locking algorithm is straightforward, though just a bit different than used in many other stabilized lasers. When the laser is first turned on, it is in "Warmup Mode" and the heater, which is wrapped around the internal bore of the laser tube, is driven to reach a fixed temperature (set by the only pot on the electronics PCB). The temperature is sensed by periodically measuring the heater's resistance. This is done by disabling the heater driver, passing a small fixed current through heater wire (for 2.56 seconds out of each 25.6 second period), and storing the resulting voltage in a sample-and-hold. Since the heater wire changes resistance with temperature, this eliminates the need for a separate temperature sensor inside the tube. Once the temperature set-point is reached (the voltage from the pot approaches the voltage on the sample-and-hold), the feedback switches to Optimal Mode and alternately samples the two polarized Zeeman split sub-mode signals with their voltage difference being the error signal in the feedback loop, which is driven to zero by adjusting the temperature, and thus cavity length. In fact, from the relative shapes of the red and blue mode cycles, it can be seen that from about the last half dozen mode cycles till just before locking, the tube is actully steadily cooling rather than heating. With the heater located inside the laser tube, the time from power on to a locked condition is typically only about 4 minutes and should also be less susceptible to ambient conditions. In fact, from the relative shapes of the red and blue mode cycles, it can be seen that during most of the time from power on (a cold start) to lock, the laser tube is heating (about 75 cycles), but it switches to steady cooling (about 6 cycles) just before locking.
Several views of a naked 5517E are shown in Agilent 5517E Laser Head With Cover Removed.
The 5517E has the most incredibly complicated Control PCB of any HP/Agilent laser I had see before finding this laser, even compared to the newer digital Control PCB (see below). (I've since found a similar control PCB in a recent vintage 5517D.) It includes a SHARC DSP, two Lattice FPLDs, and a lot of other digital circuitry, purpose unknown. They also seem to have gone back to PWM for the heater drive since there is no power transistor on a heatsink, as with the original analog and first digital Control PCBs. However, that collection of inductors visible in the lower left of the photo may be there to clean up the drive to the heater and remove the high frequency switching noise. Since the locking should be basically the same as for the other lasers, this level of complexity is perplexing unless this particular unit was designed to have much better stability - perhaps the "military calibrated" version. Unfortunately, the new Control PCB lacks all the familiar jumpers and the temperature set-point pot, and adds a couple of micro DIP-switches and connectors, purpose also unknown. Aside from the unknowns, everything else is obvious. :)
Here are a pair of closeup photos of the overall laser:
High resolution scans of the front and back of a similar digital control PCB can be found linked from the section: HP/Agilent 5517 Laser Construction.
There is also an additional resistor in series with the tube heater in the wire bundle, apparently as an afterthought since it is part of a cable extension. The heater and resistor each measure just under 5 ohms cold. The heater of other 5517 tubes measures about 8 ohms cold, so at the same current, this shorter tube which must have a shorter heater gets only slightly over half the heater power. I originally thought that this might be why it takes longer to stabilize, but then found that this was true of "normal" 5517 tubes with the fancy digital Control PCB. That shroud above the tube would make tube swaps much more tedious, as the Control PCB on the opposite side would need to come off to remove it. Then, the cable ties would have to be cut to free the wiring and tube. But at least the connector PCB is identical to the ones in the 5517B/C/D lasers, and even includes the usual appendix - the HeNe laser current adjust pot that is no longer used!
With no label on the tube assembly and that unusual plastic rear cover, for awhile I was suspecting that this might not even an HP/Agilent tube. But that style of glasswork at the back is clearly HP/Agilent even if it does differ slightly from the normal design. And everything else is normal HP/Agilent including the beam expander, HeNe laser power supply wiring/ballast, and the Newest Digital Control PCB, which, as noted, has turned up on a late model 5517D. However, there were some mica washers under the tube presumably as shims to fine tune the vertical position of the beam, so this may have been a prototype or pre-production model. While there is no manufacturing date on the laser, date codes on the ICs suggest that is from around 2003. Rework in the area of the REF out circuitry may mean this was an early version of the Newest Digital Control PCB. This is not present in my other sample.
And a further note about disassembly: To get this photo required almost totally removing the tube since the rear plastic cap was held in place by three screws with nuts and it would have been almost impossible to replace the nuts without being able to access behind and under the magnet assembly. That's when I discovered the shims. Hopefully, I got them back more or less in the proper locations. Must maintain specifications! :)
Functionally, the 5517E behaves more or less like the other 5517 lasers. The user LEDs are the same but there are 4 LEDs on the control board that I'm sure provide a wealth of information if one knows how to interpret them. My sample takes over 5 minutes for READY to start flashing. READY also stops flashing once or twice for a couple minutes, before it resumes flashing, and then locks after about 9 minutes. Whether these long times and peculiar flashing behavior are normal or indicate some problem, is also unknown. However, with a similar Control PCB and heslthy 5517B tube, the behavior is similarly strange. More on this in the sections: HP/Agilent 5517 Laser Control PCBs and Locking Sequence and Agilent 5517 Laser RS232 Communications.
The tube is somewhat low power compared to what's normal for other 5517s - about 120 µW locked. But I have no specs on 5517E minimum output power, so with the shorter tube and likely stronger magnetic field, that might be acceptable. Once locked, it's quite stable with minimal drift in REF frequency. Given the huge amount of computation power available, it may count mode sweep cycles instead of using a temperature set-point (or in addition), and might also adapt automagically to a replacement tube - or require a factory upload of tube parameters via one of those unlabeled connectors!
The high REF frequency of 6.3 MHz works fine with my home-built SGMD1 measurement display, but comes up as "Laser Fail" on a 5508A. This isn't all that surprising since 6.3 MHz is almost twice the maximum REF frequency of the 5518A for which the 5508A was intended. However, the same type of control PCB with a 5517B tube locks fine but also fails keep the the 5508A happy, so it is more likely due to wimpy line drivers or something like that. :)
These lasers consist of the laser tube assembly, potted (brick) HeNe laser power supply, beam sampler, connector PCB, and Control PCB. mounted on a an metal chassis. Any of the parts can be replaced in under 5 minutes using common tools, with only minimal or no adjustment or alignment.
Laser tube assembly:
All of these consist of the actual glass HeNe laser tube potted with some sort of rubbery material inside its Zeeman magnet, beam expander, and adjustable waveplates. The heater/cathode is attached via a 2 pin plug while the anode has its own single pin high voltage connector. The HeNe laser tube ballast resistance of about 100K ohms is conformal molded into the silicone insulated HV cable. The bifilar-wound heater inside the laser tube has a typical resistance (cold) of 8 ohms on all tubes except the 5517E. When at operating temperature, the resistance is spec'd to be 1.285 times what it is when cold. For the 5517E (and possibly the F and G models as well), the resistance (cold) is around 4 ohms.
Only 4 screws hold the tube assembly to the chassis for lasers in the small cases (all the 5517s except the 5517A, as well as the 5501B). One or two will be flat head screws which provide either a fixed axis for horizontal (pan) alignment, or self alignment (no adjustment permitted). All of these tube assemblies appear physically identical, except for the 5517E (and probably 5517F/G) which are slightly shorter. (They, of course, differ with respect to the REF/split frequency.) The larger tube assemblies found in the 5517A, 5518A, and 5519A/B mount with 3 bolts and have machined alignment pins so no adjustments are needed or possible. They, too, are physically identical except for one small area of the casting that needs to be cut away if installing a 5517A tube into a 5519A/B laser.
HeNe laser power supply:
Very old (perhaps roughly pre-1990) lasers used Laser Drive model 111-Adj-1 HeHe laser power supplies which had adjustable current via a pot on the laser connector PCB. At least most of them did. I did find one that had the same part number but no third wire. All later versions use VMI power supplies with a fixed current of 3.5 mA. However, the pot is still present on the connector PCB even on lasers made after 2006 (and probably to the present day), but does nothing.
There are three versions of the VMI power supplies used in these lasers. The two older ones (VMI PS 148 and VMI PS 217) have the same HP part number of 0950-0470. The switchover came around the year 2000. The newest one (VMI PS 373) has an Agilent part number of 0950-4459 and is found on all recent lasers, at least since 2004. I know that going from the 148 to the 217 reduced the residual current ripple from over 3 percent to less than 1 percent because I measured it. I do not know what changes were made in the 373, nor what other differences there may be between these models. VMI claims it is proprietary information. Can you believe that? :) However, I have de-potted a dead model 373 and have reverse engineered its schematic. See the section: VMI 373 HeNe Laser Power Supply.
With few exceptions, the only defective power supplies I've ever found in HP lasers were nearly all Laser Drive 111-Adj-1s. And one type of failure may result in the adjustment pot having no effect with the power supply pumping way excessive current (like 6 or 8 mA) through the tube. With luck, the ballast resistor catches fire and explodes before the tube is damaged. :( :) One exception was a VMI PS 148 that had excessive ripple, so something in its output filter had blown. But if I hadn't been checking ripple on a bunch of these power supplies, it probably would have gone unnoticed, since as long as the tube stays lit, performance of the laser probably wouldn't be affected in any significant way. However, the additional ripple would make the effective dropout current go up, so a reasonably healthy tube might start sputtering on that supply.
These consist of a first non-polarizing angled plate to sample a portion of the output beam and a second non-polarizing angled plate to take this and split it between the LCD switch, and reference photodiode. The LCD switch attaches to the Control PCB via a 4 pin connector - 2 pins for the LCD drive and 2 pincs for the photodiode behind it. The reference photodiode is actually mounted on the Control PCB and simply pokes its head into the beam sampler assembly. Beam samplers for all model lasers appear to be identical and interchangeable.
Aside from the Mil-style connector to the outside world and the 24 pin connector to the Control PCB, this has some filter capacitors; fuses for +15 VDC and -15 VDC; and the Power, Laser On (really same as Power), and READY LEDs. The one pot does nothing except for really old lasers with the Laser Drive 111-Adj-1 HeNe laser power supply brick.
Very old lasers had a case interlock switch to disable the laser tube from being power if the covers were removed, and a service switch to override this. :) Both of these switches have been eliminated, though the PCB pads and wiring for them are still present, but bypassed.
Since the introduction of the 5517A laser through the early 2000s, all the Control PCBs in these lasers were based on simple TTL logic for timing and an analog feedback loop. The one in the 5517A is physically larger and not interchangeable with those in the small lasers, but is nearly identical electrically. The main functional difference is the additiona of circuitry and a connector so it can also be used in 5518A and 5519A/B lasers. There had been virtually no change in the design over 15 years or more, except that a modification to the internal REF receiver makes newer lasers require somewhat higher optical power to lock than older ones.
But since sometime after Agilent was created, at least two versions of a Control PCB with much more complex digital circuitry appeared. These still have a substantial number of analog parts like op-amps and comparators, but FPGAs/FPLDs and/or a reasonably high performance microprocessor or DSP have replaced discrete logic. If one wants to count transistors, I bet the digital Control PCBs have 1000 times the number of transistors as the Analog Control PCB! The PCBs are also almost entirely surface mount, with parts on both sides (at least for the newest version).
The first of these digital Control PCBs (which I call the "newer digital control PCB) is electrically and physically interchangeable with the older Analog Control PCBs. It is based on a Xilinx FPGA and should be very reliable. But a failure for any reason other than an obvious problem like a blown fuse bad DC regulator with no underlying cause would likely render it non-repairable except by Agilent or an authorized service center since it's then just a black box with no real way to easily troubleshoot. A service manual may exist but I've never seen one. And even if it did, sophisticated test equipment and a surface mount rework station would be required to have any chance at repair. However, this version has all the same jumpers and temperature set pot, so normal testing and adjustment is similar to that of the Analog Control PCB. The solution would then be to simply swap in a known good board (either version).
It's not clear what, if anything, the digital controller adds to the laser, other than to make it more proprietary and difficult to service. After all, features are not being constantly changed or added, nor will there be security issues due to computer viruses - it doesn't run Windows! :) So, periodic firmware upgrades and bug fixes really aren't required. Whether this digital Control PCB is now used for all Agilent 5517B/C/D lasers or only for those with "military calibration" or OEM or special requests, I do not know. Swapping in an analog Control PCB resulted in no obvious differences in performance, which would be my suggested method of repair unless there were special requirements. And now (2009) 5517B lasers with these digital Control PCBs have been turning up from 2004, 2005, and 2006, probably removed from service in wafer fabs after degrading to the edge of Agilent specs for REF frequency, or a specific number or hours of service.
All the jumpers and their approximate locations are identical and the time spent in the major states and time to lock (READY on solid) are about the same as with the Analog Control PCB. The behavior while warming up and after locking is indistinguishable from that of the Analog Control PCB so it really isn't even possible to determine which one is inside the laser without removing the cover. The only obviously similar electronic component common to the two is the large white film capacitor for the feedback integrator, and perhaps the heater driver transistor. It's possible that the objective of this redesign was simply to eiliminate all the older SSI/MSI TTL logic and other obsolete through-hole parts, but that it is functionally identical to the Analog Control PCB with essentially the same logic inside the Xilinx FPGA and linear circuitry in SMT ICs.
More recently, another type of digital Control PCB has appeared. I first found one in a 5517E (this model laser isn't documented anywhere) and thought it might be unique to the -E version. But I have since also seen one in a 5517D-C29, manufactured in 2004. (And one of the other type of Digital Control PCB also in 2004.) This "Newest Digital Control PCB" seems to be a total redesign, with no effort made to be at all similar to the Analog Control PCB with its jumpers and test points.
Here are the three types of Control PCBs used in the 5517B/C/D/E lasers.
(As noted, the one in the 5517A differs slightly and as far as I know, only comes in the original analog version.)
The Analog Control PCB and the Newer Digital Control PCB function in a virtually identical manner, requiring about 2 minutes for the READY LED to start flashing, and another 2 minutes to come on solid. And as noted, they also have more or less the same jumpers and testpoints, as well as the temperature set-point pot. The Newer Digital Control PCB may in fact simply be essentially an emulation of the Analog Control PCB using an FPGA and more modern surface mount parts.
There are no useful indicators on the Analog Control PCB.
The third type of control PCB is not at all similar to the others. It has none of the same jumpers and several different testpoints, no pots, an RS232 port no doubt for setup and testing and almost certainly a digitally-maintained run-time meter, another unused connector, and a large unpopulated header. There are also a pair of micro-DIP switches - and a pushbutton, which I fianally dared to push, and as expected, seems to be master reset. :) There are many SMT parts on the back side of the PCB, including another Lattice FPLD.
For more, on the Control PCBs and their operation, see the sections: HP/Agilent 5517 Laser Control PCBs and Locking Sequence, HP-5517E/F, and Agilent 5517 Laser RS232 Communications.
The entire purpose of redesigning the controller more than once is somewhat perplexing. Doing it the second time with such complex digital hardware seems totally nuts, unless it was actually done for some custom high performance application, and it was simply convenient to use in late model 5517 lasers. (Given the likely relatively high cost of components including the SHARC processor and Lattice FPLDs, I find this rather hard to believe though. And it's a multilayer PCB with components on both sides, additional connectors, and other stuff.) The same very limited inputs (a pair of photodiodes sensing the modes through the relatively slow speed LCD switch and another photodiode behind a polarizer generating the REF signal) and outputs (tube heater current) are used in all 5517 lasers so it wouldn't seem to be possible to implement a significantly higher level of frequency accuracy or stabilization no matter what sort of control scheme is used. About the only thing that might be done is to actually compute the REF frequency from the REF photodiode signal and fine tune the lock position to maximize it once the basic stabilization using mode balance has been achieved. The peak of the REF frequency function optical frequency may be a more accurate means of locating the Zeeman-split gain curve center. But except for NIST-level precision, the analog method is really just fine, so even if this scheme is implement, it's not clear what customers would require it. And from my observations of the REF frequency while locking, it doesn't seem to make any effort to maximize it, but stabilizes at a point much lower than the peak, with the same sort of slow variation once locked as the analog Control PCB.
So, a combination of several explanations make the most sense:
If anyone has more information on these digital control PCBs, please contact me via the Sci.Electronics.Repair FAQ Email Links Page.
Also see the section: Common Problems with HP/Agilent 5517 Lasers.
Locking sequence with Analog Control PCB:
This also applies to the 5518A and 5519A/B since they all use the 5517A Analog Control PCB. It is also generally applicable to the 5501B though some details differ slightly.
From power-on to READY takes around 4 minutes for most 5517 lasers - all those NOT using the Newest Digital Control PCB. Even on the original Analog Control PCB, a state machine based on counters, flip-flops, and gates determines the timing. This may be true of the Newer Digital Control PCB as well, except that the state machine would be inside a Xilinx FPGA. Who knows how the Newest Digital Control PCB with its SHARC CPU implements this algorithm (which tends to take much longer than 4 minutes, reason unknown)! The following is paraphrased from the 5517A manual, which assumes the Analog Control PCB implementation. (All timing is approximate as the main clock is a 555 timer on the Analog Control PCB!):
Thus, under normal conditions, the laser will be locked and ready to make a measurement (approximately) 150 seconds after the READY LED starts flashing. Note that the only check to make sure the laser is locked is that the REF signal is present. Since this only occurs for a small percentage of the entire longitudinal mode sweep cycle, REF will not remain on for long without active feedback, so this is a reliable test. The laser will in fact continue to repeat the above sequence forever if REFON is not detected. Typically, this will occur when the output power from the laser tube has declined to below the REF detection threshold of the internal optical receiver after years of hard work. However, some marginal lasers will go through the sequence several times when powered up as the output power from the laser tube gradually increases with warmup until the amplitude of the difference frequency signal exceeds the REF detection threshold.
Locking sequence with Newer Digital Control PCB:
The Newer Digital Control PCB has 3 yellow state LEDs near the top right corner of the large Xilinx chip. These provide some information about where the controller is in the warmup process and they have a 1:1 correspondance with the major modes of the Analog Control PCB. I'm not sure the times in each state are identical for the two but they are close. Here is a rough chart of their behavior for a normal 5517C laser:
Time READY State Comments -------------------------------------------------------------------- 0:00 000 Power on (WARMUP Mode) 0:01 001 0:02 000 0:07 00X 001-000-001-000 in three seconds. 0:10 000 Remain here 3 seconds. The previous two entries repeat approximately 14 times, dependant on time to reach set-point temperature. 1:26 Blinking 010 (HEATER QUALIFIED Mode) 1:32 Blinking 01X 011-010-011-010 in three seconds. 1:35 Blinking 010 Remain here 3 seconds, The previous two entries repeat approximately 16 times. 3:03 Blinking 110 (OPTICAL Mode) 3:09 Blinking 11X 111-110-111-110 in three seconds. 3:12 Blinking 110 Remain here 3 seconds. The previous two entries repeat approximately 9 times. 3:58 ON 000 (LASER READY)
The blink rate for READY is about 1.5 Hz.
Locking sequence with Newest Digital Control PCB:
The Newest Digital Control PCB seems to go through many more gyrations during warmup than either of the others, including several times where READY flashes multiple times separated by a period of inactivity, and then finally flashing READY continuously for two minutes until it locks - the latter being similar to what the Analog Control PCB does. The entire process consistently takes much longer than the 4 minutes typical of a laser with the Analog Control PCB, up to 10 minutes or more. The behavior is not obviously different whether a weak (but functional) laser tube, or one that greatly exceeds minimun output power specs is used, though it may take slightly longer with a below-spec tube. After all this, the end result seems to be exactly the same.
Here is a chart of the typical startup behavior for a very healthy 5517B tube installed in a (previously) 5517D laser with this newest digital Control PCB:
Time READY State Comments -------------------------------------------------------------------------- 0:00 1111 Power on 0:01 0000 0:02.0 0001 0001-0010-0100 sequence in less than 0.5 sec. 0:02.1 0010 0:02.2 0100 0:03 Flash 1100 MSB LED and READY LED flash briefly. 0:04.0 0001 0001-0010-0100 sequence in less than 0.5 s. 0:04.1 0010 0:04.2 0100 0:05 Flash 1100 MSB LED and READY LED flash briefly. 0:06.0 0001 0001-0010-0100 sequence in less than 0.5 s. 0:06.1 0010 0:06.2 0100 0:07 Flash 1100 MSB LED and READY LED flash briefly. 0:08 0100 3x(Step 0, Step 3, Step 5) sent on RS323 port. 0:15 0101 "LASER" sent out RS232 port. 1:00 0110 1:35 Flash 5 X110 1110,0000,5x(1110,0110). 1:40 0110 2:14 0000 2:15 Flash 5 X110 1110,0000,5x(1110,0110). 2:20 0110 2:55 Flash 6 X110 1110,0000,6x(1110,0110). 3:01 0110 3:35 Flash 7 X110 1110,0000,7x(1110,0110). 3:42 0110 4:12 Flash 8 X110 1110,0000,8x(1110,0110). 4:20 0110 5:02 Flash 8 X110 1110,0000,8x(1110,0110). 5:12 0110 5:40 Flash 10 X110 1110,0000,10x(1110,0110). 5:50 0110 6:45 Flash 32 X110 1110,0000,32x(1110,0110). 7:17 Flash 96 X111 96x(1111,0111). 8:53 ON X000 1000,0000,1000,0000,...
The State refers to the 4 SMT LEDs above the upper left corner of the Lattice chip near the center of the PCB. The MSB is green while the three LSBs are red. All times are approximate. "Flash" is just the briefest pulse of light. "Flash n" denotes "n" flashes at a 1 Hz rate with a 50 percent duty cycle The duration for the 1110,0000 state changes in each "Flash n" sequence is relatively short (perhaps 100 ms for each of the two states). The minimum value for "n" seems to be 5, but it tends to increase as the laser warms up. (I'm not positive it's monotonically increasing though.) Once the REF frequency can be sustained by the feedback loop, it continues flashing for 32 seconds, and then switches to state 0111 for 96 seconds prior to becoming READY. Until that time, the READY LED and the MSB state bit track each other almost perfectly. But then, the MSB state bit (1000) continues to flash (but now at about a 1.2 Hz rate) while the READY LED remains on solid, And there is just a hint of the 0100 state bit flickering dimly, possibly the actual feedback loop in operation. :)
Multiple runs from a cold start may differ slightly in the number of "Flash n" sequences and the values of n, as well as other details, but always take much more than 4 minutes (typical of the analog Control PCB). The very healthy tube will lock in 7 to 9 minutes while a weak but usable one might take 11 minutes or more. In all cases where the laser successfully locks, the last two minutes will be identical in behavior to that of the other two Control PCBs with READY flashing continuously until it stays on solid. A tube that is very weak or with no detectable beat (REF) frequency will result in only occasional very short abortive flashes, and no conclusion (at least not in 15 or 20 minutes).
One annoying difference between this Control PCB and the others is that the signal level for REF and ~REF seems to be much lower - about 2 V p-p open circuit and less than 1 V (maybe as low as 0.5 V p-p) terminated, instead of more than 5 V p-p, and the 5508A display apparently doesn't accept this as a valid signal. So even if the laser comes READY within 10 minutes (the maximum allowed by the 5508A), it still comes up as "Laser Fail", which isn't recoverable without power cycling the 5508A (which means the laser as well if it receives DC power from the 5508A). However, my home-built SGMD1 display has no problemss. :) I wouldn't be at all surprised to learn that the signal levels are programmable - somehow.
Here is a chart of the typical startup behavior for the 5517E with its similar Newest Digital Control PCB. The tube is probably below the Agilent spec for minimum power, but locks without problems so the sequence of event is probably not affected singificantly:
Time READY State Comments ----------------------------------------------------------------------- 0:00 1111 Power on. 0:01 0000 0:02.0 0001 0001-0010-0100 sequence in less than 0.5 s. 0:02.1 0010 0:02.2 0100 0:03 Flash 1100 MSB LED and READY LED flash briefly. (Repeat the previous 4 events 27 times.) 1:15.0 0001 0001-0010-0100 sequence in less than 0.5 s. 1:15.1 0010 1:15.2 0100 1:16 Flash 1100 MSB LED and READY LED flash briefly. 1:17 0100 3x(Step 0, Step 3, Step 5) sent on RS323 port. 1:27 0101 LASER sent on RS232 port. 1:52 0110 2:25 0000 2:26 Mode 16 XXX0 16x(1110,0000,....,0000). 3:35 Mode 12 XXX0 12x(1110,0000,1110,0000,....,0000). 5:10 Flash 15 X110 1110,0000,15x(1110,0110). 5:40 0110 5:48 Flash 32 X110 1110,0000,32x(1110,0110). 6:10 Flash 96 X111 96x(1111,0111). 7:46 ON X000 1000,0000,1000,0000,...
The last part of the sequence is essentially identical to that of the other laser, but the initial behavior differs significantly. This one appears to keep track of the mode cycles, or at least flash the State LEDs in response to them! "Mode n" denotes "n" times where the Zeeman beat is on, at least momentarily. Also note that the READY LED only tracks the MSB State bit near the end. I assume that the AM29F040B (4 Mbit flash memory) is the firmware NVRAM but there is no version number so I don't know that they differ, but they must as everything else on the two digital Control PCBs appears identical including the DIP-switch settings.
For more on the 5517 laser Control PCBs, see the section: HP/Agilent 5517 Laser Construction.
While HP/Agilent lasers are very good for their intended metrology applications, they can't compare to the best stabilized HeNe lasers like those from Laboratory For Science, Spectra-Physics, and others. There are issues with both short term variation in optical frequency as well as long term frequency drift. The 3 most significant are probably:
Replace the HeNe laser power supply with a low noise/low ripple type or add an external ripple reducing circuit to its output. The older VMI 148 had particularly high ripple, but even the VMI 217 can stand improvement. I have not tested the older Laser Drive 111-Adj-1 or the newest VMI 373 for ripple. But the VMI-373 already has a ripple reducer built in. See the section: Reducing Ripple and Noise in a HeNe Laser Power Supply with a Switchmode Regulator.
(I'm assuming modifications to the common Analog Control PCB. I do not know if it's possible to do this easily to either of the digital control PCBs. At the very least, cuts and jumpers would be much more difficult on the dense surface mount PCB. And, since there are so many of the older Analog Control PCBs available, why would you want to!)
Remove the LCD panel and its photodiode. Drill a hole in the beam sampler PCB and mount a polarizing beam sampler (e.g., polarizing beam splitter cube and a pair of silicon photodiodes) on top of the PCB. (It might even be possible to build this into the plastic housing instead.)
The schematic for one possible modification is shown in Upgraded Electronics for HP-5517 Lasers 1. This references the part numbers found on the 5517A/B/C/D Analog Control PCB, and probably the 5518A and 5519A/B as well.
The dual trans-impedance preamp for the photodiodes generates separate S and P mode signals. These feed the "Subtracting-Sample-and-Hold" circuit modified so that when in "Optical Mode" under feedback control, it passes both straight through - no holds allowed! During "Warmup Mode", it must pass the normal heater drive signal. The added preamp can be made from any stable dual op-amp mounted on a little circuit board perhaps stuck on top of U12, the LF13331D quad JFET analog switch, and attached to the photodiodes via twisted pairs. A 1M ohm pot in parallel with a 1 nF capacitor should suffice for the op-amp feedback, providing enough gain for all but the weakest laser tubes. The op-amp, U101, isn't critical - something like an LT082 would suffice. A few cuts and jumpers will be required, but on the wide open through-hole layout of the Analog Control PCBs, that shouldn't be too difficult. An alternative would be to remove the LF13331D and install an IC socket in its place. Then, build a little PCB that plugs into that with the LF13331D and preamp circuitry on it. Add an offset pot and it will then be possible to fine tune the optical frequency. It may not end up pretty, but should work great! It may be easiest to do the modifications in two stages: First replace the LCD and its PD with the polarizing beam sampler and preamp, and confirm that the correct polarizations are selected - the system should lock normally. Then disable the LCD selection logic so that both signals are passed at all times.
I later implemented a simpler set of modifications as shown in Upgraded Electronics for HP-5517 Lasers 2. This should produce similar results but with a wee bit less flexibility:
Wiring of the lower "POWER AMP DRIVE" switch (U12B) was unchanged (enabled by "DISABLE").
See Modified Beam Sampler and Offset Adjust Circuit for HP/Agilent 5517 Laser for a photo of these modifications.
The first two sets of changes were implemented first. These worked fine with the locking characteristic after warmup, total time-to-lock, uncertainty in REF frequency, and slow oscillation in REF frequency appearing very similar to the behavior of an unmodified laser. This is actually a rather surprising and unexpected result, so more study will be required. :) A discrete time system has been converted into a continuous time system without doing anything to the loop parameters and there were no dramatic changes the system response. Interesting.... However, later I did confirm that actual locking to the modes occurred almost immediately after the laser entered "Optical" mode (about 100 seconds after READY starts flashing). I also confirmed that if the photodiode polarity was incorrect, it would repetitively pass through the lock point at a rapid rate but never stabilize there. I had expected it to lock to the opposite crossover point of the two modes, but apparently the slope there is so much different that it never latched on, so to speak. Or, possibly it would have locked there eventually but I did not wait long enough.
For the record, the laser first tested with these modifications was a somewhat high mileage 5517C with a power output of around 240 uW and a REF frequency of around 3.3 MHz, the latter being outside the spec'd range for the 5517C (2.4 to 3.0 MHz). The uncertainty in REF frequency may be 200 Hz or more. The variation starts out with a period of around 16 seconds and deviation of around 0.003 MHz. Over a few hours, it slows down and finally stops (or becomes so long as to not be obvious).
Then a 220K ohm resistor placed in parallel with the 2.15M ohm resistor. This also had no detectable effect once locked. But, while the time-to-lock didn't change that much, the locking behavior was more rapid after the initial warmup.
Later, I installed the modified control PCB and beam sampler in a certifiably healthy low mileage 5517B (510 uW, 2.32 MHz). Locking was fine and both the randomness and periodic variation in REF were still present, though subtly different. The amplitude of the randomness was slightly lower - perhaps averaging 50 Hz compared to 200 Hz. The period started at about 10 seconds and the deviation was about 0.0045 MHz. However, running this laser with its original Newer Digital Control PCB resulted in essentially identical behavior. However, the deviation as well as the amplitude of the randomness also appear to be affected by exactly which longitudinal mode pair (i.e., exact temperature) at which the laser locks. A later power-on cycle resulted in a deviation of almost 0.01 MHz. Turning my offset control too far (accidentally!) resulted in the laser losing lock and then reaquiring it after the offset was turned back toward center. But the behavior had changed! The deviation in particular had dropped from 0.01 MHz to 0.004 MHz or less. Nothing else was different other than (presumably) where it locked!
The detailed character of these artifacts remains a mystery. The randomness may in fact be a faster but lower amplitude oscillation in REF frequency superimposed on the larger slower one but it's hard to tell without actually recording it, which I'm not sure I am eager to do. :) Since other evidence suggests that there isn't a corresponding variation in optical frequency to go along with the variation in REF frequency, this peculiarity may be a fundamental characteristic of the Zeeman laser and have nothing to do with the stabilization at all. Or, they may be the result of some sort of etalon effect. The time constant of the slow down in the periodic variation in REF frequency is too long to be anything but thermal in origin. HP/Agilent laser tubes have at least 4 planar uncoated glass surfaces outside the laser cavity and these are not wedged or set at an obvious angle to minimize back-reflections. In addition, the optics of the beam expander telescope and beam sampler have several optical surfaces. Since the structures these are mounted on are all mostly temperature independent of the controlled thermal environment of the mirror spacing rod, it's possible that one or more are forming some sort of external resonator with its longitudinal modes interfering with the normal lasing process very slightly. Maybe.
And eventually, I will have to set up the dual laser setup to check the optical frequency stability.
A second order effect is external magnetic fields, but this really shouldn't be significant unless other Zeeman lasers are living nearby, or you want to run this inside an MRI machine. :) And for the purist, air pressure and seismic activity also affect optical frequency, but the three modifications described above should reduce the short and medium term (up to days, probably not years) variation by more than an order of magnitude. Long term drift of optical frequency will be dominated by changes in the laser tube gas pressure and fill ratio from use, and this can't be easily controlled. But periodic diddling with the offset pot can compensate for those. :)
And, no, there is nothing labeled "RS232 Port", even on the newest digital Control PCB. But, there was a header a with a suspiciously appropriate number of pins (10) near the SHARC chip, so I started looking at voltages, and sure enough, pin 3 on the header had -9 VDC on it, and was occasionally pulsing to 0 V. So, I made up a cable to my ancient Kiwi laptop, and sure enough, there was ASCII being spit out at 9600 baud! :)
Header Pin DB9 Pin Signal ---------------------------------------------------- 3 2 Data from 5517 (transmit) 5 3 Data to 5517 (receive) 9 5 Ground
The "DB9 Pin" is the result of using an IDC cable wired directly to a DB9 connector. These may be salvaged from old PCs as they are often used to attach the mainboard to the rear panel. The pin numbers will be the same on the PC (not swapped). It's 9600 baud and full duplex (the laser echos characters typed). I have no idea about start and stop bits and parity, but suspect they don't much matter.
The few interesting things I've discovered so far are:
The complete sequence of what's sent from the RS232 port with a working tube (for either controller) from start to finish is:
Step 0 Step 3 Step 5 Step 0 Step 3 Step 5 Step 0 Step 3 Step 5 LASER READY
That's it! See, I told you it wasn't very exciting. :)
Perhaps flipping one of those DIP-switches will put it into Verbose mode, but picking the wrong one might erase the Universe, so I'm not willing to risk that - just yet. :)
At the very least, the runing time is probably maintained in NVRAM and it would be nice to know how to access that!
If anyone has more information on these digital control PCBs and their RS232 or other diagnostic port, please contact me via the Sci.Electronics.Repair FAQ Email Links Page.
More to follow, maybe, but this "First Contact" is encouraging. :)
HP-5517 power/reference connector J2
Pin Function ------------------------------------------------------------------ A No Connection on 5517 (MEAS signal level on 5508A) (1) B ~MEAS (Not used on 5517) (2) C MEAS " " D Signal Return (MEAS) E ~REF (Zeeman beat signal from internal optical F REF receiver's differential line driver.) G,H Ground J +15 VDC Sense K +15 VDC L -15 VDC M +15 VDC N,P Cable Shield R Signal Return (REF) S Ground T +15 VDC U Cable Shield
The 5517 laser head connector looks like a standard MIL-style bayonet lock type but apparently is special built for HP by Amphenol. The part number may be PT06A-14-18PZ but this has not been confirmed. It's possible the mate to these is available direct from Amphenol or a distributor, but probably not a standard item at an electronics distributor. It's actually the same connector as used on the 5500C laser head (PT06A-14-18P), but the keying is rotated 270 degrees, not that that helps much. I did check a Mouser catalog and indeed, the standard connector has the keying rotated by 270 degrees like the 5500C. One mating connector from an original HP cable is labeled: 97 USA/CTI 26SOU 851-06P14-18PX50-44. Searching for any of these part numbers only seems to result in non-stock items with no listed prices - "Ask for Quote". You know you're in trouble when this is involved! :) Used cables for the 5517 are available for around $100 from various surplus dealers and often on eBay. But the standard cables may be 10 or 20 feet long and much more than is needed if the other parts of the interferometry system aren't being used. For power and the reference signal, the mating connector and a few wires should suffice. It would be a pity to chop up an expensive high quality cable simply for the connector. However, it is possible to modify a standard PT06A-14-18P See the section: Making HP Interferometer Cables.
The pinout for HP cables with a 5 pin power connector (looks like an old PC AT keyboard connector) is as follows:
Pin Function 5 Pin DIN Female ------------------------ 2 o 1 Ground 4 o 5 o 2 Ground 3 No Connect? 1 o 3 o 4 -15 VDC ___ 5 +15 VDC [KEY]
A missing or low -15 VDC supply will not prevent the tube from operating but the laser will never lock.
The tube and HeNe laser power supply in the 5517, 5518, 5519, (and 5501B) can easily be tested using a 15 VDC power supply without powering the rest of the laser. The power supply must be regulated and capable of a current of at least 1 amp:
There still can be problems once power is applied to the heater inside the tube but at least this test proves that the tube isn't dead. Also note that if output power is measured in this unlocked state, it may be 50 percent or more greater than once locked.
There is no practical way to boost output power from a weak tube. But note that the power may increase 10 percent or so with an hour warmup compared to what it is just after locking. So, the situation may not be quite as bad as it appears initially.
However, a laser with power below the 180 µW HP/Agilent specifications may still be very usable, especially with simpler interferometer setups or fewer interferometer axes. The optical receivers are quite sensitive and only a few µW is sufficient for a stable beat frequency signal. With my crude setup using a 10780A optical receiver at its default threshold setting, 12 µW from the laser resulting in 8 µW at the receiver is sufficient power. Adjusting the 10780A's threshold setting or using a 10780C optical receiver which is more sensitive would require even less power.
But a low power laser may have other problems. See below.
All Analog Control PCBs made after somewhere around 1990 are wired this way yet the PCB artwork was never updated. I don't know whether this modification was done to prevent locking where the REF signal is so low that it might be corrupted by amplitude ripple resulting from HeNe laser power supply current ripple, or simply to sell more lasers since they will fail to lock sooner. :)
For late model 5517s, the Newer Digital Control PCBs will lock below 25 µW if the "REF" jumper is set to "LO" and they will generate a REF signal, though it's not known what threshold is needed to operate normally. The corresponding parameters for the Newest Digital Control PCBs are not known at all. See the section: HP/Agilent 5517 Laser Construction.
There are two other work-arounds for an inability to lock due to low power where you don't want to modify the Control PCB:
The first of these is probably preferred as it doesn't require permanent modifications to the laser. An external 5 or 10 percent beamsplitter and optical receiver are required but the beamsplitter can probably just be a microscope slide at 45 degrees and any of the optical receivers would be suitable for generating the reference since it operates at a fixed frequency not greater than 4 MHz regardless of the laser model. The obsolete 10780A or 10780B can be obtained very inexpensively.
It may also be possible to change a component in the reference detection circuit but I have so far been unable to obtain a datasheet for the actual IC that is used there - HP part number 1826-0775 or the manufacturer's part number 1DA7Q (assuming this isn't simply some random collection of characters that was never updated since Google has no clue about it!).
There is most likely nothing really wrong with the laser but the tube is simply high mileage or one that has difficulty starting. Some tubes become like this. Even the Ph.D. types at a major laser company really don't know why. Aside from time wasted twiddling one's thumbs, a tube that takes a long time to start is hard on the HeNe laser power supply, but the ones in these lasers seem tough. Do check the DC voltages, particularly for the +15 VDC supply, which is what powers the HeNe laser power supply. While the rest of the laser may run on 12 VDC or below, lower DC voltage means proportionally lower starting voltage for the tube. And lower voltages may be more stressful both on the HeNe laser power supply and other parts of the laser, as well as being more likely to take a long time to start with an uncooperative tube.
In several instances, I found that shining a light on the *back-end* of the tube would promote starting in an otherwise uncooperative laser. Electrical discharge initiation is known to be sensitive to light and radioactivity, so this effect isn't entirely surprising. A radioactive source would work but putting a radiation warning sticker on the laser might invite a visit from Homeland Security. So, I opted for LEDs instead. :) For some tubes, a high brightness red LED shining on the glass extension at the back of the tube is sufficient to reduce the starting time from a minute or more to a couple of seconds. A blue LED seems to be even more effective especially for particularly uncooperative tubes. The LED can be conveniently wired to the HeNe laser power supply on the back of the connector PCB. I now routinely install an LED if there is any hint of slow start. It can't hurt and so far, has seemed to help significantly. For a higher level of sophistication, add a circuit to turn off the LED once the tube starts!
It's also possible that electrical leakage is reducing the effective starting voltage. If there is a smell of ozone while the tube is trying to start, then corona is present from the anode terminal.
CAUTION: Take care to avoid stressing or bending the wire connections to the tube terminals when removing the old foam and cleaning and insulating the anode, as well as reinstalling the tube assembly in the laser. Avoid applying force (especially side-ways) to the cathode/heater terminals and their glass-to-metal seals, and avoid repeated bending of the anode wire since re-attaching that should it break off could be challenging.
Attach a multimeter on DC Volts across the 1K ohm resistor. The reading will be 1 V/mA. Power up and start at the default current setting for the internal HeNe laser power supply of 3.5 mA. If increasing the current results in a stable output, then the problem is almost certainly the dropout current as noted above. The current will need to be slightly beyond where the laser is stable. 3.75 or 4 mA shouldn't hurt it or significantly reduce life expectancy. There's no choice anyhow as this may be the only practical way to get these tubes to stay lit! If the tube is unstable even at 4 or 4.5 mA, then the problem may be the power supply, or the ballast resistor attached to the tube (quite unusual).
Limited anecdotal evidence suggests that a laser repaired in this manner will run continuously with useful power for several months. And, of course, if only turned on when needed, for much longer.
Adding an anode ballast resistance without increasing the laser tube current may work in marginal cases. But in my tests, even as much as 35K ohms only reduced the dropout current by 0.1 or 0.2 mA. So, it alone is probably not a reliable solution for a tube that doesn't stay lit. But adding some modest anode ballast resistnace (10K to 20K) is worth doing to reduce the chance of amplitude ripple as discussed below.
CAUTION: DO NOT allow a laser to continue sputtering for a long time. This may damage the laser tube and destroy the power supply. I've had 5517 lasers where the HeNe laser power supply had been blown due to unattended sputtering, though it's not clear if there was any damage to the tube. But, on two 5501A tubes I tested, rapid sputtering as a result of a defective HeNe laser power supply caused the output mirror inside the tube to literally have a hole blown in the exact center of its coating, rendering the tube useful only as a magnetic paper clip holder/desk ornament or paperweight! One tube had a hole just about the size of where the beam would have been as can be seen in Hewlett Packard 5501A HeNe Laser Tube with Missing Coating in Center of Output Mirror. But the other had a clear hole in the coating over 2 mm in diameter!
Note that in a 5501A or 5500C, sputtering may either be due to a defective HeNe laser power supply (probably the potted module), the laser current being set too low, or the tube itself being unable to stay lit at any current setting. With the current setting being under user control, it's critical to set it so that the tube will stay lit. The current should be set according to the recommended value on the label (if any), but subject to the constraint that it be at least 0.2 to 0.3 mA higher than the dropout current after a 1 hour warmup to assure reliable operation in the long term. If there is no value listed, then assume 3.0 mA or adjust for maximum output power when locked between 3.0 to 3.3 mA, but subject to the same constraint. (The nominal operating current may range from 2.6 to 5.1 mA, according to the 5501A manual. But most are between 3.0 and 3.3 mA, so if there is no value listed, it's safer to keep it within this range if possible.) The current may either be measured by installing a mA meter between the tube cathode post (on the side of the large glass bulb of the tube) and its connecting wire, or by measuring the voltage on the laser current testpoint, which is series with a 390 ohm resistor to ground. So, the current will be V/390. The testpoint is accessible on the left side of the rear connector PCB, just above the laser current adjust pot, R11. Leave the right side cover in place to activate the interlock switch that enables the laser to turn on.
Aside from problems in the electronics, a very unusual cause might be an intermittent connection *inside* the HeNe laser tube. Broken welds are possible, but what's more likely is just bad contact with respect to the cathode terminal, a pressed-on slide fit in the 5501B and later lasers, and simply a tab pressing on the through-glass terminal on the 5501A. As the parts expand, the result is momentary loss of contact. I've never actually confirmed this in an HP/Agilent tube though I do have a 5501B that is suspect. While not that common, it does happen with conventional tubes. Needless to say, there is no truly guaranteed practical fix other than installing a replacement tube. However, such events are much more likely before the system reaches thermal equilibrium. So, simply running the laser for awhile before use may be sufficient to reduce the frequency of occurrence of these glitches to zero. And thus lasers run 24/7 may never experience them after the first few hours.
One time, I was 100 percent sure that a bad internal connection was the problem with a 5501A. But it turned out to be much simpler. The tube would drop out at random times anywhere from a few seconds to hours apart (generally less frequent after warming up). The tube was absolutely healthy in all other respects - great power, instant start, and stable over the full range of laser current adjustment. But the symptoms always remained with the tube when it was installed in two known good laser chassis. The tube was even connected to a stand-alone HeNe laser power supply and then, tapping on the tube would sometimes induce a dropout. However, I was suspicious of the anode contact as jiggling the HV wire would also tend to cause dropouts. And, indeed, with the front optics assembly removed, the anode terminal was found to be only a short stump (probably original) flush with the glass. And there were also bits of RTV Silicone stuck to it (origin unknown). I had tried to clean that terminal early on in this saga with no change in behavior, but RTV Silicone bits don't come off easily, especially if they were visible! So, they were preventing the spring contact from seating firmly against the terminal. Or something. :)
Viewing the beat frequency from an optical receiver on an oscilloscope will show all edges except the one used for triggering the scope to be fuzzy as the bogus signal modulates the position of the zero crossings, rather than the clean waveform that is expected. (But check and touch optical alignment as poor alignment can also result in a fuzzy signal.) The oscillation itself will show up when only one polarized Zeeman mode is presented to an optical receiver (e.g., by blocking the return beam from the interferometer) or by using a photodetector and oscilloscope. In the latter case, it will be seen as a sinusoidal waveform that is present *without* a polarizer. With a polarizer, the beat frequency signal will be riding on top of the ripple. A typical ripple amplitude is 10 µW but this can vary greatly. It will also appear by itself at the output of an optical receiver while the laser is warming up between those times when the normal beat frequency signal is present.
The exact cause of the bogus signal is not known but it probably has to do with the tube's negative resistance. The ballast resistor for these tubes is located 5 to 6 inches from the anode (which is much longer than the 2 to 3 inches usually recommended for HeNe lasers) and the wire between the resistor and tube may run close to the grounded chassis, adding capacitance. So this certainly makes such problems more likely.
For tubes that meet HP specifications for output power (180 µW for most models), it is probably not necessary to do anything about this oscillation unless specific measurement issues can be directly tied to it. In fact, I've seen it in lasers that appeared to be virtually new in all other respects, so it may simply be considered normal!
There are two ways of eliminating the amplitude ripple:
Adding a cathode ballast resistor would probably eliminate the oscillation as well but this is not an option with 5517/18/19 or 5501B lasers since the cathode is attached to the heater used for thermal tuning inside the tube and it must be near ground potential. A cathode ballast resistor should be acceptable on the 5501A.
I don't know for how long these cures will be effective or whether they work in all cases. And sometimes, both will be required. If the increased current is needed to fix a tube that won't stay lit, try that method first.
I have found another oscillation at around 100 kHz with some lasers that's probably even lower level. Adding ballast has little or no effect on this so it may be from some other cause like the switching frequency of the HeNe laser power supply. I haven't seen a laser where this was of any consequence once locked though - the beat waveform is clean, especially with respect to the full cycle. If the scope is triggered on the rising edge, then there may be some fuzz on the falling edge, but not subsequent rising edges. And, it's generally very small.
The following are causes that may produce a variety of symptoms:
Variations in the beam sampler behavior even among units considered to be good can result in a shift in the optical frequency of 5 MHz or more. In fact, from my admittedly limited tests of 3 supposedly good beam sampler assemblies, this may be the dominant factor affecting optical frequency in lasers with a similar number of hours on the tube. However, I do not know if it is due to variations in the LCD panels, or the other optics of the beam sampler. While the laser will still easily meet specifications (5 MHz is only about 0.01 ppm), this is an annoyance in the elegance of these systems department. :) But, without comparing the laser's output to a reference, the only symptom may be a larger than expected mode imbalance, though a visual inspection and electrical testing of the LCD panel as described below may identify marginal units.
Some LCDs also seem to cause a continuous or intermittent hunting behavior of the optical frequency with a larger deviation than is normal. All HP/Agilent lasers exhibit a slow variation in optical frequency with a period of around 2.56 seconds and a deviation of 100 kHz or so. The frequency deviation in some lasers may be up to +/-0.5 MHz or even more. Since this amount of wobble in the optical frequency is well below HP/Agilent specifications, this should probably be filed under the "well that's interesting department" rather than considered a serious issue. The only way to detect it would be to beat (heterodyne) two lasers together and look at the difference frequency. The cause is a direct result of the stabilization loop implementation using an LCD to alternately select each of the two polarized modes, rather than using the more conventional polarization beam sampler and a pair of photodiodes.
The LCD panel is inside the beam sampler assembly following the output end of the laser tube. Remove the two screws on top holding the beam sampler in place and at least the two mounting screws for that end of the main PCB to allow it to be pushed away from the laser body slightly. Then, it should be possible to pop the beam sampler off of the laser. Working over a soft pad should the LCD panel fall out, remove the small PCB with the photodiode, and use fine tweezers to pull out the elastomer ("Zebra Stripe") connector padss. Then, the LCD panel should slide out of the plastic housing. Simply rotating a polarizer behind it while looking through from the front may reveal an obvious problem like two or more sections that have different polarization orientations. The entire panel should be the same polarization. If this simple test doesn't turn up anything, then apply a 50 Hz AC (balanced) 5 p-p V squarewave across the electrodes on either side of the LCD panel (where the elastomer pads were pressing) while looking through it and the polarizer. (It can be placed on top of a polarizing filter on a white piece of paper.) When applying the 50 Hz signal, the density should change dramatically and uniformally. CAUTION: Make sure the signal is AC with no DC component of applied for more than a few seconds as prolonged DC current through the LCD can damage it. However, momentary DC won't hurt.
Since the elastomer may tend to stick to the LCD, even easier is to test the LCD in place by removing the sampler PCB and covering the reference PD port (side) and output window (front) with pieces of black tape. Put a polarizer over the input port (back) and look through the sampler port (top) with the input port facing a brightly illuminated white surface. Then, one orientation of the polarizer should show a uniform bright field while the orthogonal orientation should show a uniform dark field. Apply the voltage and these should flip states.
Or, just swap in an LCD panel from another laser and see if it now locks! :) The 5501B and all 551X lasers appear to use the same LCD panel and probably the same beam sampler optics assembly (the black plastic housing and the optics inside it), the top PCB with the photodiode monitoring the LCD output is larger for the lasers in the large cases (5517A, 5518A/B, and 5519A).
Where an adjustable power supply brick is found to be bad, it can be replaced with a fixed current supply as long as the tube was being run at the default current of 3.5 mA - which would be in the vast majority of cases unless the tube was high mileage and would only stay lit on higher current. I had to do this for a 5517A. The original supply would start and run long enough for the laser to lock, but would then sputter a few tims and blow the internal fuse. Swapping in a supply from a 5517C repaired that.
For detailed service information, see the section: Additional HP/Agilent Resources. While there is nothing on the 5517 laser specifically, the electronics of the 5518A (part of the 5528A Measurement System) and 5517A is identical except that the 5518A has an additional PCB (the internal optical receiver). And the electronics of 5517B/C/D lasers using the Analog Control PCB is close enough to that of the 5518A to be useful for troubleshooting and repair.
There are two apertures at the output-end of the laser. The top one is the normal laser output, with the usual control wheel for a large opening (normal), small opening (alignment), and closed. It is also the return port for straightness measurements only. A second aperture below it is for the optical receiver. This aperture is used for all measurements except straightness. It has a control wheel for large (normal) and closed (which then has an alignment target printed on the exposed surface). A large Turret Ring behind the apertures has two positions: Straight and Other. For straightness measurements, it inserts optics in the normal laser output aperture to direct a return beam there to the optical receiver, and a microswitch is activated to change the gain of the optical receiver. (The laser output power is also reduced somewhat in this position, so the optical receiver needs to be more sensitive.) There are also "Laser ON" and "Signal" LEDs on the front bezel. Laser On is the same as the LED on the back panel. Signal is lit when there is enough of a return beam to the optical receiver to be useful.
Several photos of a 5518A laser head can be found in the Laser Equipment Gallery (Version 2.42 or higher) under "Hewlett Packard HeNe Lasers".
The 5518A I have includes one additional component, not present in any other HP/Agilent laser I've seen, and that is a shield or cover surrounding the area of the beam expander, purpose unknown.
I have reverse engineered the schematic for the Optical Receiver PCB shown in Photos of HP-5518A Optical Receiver PCB. See HP-5518A Optical Receiver Schematic. Most of the component designations are arbitrary since very few had anything on the artwork. Although it performs a function similar to that of external optical receivers like the 10780C, the circuit is considerably simpler and nearly identical to that of the reference receiver on the Control PCB. The built-in photodiode can be seen below the hole through which the output beam passes. The two pin header attaches to the microswitch in the current assembly that selects gain based on whether it is set for "Straight(ness)" or "Other". The gain is increased in Straightness mode since the outgoing beam passes through a non-polarizing beam-splitter and the return beam reflects off of it
I would also like to find the non-HP equivalent of the receiver IC U1, HP part number 1826-0775, listed as 1DA7Q on the HP schematic of the 5517B laser, which (among others) uses the same IC. If anyone has a standard part number and/or datasheet, please contact me via the Sci.Electronics.Repair FAQ Email Links Page. Of course, maybe 1DA7Q was just a random text string intended to be replaced by the actual part number and that never happened! :) A different revision of the schematic shows the manufacturer part number as 1826-0075 which could be another typo.
The case style of the 5519A/B (and 5518A) is similar to that of the 5517A and the tube assembly is very nearly physically interchangeable. The "very nearly" means that a small piece of the tube assembly casting needs to be cut away to provide clearance for the internal DC power supply, not present on the 5517A or 5518A. So, where the higher REF is not needed, a 5517A tube assembly can be installed relatively easily. Presumably, the modification wouldn't be needed for the 5518A since it uses external DC power like the 5517A.)
The control electronics of the 5519A/B laser heads is functionally the same as all the other thermally tuned HP lasers. But the main PCB is probably not interchangeable among the 5519A/B and the 5517A and 5518A. The 5519A/B seems to have done away with any need for -15 VDC unless it's generated on the main or connector PCBs as the built-in DC power supply only provides +15 VDC. And the 5517A may not have the needed connections for the optical receiver of the 5519A/B.
Many photos of a 5519A laser head can be found in the Laser Equipment Gallery (Version 2.31 or higher) under "Hewlett Packard HeNe Lasers".
The 10780A is used in interferometry systems using the 5501A, 5501B, 5517A, or 5517B laser heads. It contains a silicon photodiode behind a focusing lens and polarizing filter oriented at 45 degrees, a preamp, a comparator to generate a digital signal from the heterodyne beat of the two polarized modes of the Zeeman-split lasers, and a differential line driver. The primary output is called called "MEAS" and its complement "~MEAS". There is also a Beam Indicator LED which will be lit when there is enough power to produce a reliable beat frequency signal. (This threshold is adjustable.) The 10780B appears substantially similar to the 10780A except that the threshold pot is accessible without removing the receiver cover.
The pinout of the main connector (J1) is:
BNC Pin PCB Pin Function ----------------------------------------------------------------- 1 (LL,F) 1 ~MEAS (Zeeman beat signal pair from 2 (UL,F) 2 MEAS differential line driver.) 3 (LR,M) 3 Return (also BNC shell and receiver case). 4 (UR,M) 4 +15 VDC
The PCB pins are counted from the edge of the board. I don't know the official designations of the pins on the funny bi-sex 4 pin BNC connector. LL (Lower Left, etc., F for female and M for male) reference the connector with the receiver oriented vertially - with the optical input and Beam Indicator LED at the top. (Rather than buying the way overpriced mating cable, I fashioned a 2 pin female header for power that fits only one way into the male pins, and a separate 2 pin male header for the MEAS signal. These were then glued into a BNC shell. It's not as pretty as the original but it works. I have not found any supplier for the mate that sounded like the cost would be less than an arm and two legs for each one.
(Note that the 10791, one of the types of cables that is used to connect the 5517 laser heads to DC power and the measurement electronics, has a 4 pin BNC plug like the one that mates with the optical receivers. The REF outputs of the laser are on the male pins with +15 VDC and GND on the female pins. This connector should normally NOT be attached to the optical receiver!)
There is also an external test-point called "Beam Monitor" on a feed-through pin sticking out above J1. This is the intermediate rectified and filtered signal used for the threshold detection.
The case should not be connected to the optical metal chassis or Earth ground (I assume for single point grounding noise considerations). Use Nylon screws through the plastic insulated mounting holes at each end.
The 10780A and 10780B are now considered obsolete as they are not guaranteed to work with 5517C/D and later interferometer lasers over the full specified velocity range since the spec'd upper cutoff frequency is too low (5 MHz). However, HP/Agilent specs are often very conservative. A 10780A I tested using a function generator and LED operated from below 40 kHz to over 8 MHz. It actually would probably be usable down to around 10 kHz but the waveform was somewhat distorted below 40 kHz. The sensitivity as determined by the voltage on the Beam Monitor test-point was down to about 50 percent of what it was at 5 MHz, but some of that fall-off might have been due to my LED/driver. The replacements are the 10780C (free space optical input) and 10780F (fiber optic input, though some of these may actually have the 10780C model number and/or be designated 10780U). The 10780C and 10780F have a guaranteed frequency range from 100 kHz to 7.2 MHz. But for experimental use, when using a single interferometer, or when not requiring high velocity in one direction, the 10780A or 10780B should be fine and typically much less expensive on eBay. :-)
Any of these HP receivers make good general detectors for optical heterodyne beat signals within their frequency bandwidth since they will operate over a wide range of input optical power from a few µW to 1 mW or more without adjustment. They will also operate with similar optical pulsed signals and work fine to detect the chopped drive of some of my LED flashlights! :) However, note that although the 10780F/U can be used with free-space input, to do so will require a polarizer at 45 degrees to be added, and since there is no lens to focus the light onto the small area photodiode, the maximum sensitivity is much lower than for the other optical receivers.
5517/5508A adapter pin-out
Mil DB25 Pin Pin Function ----------------------------------------------------------------------------- A 1 MTR (MEAS signal level to meter on 5508A) B 2 ~MEAS C 3 MEAS D 15 Signal Return (MEAS) E 5 ~REF F 6 REF G,H 7,10 Ground J 11 +15 VDC Sense K 12 +15 VDC L 8 -15 VDC 9 -15 VDC Sense 20 -15 VDC 21 -15 VDC M 23 +15 VDC N,P 13,16 Cable Shield R 18 Signal Return (REF) S 19 Ground 17 Ground 22 Ground T 24 +15 VDC U 25 Cable Shield 4 NC 14 NC
MTR ~MEAS MEAS NC ~REF REF GND -15 -15S GND +15S +15 CSHLD 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 NC MGND CSHLD GND RGND GND -15 -15 GND +15 +15 CSHLD
The Power, REF, and MEAS signals are also brought out to terminal blocks so they can be monitored or easily attached to test equipment like a frequency counter or oscilloscope, as well as the 5501A reference connector.
For testing 5517s (all versions) and 5518As, the 5508A is used directly with the DB25 adapter. For testing 5501As, a separate DC power supply is used but with the 5508A powered and fed with REF and MEAS via the terminal blocks.
Since the circular MIL-Spec connector for the 5517/5508A is non-standard, the connectors available from electronics distributors need to be modified. The part number is PT06A-14-18P-SR with strain relief. Mouser has them for around $20 in single quanity, which is actually quite good as these things go. The modification turned out to be easier than I had anticipated. The pin block is made of rubber and can be pushed out with a piece of 1/2" copper pipe in a drill press. (Though a very slightly larger cylinder would be a bit better.) First, go around the periphery from both ends with a thin blade which will free most of the rubber from the adhesive used to secure it in place. The pipe fits around the pins without mashing them and only contacts the rubber. Push in increments, making sure the rubber doesn't get too misshapened or skewed in the process. The screw-on strain relief (if present) or some other suitable spacer with a hole in it will be needed under the connector to allow the rubber block to be pressed clear of the shell. Then reinstall in a similar way after aligning with respect to the 5517 or 5508A connector. There will be some damage to the rubber, but it should not affect anything unless you're a purist. Even without any adhesive, it's really snug enough but won't be a Mil-Spec connector that's waterproof. :) It would also be straightforward to fabricate a "punch" that matches the pin pattern. That may reduce collateral damage, but doesn't seem to be worth the effort unless 1,000 of these connectors need to be modified.
And a note about trying to salvage HP cables if all the required connections aren't already present: Forget it. The cover on the laser-end connector consists of a thick rubber boot on top of a hard plastic conformal molded inner core. While the boot can be slit from end-to-end and peeled off, I doubt it is realistic to remove the core without damage to the connector and pins. I gave up after seeing what would be involved since I didn't have any TNT handy. :) So, for example, an ET-319283 adapter cable which has the 5517 connector at one end and a 7 pin LEMO at the other, possibly intended to connect a 5519A/B to a 5508A Measurement Display isn't useful to power a laser since the DC power connections are not present. (The 5519A/B has a built-in switchmode power supply that runs off the AC line.)
The 5501A and 5501B use a pair of 4 pin circular connectors. The power connector is standard with a suitable mate being Amphenol PT06-8-4P-SR. (This is also available from Mouser, but is more expensive than the 18 pin connector!) The reference connector has the keying rotated 45 degrees but a similar push out and reinsert approach works, though more care is needed to assure that the rubber doesn't get destroyed. The diagnostic connector (present only on the 5501A) mates with the standard PT06-14-18P-SR. Unless you're into automated monitoring, building a cable for that is probably not worth it. See the sections on the 5501A/B, above, for pinouts.
Scans of original product brochures for the Model 200, 220, and 260 lasers, and html versions, as well as general desciptions and a price list can be found at Vintage Lasers and Accessories Brochures under "Laboratory for Science". The brochures include a nice description of the principles of operation and applications considerations in addition to the specifications.
The following brief descriptions include extensive contributions from David Woolsey (http://www.davidwoolsey.com/).)
There were three Laboratory for Science stabilized HeNe lasers known to have been produced and sold:
All three models had the same size power supply/control box but the laser head for the Model 260 was longer than those for the models 200 and 220. The user controls and general operating procedures are also basically the same for all models.
A number of features and attention to detail set these lasers apart from most other commercial stabilized HeNe lasers that are or have been available. These are described with respect to each model in the following sections. Unfortunately, clever ideas and implementation are often not the most important factors in determining the success of a product or business.
Even with the superb technology, not many of any of these lasers were ever sold. The total production run for all the years of the product line from the early 1980s to sometime in 1995 was soemthing like: 300 for the Model 200, 60 for the Model 220, and only 10 for the Model 260. There are references to other models ranging up to 280 in the product literature, but someone who actually worked at Laser for Science throughout the years of ultra stable laser production never heard of them going beyond the discussion stage.
Ironically, the extensive discussion of retro-reflections in the product brochures may have scared off potential buyers. Nearly half the text in the brochures for the LFS-200, LFS-220, and LFS-260 is related to the effects and mitigation of retro-reflections which some people might interpret as a deficiency with these lasers. Retro-reflections are a problem with all lasers, but especially with lasers designed to have the best stability performance. Other manufacturers tend to simply mention retro-reflections in the operation manual - not the product brochures! - as something to be avoided, but even there, they don't dwell on it.
Even experienced laser jocks find it hard to understand how reflected light with a power level 1/100,000,000 or less compared to the intra-cavity power can have an effect on the behavior but it definitely can with these type of lasers.
If anyone has schematics, a service manual, or other detailed documentation for any of the Laboratory for Science lasers (or an actual Laboratory for Science laser!) stached away they no longer need, please contact me via the Sci.Electronics.Repair FAQ Email Links Page.
The LFS-200 and LFS-220 are described in the order in which I acquired samples.
Although LFS is now out of business, other companies do offer transverse Zeeman stabilized HeNe lasers. One example is NEOARK (Japan).
Among the features and attention to detail that sets the Model 220 laser apart are:
The well known commercially available HeNe lasers I'm aware of implement very few, if any of these. And note that many duel frequency Zeeman like the 5501A/B, 5517, and others, use simple dual polarization mode stabilization techniques despite their being Zeeman lasers.
Scans of an original product brochure for the LFS-220 can be found at Vintage Lasers and Accessories Brochures under "Laboratory for Science". A much more compact html version is at Model 220 Ultra Stable Laser Brochure. The brochure includes a nice description of the principles of operation and of course, the specifications.
The first Model 220 I (Sam) acquired on eBay - serial number 51 - has IC date codes and PCB fab dates between 1981 and 1986. But if the serial numbers started at 1 (or even 10 as has been suggested) rather than 50 and only 60 lasers were ever built, it may be much newer than 1986, possibly between 1988 and 1992. So what if the chips are a bit moldy, they haven't changed in any way other than dropping in price by 1 or 2 orders of magnitude since 1981. :) Maybe LFS bought their chips from PolyPacks (a popular surplus outfit for cheap chips that also no longer exists). ;-)
While the tube in this laser is weak - around 0.8 mW on a good day which is about half the minimum power spec - this is more than adequate to provide a stable beat frequency signal. Originally, the laser was going through what appeared to be normal warmup, but would not lock after the warmup period and the Lock Level indicator came on. The Model 220 has a headphone jack to permit listening to the PLL error signal (as do the other models as well) and a knob to adjust the PLL gain. And while the knob affected the sound in the headphones, there was little correlation with anything else. It was like a bad SciFi movie sound track! I was thinking there must be electronic problems preventing a stable lock from being achieved. Fortunately, all ICs are standard 4000-series CMOS and common analog parts. Unfortunately, it's not likely that a schematic will be available given how few of these were probably built. Google gas been totally incapable of finding much of any useful information beyond the brochure for the Model 200 on David Woolsey's Web site and a few journel references citing the use of Laboratory for Science lasers for such-an-such research.
However, a miracle happened. Someone sent me the user manual for the Model 220 and lo and behold, that empty socket under the controller I had been pondering since acquiring the laser needed a jumper plug to complete the internal signal paths! It provides access to all the critical input and outputs of the internal architecture of the Model 220 controller with the intent to permit the use of an external frequency reference, remote control and monitoring, and other advanced functions. The jumper block must have either fallen out in shipping, or the previous owner had been using the remote hookup and kept the cable. No wonder it didn't work. Nothing was connected together! So there were electronic problems of sorts. :)
With the default jumper plug constructed and installed, *everything* started working in a manner that actually made sense. The Mode LED went on and off as the modes cycled in the HeNe laser tube during warmup and the headphones produced a satisfying chirp a couple times during each mode cycle. When the 30 minutes or so warmup time was completed, the laser locked instantly!
The sound from the headphones is nearly pure white noise and the beat frequency appears rock stable on an oscilloscope and around the 425.8317 kHz it should be based on the PLL synthesizer BCD switch settings of 511. (The frequency is: 3.4 MHz*m/M where M is the switch setting and m is 32, 64, or 128, preselected based on the laser tube to provide the maximum number of possible discrete Zeeman frequencies.) I intend to check it on a frequency counter but have little doubt that will also show the correct frequency with crystal accuracy. Unfortunately, I don't have a spectrum analyzer or an iodine stabilized laser to check it more precisely. The stability should increase is allowed to warm up for longer - 90 minutes is the time to reach spec'd performance. Originally, I thought it might not be working quite correctly due to the sound from the headphone jack having rumbling and other non-white noise components, but I now believe that may have been due to acoustic feedback since I was actually listening using a stereo amp.
Here are some photos:
The HeNe laser tube construction is nothing special, at least on the outside. Like the two mode stabilized HeNe lasers, a Spectra-Physics 088-2 or similar tube would work. But the actual tube used by LFS was apparently custom built though. Some, if not all, were filled with isotopically pure Ne20 or Ne22 to provide the narrowest linewidth and/or to select the precise line center, and possibly He3 as well. Later ones were made with a special bore support spider that eliminated the "slip-stick" behavior during warmup of some other designs.
The waste beam from the HR-end of the tube is used for the reference beat tone. It has a polarizing filter between the tube and the photodiode and a glued-on wedge to make sure the waste beam can't reflect back into the bore. There is an AGC circuit of sorts for the photodiode so that a usable signal can be obtained as the tube ages regardless of a (reasonable) decline in tube power.
The tube is rather elaborately suspended as can be seen in the photos. The suspension provides some degree of vibration isolation and there is even a fine thread screw (visible on the top of the laser head) to rotate the tube by a few degrees. The complex suspension was designed to minimize stress in the glass envelope and eliminate stick-slip noise due to length changes of the overall tube. It also allowed the tube to be rotated by a 100 pitch screw adjustment without twisting the tube at all. This was desirable to align the tube's birefringence axis (mode orientation) precisely with the magnetic field.
The entire laser head is thermally regulated by a temperature controller which is the circuitry on the lone PCB inside the head. The temperature set-point can be adjusted via a pot accessible from underneath the laser head. Power resistors attached to the baseplate on which the tube and magnet assembly is mounted provide the heating and an LED on the rear panel of the laser head shows the amount of power to the heater by its intensity. The baseplate bolts to the outer aluminum case with close-fitting end-plates. Although perhaps not obvious from the photos, the wall thickness is much greater than that of most other HeNe lasers.
There is also a rather elaborate transducer attached to the tube. While serving a similar function to the heaters on many mode stabilized lasers, the design was optimized for fast response. Power to this heater is what is controlled by the PLL responsible for locking the laser.
The transducer consists of a dense "zig-zag" run of copper wire about 3.5 inches long Epoxied directly to the outside of the glass tube envelope. The wire is oriented (back and forth) along the long axis of the tube, *not* as a helix or coil (it is not an inductor). When a current is run through the winding the wire heats up and immediately pulls (stretches) the glass with it. The response bandwidth is something like 10 kHz since the length change between the mirrors did not have to wait for the glass to heat up. With the wire arranged along the tube axis all of its change in length was in the intended direction - unlike with a the the more common coil arrangement.
With a simple coil, the initial change in dimension when current is applied is an increase in winding *diameter* which pulls the glass with it (expands the tube diameter) and causes an initial *shortening* of the tube. The shortening is followed by a lengthening as the heat from the transducer diffuses into the glass. This is not a good way to make a fast feedback loop. Also unlike other heater schemes generally used, with the wire directly attached to the tube glass, there is nothing in between to limit the response as with taped on thin-film heaters.
On the anode-end of the HeNe laser tube (the front of the laser head and output) is a collar with two LEDs on it and a trim pot. Only the anode wire connects to this collar. One LED is lit when the tube is first turned on. Inside the collar is a temperature regulator for the output mirror. There is a small amount of internal reflection in the mirror that gets back into the laser cavity and this is the way it was tamed. There is a thermistor regulated heater in there that uses the laser discharge current for power. The voltage drop across the heater box will vary, but the current through it is held constant. So, the mirror temperature is regulated so that the etalon formed by its front and rear surfaces has a peak covering the neon gain curve resulting in a constant transmission without retro-reflections. For the approiximately 5 mm thick mirror - 7.5 mm optical length - the FSR is 40 GHz, compared to 1.5 GHz for the Doppler-broadened neon gain curve. So, the peak is rather broad in comparison, but keeping it centered helps long term stability.
The rear mirror had a simple prism made of cover glass that was Epoxied onto it so that the internal reflection was removed by putting it off axis. The Epoxy was made to be thicker at one side than on the other by supporting one side of the cover glass with little tabs of tape. This method couldn't be used on the output mirror.
When the output window is under proper thermal regulation both of the LEDs on the thermal regulator enclosure should be half lit. The upper one lit means heating and lower one lit means cooling. The pot adjusts the temperature set-point.
And note that neither anode or cathode is at ground potential! Don't ask how I (Sam) found this out. :( :) This was apparently for noise suppression. Grounding one end of the tube will risk inserting some 60 Hz hum onto the tube current through ground loops and such. Talk about paying attention to every last detail!
The HeNe laser tube is driven by a linear power supply with totally exposed components once the controller cover is removed. Not even a plastic shield! It is the typical voltage doubler with parasitic voltage multipler for starting. Four power transistors provide current regulation in the cathode return. While at first glance it looks similar to many other linear power supplies of the early 1980s, it was designed to put out 5 mA at 1,200 Volts with a supply ripple of about 1 mV! That gives it a SNR of around 127 dB. This was necessary in order to reduce the very small fluctuations in laser power output due to supply ripple, and their corresponding phase noise, to a minimum. This was somewhat tricky to do back then. Specifically, the current regulation control circuit has better components and additional filtering compared to common commercial HeNe laser power supplies. The PCB traces were also apparently arranged to minimize pickup of hum and noise from the nearby power transformers. A partial schematic I traced of the Model 200 HeNe laser power supply can be found in the section: Laboratory for Science Model 220 Laser Power Supply (LS-220). I still need to determine the details of the current regulation circuit (lower right in the schematic) but it's diffiult to make out because the PCB can't easily be removed from the controller case.
And speaking of details. There are some zener diodes in the power supply. If they are clear glass, room light getting in via the ventilation slots will end up modulating the power supply current, so they should be painted or replaced! Mine has the silver painted variety so I guess it's OK.
The controller has two PLLs. One is used as a frequency synthesizer to produce a highly stable reference derived from a 3.4 MHz crystal. The reference frequency may be set via 3 rotary BCD switches accessible through holes in the case. The other PLL then locks the Zeeman beat to the reference once the laser has reached operating temperature (about 1/2 hour). Thus, the reference determines the exact place on the neon gain curve where the laser will operate. (A little typewritten note on the unit I have states that the center of the Ne20 lasing line corresponds to a setting of 511.) So, maybe my laser tube is filled with isotopically pure gases.
There are 3 indicators on the front panel. The "Lock Signal" lamp on the right shows by its intensity, the approximate power to the heater transducer attached to the tube. The indicator on the left is called "Reference" and is on all the time at relatively low intensity. It is a power indicator but at a reference brightness that should be similar to the "Lock Signal" indicator when the laser is optimally stabilized. The LED at the top is called "Mode" and goes on and off during warmup as the modes cycle. When locked, it will be on at partial brightness.
A switch on the rear panel can be used to override the PLL output and select heater at max or off, to adjust the lock temperature, either because the tube is at too high or too low a temperature for stable locking, or should it lock onto a "bad" point of the Zeeman frequency response function.
The headphone jack is used not only to check on the laser during warmup and to confirm that stabilization has occurred, but also is a sensitive detector of back-reflections, which may be a destabilizing influence. Effects of optics resulting in back-reflections will be heard as transient tones in the headphones. (The headphone output may also be connected to the "Line", "CD", or "Tape" input of an audio amplifier.) Waving anything in front of the laser is audibly detectable, as are any sort of vibrations including gently touching the laser or even the table it's on, or walking across the floor. If the output is piped through loud speakers, having the volume above a very low level will result in acoustic coupling into the laser tube and a very noticeable increase in audio level as well as a change more toward non-white noise.
There is also a calibration jack which provides a beat frequency signal and DC power source for the Model 225 Zeeman Beat Frequency Range Register, whatever that is. :)
For an overview of the operating principles, which seem to track the actual implementation quite closely, see the following patents. (For the model 220, the main patent of interest will be #4,468,773.)
And a non-LFS patent for a green (543.5 nm) transverse Zeeman laser (though I don't know if it actually uses that term):
The patents also include a number of relevant references.
About two months after snagging the first LFS-220, I obtained another one, also on eBay - serial number 36. Its tube is a bit hard starting but has slightly higher power than the first - about 1.1 mW. After replacing 2 transistors and a diode which may have been bad or may have been killed when I accidentally shorted the high voltage to the Mode light bulb socket (don't ask!), it also works quite well. Internal construction appears virtually identical to SN 51.
At some point in the future, I plan to combine the beams of the two LFS-220s and record and plot the frequency of the beat signal to determine the actual stability. I'll have to complain to the LFS QC department if they don't meet published specifications!
I have also built an experimental setup using a normal barcode scanner tube in a transverse magnetic field. While turning this into a stabilized transverse Zeeman laser is unlikely to occur, I have captured some plots of it's behavior. See the section: Two Frequency HeNe Lasers Based on Zeeman Splitting.
I have acquired a scan of the operation manual for the Model 220 laser but have not gotten permission to make it public as yet. However, much of the same technical information with respect to theory of operation can be found in the brochures at Vintage Lasers and Accessories Brochures and in the patents. In fact, the block diagram in the operation manual is taken directly from Fig. 1 of Patent #4,468,773.
Among the features and attention to detail that sets the Model 200 laser apart are:
The well known commercially available HeNe lasers I'm aware of implement very few, if any of these except for tube testing, which would be essential.
At the same time, the electronic implementation (see the schematics) is a bit too simple and could benefit from a few things like an integrator in the feedback loop and bypass capacitors!
There were also a few LFS-210s which added a 10 turn pot on the back that provided adjustment of the mode position on the neon gain curve and thus the optical frequency.
Scans of an original product brochure for the LFS-200 can be found at Vintage Lasers and Accessories Brochures under "Laboratory for Science". A much more compact html version is at LFS Model 200 Ultra Stable Laser Brochure. The brochure includes a nice description of the principles of operation and of course, the specifications.
Both the laser head and controller for the LFS-200 are superficially identical to those of the LFS-220 except for the lack of a tube rotation knob on the laser head. Operation is generally similar as well, including the use of the audio headphones for locating back-reflections. However, the tube lacks the heated OC mirror and of course, the additional rotation hardware. The shutter lever on the laser head selects among NP (Non Polarized), off, and LP (Linear Polarized). (This contrary to the manual which says the latter is CP (Circular Polarized).
The interior of the laser head also differs in a number of ways. The HeNe laser tube appears to be a bit shorter than the one in the LFS-220 and the anode is at the HR-end. The mode pickoff optics and photodetectors are in a little box behind the HR mirror with their premap mounted on the side. There is an offset trimpot for the mode position accessible from under the laser head. The heaters and temperature controller are mounted on the baseplate as with the LFS-220.
The controller box is arranged roughly the same way as for the LFS-220 but the locking circuitry is substantially simpler having a total of three 8 pin DIPs: LF412 and LM358 op-amps, and an LM2905 timer, presumably for the warmup delay. But there are 6 pots for adjustment (in addition to the user accessible "volume control" servo gain knob). The HeNe laser power supply is similar to the one in the LFS-220 but several additional high voltage filter capacitors have been added on the Control PCB to zap the unsuspecting. There is also an additional pot, as well as an unidentified object in the vicinity of its control circuit, purpose unknown.
Here are some photos:
There were several versions of this power supply, at least two without the additional pot and its circuitry, and another with the 4 extra capacitors mounted on the same PCB.
The designers at Laboratory for Science appear to be more obsessed with retro-reflection or back-reflection (same thing) than at any other stabilized laser company. This is understandable considering the higher level of performance that is being achieved with the higher bandwidth servo system more sensitive to cavity perturbation. For example, while other stabilized HeNe lasers will simply use a polarizing beam splitter or two to separate the modes making sure to angle all reflective surfaces to prevent back-reflection, the LFS-200 has added the QWPs after the polarizers. The optics stack sandwich for each mode visible in the photo of the HR-end of the LFS-200, above, is something like:
Plexiglas back-plate | Amber filter | Polarizer | QWP | Plexiglas front-plate -> PD
Two passes through the QWP (out and back) result in a 90 degree rotation of the polarization axis so any reflected light is blocked by the polarizer.
There is also a significant amount of electronics in the laser head including the laser head temperature controller and photodiode amplifiers. Reverse engineering those would require ripping apart a laser head - something I'm not planning on doing any time soon.
Note that there were many engineering changes over the course of manufacturing relatively few lasers, with little if any documentation or revision numbering on the PCBs. So, don't be alarmed if there are discrepancies between the schematics and the PCBs in your laser!
The tube in the 260 was 15 inches long. It lased on three modes, giving it a more complex inter-combinational beat frequency pattern. About 50% of the power was in the central mode and a polarizer could be used to discard the other two modes since they were polarized orthogonally to it. This would get rid of the beats.
Some of the tubes were filled with single isotopic neon. Most were not. The isotopic mix did not depend on the model type though.
The tubes used in some of the later lasers were custom made by Shasta Glass (R.I.P.). These tubes had a specially designed capillary support "spider" that produced no "stick-slip" noise as the tube changed length under regulation. Other than that, there was nothing any different between the tubes used in the Laboratory for Science lasers and the tubes used in supermarket barcode scanners. We did exploit mirror defects that were typical of the type of laser tube though. Some types of sputtering artifacts can make a laser less prone to mode hopping. Also, since the mirrors were imperfect, there was a small amount of birefringence in them that we exploited as well. They were cheap tubes, but with lots of sorting and characterizing. We used about 2/3 of the tubes we bought.
The transducer was one of the fundamental, and patented, ideas that made the Laboratory for Science lasers better than any others. All the lasers used the same transducer system. One of the other patents was related to the phase locked loop electronics on the Model 220. (See the patent list above.)
A Model 220 was used by IBM in the first Atomic Force Microscope (AFM). The Model 220 could be used to measure distance changes on the order of 1/20 of an Angstrom right out of the box. Compare that to what the "competing" HP laser could do ("Position/distance resolution down to better than 10 nm") and then compare the price tags.
NASA bought a 260 for the robot that they made to test the tiles on the Space Shuttle. The robot had a YAG laser to hit the tile with a high power pulse that, due to the resulting thermal shock, would make the tile ring. The 260 was used to detect the ring modes. All this was done without contact or close proximity to the surface.
If you need a tube replacement, the right thing to do is contact Dr. Seaton. He may be able to supply you with one (even though the Lab is nominally out of business). It'll cost a bit over $1,000 installed, I would guess. There are quite a number of subtle things about tube replacement and it is best left up to someone who has done it before (unless you consider your time to be of very little value).
Why aren't there other lasers like these available today?
There are much simpler solutions available now for lasers with a coherence length of a few hundred meters. Distributed FeedBack (DFB) diode Lasers can have coherence lengths of a couple hundred meters, power outputs of many times what the Model 200 put out, cost much less than the Model 200, turn on and stabilize quicker, and don't die as easily when abused. (However, DFB lasers do not provide a self-referenced absolute frequency, as do stabilized HeNe lasers. --- Sam.)
As for the Model 220, I am not quite sure why nobody is making an equivalent system now. I suppose that there is just no significant demand for 1 mW of optical power with 20 km of coherence length. Also, there is only so much that modern manufacturing will get you in this case because there is just too much "hand tweaking" that went into these lasers.
LFS could have charged 2 or 3 times as much as they did and not lost sales. There was no place else to turn, short of much more complex and expensive iodine stabilized lasers and such, for the 220 and 260 levels of performance. The Lab almost got involved in making an iodine stabilized system. I think I recall Dr. Seaton claiming that it would have something like 0.01 Hz stability.
The RB-1 consisted of two pieces. The first RB-1 I saw had laser head SN# 1 and controller SN# 2, so at least two of these systems were built and I had mismatched pieces. However, I have photos (below) of RB-1 SN# 8 with very similar construction, which still looks like someone's science fair project. :) The thing clearly wouldn't be caught dead going out to a paying customer, though it's likely that the RB-1 or its successor eventually morphed into the Newport NL-1 (maybe "Newport Laseangle 1"?) as a result of a merger or buy-out. However, I've yet to see an actual NL-1 (or production RB-1 if there ever was such a thing).
The RB-1 laser head contains the HeNe laser tube, with wrap-around heater, a beam sampler assembly that diverted all of one polarization to a photodiode and part of the orthogonal polarization to another photodiode, and preamps for the photodiodes. The base is a 3/4 inch thick aluminum slab with a 1/8 inch aluminum cover sealed with foam rubber.
The HeNe laser tube was from Uniphase, a garden variety model with a length of about 8 inches, which is somewhat unusual, probably rated around 2 mW. A tube length of 6 or 9.5 inches being more common, at least today.
The beam sampler includes a polarizing beamsplitter cube to extract one of the mode signals and prevent it from reaching the output at all, and a separate angled plate to extract a portion of the orthogonal mode. A pair of EG&G SGD-100A photodiodes (may be similar to the Perkin Elmer FFD-100) fed LF356 op-amps.
The controller houses a linear DC power supply, standard Laser Drive HeNe laser power supply brick, feedback circuitry, and heater driver. There were controls on the front clearly not for an end-user, like 8 or 10 gain settings and a fine gain control for one of the op-amps, selection of which mode signal to pass to an output, a current meter for the heater, and so forth. People who typically use these things would have no clue of what to do with the knobs and switches. I've yet to see a user manual for the RB-1.
While the mounting of the HeNe laser tube is somewhat overkill and the beam sampler is a nice solid unit with an adequate number of adjustments, the electronic construction of both the laser head and controller are, to put it politely, a disaster. Everything is on those copper strip prototyping boards, with capacitor upon capacitor added in various places no doubt to tame noise pickup or instability. (Someone must have had stock in a capacitor company!) The designers must have had a goal of using strange and hard to find connectors wherever possible which they did for the separate cables of the photodiode signals (blue multipin) and heater drive (microphone two pin). Power for the HeNe laser tube in SN#s 1 and 2 came from a standard Alden on the controller but at the laser head had both the medium voltage BNC on top for the positive and the normal BNC on the bottom for the negative. In SN# 8, the high voltage cable is hard-wired into the laser head. Maybe the engineers were getting zapped too often. :)
Here is a composite photo of SN# 1/2:
Here are some photos of SN# 8 courtesy of eBay seller: rdr-electronics.
The 7900 is a dual mode polarization stabilized laser essentially similar to the Coherent 200, Spectra-Physics 117/A, and others. It consists of a rectangular laser head which contains the controller and HeNe laser power supply, and a separate box with DC power supplies and possibly a status indicator.
The specifications and a photo for the 7900 can be found at Mark-Tech Model 7900 Frequency Stabilized Laser. And the 7910 at Mark-Tech Model 7910 Frequency Stabilized Laser
The HeNe laser tube appears to be a Uniphase 098-2 or similar, 2 to 3 mW. It uses a Laser Drive power supply.
The one interesting difference between the 7900 and most other similar lasers is that the heater to control the length of the HeNe laser tube is painted or coated on the outside of the tube, rather than being a thin film heater or wound with wire. This should potentially have a more predictable response and thus lower frequency/phase noise once locked.
During initial warmup, the controller runs the heater at rather high power until a reference temperature is reached, and then closes the feedback loop. Since it doesn't need to wait for the temperature to reach equilibrium, this greatly reduces the lock time to under 5 minutes. This is similar to that of the HP/Agilent lasers which use custom and expensive HeNe laser tubes which have an internal heater wrapped around the bore. Most other stabilized HeNe lasers using off-the-shelf tubes take 10 to 20 minutes to lock. The tube does run rather hot though, but this is probably normal.
The stabilization feedback is implemented in 2 op-amps with some other stuff to monitor the heater temperature, do the switchover from preheat to feedback mode, and generate status signals.
The output power when locked on the sample I have is about 0.9 mW. (The spec'd minimum locked power is 0.5 mW.) So, this one appears to be basically in like-new condition even though it has a manufacturing date of 1984 making it 24 years old, with a serial number of 112. And I bet they started at SN 100! :)
In addition to connections for +/-15 VDC and ground, there are a pair of status signals from the laser head. One goes high (around +12 VDC) a few seconds after the laser locks. The other is open collector, and turns on at the same time. However, this signal will start pulsing if the lock is interrupted - for example, if the beam to the photodetectors is momentarily blocked. The pulsing continues even after lock is re-established. There is a third signal, also open collector, that is always on. I have no idea what that does.
Here is the pinout for the circular connector (J5, mating connector is AMP/Tyco part number 206434-1 with possible pin part number 66507-9). The same pin numbering is also used on the internal PCB header:
Pin Function Commecnts ----------------------------------------------------------=------------------- 1 Ground 2 Ground 3 +15 VDC Direct to HeNe laser PSU, +12 V reg elsewhere. 4 -15 VDC 5 Ground 6 Lock/Error (Blink) OC, on when locked, 2 Hz for loss of lock. 7 Lock Low initially, +12 VDC when locked. 8 Unknown OC, on all the time.
(There are two similar circular connectors on this laser but only one of them is wired to anything internally.)
The same case seems to be used for a fancier Mark-Tech laser as there are obvious tapped hole locations for mounting additional optics and other stuff. This may be for the model 7910 which appears to be an interferometer laser used in measurement/calibration systems. However, unlike those for similar applications from HP/Agilent and Zygo, it is probably NOT a two-frequency Zeeman laser but simply a model 7900 with an internal optical receiver using simple quadrature A/B fringe counting in the interferometer. See Model 7910 Interferometer System.
Here are two photos of the interior:
Here are the optical and stabilization specifications for the 05-STP-901 (from Melles Griot):
Optical Specifications ---------------------------------------------- Output Wavelength: 633 nm Output Power: 1 mW M2: <1.1 Beam Diameter (1/e2): 0.5 mm Far-Field Divergence (1/e2): 1.60 mrad Polarization: Linear, >1000 Mode: TEM00 Stabilization Characteristics - Frequency Stabilized Mode ------------------------------------------------------------- Frequency Stability (1 min/1 hr/8 hr): +/-0.5/2.0/2.0 MHz Power Stability (1 min/1 hr/8 hr): 1.0% rms Frequency Offset: +/-150 MHz Temperature Dependence: 0.5 MHz/°C Stabilization Characteristics - Intensity Stabilized Mode -------------------------------------------------------------- Frequency Stability (1 min/1 hr/8 hr): +/-3.0/5.0/5.0 MHz Power Stability (1 min/1 hr/8 hr): +/-0.1/0.2/0.2% rms Frequency Offset: +/-50 MHz Stabilization Characteristics - General --------------------------------------------------------------- Noise: 0.05% rms Lock Temperature Range: 10 °C to 30 °C Time to Lock: <30 minutes
The specifications for the SP-117A should be similar.
The unit I acquired is of relatively recent manufacture (as these things go) - 1996. The only major problem I found with it was a dead HeNe laser power supply brick - a Laser Drive unit rated 4.5 mA at 1,600 V, similar to the one in the SP-117. It appears to be a standard model except for a hand-printed label with "0.03 percent noise". So, it's either built with better filtering or is specially selected for this application from standard units. Using an external HeNe laser power supply temporarily allowed the controller to be tested. However, it appears to be much more finicky than the original SP-117 in Frequency Mode and would only stabilize with one of my SP-117-compatible laser heads. It basically ignored one that had a slightly leaky photodiode and my home-built clone, simply turning on the "Locked" LED but not actually doing anything. All three of these laser heads stabilize reliably on the much older SP-117 controller. I suspect that an adjustment of the gain of the photodiode preamps would take care of this - probably just turning it up all the way. So, perhaps I shouldn't be so hard on it. :)
Switching to Intensity Mode at first resulted in the heater simply turning on. The offset pot had to be adjusted to get the mode signals to be in the required range for the locking circuitry to operate, but then it would lock with perhaps 30 seconds required to settle down.
Switching from Frequency to Intensity Mode or back caused the Locked LED to flash and the Locked relay to chatter for a few seconds. This seems to be normal behavior, as it happens on every one of these lasers I've seen. (The Locked relay provides a set of SPDT contacts that can be used to control auxiliary equipment, though there is no external connector for it. But a cable could be wired to the PCB pads and snaked out through the ventilation slots on the bottom of the case.)
Monitoring the heater drive signal on an oscilloscope shows how sensitive this feedback scheme really is. Even playing music at a moderate level evoked a detectable response. Tapping on the concrete floor resulted in an oscillation that took a few seconds to die out.
The electronic design of the SP-117A and 05-STP-901 clearly has it roots in that of the SP-117 (no A) but in addition to the added circuitry for the intensity mode, the frequency control loop has been upgraded to a pure integrator (with a few additional op-amp stages). The PCB layout is completely new and TL084s have replaced LF347s. But much of it looks like it is unchanged from the SP-117 design. The heater drive is a crude pulse width modulator rather than linear pass transistor. Considering the care with which the PCB is laid out with separate analog and digital grounds and linear everything else, it seems strange to have this source of high level digital noise. And, in fact, the varying heater current results in either thermal or magneticly induced vibration of the tube. This can be detected in the spectrum of the laser output if one looks hard enough. For example, if it is heterodyned with another clean laser. (An upgrade is described later to eliminate this.)The input is +12 VDC from a linear power supply. A source of +9 VDC is provided by a LM317 linear regulator. A 555 timer generates the PWM clock and also the -9 VDC power via a charge pump. These power most of the analog circuitry. Timing delays are implemented using several CMOS monostables. Grrrr. :)
Also see the sections starting with: SP-117 and SP-117A Stabilized Single Frequency HeNe Lasers for more details of the electronics including complete schematics for both the SP-117 and SP-117A/05-STP-901.
For best performance, the controller should be adjusted to match a the specific laser head. (Intensity mode may not work at all if swapping heads without readjustment.) There are only three pots inside, so this isn't that complex a procedure! Remove the cover by taking out the 4 Philips-head screws on the bottom near the feet at the edge of the case.
First, power up the laser and check the 12 VDC power supply at the mainboard PCB connector. There are two pots on the power supply PCB. The one closer to the HeNe laser power supply-side is the voltage adjust. (The other one is unlikely to need adjustment.) You'll need a tiny right angle flat blade screwdriver or bent flat hairpin turn the pot. Set the voltage for 12 V +0.25/-0.0 V.
Adjustment procedure for MG-05-STP-901 and SP-117A:
Pots R9 and R10 (500K ohms) set photodiode preamp gain while R13 (50K ohms) sets balance in INTENSITY mode. All measurements should be made with respect to AGnd (TP7). An oscilloscope is desirable for the INTENSITY mode adjustments but not essential.
It's possible that an earlier PCB revision of the MG-05-STP-901 or SP-117A may have different parts designations. In particular, the SP-117 - no "A" - has other part numbers but it should be obvious which pots and test points to use.
The LOCKED LED should not be on at this point (even if it was before the adjustments were made).
The system is basically working at this point but for the final adjustments, let it remain this way for at least another hour to give everything time to warmup fully.
Use the PBS rhombus or any polarizer to compare the power in the two modes. (There should be a mark on the laser head case to indicate the plane of polarization.) If it isn't adequately equal, adjust R9 or R10 to make it so. Only a series of small adjustments should be needed, giving the system a few seconds to reach the stable position.
This adjustment can be done without electronic test equipment but it's much quicker with at least a DMM and easier with an oscilloscope:
Here is the procedure without electronic test equipment, but a laser power meter (or calibrated eyeball) and polarizer will be required:
Further adjustment is left as an exercise for the student and your mileage may vary. :)
Adjustments procedure for SP-117:
The SP-117 only has FREQUENCY mode, which is functionally identical to FREQUENCY mode of the SP-117A. There are also 3 pots. R11 and R13 (the two 500K pots) are equivalent to R9 and R10 of the SP-117A with R12 (the 50K pot) adjusting the position on the gain curve. I do not know why a similar function to this last pot isn't included in the SP-117A. Or, perhaps it is and my schematics have errors. Errors? Nah. :)
The LOCKED LED should not be on at this point (even if it was before the adjustments were made).
The system is basically working at this point but for the final adjustments, let it remain this way for at least another hour or so to give everything time to warmup fully.
There are two sources of periodic modulation of the optical frequency in the SP-117 and SP-117A:
An external ripple reducer circuit can be added to the output of the HeNe laser power supply. See the section: Reducing the Ripple and Noise in a HeNe Laser Power Supply.
SP-117A Stabilized HeNe Laser Linear Heater Drive Modification shows a simple circuit that should eliminate this source of FM. It consists of an RC filter (including the 15K ohm resistor that that feeds Q1) to convert the PWM to a linear drive signal. The same fancy transistor can be used by removing it from the controller PCB and remounting it on an isolated heat-sink. The gain pot is set so that the sensitivity is about the same as with the PWM signal.
As noted, the L-109 does NOT employ a two frequency technique with orthogonal polarization and external interferometer optics as in the Zygo or HP/Agilent lasers. Rather a Bragg cell (basically an Acousto Optic Modulator or AOM) adds a fixed frequency offset to the return beam from the remote target which adds to the Doppler shift due to target motion. The two beams are then heterodyned by the built in optical receiver shielded with copper foil that feeds the processor. It must get its power via the coax since there are no other connections. Either that, or it's simply a photodiode. The signal out of the optical receiver is equivalent to MEAS in the HP/Agilent lasers. Because the Bragg cell introduces a frequency shift for both the outgoing and return beams that add, REF is double the frequency of the RF signal used to drive the Bragg cell. One interesting twist is that the output mirror of the HeNe laser tube is also one mirror of the Michelson interferometer so that wavefront alignment of the outgoing and return beams is assured as long as the two beams are coincident on the photodiode of the optical receiver. The residual beam entering the laser tube through the mirror does not destabilize the laser because it has been shifted in optical frequency by the Bragg cell. In principle, the resolution and accuracy of this approach should be similar to that of the orthogonally polarized two-frequency lasers with external interferometer optics.
The most detailed explanation of how this all works seems to be in U.S. Patent #5,116,126: Interferometer requiring no critical component alignment. What's on the Optodyne Web site itself is rather sparse. However, some information is available on the "Downloads" page.
The laser head is much smaller than it appears on the Optodyne Web site, only about 2x2x9 inches. The "R" seems to refer to the normal 9 or 10 mm beam diameter. There is also an L-109N with a 0.5 mm beam and normal 1.7 mR beam divergence with no beam expander! And, there is a 20 mm version, probably with an external beam expander (or additional beam expander).
The DC power input is 15 VDC based on measurements of a P-108L controller, though the laser seems to be happy on as low as 10 VDC. Older versions of the L-109 laser head have a funky round LEMO-style connector with 3 mini-coaxes (1 of these appears to be unused) and 2 small female pins for power. The L-109 photos on the Optodyne Web site seem to have separate connectors for the two signals (RF and MEAS) and power.
Here are some photos:
The connector pinout is as follows (view with laser head labels on top):
RF In (Coax) O DC- * * DC+ MEAS Out (Coax) O O NC (Coax)
The first L-109R that I acquired behaved rather strangely. I was only powering the laser and its stabilization electronics as I did not have an RF source for the Bragg cell or processor to do anything with the MEAS signal, but I don't think that is the problem. It did lock to a single mode after 10 or 15 minutes. However, after that there is a 2 or 3 minute cycle whereby lock point slowly drifts part way over the gain curve, sits there for 10 or 20 seconds, at which point the green LED on the back of the laser head gradually increases in brightness. Then it moves back the way it came much more quickly and abruptly re-locks at a different location on the gain curve at which point the LED goes out, and the cycle repeats. It's almost certain that it remains locked during the slow drift, just that the lock point is moving for some unknown reason, possibly a fault in the electronics. This can't be normal behavior, though the limited optical frequency/wavelength variation may not materially affect the measurement performance.
But a second L-109R run from a P-108L controller (described below) locked to a stable constant mode position after about 15 minutes, though it did go through several false starts similar in behavior to the first laser head. I retested the first laser on the P-108L with no obvious change, so there would seem to be a real problem.
The output of the first laser when locked (or what passes for being locked!) is about 250 µW, which is probably normal as at least one half the power from the tube is lost in the beamsplitter optics forming the internal part of the Michelson interferometer. For testing without RF drive to the Bragg cell, I adjusted the alignment of the beam expander to output the un-deflected beam. Otherwise, there is very little output power, just a couple of splotches totally perhaps 40 µW! So, depending on the efficiency of the Bragg cell, the locked output could be lower than the 250 µW I measured as it will depend on what portion of the beam is actually deflected. When I tested the first L-109R laser head, I thought it was broken. But after reading the patent, have concluded that this is normal without any RF drive. I do not believe that the altered alignment is the cause of the peculiar behavior as I tested to confirm that back-reflections weren't confusing stabilization by blocking the outgoing beam and monitoring the mode sweep from the waste beam out the other end of the laser tube. There was no change. One interesting tid-bit though: This tube is a flipper for the first 5 minutes or so as it warms up, and then reverts to normal behavior.
There's one other mystery associated with the first L-109R: A skinny blue wire (like wire-wrap wire) is connected to the electronics PCB and runs the length of the laser but is not attached to anything at the other end. All the other wires are fatter. This sort of thin wire is only used elsewhere in the laser head to as cable ties. These can be seen in Optodyne L-109R HeNe Laser Head - Closeup of Interferometer Optics. The unconnected wire runs along the top of the photo and then curves down on the left. Cable ties using similar red and blue wires are also visible in the upper center and surrounding the ballast case at the upper right. It's not a ground wire as there's 1K ohms or more between it and ground, and it doesn't appear to have a counterpart in the second L-109R laser head, though I have not removed the chassis from the case to examine its entire length. At first I thought the wire had ripped out of someplace when I pulled off the rear cover, but I don't see any evidence of that. It looks like it was cut clean. Of course if it was supposed to go somewhere, that could explain the peculiar locking behavior. :)
If anyone has more technical information on Optodyne lasers, please contact me via the Sci.Electronics.Repair FAQ Email Links Page.
The Optra two-frequency HeNe lasers are not particularly unusual except for their stabilization technique, which uses a heater wound around the restricted portion of one mirror mount stem to control cavity length by changing the length of the mirror mount itself, rather than of the entire tube. The only technical references I could locate were two Optra patents:
Due to the low split frequency, long conventional tube, whimpy magnet, and PLL locking technique, I originally thought that the Optralite was a transverse Zeeman HeNe laser. Transverse Zeeman HeNe lasers normally have split frequencies in the hundreds of kHz range, not the MHz range as with most other axial Zeeman HeNe lasers. However, after finding the raw output from the HeNe laser tube to be circularly polarized and not linearly polarized, I checked the magnetic field orientation more carefully with my trusty cereal box compass and found it to be pointing along the axis of the tube, and uniform around the perimeter of the ring magnets - the definition of an axial magnetic field! That they are ring magnets and not bar magnets on either side of the tube should have been the tip-off, but I guess I wasn't paying attention! The tube in an axial Zeeman HeNe laser produces a pair of left and right circularly polarized components while the tube in a transverse Zeeman HeNe laser produces a pair of orthogonal linearly polarized components.
However, given the very limited coverage of the magnetic field of perhaps 25 percent of the gain region (the bore discharge) - and probably not being particularly uniform as well - it's not clear how the neon gain curve will be split and what will happen when more than one longitudinal mode is lasing. In addition, the near-hemispherical cavity this tube likely has will result in a tapered intra-cavity mode volume with a corresponding variation in gain. This seems to be leading up to a messy integral better left for the advanced course. :) But ignoring the mode volume issue and simplifying the situation to two parts:
These two pairs of functions could be summed resulting in a combined Zeeman-split neon gain curve. Each portion would look like a lop-sided Gaussian weighted a bit away from the center. (The gain curve diagrams shown in the patent do not represent reality.)
The Optralite uses a 7-1/8 inch (181 mm) random polarized Aerotech HeNe laser tube of conventional design, but with an additional optic glued to the HR mirror, possibly to add wedge and/or an AR-coating to minimize back-reflections and etalon effects. Whether it is specially modified in any other way is not known but all indications are that it is a standard model. And according to the second patent, above, the tube may be an Aerotech LT05R which uses natural Ne but isotopically pure 3He. The tube used in the Aerotech Syncrolase 100 is physically identical - perhaps this is more than a coincidence. A tube of this length would almost certainly result in the production of rogue modes if mated with the type of strong full length magnet found in the HP/Agilent lasers. Normally, when the laser is locked, there should be a single Zeeman-split lasing mode consisting of the two frequency components F1 and F2. Rogue modes are undesirable lasing modes that are present due to additional longitudinal modes fitting under the Zeeman-split neon gain curve. The tubes in HP/Agilent lasers have a cavity length of only around 5 inches (127 mm) to suppress rogue modes, and axial Zeeman HeNe lasers from other manufacturers use tubes with shorter cavities - some are 4 inches (102 mm) or even less. But with the Optralite, this relatively long HeNe laser tube, which has a cavity length of around 6-3/4 inches (172 mm), can be used without the risk of producing rogue modes because the two parts of the Zeeman-split neon gain curve are not spread very far apart by the small low strength magnet.
As with the HP/Agilent lasers, a quarter waveplate at the output is used to convert the circular polarization to linear polarization, with the same adjustments - rotation and tilt. But the mount is a real pain to deal with as the waveplate itself is glued to the hub of a spherical bearing that can rotate around its axis in an outer shell. This entire affair is simply clamped via a set-screw so it's almost impossible to independently adjust rotation and tilt. However, there are a pair of holes on the front side of the inner hub so it may be possible to make a special tool (like a bent paper clip!) for this purpose. Perhaps that's what they used at Optra. :) There is no half waveplate like the one present in *all* HP/Agilent lasers. It's possible that since the laser tube can be optimally oriented in its mount during final assembly and alignment, the half waveplate may be unnecessary. (This is not possible with HP/Agilent lasers since their tubes are embedded in potting compound.)
Feedback is provided by a single photodiode behind a polarizer which monitors the waste beam from the rear of the tube. A CD4046-based PLL locks the Zeeman beat to a reference oscillator using thermal control of the cavity length of the HeNe laser tube. While a PLL can be used with any laser where beat frequencies are present and correlate with lasing mode position, the only other Zeeman HeNe laser I'd seen that used a PLL was the Laboratory For Science model 220, a transverse Zeeman laser which also used a CD4046. (The LFS-260 had a PLL for stabilization as well, but it is not a Zeeman laser.)
There are two heaters for cavity length control. One covers a bit over an inch of the length of the tube near its center to maintain the overall temperature of the tube - DC and low frequency response. The other is wrapped around the OC mirror mount stem to provide high frequency (well relatively speaking!) response. The tube heater is a bit strange, at least compared to those on every other stabilized laser I've seen. It's not a low voltage thin film Kapton heater, but is covered with some sort of high temperature fabric, has a resistance of around 1K ohms, and is powered from 115 VAC via a DC-controlled solid state relay! During warmup, it is run at full voltage but once the system locks, it seems to be driven by a bang-bang-bang control loop with only three states - off, 1/2 voltage, and full voltage, swinging wildly between these even when the laser is locked and stable, and more so if the optical feedback or thermal environment is disturbed. (However, I do not really know if this is how it is supposed to work and the 1/2 voltage state may just be an illusion due to very rapid switching.) At the same time, the mirror mount stem heater has a nearly constant voltage of approximately 7.5 VDC on it when the laser is locked and stable, but which varies in a more continuous manner if the optical feedback or thermal environment is disturbed. The actual drive to the mirror mount stem heater is a filtered version of a PWFM (Pulse Width Frequency Modulation) signal - a pulse train that varies in frequency from about 50 Hz to several kHz and whose duty cycle also changes. Its derived from the PLL somehow but is clearly not directly out of the CD4046. A TIP121 Darlington power transistor configured as an emitter-follower provides up to approximately 10 V to the heater.
Interrupting the beam to the feedback photodiode will result in wild swings of both heater voltages and if done repeatedly, will eventually cause the system to lose lock and return to the warmup state for a few minutes before re-acquiring lock. Touching the tube (which affects its temperature) or blowing on it will produce a somewhat more muted response.
The Zeeman magnetic field is produced by a pair of permanent (ferrite) ring magnets each about 1/4 inch thick and not quite touching, attached to the tube and frame with globs of RTV Silicone. They are approximately centered on the discharge of the tube (the bore), although they only actually overlap perhaps 20 percent of it. The field strength is probably a few hundred gauss.
Several photos of the Optra Optralite laser can be found in the Laser Equipment Gallery (Version 3.02 or higher) under "Optra HeNe Lasers". Four interesting ones are included here:
The Optralite I have is in good physical condition and seems to operate normally - at times (more below). It locks in about 9 minutes from a cold start, with an output power of about 1.35 mW. Although this may be a bit low for a typical 7 inch HeNe laser tube when new, it is above the CDRH sticker rating of 1 mW!
There are a pair of BNC connectors on the rear panel (see photo, above). The one on the left is for the internally generated REF signal and the one on the right is for the output of the optical receiver which Optra calls SIGNAL (but HP/Agilent calls MEAS). Both outputs are simply amplified from their respective photodiodes with no clipping or conversion to digital signals. So, the amplitude of REF depends on the output power of the laser tube and the amplitude of SIGNAL depends on the strength and alignment of the return beam.
From a cold start at around 65 °F (18.3 °C), the laser goes through around 196 complete mode sweep cycles and then a partial one as the PLL kicks in after about 9 minutes. The first cycle takes about 1 second and the last about 8 seconds. However, each cycle represents a change in cavity length of only 1/2 wavelength rather than the full wavelength of a laser with orthogonal linearly polarized modes, so it's equivalent to 98 of those. But this still seems a bit high.
Interestingly, there is less than a 20 percent variation in the output power of F1 or F2 during mode sweep. This must mean that there are multiple longitudinal modes present throughout almost its entire range. Although the tube is long by axial Zeeman laser standards, it's short enough that under normal conditions with no magnetic field, the mode sweep would be very large, possibly even 100 percent.
At the same time, REF exhibits a wide frequency variation from below 60 kHz to over 310 kHz due to mode sweep. Over most of this range, REF is a nice sinusoid. But near the minimum frequency, the waveform becomes quite distorted, though the beat never disappears entirely as it does over much of the mode sweep in HP/Agilent lasers. The lock point is at 250 kHz - just below the peak. And the PLL maintains REF at precisely 250.000 kHz +/-0.5 Hz averaged over 1 second.
However, there is definitely a problem with this laser. Although it sometimes remains perfectly stable for several hours with the cover removed, it lost lock after a short while when restarted with the cover installed, seemingly unable to increase the tube temperature even when RESET. That 8 second mode sweep cycle just before lock already means that the tube is nearly as hot as it can get since the heater has been running at full power. Assuming that some sort of timing criteria is used to determine when to lock (as there doesn't appear to be any temperature sensor for the tube), perhaps the heater is not supposed to be running at full power during warmup or it should be locking earlier when the mode sweep cycle is much shorter than 8 seconds.
In addition, if disturbed in any way (or perhaps even if not), that remarkable stability is lost as though there is something trying to push it up by a few kHz. REF would go down to 250.000 kHz, then after a few seconds start creeping up until it gets forced back down to 250.000 kHz, and the cycle repeats. It might go up to 255 kHz or higher, or only to 250.1 kHz depending on how it feels. However, given that it is possible under some undefined conditions to maintain 250.000 kHz continuously apparently forever, this suggests some sort of intermittent problem in the electronics. Perhaps an op-amp is latching up. But it could even be noisy power. Or, maybe that spastic drive to the tube heater should really vary more continuously. A schematic would be *really* helpful about now. :-)
The maximum beat frequency during mode sweep would actually correspond to the condition where the optical frequency is centered on the Zeeman-spilt neon gain curve and be the most stable in terms of absolute optical freqency over the long term. So, since the laser locks at 250 kHz which is offset by about 60 kHz suggests that as the tube ages and its power declines, there will be some additional drift in optical frequency, not present with the HP/Agilent lasers that automatically keep the lock point centered.
Since the range of beat frequencies during mode sweep is rather wide, it might be possible to change the strapping of the SPG8640BN to 200 kHz, 166.67 kHz, or perhaps even 100 kHz if a different REF frequency were desired for some unfathomable reason. Hey, another programmable two-frequency laser! :) (The LFS-220 PLL has BCD switches for this purpose, but with enough resolution to actually select the optical frequency to a very high precision.)
As expected, the output from the laser consists of two linearly polarized components, F1 and F2, separated from each-other by the split frequency, and orthogonal to each-other lined up with the laser's X and Y axes. Thus, as with other similar lasers, the beat signal amplitude with an external photodiode is a maximum with a polarizer oriented at 45 degrees. The optical power of F1 and F2 is nearly equal when the laser is locked, but there is relatively little variation even during mode sweep. It is not yet known whether F1 (the lower frequency component) is horizontally or vertically oriented. There is a shutter with positions for "OPEN/NORMAL" (with polarizer at 45 degrees in front of optical receiver photodiode), "CLOSED" (output beam is blocked), and "OPEN/FDL" (with no polarizer in front of optical receiver photodiode - no idea what "FDL" means).
The green "LOCKED" LED comes on once the laser has locked (about 10 minutes from a cold start, possibly once a complete mode sweep cycle exceeds about 8 seconds). It is apparently monitoring the PWFM drive to the mirror mount stem heater as it is a pulse train with a 70 or 80 percent duty cycle and also flickers with back-reflections or when the laser is disturbed in another way such as by blowing on the tube. Pressing the "RESET" button while locked causes the laser to go through a few minutes of behavior similar to the mode sweep during warmup, at which point it then snaps back into lock. I'm not sure under what specific conditions the red "UNLOCKED" LED is supposed to come on. I forced it to light up momentarily by blowing on the tube so lock was lost and then reacquired. It is not on during warmup or RESET. (Originally, UNLOCKED never came on under any conditions, which seemed suspicious. It turned out that the LED was dead. Surely the laser hadn't been unlocked for long enough to wear out an LED!)
Only pins 1 to 9 of the "REMOTE" connector are wired to anything. Here are their functions:
Pin Name/Description Locked Condition (if applicable) ------------------------------------------------------------------------- 1 REF 250.000 kHz sinewave 2 Ground 0 V 3 PWFM, 50 Hz to several kHz ~65 Hz, 70 to 80 percent duty cycle 4 Similar to pin 3 5 RESET- To RESET button via diode (cathode) 6 Heater transistor drive +8.3 V (filtered version of pin 3) 7 Ground 0 V 8 SIGNAL 250.000 kHz + Doppler shift 9 Ground 0 V
Pins 10 to 15 are not connected to anything.
The laser may be RESET by pulling down on pin 5. While there are three signals relevant to the mirror mount stem heater (pins 3, 4, and 6), the tube heater doesn't seem to be represented at all. Surprisingly, digital status information - like whether the laser is locked - is not present on this connector, at least not in a simple 0 or 1 form. But it can be inferred from the PWFM and heater drive signals.
This laser has no electronic adjustments of any kind as there are exactly *zero* trimpots on the control PCB. But perhaps some of the component values are hand-selected at the time of manufacture based on the measured output power of the HeNe laser tube and other characteristics once the tube assembly with the Zeeman magnet has been installed. On the other hand as long as the range of beat frequencies during the mode sweep includes 250 kHz regardless what happens as the tube ages and the feedback can handle the normal decline in waste beam output power, there really shouldn't be any need for adjustments!
If anyone has additional info for the Optralite or other Optra HeNe lasers including brochures, specifications, and operation and service manuals, please contact me via the Sci.Electronics.Repair FAQ Email Links Page.
This system had a sticker price of just under $5,000 but since around the time of the acquisition of Spectra-Physics by Newport, is no longer in production. It consists of a cylindrical laser head and power supply/controller. Go to Spectra-Physics/Newport and search for "117A".
The 05-STP-901 from Melles Griot appears to be the same system as the SP-117A except for the front panel decor and color scheme as it has the same specifications, controls, indicators, and connectors - including the strange three-pin Spectra-Physics HV connector not found on any other Melles Griot lasers. In fact, the PCB inside the 05-STP-901 case has the Spectra-Physics logo on it and "Fab" and "Assy" numbers that begin with "117"! As if this is not enough, the 640 MHz mode spacing of the 05-STP-901 listed in the Melles Griot catalog is the same as the Spectra-Physics 088-2 HeNe laser tube used in the SP-117. And, Melles Griot *does* have an 05-LHR-088 tube which matches the 088 physically and has a mode spacing of 641 MHz. Coincidence? I don't think so. :) (However, inquiries to Melles Griot have not indicated any acknowledgement of this, nor the option to actually buy such a tube without the attached laser.) Thus, a new SP-117A would probably have a Melles Griot tube inside since Spectra-Physics has been out of the HeNe laser tube business for some time. See the section: Melles Griot Stabilized HeNe Lasers. However, at least some of these tubes have an AR coating on the HR mirror (likely in addition to being wedged) as well as on the OC mirror. The purpose would be to futher minimize backreflections and to get the most power from the waste beam since that's used for the mode sampling photodiode pickup. The minimization of backreflections is by far the most important for two reasons:
So, the existence of an AR coating on the HR is one way of determining that the tube you have is from or for a SP-117/A or 05-STP-901 stabilized laser, rather than a barcode scanner. Common non-stabilized HeNe lasers - even very expensive ones - do not have this. And, I've even seen what I believe to be standard Uniphase 098 tubes (no AR coating) in these lasers. Go figure. :)
A search of the Spectra-Physics Web site used to return a link to a model 117 with some confusing text about the model 117, 117A, and 117B, but that 117 isn't the same as the original SP-117 or SP-117A described below, and the page seems to have disappeared. It even mentioned Brewster windows but none of these lasers ever had Brewster windows! The SP-117B was probably an OEM version of the SP-117A, possibly superseeded by the SP-117C, covered later. There may also have been an SP-117D, similar to the SP-117/A, also for OEM applications.
The following description applies to both the SP-117 and SP-117A (and MG 05-STP-901) unless otherwise noted. The laser head I dissected was an SP-117, though I expect the newer ones to be very similar. A typical SP-117 is shown in Spectra-Physics Model 117 Stabilized HeNe Laser System. The SP-117A is in an similar package with the Frequency/Intensity mode keylock switch and Locked LED added. The PCB is a completely new layout to accomodate the added circuitry for Intensity mode but everything else is similar. Why change a good thing?
The HeNe laser head is powered from a HeNe laser power supply brick (approximately 1,700 V at 4.5 mA) via the usual strange Spectra-Physics screw-lock HV connector, with a separate cable with a DB9 connector for the photodiode signals and heater power. The only thing non-standard about the brick may be a lower p-p ripple and noise specification but there is no special external regulation of this power supply. However, for it to turn on requires that the interlock plug be present on the back of the controller, that the microswitch inside the HV connector be depressed by a plastic pin in the HV plug, and that pins 2 and 7 on the signal connector be jumpered.
The cylindrical laser head contains the tube, output optics, and beam sampling assembly. A view of the parts after disassembly is shown in Spectra-Physics Model 117 Stabilized HeNe Laser Head Components. Sampling is from the waste beam at the HR-end - simply a polarizing beamsplitter (inside the black cylinder, upper left) feeding a pair of solar cells/photodiodes (glued to the metal bracket attached to it). Since this is at the HR-end of the tube, it doesn't reduce the output power. The entire guts can be pulled out by loosening a bunch of setscrews. Disassembly to the state of affairs in the photo took about 10 minutes, all completely reversible except for cutting small blobs of black RTV silicone holding the laser tube in place in the cut out aluminum cylinder at the top of the photo..
The HeNe laser tube itself looks like it is some version of a Spectra-Physics model 88 - the same type that used to be found in barcode scanners. However, note that the version used in the SP-117/A has cathode-end output, unlike the anode-end output of the barcode scanner tube. A sample I have produces over 3 mW, so it's probably a higher power version, perhaps an 088-2. As noted, those of more recent manufacture may use the Melles Griot 05-LHR-088. I have no idea if the tube is special in any other way, like having been selected for no more than two longitudinal modes or filled with isotopically pure gases or blessed by the Laser Gods. :) There were no markings of any kind on this one and at this point, I rather think that the only special requirement is that the tube not be a "flipper" - one where the polarization state of the modes switches abruptly rather than remaining fixed. In fact, I rebuilt an SP-117 laser head with a surplus 088-2 tube and it would seem to work fine. See the section: Transplant Surgery for Two Sick Spectra-Physics Model 117 Stabilized HeNe Laser Heads.
The HeNe laser tube has the multiple layer aluminum foil covering Spectra-Physics is so fond of. There may even be more layers than normal, covering a larger portion of the tube than normal. A thin film heater (copper-colored cylinder) is wrapped much of the way around the tube on top of the aluminum foil, and glued in place. Finally, here is an application where the aluminum foil actually might make sense to distribute the heat uniformly. :) The short black cylinder on the right holds a (PBS) Polarizing BeamSplitter (with the reflected output blocked) to select one of the two orthogonally polarized modes of the laser, possibly optional since it was not present on one of the SP-117 laser heads, or an SP-117A laser head that I've seen. So, either, whoever originally had these things salvaged the PBS as the only remaining useful part before selling them, or that is an option not present on all units. On the SP-117, it's a normal PBS cube mounted such that the laser guts have to be slid out of the laser head cylinder to access the set-screw holding it in place. On the SP-117A, it's a PBS that has a rhombus cross-section - a slightly distorted cube - presumably to minimize backreflections. And on the SP-117A, the PBS is secured by two screws accessible from the front once the bezel is removed
Prior to assembly at the factory, the tube must be tested to determine the best orientation for maximum signal change of the two polarized modes since there is no adjustment for this once the tube is mounted with RTV silicone. The specific orientation is determined by slight asymmetries in the tube construction - random factors like mirror coatings and alignment - but should not change with age or use.
After an initial warmup period where the heater is run continuously, the controller enables the feedback loop which monitors the two outputs of the beam sampler and maintains cavity length using the heater typically so that the two orthogonally polarized longitudinal modes are equidistant on opposide sides of the neon gain curve (for best stability when running frequency stabilized) or where one mode is closer to the center of the gain curve (which may be desirable when running amplitude stabilized to get a bit more output power in that mode). Although I haven't measured it, there are probably around 75 complete mode cycles before locking. The switchover occurs when the period of a full mode sweep cycle exceeds about 18 seconds. (If rebuilding one of these lasers, that time may have to be adjusted due to differences in the thermal characteristics of the tube, and the thermal coupling and insulation of the heater.)
The user controls on the SP-117A consist of a switch for power and a switch to select between frequency and amplitude stabilization. There are indicators for AC power and Stabilized. (The SP-117 is physically identical but lacks the mode select switch.) After a warmup period of about 15 to 20 minutes for the laser head to reach operating temperature, the Stabilized indicator will come on and may flash for a few seconds, and after that should remain solidly on. This really indicates only that the stabilization feedback loop is active, NOT that the laser is actually stabilized and meets specs - that may require another minute or so. For the SP-117A and MG 05-STP-901, the behavior is similar in Frequency or Intensity mode. (The SP-117 has no Intensity mode.) In fact, the way they are designed, everything is identical in both modes until the Stabilized indicator comes on, then it switches to the Intensity signal for locking. If power is cycled, the delay to Stabilized is much shorter, so no actual counter delay is involved, just some circuit watching for the mode changes to slow down below that 18 second threshold. Indeed, if the photodiodes are disconnected, Stabilized will come on in under a minute even though the modes are varying wildly. Stupid electronics. :)
The internal circuitry of the controller box is relatively simple and includes some CMOS logic including several monostables (!!) for timing the warmup period, multiple op-amps and comparators, a 555 timer, voltage regulator, and switching transistor for the heater - all standard stuff. A linear power supply feeds the HeNe laser power supply and the control electronics,
Here is the pinout of the DB9 control connector as determined by my measurements. There may be errors.
Pins Function Comments ---------------------------------------------------------------- 1,6 Heater ~19 ohms, 12 V source on pin 6. 2,7 Interlock Shorted. 3 Ground May not be connected on some versions. 4,5 Photodiode 1 Anode is pin 4. 8,9 Photodiode 2 Anode is pin 8.
Although, it's very likely that any reasonably healthy SP-117/A laser head will lock on any SP-117/A controller, adjusting the controller for optimal mode signal swing will result in best stability and permit the location of the two orthogonal modes on the neon gain curve to be fine tuned. If anyone has information on the official internal adjustment procedure for the SP-117 or SP-117A controller and/or a service manual or schematics, please contact me via the Sci.Electronics.Repair FAQ Email Links Page.
There are three sheets:
The PWM ramp generator ends up on the analog page only because it is actually in the same chip as the amplifiers.
Note that older versions of the SP-117A differ slightly in their design. The most obvious one being that they totally lack the two relays found in the 05-STP-901 and appear similar to the SP-117 in other respects like the use of the ICL7660 rather than 555 for generating the clock. But I haven't examined one in enough detail to identify other differences.
There are three sheets:
Some of the circuitry of the SP-117 and SP-117A is similar and the designer of the SP-117A clearly had access to the SP-117 schematics. However, substantial portions have been totally redesigned rather than simply tacking on the amplitude stabilization and modulation input as a minor addition. The low voltage power supply including the on-board regulation, the heater driver, the HeNe laser power supply, and at least some of the logic/timing circuitry are very similar. But in the SP-117A, the negative voltage source is provided by an a 555 rather than an ICL7660, different op-amps types are used in the analog sections, and the basic design of the control loop has also changed. Of most significance in terms of performance, unlike in the SP-117A, there appears to be no pure integrator stage in the analog chain of the SP-117. Note how much simpler the control loop (analog schematics pages) is for the SP-117 compared to the SP-117A! So there will always be an offset due to the finite and relatively modest gain between the photodiodes and heater response. This means that the exact position of the locked lasing line on the neon gain curve will have a dependency on ambient temperature and initial conditions when the SP-117 is turned on or restarted. And, it will drift somewhat until thermal equilibrium is achieved, which may take considerable time after the Locked indicator comes on. Exactly this behavior has been observed with the SP-117 and wouldn't make sense had the design been similar to that of the SP-117A with a full PID control loop. These anomolies were bothering me and in fact were the original reason I decided to trace the SP-117 circuit. My confusion was your gain. :)
This laser seems to be interesting in another respect: While for the typical ordinary HeNe laser, the modes roughly follow the profile of the gain curve as they traverse it, with this tube, the mode on one side will tend to disappear and reappear on the other side of the gain curve relatively abruptly. I don't know whether this behavior is a peculiarity or a feature but it seems like it could be beneficial. ;-)
The laser is a single unit mounted on a solid baseplate (with exposed high voltage!). It is designed to install in a cabinet painted with the decorator colors of your choice. :) See Spectra-Physics Model 117 OEM Stabilized HeNe Laser Assembly. On my sample, there was a separate box with a +/-12 VDC switchmode power supply and lighted power switch as the only user control. The SP-117C has no output polarizer so that must be supplied by the user. Its operational behavior is similar to the other SP-117s, though the warmup is faster - under 10 minutes. Locking is then abrupt with no overshoot or ringing. Locking following a power interruption of a few seconds occurs in under 1 minute. When to switch from warmup to lock mode is probably detected by a complete mode cycle taking more than around 15 seconds.
The HeNe laser tube looks like the same 088 used in the other systems. It's probably from Melles Griot by its relatively thin-walled construction. The 12 VDC input HeNe laser power supply brick is hidden underneath. The PCB generally resembles the one in the SP-117A and 05-STP-901 controllers with many of the same part numbers, though there are also many differences and it has clearly been substantially redesigned. The timing is now done using 12 bit binary counters instead of multiple monostables. The majority of the discrete resistors have been replaced with resistor packs. There is also a pair of resistor packs in sockets for reasons unknown. The input is +/-12 VDC (rather than just +12 VDC), supposedly being regulated to +/-9 VDC on-board according to the test point labeling. But the resistor that sets the voltage on the sample I have has been selected to produce +/-8 VDC instead of +/-9 VDC and that works fine. There is no 555 oscillator to generate the negative voltage of the SP-117 and SP-117A controllers, so the associated PWM clock must be produced in some other way. There are pads for four large series diodes with jumpers in their place. These would be used to reduce the DC voltage to the HeNe laser power supply if more than 12 VDC were used for the positive power supply. Small MOSFETs are used to control the Enable line of the HeNe laser power supply and the Locked signal, as well as some internal signals. And, in case you're wondering, I have absolutely no intention of reverse engineering this unit the way I did the SP-117A! But I have determined most of the external connections to the 14 pin header visible in the upper left corner of the above photo based on how it is wired and the obvious PCB traces:
Pin Function -------------------------------------------------------------------------- 1 +Va - Positive analog power, +12 to +15 VDC. 2 +Va? 3 Mode control?? Input to NOR gate, pulled high. 4 Analog Ground 5 Locked Status (+Va V: unlocked, 0 V: locked, will sink 0.6 A). 6 Reset? 7 -Va - Negative analog power, -12 to -15 VDC. 8 -Va 9 Digital Ground 10 Digital Ground 11 Digital Ground 12 Digital Ground 13 NC 14 +Vd - Digital power, +12 VDC (+15 VDC with all diodes installed).
Gary Turner says that Pin 6 is connected to a front panel "Reset" switch on his controller. The manual says that you can use the Reset to clear a flashing "Lock" status, which can occur if there is a temperature/power issue.
The Locked signal originates from an IRFD210 MOSFET which can sink 0.6 A, more than enough current to drive an LED - or a bank of them. :) The +/-12 VDC for the unit I have comes from a small switchmode power supply in a separate box. The analog and digital positive voltages (+Va and +Vd) are the same. I added a Locked LED there and will install a switch if the unidentified control signal does something useful. I guess a reset function could be useful, but then neither Gary or I have ever seen Lock flashing!
I am in need of a user manual for the SP-117C including info to confirm what I have on the 14 pin header is correct and to determine the function of the unidentified control signal. (Grounding it either during operation or prior to power-on has no obvious effect.) If you have any info, please contact me via the Sci.Electronics.Repair FAQ Email Links Page.
The spec'd output power for the original SP-119 is a whopping 70 microWatts (µW) at 632.8 nm, the normal red HeNe wavelength! With only temperature control, the frequency stability was spec'd to be +/-75 Mcps/day. This is before Hertz was used for cycles per second (cps)!. With the optional Servo Option 259-002 for active cavity length control using the Lamb dip for stabilization (which is actually more sophisticated than modern mode-stabilized HeNe lasers) it goes down to +/-5 Mcps/day.
And the price loaded for the 119 laser head and 259 exciter with the 259-002 Servo Option in 1964: $5,775.00. That's about the same cost as a his and hers set of Chevy Impalas. :)
The specifications for output power and stability were subsequently upgraded to 100 µW (WOW!) and +/-1 Mcps/day. The latter IS impressive and represents a stability of +/-2 parts-per-billion (ppb), which is better than most modern stabilized HeNe lasers.
Here are the specifications for the SP-119 laser head with the SP-259B exciter from the Spectra-Physics 119 Gas Laser Operator Manual. I've tried to use the original units and terminology:
SP-119 Laser Head --------------------------------------------------------- Output Wavelength: 6,328 A Output Power, uniphase, single freq. >100 µW Beam Diameter (1/e2) at laser aperture: 1 mm Beam Divergence (1/e2), no beam expander: 10 mrad with 3 mm beam expander: <0.3 mrad with 6 mm beam expander: <0.15 mrad Long Term Frequency Stability - deviation from neon-20 emission center after warmup assuming a maximum ambient temperature change of +/-1 °C. Without servo control: +/-75 kc/day With servo control: +/-1 kc/day Warmup from Cold Start: Without servo control: 3 hours With servo control: 45 minutes Warmup from Standby: Without servo control: 30 minutes With servo control: none External Modulation: Maximum Deviation (10 to 3,000 cps): 1,200 Mcps p-p (20,000 cps): 200 Mcps p-p Sensitivity: 12 Mcps/V Servo FM Deviation: <5 Mcps p-p Servo FM Frequency: 5 Kcps Laser Head Dimensions: 8-1/2" (D) x 6-3/4" (W) x 4-1/2" (H). Laser Head Weight: Approximately 10 pounds. SP-259B Exciter ------------------------------------===----------------------- Plasma Tube Current: 4-10 mA at approximately 2,000 V. Current Regulation: 0.1% for +/-10% line or load changes. Exciter Dimensions: 12" (D) x 16-3/4" (W) x 5-1/2" (H). Exciter Weight: Approximately 25 pounds. Input Power: 115/230 VAC, 50/60 cps, 250 VA maximum.
The Lamb dip isn't something that goes with mint sauce or cheese and crackers. :) It is a depression at the center of the gain curve that may occur under the proper conditions in a laser with an inhomogeneously broadened gain medium and is the result of hole burning or depletion caused by the lasing process in a standing wave cavity. W. E. Lamb was an early laser researcher who first predicted its existence in: "Theory of an Optical Maser", Phys. Rev. 134, pp. 1429 (1964). A. Szoke and A. Javan described it in a more useful form in: "Isotope Shift and Saturation Behavior of the 1.15-µ Transition of Neon", Phys. Rev. Lett., vol 10, issue 12, pp. 521-524, June 15, 1963. (Letters are published much more quickly than full papers which is why I assume this seems to be acausal.)
In a HeNe laser, the Lamb dip is the result of hole burning or depletion in the inhomogeneously Doppler-broadened neon gain curve. It isn't present in all such lasers, but under some conditions, mostly determined by the design of the laser, it will appear as a small depression at the exact center of the neon gain curve where its peak should be, as shown in The Lamb dip in a Helium-Neon Laser. (The Lamb dip is not something that comes and goes, though the health of the laser tube and thus its gain does affect it.) In a nutshell, the explanation is as follows:
The Doppler-broadened neon gain curve reslly respresents a distribution of atomic velocities, with zero velocity being at the center. Atoms moving toward a photon traveling down the (Z) axis of the laser have a higher frequency at which stimulated emission can occur. Atoms moving away from a photon traveling down the (Z) axis of the laser have a lower frequency at which stimulated emission can occur. So, a photon traveling in the +Z direction will only be able to produce stimulated emission if an excited atom that it encounters is moving with the specific velocity needed to Doppler shift the neon gain center frequency by the appropriate amount so it equals (within the natural line width) the photon's frequency. If the photon's frequency is above the neon center, then the atom must be moving toward the photon with a velocity of, say, -V. However, the exact same conditions will be met by a photon traveling in the -Z direction and an atom traveling with a velocity of +V. And this is exactly the same offset from the center on the opposite side of the Doppler-broadened neon gain curve. Thus, the result is two depressions until the cavity tuning is such that the F-P resonance condition is at the very center and the lasing process is drawing on the zero velocity population.
In the diagram, the cavity tuning is increasing in frequency (the cavity is getting shorter) clockwise for each successive residual gain curve (1 to 5). The "Unsaturated Gain" is present when there is no cavity to enhance stimulated emission, or below the lasing threshold. The "Saturated Gain" is present while lasing. And the "Output Power" is the useful beam from the Output Coupler (OC), the partially transparent mirror. (The amplitude of the dips in the diagram are somewhat exaggerated compared to what is typical in practice, and nothing is guaranteed to be to scale!) Though the dip pairs can't be observed in the output of the laser, they would show up if the single pass gain were measured using a probe beam from a tunable laser passed down the bore of the tube. Energy is being extracted from the gain curve to produce the intra-cavity (and output) beams through stimulated emission. So, what's left will be reduced in the areas where this takes place. (And there's an entire scientific field called "Lamb dip Spectroscopy" that involves the finer points of this phenomenon.) (A more detailed explanation of the Lamb dip can be found at the end of this section.)
Under the proper conditions (more below), the Lamb dip will result in a very pronounced variation in laser output power as the cavity is tuned, and this can be used to accurately lock the laser to the center of the neon gain curve. The locking technique is actually quite simple: The cavity length - and thus the optical frequency - is periodically varied, or "dithered" by a small amount (a few MHz) via a PZT to which the HR mirror is attached. This results in a corresponding variation in laser output power based on the profile of the neon gain curve on either side of lasing mode location. The optical power is sensed via a photodiode monitoring the waste beam from the HR mirror. The lasing line can then be maintained at the exact center of the neon gain curve where the Lamb dip is located by satisfying two conditions:
(1) locks to an inflection point and (2) forces it to be a local minima.
The electronics is designed to generate a low speed correction signal to drive the PZT to maintain these conditions. In modern terminology, the circuitry to do this would be called a lock-in amplifier, phase-sensitive detector, or synchronous demodulator. Even using 1960s technology, it isn't very complex.
The SP-119 system can still be used with manual control of wavelength (Lambda, or equivalently, optical frequency), which varies the DC voltage on the PZT via a 10 turn pot on the front paenl of the SP-259. After a short 3 hour warmup :), the wavelength (and thus optical frequency) will be fairly stable as a result of the thermal regulation of the resonator. But once the approximate gain center has been found by monitoring the servo error (on the built in meter), switching to the "Lock" position should maintain the laser on the center of the neon gain curve (center of the Lamb dip) indefinitely. Should the system lose lock for any reason, even momentarily, an "Error Alarm" indicator will latch on.
A number of requirements must be satisfied to result in a pronounced Lamb dip (or any Lamb dip at all) whose center frequency is relatively independent of tube current and output power, and it's not generally observed in common commercial HeNe lasers. The most important of these are probably:
For the HeNe laser, the homogeneous linewidth is about 100 MHz compared to the 1.5 to 1.6 GHz for the full inhomogeneously Doppler-broadened gain bandwidth (FWHM) of neon.
The SP-119 uses isotopically pure He3 and Ne20 which results in a Doppler-broadened neon gain curve where zero velocity actually corresponds to the very center or peak. This is required for the symmetry condition as noted above. With mixed isotopes and a smeared out gain curve - or one with multiple peaks - the merging of the symmetric dips would not be distinct or coincide with the neon gain curve center
The SP-119 has a conventional linear Fabry-Perot (standing wave) resonator.
The SP-119 uses a (nearly) hemispherical resonator configuration with the rear (HR) mirror being planar and the front (OC) having a RoC just slightly greater than the cavity length (to guarantee resonator stability). This results in the diameter of the intra-cavity mode volume near the HR mirror being very narrow which fully saturates the gain in that region and results in a reliable Lamb dip regardless of overall gain, which changes as the tube ages.
The SP-119 resonator is just under 10 cm long corresponding to an FSR of over 1.5 GHz. So, the nearest adjacent longitudinal mode is 1.5 GHz away and well into the tail of the neon gain curve.
In fact, even very short barcode scanner HeNe laser tubes like the Melles Griot 05-LHR-007 (mirror spacing of 110 mm, 1.36 GHz FSR) show no evidence of a Lamb dip. Though these tubes typically have a long radius hemispherical resonator and satisfy most of the other requirements, they may not have isotopically pure gases.
And note that while longer tubes like the SP-088 (or Melles Griot 05-LHR-088) may produce a very distinct valley when one mode is near the center of the neon gain curve as shown in Plot of Spectra-Physics 088 HeNe Laser Tube During Warmup (Detail), this is NOT the Lamb dip, but simply a consequence of the relative amplitudes of the other modes that are oscillating. However, the SP-259 should have no trouble locking a longer laser tube to that valley. I may try that using a one-Brewster HeNe laser tube with the OC mirror on a PZT, configured to be about the same length as an 088. But, this would not result in a single frequency laser unless the cavity were somewhat shorter as there would probably be two other weak modes lasing on the tails of the neon gain curve. Since all modes have the same polarization orientation due to the Brewster window, it's not possible to suppress these unwanted adjacent modes. See the section: Using the SP-259B to Control Some Other Stabilized HeNe Laser.
More on the Lamb dip:
The following much more detailed explanation of the Lamb dip was excerpted from Professor Tony Siegman's book, LASERS, Chapter 30, page 1,199: "Hole Burning and Saturation Spectroscopy".
(Forwarded by: Confused2.)
At the beginning of the chapter, he starts with saturation of homogenous media and notes that saturation at a particular frequency within the atomic linewidth results in a reduction or shrinkage everywhere of both the real and imaginary susceptibilty curves. He emphasizes this with "a strong saturating signal even well out in the wing of the atomic transition will, if strong enough, saturate the entire transition uniformly across its line shape." Thus, hole burning cannot exist in homogenous media.
In section 30.6: "Inhomogeneous Laser Oscillation: Lambs Dips", he goes on: "In an early and widely read analysis of the gas lasers, Willis Lamb predicted, and experimenter soon confirmed an unexpected aspect of Doppler-broadened gas laser oscillation. If we tune the resonance frequency of a single oscillating cavity mode across a Doppler- broadened gas laser transition, the curve of oscillation power output versus cavity frequency shows a comparatively sharp and narrow dip in output power when the oscillation frequency coincides with the center of the Doppler broadened line." He notes that the Lamb dip only occurs in standing wave cavities and "is a consequence of saturation and hole burning effects in the Doppler-broadened line, caused by two oppositely traveling waves in the cavity."
"Physical explanation: The signal field inside a standing wave cavity can be divided into two oppositely directed traveling waves which we have referred to as +z and -z waves. Any single atom with axial velocity v thus sees two opposing traveling waves, for which it has equal and opposite Doppler shifts, even thought the two waves are at the same frequency. This leads to double whole burning effects" and thus a Lamb dip. "Consider a laser with an inhomogeneous Dopper-broadened transition oscillating in a single frequency standing-wave axial mode resonance, with the frequency w of this resonance detuned from the atomic line center by several inhomogeneous linewidths or whole widths.... The traveling +z wave component of the standing wave cavity fields will interact with and burn a hole in only those atoms in the velocity class given by v/c-w0-w/w0; while at the same time the fields in the - z traveling wave component will burn an equal and opposite hole in the symmetrically located velocity class at opposite value of v/c. Whenever the cavity frequency is well away from line center on either side. Therefore, two symmetric holes are burned, and in essence the laser is able to extract power from two separated set of atoms or velocity classes in the atomic velocity distribution8on."
"Velocity class" is jargon for a small range of velocities (never thought about why the term "class" was used, but that was the usual terminology from the beginning). As usual when considering continuous distributions, "small" is not precisely defined, but is to be taken as meaning a range small enough that all the atoms within it act more or less the same - in other words, somewhat smaller than the velocity spread that would create a Doppler frequency shift larger than the atomic linewidth of those atoms. You can start off thinking of a discrete number of velocity classes, which you sum over; then make these classes narrower and more numerous, until you're really integrating rather than summing, in which case each velocity class takes on in fact a differentially small range."
"If, however, the cavity frequency is tuned exactly to the line center, both the +z and -z waves can interact only with the v=0 velocity class in the Dopper distribution. This velocity class is therfore saturated twice as heavily as either of the separate velocity classes in the off- resonance situation because it sees two signal instead of one. But this means that the laser need only oscillate roughly half as hard to produce the same degree of saturation needed to reduce the gain to equal the cavity losses. In essence the two symmetric holes coalesce into one, and the laser power is taken from the single velocity classes v=0. In the inhomogeneously broadened single frequency laser this results in the slight, but definite dip in laser power at the line center known as the Lamb dip."
Since the SP-119 locks on the center of the gain curve using the Lamb dip rather than on one side of it, the resonator needs to be somewhat shorter than those of dual polarization mode stabilized lasers to guarantee that adjacent longitudinal modes - which have the same polarization and thus can't be separated from the middle one - are are far enough away that they don't have enough gain to oscillate within the 1.6 GHz Doppler-broadened neon gain curve. The active discharge length is 7 cm while the distance between the mirrors is 9.7 cm corresponding to an FSR of a bit over 1.5 GHz. This very short cavity length is intended to guarantee that at most, only a single mode can lase. However, as a practical matter, it's quite possible a somewhat smaller FSR would be adequate, and provide more output power in the lasing mode. In fact, with such a large FSR, there may be no lasing at all over a portion of the mode sweep cycle - where two adjacent modes are on the tails of the neon gain curve. One paper that mentions the SP-119 suggests that 10 cm RoC mirrors are used to assure a single spatial mode (TEM00) beam profile. (But I don't know if these are exactly the RoC used in the standard SP-119 laser. See "Pressure Shifts in a Stabilized Single Wavelength Helium-Neon Laser", A. L. Bloom and D. L. Wright, Proceedings of the IEEE, vol. 54, no. 10, Oct., 1966, pp. 1290-1294.) With the very short cavity, the output power varies significantly over a mode cycle and the beam may disappear entirely over a portion of it, especially with a weaker tube.
There are no real adjustments for mirror alignment as this is determined by the precision ground resonator assembly. (Though, tightening the various screws that hold the resonator sections together have some effect!) However, there are bore centering adjustments, which essentially align the bore to the mirrors. :) The HR mirror is on a PieZo Transducer (PZT) for cavity length control. The active part of the tube (the bore and Brewster windows), mirrors, and PZT, as well as a heater and sensor for thermal control, are all enclosed in a Mu-Metal cover to shield the gain region from stray magnetic fields.
I first acquired an SP-119 laser head (Model 119, SN# 3143-532) but no controller. It appears to be mostly complete though the output beam expander was missing. That was probably the only thing the previous owner considered useful after completing experiments with the laser!
The wiring is rather overly complex with 3 separate cables that run between the laser head and controller. (In fact, the fancier version that goes with the servo lambda control unit has 4 cables; this one lacks the cable for the photodetector.) The cable and connector for the external HeNe laser power supply is HUGE, a bit excessive considering that it's basically a sub-1 mW-class tube! This is pre-Alden though.
I had absolutely no doubt that this tube (call it #1) would be completely dead and up to air. However, upon removing the top cover, I was almost dazzled by the getter spot, which was huge and nearly like new. Something wasn't right here. I'd expect that with a modern hard-seal tube, but not something presumably from the 1960s. With a modern HeNe laser power supply hot-wired to the tube directly, it lit instantly with a bright stable discharge, but no sign of an output beam. See Spectra-Physics 119 HeNe Laser Tube 1 With Good Complexion. The large silver/black getter spot with just the slightest evidence of contamination around its periphery is visible near the top of the photo. The white block contains the ballast resistor normally used with the SP power supply. The cylinder on the left is the Mu-Metal and thermal cover for the tube bore, optics, and PZT.
Then I noticed a sticker that had fallen off the side of the tube:
HeNe 9:1 @ 4.0 Torr 2-14-86 Isotopes 3 & 20 Mfg. by El Don Engineering
El Don Engineering is apparently a company founded by the brother of the owner of Jodon, Inc., but now defunct as Google could find no reference to it.
So, this was a rebuilt or replacement tube manufactured in 1986, and with a decent sealing technique as there has been almost no leakage.
There are photos of the Spectra-Physics 053, the tube that was probably the original one from the SP-119 laser in the Laser Equipment Gallery under "Spectra-Physics HeNe Lasers". Tube #1 in my SP-119 laser head doesn't have quite as large a gas reservoir but is otherwise identical with the short two-Brewster bore mounted on the side far away from the gas reservoir. The main bore and Brewster windows are hidden by the Mu-Metal cover but there is little doubt that it they are similar. Note how short the bore is - the 7 cm active gain region is similar to what is in a 1 mW tube with internal mirrors, and there will be significant losses through the Brewster windows, so the output power of this tube will be lower. (As noted above, the spec'd output power is only 70 µW.)
Why wasn't it lasing? I would have expected it to be burning holes in the wall. :) Sure, alignment could be messed up and/or the optics could be dirty after 20+ years. But then I carefully looked in the ends) and at first thought there were no mirrors! The discharge was clearly visible and bright with no hint of the blue coloration or reflections that would be present with 633 nm mirrors. Someone ripped the mirrors out for other projects? Ridiculous! That didn't make any sense. Not only would it be rather substantial effort to get to the mirrors to remove them and then put everything back in place without even any missing screws, but why bother? Aren't they just ordinary red HeNe mirrors?
Then something occurred to me: That 4.0 Torr is a rather high pressure for a 632.8 nm HeNe laser and the discharge was rather bright and more orange than usual, which would be consistent with the higher pressure. Normally, it should be 2 to 3 Torr for a visible HeNe laser. And a 9:1 He:Ne ratio is also rather high - 5:1 to 7:1 would be more typical. The use of isotopically pure gases is also something that wouldn't be used routinely in common HeNe lasers, at least not back then. But all three of these would make sense if someone wanted to do experiments with an IR stabilized HeNe laser! So I dug out my IR detector card. At first I didn't see anything. But in a dark room, there was just the faintest evidence of a lasing spot. Yikes! A 20+ year old tube still lasing on a (likely) low gain IR line in a 40+ year old laser! Not only is this laser still functional, it is a most unusual specimen!
Next to determine the wavelength. There are only two HeNe IR lasing lines that are likely: 1,152 nm and 1,523 nm. I suppose 3,391 nm might also be possible but I don't think my IR card would show it. There are more than a half dozen other near-IR HeNe lasing wavelengths, but their gain is much lower and I've never heard of anyone doing anything with them except to prove they are possible. A thermal laser power meter barely registered anything, perhaps 20 µW. But the silicon photodiode-based power meter I use for testing HeNe lasers also had a barely detectable response. If the wavelength had been longer than 1,100 nm or so, a silicon photodiode would have been totally blind. So, the lasing wavelength is most likely 1,152 nm. That is also consistent with the color of the mirrors (or lack thereof). Mirrors for 1,523 nm tend to have a slightly pink or tan appearance in transmission, and green appearance in reflection. Mirrors for 3,391 nm also have rather pronounced characteristics - possibly clear for the OC but often totally opaque for the HR.
After playing with the mirror adjustments for awhile (including losing lasing and having to use an external alignment laser to get it back!), I was able to increase output power by a factor of 2 to 3, to somewhere between 50 and 75 µW. For a two-Brewster tube this short and optics that were probably last cleaned over 20 years ago on the weak IR line, that is certainly acceptable. :) I don't know what the HeNe gain is at 1,152 nm but if it's similar to the gain at 1,523 nm, a tube that produces 5 mW at 633 nm will only produce 0.5 mW at 1,523 nm. The tube in the SP-119 is at best good for 0.5 mW at 633 nm if it had internal mirrors. It will be less with the two-Brewster window tube. The SP brochure only specs 70 µW at 633 nm! So a similar output power at 1,152 nm is truly amazing.
Now the question becomes: What do I do with this laser? Retain it in its present form as a something unique in the Universe, and also probably rather useless for anything I'd want to do? Or, replace the mirrors with normal HeNe 633 nm mirrors and have another 0.5 mW laser, but one that's consistent with the original SP-119? This is the dilemma I face! ;-) One thing is certain, I won't attempt any mirror transplants until I've had a chance to examine another (probably dead) specimen of this laser to determine the required technique that would minimize exposure of the Brewster windows since cleaning them is not likely to be a pleasant experience.
I later acquired another SP-119 tube (call it #2) including heater jacket (thanks Kevin!), but no laser head OR controller. :) (At least I assume it's an SP-119 tube since I am not aware of other lasers that used one that is similar and can't imagine that there are any.) See Spectra-Physics 119 HeNe Laser Tube 2 With Good Complexion. The distance from Brewster tip to Brewster tip is about 3-5/8" (9.2 cm). The actual bore is enclosed by the black cover. I had assumed this was a heater jacket used for thermal control in the SP-119 laser, but it seems rather odd to put it only around the bore. On this sample at least, it is Epoxied in place and definitely non-removable. A pair of wires goes inside with a resistance between them of around 300 ohms, kind of high for a heater in the feedback loop, but I later found this to be used during Standby where the laser tube is turned off. It provides a power dissipation similar to that of the bore discharge, so that the time to stabilize after coming out of Standby is greatly reduced. What's interesting is that for this wimpy tube, the output power reaches its stellar value of 15 µW or so whether the tube has run or simply has been in Standby and thus at a similar temperature.
This tube (#2) had no sticker on it but the glassowrk is the same as that of the geniune Spectra-Physics 053 tube so it is probably original, or a non-El Don exact copy. (Stickers don't generally fall off of SP tubes!) It may have never been used and has a large portion of the getter spot remaining. As can be seen, it lights up nicely and the bottom photo is especially spectacular with subdued lighting. :) The operating voltage is nice and low, it starts instantly, and runs stably at a very low current - down to 3 mA or less. The discharge in the expanded tubing doesn't appear quite as orange as the other one, so it may be filled at lower pressure for the normal 633 nm (red) wavelength, but it's hard to really tell without seeing the exposed bore, and that isn't going to happen. :)
The laser head is Model 119-3683, SN# 578, and the controller is Model 259-3664, SN# 579. Overall, the system is in very good condition for equipment at least 36 years old. The interior of both the laser head and controller could pass for new, with only minor signs of wear on the exteriors, some rotted rubber grommets in various places, and decayed foam pads cushioning the tube. They must have been well stored as there is even very little dust inside.
The power supply/controller (what SP called an "exciter") for the SP-119 laser head is the SP-259. This one is labeled 259B. I'm not sure what the difference is between the "B" and straight 259 or 259A, if there is one (though the improved specifications seem to be associated with the B version). The only obvious difference is that the 259B has a three position switch for Lambda (frequency) Modulation - Off, 60 Hz, and External, while the original 259 only has a toggle for Off or On, with the External BNC. The modulation (input) bandwidth is at least 20 kHz, though the p-p optical frequency excursion does drop off from 1.2 GHz between 10 Hz and 3 kHz, but only 200 MHz at 20 kHz.
To emphasize how ancient the design of this system really is, the SP-259B uses vacuum tubes in the HeNe laser power supply. A 6GJ5 high voltage beam power tube is the current regulator, controlled by a a 12AX7 used as a differential amplifier with a 0A2 gas tube (basically a big glass 150 V zener diode) as the voltage reference. And the main power supply uses four more 0A2s. Based on the date from the SP brochure (above), the original SP-259 was available in 1964. Or at least SP starting testing the market for the SP-119 laser in 1964! However, my SP-259B has what appear to be date codes on the main electrolytic capacitors of 1973 if I'm interpreting the labeling correctly (and assuming they are original). The latest date of the Operator Manual is 1966.
The SP-259 provides the following functions:
Before applying power, I checked the ESR of most of the electrolytic capacitors and they all were reasonable. Even those orange Sprague Atoms showed very low ESR, so I don't know why one of them seemed to have been unhappy in the past. It's also not entirely clear why they need to be rated at 600 V as I only measured about 400 V on them. So, this gave me confidence to actually apply power to this thing. Only the HeNe laser tube connector is plugged in so far.
And, you're not going to believe this, but the laser works, sort of. I'm getting a maximum output power of a whopping 15 µW from this tube (call it #3). Despite the cloud of death getter, the discharge color doesn't look all that bad. I wouldn't be totally surprised if it had been regased, without replacing the getter. Just a chop, suck, and fill job, but better than nothing. I can't say it's perfect color, but certainly not dead. However, that decayed foam suggests that tube might be original. The beam (I'm being generous here!) slowly goes on and off as the very short cavity (which is not yet temperature controlled) expands, and it only lases when a single mode is near the center of the neon gain curve. The on-off behavior is probably normal due to the large FSR of the cavity, close to the width of the neon gain curve, though the low gain exaggerates the effect. I suspect that at least part of the lack of power may be due to contamination on the Brewster windows or mirrors. However, the major cause may still be an old, well used, tired tube. And even if it is partly due to contamination, this doesn't help that much - cleaning optics on this thing will be a real treat! What, contamination after 36 years? No way. :)
The power-on of the HeNe laser tube is itself interesting. Since the current regulation is via a vacuum tube, and that needs to warm up to conduct, the laser tube comes on and sputters for a few seconds, then appears to stay on dimly and gradually increases up to a normal discharge brightness. According to the Operator Manual, the current is adjustable from about 4 to 10 mA, with the normal range between 4 and 6 mA. For some reason, this one refuses to go below about 5.5 mA, and sort of doubles back. At first, I assumed it was a problem in the circuity since during the initial warmup, the tube starts out at much less than 6 mA and seems to remain on steady as the current ramps up. But, perhaps it's really flickering too fast to see. Geez, after 36 years, a bad part, no way! So, initially, I set it at 6 mA and proceeded to other checks.
Aside from the wimpy output power, the only thing I found wrong so far is that the power neon indicator lamp was burnt out, no doubt from the system being left on 24/7 for a 100 years. So I replaced that. :)
Then, figuring, "what the heck", I plugged in the other 3 cables and after actually reading the Operator Manual (what a concept?!) proceeded to go through the power up checklist, checking the meter readings for voltages, that the heater seemed to be working, and that adjusting the 10 turn Lambda pot actually changed the cavity length. All were satisfactory. So, then I switched to "Servo Null", the active stabilization mode. And, would you believe it #2, the thing actually locks, even with the very low output power! See Spectra-Physics 119 Laser Head with 259B Exciter - Locked. It's very twitchy as it warms up and won't stay locked for long because something is drifting, but that is truly amazing. However, I didn't wait the three hours as stated in the manual. Oh, and the "Lock Alarm" lamp, a GE-334, was also burnt out, so I stuffed a GE-327 into the socket (same electrical specs, slightly larger diameter, only requiring a nano-crowbar to make it fit). Eventually, that may become an LED. Why can't designers learn to run incandescent lamps at reduced voltage?!
Later, I returned to the laser tube current peculiarity where adjustment of the current pot does not result in a monotonic change in current, but has a fold-back characteristic with hysteresis. When first powered on, it could be pulled down to about 5.5 mA before abruptly jumping to 7 or 8 mA, and then going only down to 6.5 mA or so even fully counterclockwise. After being on for awhile, that minimum increases to close to 6 mA. When turning the pot clockwise, it must go past the point where the minimum would have been, and then abruptly jumps to a high current. And, if set at close to 6 mA and powered off for awhile, when powered back on, it might not "catch" and end up at 7 or 8 mA. After trying both tubes #1 and #2 (above), I am virtually certain that this anomoly is associated with the laser tube and not the power supply. (Or, at least, is the result of the I-V characteristics of laser tube #3.) There were no problems adjusting the current on those tubes from 3.5 mA to more than 9 mA with no kinks and no hysteresis. So, suspecting that this tube has problems staying lit below about 6 mA (not exactly surprising for a high mileage tube), I powered it from my test supply, and sure enough, it wouldn't stay lit below about 5.5 mA. Perhaps the power supply does funny things when a current is dialed in that's below where the tube will stay lit, either by chance or by design to prevent continuous restarts or sputtering, which can damage both laser tubes and power supplies. Adding some ballast resistance closer to the tube anode might help some as the main 70K ohm ballast is at least 6 inches away. But there is little point since (1) 6 mA is still an acceptable current and (2) the output power will be even lower at reduced current - power continues to increase to well beyond the 9 or 10 mA maximum!
Then I tackled the drift of servo settings, which resulted in the sensitivity of photodiode output declining and the set-point changing as the system warmed up. Shortly after power-on - in fact about as soon as there's a visible beam - it was possible to lock reliably with only a few µW of output power. Only after the system had been on for awhile did locking become more problematic, even though the laser output power had increased substantially. (Well to 12 or 15 µW!) In addition, the meter didn't respond in the "Servo Null" position, though that function appeared to continue to respond. I assumed that the servo unit was full of germanium transistors and it was all too possible that one or more of them was being affected by heat.
However, after poking around with an oscilloscope, the first problem was that the adjustment of the frequency and symmetry of the multivibrator that generates the dither signal wasn't behaving as expected. I replaced the ancient 2N697s (actually silicon transistors!) with 2N3904s and that helped somewhat for no really good reason, since the specs are similar to the 2N3904. But then I noticed a rogue oscillation at around 10 kHz that appeared only after the system had warmed up. This signal was present everywhere, and even showed up across perfectly healthy filter capacitors. That didn't make any sense. There is not supposed to be any legitimate 10 kHz source and this signal was coming from somewhere other than the servo unit since grounding the input to the photodiode preamp made it go away. On a hunch, I figured that perhaps the cause was plasma oscillation in the HeNe laser tube feeding back to the power supply, or even showing up in the optical signal to the photodiode. I knew that the tube was running just barely above the dropout current, which is where such things tend to happen. And, sure enough, increasing the tube current by 0.5 mA to 6.5 mA made the 10 kHz oscillation totally disappear. Adding ballast resistance near the tube might cure this as well as increasing the dropout current, but that's for the future. And 6.5 mA is still acceptable, and now the laser can be left On or in Standby indefinitely with no noticeable drift. In fact, even with the miniscule amount of laser output power, it's now possible to adjust the electronics for normal meter deflection when adjusting the Lambda pot with the servo unit in the Null position.
So, aside from the wimpy output power, there appears to be nothing wrong with the entire system.
The only other circuitry on the main PCB, mostly hidden under the 259-002 on the right, is the all solid state laser head heater controller.
The larger transformer is for all the high voltages and vacuum tube filaments, while the smaller one is for the heaters and low level servo circuits. In Standby mode, only the latter is powered.
Aside from the pots for HeNe laser tube current and Standy heater power accessible from the front panel, the only other electronic adjustments in the entire system are two trimpots visible at the bottom right corner (dither frequency and symmetry) and the one labeled "Inc" (Servo gain), a 10 or 20 turn trimpot accessible through a hole in the 259-002 cover.
This relates to the second SP-119 laser head (with tube #3) and the SP-259B exciter as described and shown above:
If anyone has another SP-119 laser head and/or controller gathering dust that they'd like to contribute to the cause, or other information in this antique laser, please contact me via the Sci.Electronics.Repair FAQ Email Links Page.
DC OUT - HeNe laser tube and AC interlock:
This is a large circular bayonet-lock connector with 7 pins:
Pin Function Comments ---------------------------------------------------------------------------- 1 Interlock Pins 1 and 2 are jumpered in cable, and are in 2 Interlock series with main power. 3 Heater Return 4 Spare? Second black pin jack in laser head, no connection. 5 Laser Tube- First black pin jack in laser head for tube cathode. WARNING: As much as -5,000 V when starting! 6 Laser Tube+ Red pin jack in laser head for tube anode, via 70K ohm ballast resistor from cable. Tube current may be adjusted from about 3.5 to 9 mA (same reading on meter) via recessed pot below "On" switch marking. 7 Heater Drive From temperature regulator for 18 ohm laser tube heater jacket.
J101 - PZT/Standby Heater:
This is a small circular screw-lock connector with 6 pins.
Pin Function Comments ---------------------------------------------------------------------------- 1 PZT Shield/Return 2 Standby Heater Bore heater on during Standby. Heater is 3 Standby Heater 300 ohms between pins 2 and 3. Adjustable from 32 to 42 VAC RMS (65 to 90 on meter) via recessed pot below "Standby" switch marking. 4 External Oven Null Measured +16 V (Meter is 20 V full scale). 5 Meter Return 6 PZT drive +10 to +215 V via 10 turn Lambda pot with Servo set to "Off" (or manual).
J106 - Thermistor:
This is a small circular screw-lock connector with 3 pins.
Pin Function Comments ------------------------------------------------------ 1 Shield 2 Thermistor 10K ohms between pins 2 and 3 at 3 Thermistor room temperature.
J201 - Photodiode:
This is a small circular screw-lock connector with 3 pins. The photodiode mounted behind the HR mmirror and cable is only present on the laser head if the 259-002 Servo Option is installed.
Pin Function ---------------------------- 1 Shield 2 Photodiode Anode 3 Photodiode Cathode
The 1-B tube I would use is the Melles Griot 05-LHB-270, which has a narrow bore and is only 222 mm in length (just under 9 inches). With an OC mirror mounted on a piezo beeper, the total cavity length would still be only about 9 inches, similar to an SP-088. A photodiode behind a small aperture (to block bore light) would be mounted behind the HR, well insulated from the anode voltage!
I have already done experiments with a similar setup as shown in One-Brewster HeNe Laser Tube with External OC Mirror on PZT and it does have a very nicely shaped output power versus mode sweep as shown in Effect of Mirror Alignment on Scanning Cavity HeNe Total Power Display. This set of photos was taken for another purpose, but they do clearly show the very distinct valley, also similar to that of the 088 tube. One uncertainly is what the response of the piezo beeper will be at the 5 kHz dither frequency of the SP-259B. However, it is also 5 to 10 times more sensitive than the SP-119 PZT, so a simple filter network may be able to compensate peculiarities in its response. At the very least, the DC sensitivity will need to be reduced.
A separate HeNe laser power supply might be required as the 05-LHB-270 requires considerably more operating and starting voltage than the SP-119 tube. (The latter is probably what would really be the problem.) To keep the internal HeNe laser power supply happy, a short tube or dummy load could be connected, or the vacuum tubes could simply be removed. :)
There should be no problems with the photodiode, but if the gain adjustment on the SP-259B servo unit doesn't have enough range (because the power in the waste beam may be higher than allowed for), a neutral density filter or other means can be added to reduce it.
Note that with a laser based on this length 1-B tube, a pure single frequency output will not likely be possible as two weak modes will probably lase on the tails of the neon gain curve. Since all the modes have the same polarization, there is no way to suppress these. However, if a one-perpendicular window (1-W) tube were used instead (very rare), then the two weak modes would have the orthogonal polarization, and a simple polarizing filter would eliminated them. The closest modes with the same polarization would be around 1.5 GHz away and would have no chance of lasing. It might even be possible to use a somewhat longer cavity and still achieve single frequency operation with this setup.
An alternative to the 1-B or 1-W tube would be to use just the glass tube from a Hewlett-Packard 5501A (without the magnet and optics). This has a relatively short cavity with an internal PZT. The 5501A tube does appear to have a mode shape with a Lamb dip, though I don't know for sure if that's the cause. However, to use a 5501A will require a HV amplifier for the PZT as it needs about 1.5 kV to go through more than two FSRs, and a beam sampler at the output of the tube since the waste beam is blocked by the PZT. And, it's a total joy to remove the glass tube from the magnet assembly! The tube I tested also had a peculiar mode flipping behavior whereby it tended to be polarized in one plane on the forward stroke of the PZT, and the orthogonal plane on the reverse stroke of the PZT, even across multiple FSRs. However, a relatively weak external magnetic field had an effect, and with care placement and orientation of a weak magnet, it could be convinced to act normally. The Lamb dip can be clearly seen in Modes of HP-5501A HeNe Laser Tube 1 With No Magnetic Field along with the mode flipping anomaly, as well as some hysteresis and non-linearity in the PZT response. The flipping quirk wouldn't matter as far as Lamb dip locking is concerned since only the output power is used, but the actual beam polarization once locked might depend on, well, the flip of a coin. :)
(Mostly from: Skywise (firstname.lastname@example.org).)
This is a Teletrac 1 mW stabilized HeNe laser with built in interferometer receiver. Going to Teletrac, Inc. redirects to Axsys Technologies, which only has information in their quarterly earnings reports referencing the sale of the company. But I found a user manual for a later model, but similar laser at Teletrac Stabilized Single Frequency Long HeNe Laser or Sam's Backup of Teletrac Stabilized Single Frequency Long HeNe Laser. (This manual appears to be for another Teletrac laser which may be similar to the ones described in the sections: Teletrac Model 150 Stabilized HeNe Laser 2 and Teletrac Model 150 Stabilized HeNe Laser 3. The general information and theory of operation should be similar though.
The electronics for the receiver are totally independent of the rest of the laser and are powered through its connector.
The HeNe laser tube itself has no markings. It's about 8 inches mirror to mirror. According to the user manual I found on-line it's manufactured by Zygo. Output is polarized. (Tubes in some other Teletrac lasers are made by Zygo but this one is Melles Griot. --- Sam.)
The output of the OC-end goes through a collimator to get a 1 cm low divergence beam. And it is LOW divergence. I once shot this thing out my window to a brick wall about 1/4 mile away, took a walk and found the beam to have barely grown, if at all.
The HR-end has what is obviously a mode detection assembly, but it's all covered in shrink tubing.
A two terminal device (probably an LM335) is glued face down onto the glass of the tube near the cathode end for temperature sensing.
There are two low wattage filament lamps under the tube for heating.
The HeNe laser power supply is a standard brick made by Power Technology, Inc.
Temperature regulation is done by two fan blades that vibrate, driven by a piezo. The vent is on the bottom of the laser so I have to make sure the 'tail' is sticking out in free space or it overheats and the fan blades really start clattering.
From a cold start the laser reaches mode lock in about 11 minutes.
The receiver electronics are dirt simple. Just 3 good op-amps (2 LM6361N and 1 LM353). Everything else is just caps, resistors, and two trim pots. The board has space for two other 16 pin ICs but the spots are empty with no labeling to infer their function. It looks like the outputs are all analog. On the board the wires going to the detectors are labeled SIN, COS, and INT. I think the SIN and COS imply quadrature output, but have no clue what the INT is. That signal goes to the chip that got really hot. The other two signals go to the other op-amps, and I'm seeing signal there on their two test points.
Here's page with 31 photos and 1 Quicktime movie: It's under the reference section of my Lasers Page but here's a direct link: Teletrac Interferometer Laser.
The HeNe laser tube is from Melles Griot, regardless of what that manual says. It may be a 05-LHR-120 or similar tube, possibly selected to for specific characteristics to optimize it for use in this application. Some older Teletrac lasers like the ones described later in this chapter did use Zygo tubes but not this one.
SIN and COS are the quadrature outputs from the optical receiver. INT is the "intensity" which is proportional to the total output and would be used to compensate for variations in optical power due to tube aging and/or interferometer alignment and losses.
The LED on the back of the laser that changes from red to green as the modes cycle during warmup and then goes out when locked is a nice touch and is present on all subsequent Teletrac (and Axsys) stabilized lasers.
I'm impressed with how simple and clever this system is, though some might describe it in another way - a kludge. :-)
There is more on the likely way this laser is used in the section: Teletrac Model 150 Stabilized HeNe Laser 3. That laser uses an external interferometer and optical receiver which are implemented in the same way.
I've since acquired an operational sample of this laser. It's definitely an interesting piece. :)
It has the same Melles Griot tube as the Teletrac laser (above). The actual model number is 05-LHR-219 with some -dash number. Its output power exceeds 4 mW (!!) but the output power of the laser is only a bit over 1 mW. Even this wimpy output power violates the safety sticker maximum rating of less than 1 mW! There is a polarizing filter at the input to the beam expanding telescope at an angle of +45 degrees looking toward the output of the laser, and a 1/4 waveplate at its output. So, the actual beam from an umodified Teletrac laser is a single mode that is circularly polarized. So this can go directly into the interferometer optics without worrying about orientation as the polarizing beam-splitter will separate it into two linearly polarized (REF and MEAS) as required. I was surprised that the polarizer was a cheap filter and not a polarizing beam-splitter cube, which would be of much higher optical quality and have lower losses, resulting in much greater output power. In fact, initially, because the the output had no obvious polarization axes due to the quarter waveplate, I thought the polarizer was simply a neutral density filter to cut down on the output power to satisfy the safety rating - and then to enable the output power to be easily readjusted upward (at great expense to the owner!) as the tube aged. :) The latter I found out quickly to be bogus since removing that polarizing filter isn't fun. Without the polarizer, it produces a beam with right and left circularly polarized modes 687 MHz apart (the longitudinal mode spacing of the 05-LHR-219 tube). Or, by removing the quarter waveplate, one or two linearly polarized modes depending on whether the polarizer is present. In the latter case, the total output power would be almost 4 mW. Thus, regardless of the original intended application, it could be set up as a nice general purpose stabilized HeNe laser.
The stabilization system uses a conventional Minco thin film heater wrapped around the tube, rather than the light bulbs and piezo fans. The control algorithm is implemented digitally with a PIC, quad digital pot chip, and some other stuff. :) A serial EEPROM/NVRAM stores the calibration information unique to each laser. Unfortunately, this basically means there is no easy way of making adjustments that may be required as the tube ages, or if it is replaced. And troubleshooting and repair of the Control PCB with no accessible (analog) signals and all its SMT components would not be fun. I really can't imagine that the possible flexibility of the digital control scheme has any functional benefits. And, in fact, digital control may not work as well as a garbage LM358 op-amp implementation. But it probably does enhance the job security of the designers!
After power-on, the controller appears to first check that the mode or modes from the polarizing beam sampler at the HR-end of the laser tube are present and of adequate power. Only then does it turn on the heater at full power (approximately 10 V across a 5 ohm resistance or about 20 W). And based on one sample that refused to do so quickly, the threshold (presumably contained in the NVRAM) must be set within 15 to 20 percent of the tube's power when new. That laser refused to turn on the heater until a couple minutes after a cold-start, and this was long enough that it gave up and flashed an error code. Then, power cycling would usually enable it to start up successfully within a few seconds after that.
Once the heater is powered, warmup is quite rapid, with locking occuring in around 5 minutes. It probably uses the resistance of the heater as a temperature sensor, switching to feedback control once it has increased enough to exceed a stored reference value.
Once locked, the short term stability is quite good, but there is a slow periodic variation in locked output power of perhaps 10 percent p-p. This settles out in several hours once the laser has reached thermal equilibrium. I suspect the cause is insufficient or lack of wedge in the HR mirror and/or lack of AR coating on the HR mirror. This results in an etalon effect, causing variations in both waste beam and main beam power, both intrinsic to the laser tube, and amplified through the feedback since the power of the waste and main beam are no longer in a fixed ratio. And the relative power of the two modes would also vary slightly.
I was hoping to repair the Control PCB rather than simply salvage the almost new tube for use in some other stabilized laser like a Coherent 200 (which appears to use the same tube, or one close enough). I suspected that the AD8304 quad digital pot chip was bad as the "red" mode input is stuck high. Of course, it could have been something else like bogus data read from the serial NVRAM. The HeNe laser power supply was also dead, and I suspect that my testing with a substitute power supply is what actually damaged the Control PCB, though I'm not sure how. However, an arc from the anode of the HeNe laser tube to the red mode photodiode could conceivably have been the cause. Really? :)
CAUTION 1: Be extremely careful around the anode area of the HeNe laser tube, especially if there is a need to remove the heat-shrink insulation. While contact with the HV probably won't be lethal to you (just a shock and the smell of burning flesh), it is very likely to jump from you to one of the conveniently located cables in the vicinity and kill the controller. It's possible that arcing to the chssis (as with a hard-to-start tube) might even do this. Almost everything on the controller PCB is surface mount which along with the PIC (Microchip PIC16C73A-20/SP) and its serial NVRAM (Xicor X24C44P), makes troubleshooting virtually impossible. I've managed to screw up the controllers on two separate lasers! :( The symptoms then seem to be that the mode LED remains red and no longer responds correctly, with a significant offset and difference in gain (or more), the heater never turns on (even though its LED says its on), and the firmware gives up after a few seconds and flashes the right-hand green LED forever. I suspect that at least part of what's really happening is that the perhaps the NVRAM has gotten erased since removing the NVRAM from its socket results in no noticeable difference in behavior. Or, possibly there's nothing wrong with the NVRAM, but the PIC is unable to communicate with it and/or the quad digital pot chip that connects to the PD inputs (among other things). So key parameters never get initialized correctly. I did reaplace the quad digital pot chip with no change in behavior, but it's quite possible I screwed up the SMT soldering! The PIC remains slive though. And with enough light into the beam sampler to get the modes to switch back and forth from red to green repetitively, the onset of it giving up can be delayed indefinitely. So it's still looking at the inputs, for whatever good that does!
CAUTION 2: When using the X-Y adjustment screws to center the output from the HeNe laser tube in the beam expander, DO NOT use a tool, only finger rotation. They are made of metal and press against the glass of the laser tube, separated from it only by the not very compliant Minco heater. Too much force WILL crack the tube. I also found this out the hard way - after realigning the mirrors on a non-lasing tube to like-new specifications. :(
I did eventually try replacing the AD8304 (quad digital pot) on a bad controller PCB with no obvious change in behavior. My surface mount rework skills are somwhat lacking, but I don't think that was the problem since at the very least, behavior before and after were identical. I later found out that the X24C44P (NVRAM) was dead (or erased). In fact, I now have two of these lasers with bad controllers, due to zapping. One controller PCB seems to be fully functional except for a bad or erased NVRAM. It works normally with a known good NVRAM except that the switchover threshold to feedback control appears to be several degrees higher then another known good controller PCB with the same PIC and NVRAM installed, possibly simply due to normal component tolerances. (The NVRAM in each controller would be programmed for its specific tube, so swapping controllers might not be an acceptable repair technique!) The other seems completely dead even with a known good PIC and NVRAM, except for turning on the red status LED and the heater!), but the mode LED never comes on and the laser never locks.
If someone has more information, a working laser whose X24C44P NVRAM they would be able to copy or a dead (or alive!) laser like this they would be willing to offer to the cause, please contact me via the Sci.Electronics.Repair FAQ Email Links Page.
An operation and service manual for a similar laser can be found at Teletrac Stabilized Single Frequency Long HeNe Laser or Sam's Backup of Teletrac Stabilized Single Frequency Long HeNe Laser. It isn't quite identical though as the output of the laser in the manual is linearly polarized at 45 degrees, while this one is circularly polarized.
With the separate temperature sensor and likely different heater resistance and thermal response, the Axsys and Teletrac controllers are not interchangeable, and fortunately for those wanting to try, the connectors differ sufficiently to make it difficult (though not impossible!) to do something bad in the process.
Naturally, since this version is emminently repairable, there would be nothing seriously wrong with the sample I acquired. It only had a smashed on/off switch and polarizer with excessive scatter! The tube is like new - instant start, stable run, locks in 15 minutes or so, and well aligned producing over 3.2 mW total out the beam expander in both circularly polarized modes since I never actually replaced the polarizing filter.
That lock time of 15 minutes is somewhat longer than for the Axsys equipvalent, probably due to the fact that the thin film heater occupies a very small portion of the Zygo tube (less than 2 inches) compared to most of its length for Melles Griot tube. However, I bet the life expectancy of the Zygo tube, typically 50,000 hours, is more than double that of the one from Melles Griot.
Why do manufacturers redesign a perfectly functional easy to manufacture low cost PCB for no obvious reason other than to make it more proprietary? This is the third example in this chapter alone (Agilent and Zygo being the others I've seen so far). There's no evidence that the digital controller has any benefits in terms of specifications since they haven't changed. It would be hard to believe that it is cheaper to manufacture or test. But there is no doubt that it is more difficult or impossible for anyone other than the original manufacture to repair or adjust!
An operation and service manual for a similar laser can be found at Teletrac Stabilized Single Frequency Long HeNe Laser or Sam's Backup of Teletrac Stabilized Single Frequency Long HeNe Laser. However this laser is probably the "Short" version since (1), it IS short, and (2) the spec'd output power has to be much lower due to the much smaller HeNe laser tube it uses.
The reason I think this was an OEM special is that the beam exits out a large hole in the side of the case via a 45 degree mirror on an adjustable mount. However, the front plate is clearly original but has no aperture, so this was almost certainly done by Teletrac, not and end user. The approximate arrangement of components is shown in Teletrac 150 Laser and Optics. From left to right: Teletrac 150 laser (version 2) with right angle output, optical receiver, linear interferometer optics, and retroreflector on rotary mount. The 4 holes in the top of the optical receiver are for the four (4) adjustment pots. The linear interferometer is basically a miniature version of the HP/Zygo units. The retroreflector is a cut-off cube-corner RTV'd into a bracket that clamps to a ball bearing shaft. See Teletrac Retroreflector on Rotary Mount for a closeup. So, the total travel would have only been a few cm.
The principle of operation for a positioning system using this laser would be similar to that of one using the HP/Agilent or Zygo lasers found elsewhere in this chapter. However, it is what's known as a "homodyne" system since it uses baseband fringes, rather than the "heterodyne" system using the difference (split) frequency of the two-frequency laser. The general approach is shown in Interferometer Using Single Frequency HeNe Laser. The linear interferometer is placed between the laser and the retroreflector on the remote "tool". The optical receiver goes between the laser and the interferometer. It has a single aperture on its input (from the laser) side and a pair of apertures on its output (from the interferometer) side, so it intercepts the return beam but passes the outgoing beam unaffected. (The only reason for both beams to pass through the optical receiver is one of practicality - the beams and spacing between the two beams is about half what it is with the HP/Agilent or Zygo systems.) But here there is only one frequency, so the measurement is based on simple fringe analysis looking at fringes in quadrature to determine position change and direction. The beam enters the linear interferometer polarized at a 45 degree angle with part (polarized vertically) being reflected via the attached "reference" retroreflector back to the laser. This is called the reference beam or REF. The rest (polarized horizontally) goes through to be bounced off of the remote "tool" retroreflector. This is called the measurement beam or MEAS. The linear interferometer combines the two return beams and passes them to the optical receiver. (REF and MEAS should not be confused with signals of the same names used in the heterodyne systems.)
The optical receiver module contains a non-polarizing beam-splitter in the path of the combined return beam feeding a pair of photodiodes. Each PD has a polarizer in front of it but one PD also has a 1/4 waveplate before the polarizer that shifts the relative phase between REF and MEAS for its PD by 90 degrees. The outputs of the PDs thus vary sinusoidally with respect to the relative phase of REF and MEAS. These are the "cos" and "sin" (quadrature) signals required to sense both position change and direction. So, in the same way that a rotary encoder creates quadrature sin and cos outputs, the optical receiver produces similar signals as a function of position (or more accurately, displacement or change in position). This raw quadrature output is what is often needed to interface to a generic machine tool's processor, which then does conversion to whatever units are required. A third photodiode labeled "Intensity" is also present which is insensitive to phase and is used to compensate for the change in laser power over time.
The optical receiver electronics consists of an LF353 (dual op-amp, but only one section is used) and a pair of LM6361s (single op-amps) with only a couple hand-fulls of discrete parts. See Teletrac 150 Optical Receiver 1 Assembly and Teletrac 150 Optical Receiver 1 Schematic. The back of one photodiode can be seen under the PCB. The PD/beam-splitter assembly is similar to the one in the Teletrac laser which included an built-in optical receiver, in section: Teletrac Stabilized HeNe Laser 1. The LM6361s are preamps for the sin and cos PDs. There are 4 externally accessible pots to adjust! Those for the cos channel are labeled "H Gain" and "H Offset" and those for the sin channel are labeled "V Gain and "V Offset". Perhaps the H and V refer to the polarization orientation of the PDs but that doesn't make a lot of sense since I would expect those to be oriented at +45 and/or -45 degrees with respect to the base. And why didn't they simply stick with sin and cos! The LF353 op-amp is the preamp for the Intensity PD. Its output is the reference voltage for the offset pots and is thus in effect multiplied by the offset pot settings to shift the DC output levels. That reference voltage also goes to the cable, along with the sin and cos (or V and H!) outputs.
The tube in my sample started sputtering shortly after power-on, but the Power Technology HeNe laser power supply brick has a current adjustment pot, so a quarter turn clockwise and presto! - no more sputtering. The tube is clearly high mileage with some unsightly brown bore crud and is perhaps somewhat lower in power than when new (around 0.6 mW or 600 uW peak out of the beam expander), but still starts instantly and locks just fine with a locked output of around 300 uW. For a single axis, the relatively low output power of 300 uW (compared to other garden-variety stabilized HeNe lasers) would be more than adequate. In fact, most HP/Agilent lasers have a minimum output power spec of 180 uW or less and they support multiple axes. But more power would be necessary with the homodyne system.
The laser came with a bunch of other components including a compact linear interferometer (a polarizing beam-splitter with attached retroreflector), a remote retroreflector on a ball-bearing mount, and an optical receiver for the return beam. (The linear interferometer is the same size as the HP/Agilent single beam interferometer, but with a slightly larger aperture.) All this may have been part of the angular positioning servo for a hard drive or CD/DVD mastering system. A basic system like this would probably be adequate for such an application due to the relatively short travel (a few cm or less).
All models use HeNe laser tubes of conventional design that are custom made by Zygo, along with an external heater for cavity length control. Their 7705 is a Zeeman-split two-frequency laser with specifications similar to those of the HP/Agilent 5517D, but uses a very short HeNe laser tube with a total length of around 4 inches. The 7701 and 7702 are dual mode polarization stabilized lasers based on 9 to 10 inch long tubes similar to those in laboratory frequency and/or intensity stabilized single frequency lasers like the Spectra-Physics 117A (as well as many common unstabilized HeNe lasers). (The 7701 used Aerotech laser tubes at some point in the past though.) These lasers utilize an Acouto-Optic Modulator (AOM) rather than Zeeman splitting to generate a second component 20 MHz away. The split or REF frequency for all Zygo lasers except the 7705 is thus 20 MHz, which is crystal controlled rather than somewhat random. :) The 7712/14/22/24 lasers use similar methods for stabilization and generation of the split frequency, but with more sophisticated techniques to minimize the effects of back-reflections by shifting the frequency of any return beams away from the lasing line. They also have better specifications for stability (probably because they are water-cooled rather than air-cooled) and produce higher power by more fully utilizing one (7712/14) or both (7722/24) lasing modes. The 7722/24 are fiber-coupled with a remotely located "delivery module".
For more information on alternatives to purchasing new Zygo lasers and critical issues in their selection and testing, see the companion document: Considerations in Evaluating Used or Rebuilt Zygo Metrology Lasers.
This scheme is more complex in terms of external optics and electronics than the Zeeman lasers, but doesn't require a special expensive tube assembly. Moreover, it allows for a larger difference frequency (and thus higher measurement speed) and precise control of its actual value and quality in terms of jitter/phase noise (which permits higher measurement precision, or so Zygo claims). Zeeman techniques are limited to a maximum of about 3 to 4 MHz easily, perhaps 8 or 9 MHz with difficulty. And the exact frequency is determined by physical characteristics of the laser tube and strength of the magnetic field, (and stray magnetic fields from other equipment), as well as the age of the tube - parameters that are hard to control precisely. (However, note, that the the exact value of the difference frequency is not critical only affecting the maximum rate of change of position that can be measured. And slow drift of the difference frequency is generally of no consequence since measurements are based on the difference between the two difference frequencies - the reference and the return beam from the item that moves.) But the 20 MHz reference frequency means that the rate of position change can be 5 times or more greater for the Zygo system than for most of the HP/Agilent systems, though at the possible expense of higher speed electronics. But the signal processing can take advantage of the precise crystal controlled reference, something not possible if it can vary 100 percent or more as with HP/Agilent systems depending on the type of laser head is used.
However, one consequence of this scheme is that the optics are more critical. While the two frequency components of a Zeeman laser are inherently precisely co-linear, this is not true with the AOM implementation. Careful optical design, fabrication, and alignment are needed to assure co-linearity for the Zygo lasers. A prism following the AOM diverts the desired horizontal and vertical components such that they appear to originate from a common point and thus will be co-linear following the output beam expander telescope. Errors can be quite evident in a laser whose alignment has shifted or when adjusting the orientation of the AOM for optimal performance. Although the horizontally and vertically polarized components won't actually shoot off in different directions, their central peak will vary in location across the overall beam profile.
For general info on the Zygo implementation, check out the following Zygo Patents:
The guts of a typical Zygo 7701 laser are shown in Interior of Zygo 7701 Stabilized HeNe Laser. A Power Technology HeNe laser power supply brick drives the Zygo laser tube which is in a metal enclosure (lower right). A resistance heater consisting of a bifilar-wound copper wire coil surrounds tube as shown in Zygo HeNe Laser Tube with Bifiler-Wound Heater. Some Zygo lasers use a thin-film heater as shown in Zygo HeNe Laser Tube with Thin-Film Heater but I do not know if any 7701s or 7702s use this approach, which would clearly be much less labor-intensive, though the parts cost would be greater. The heater is used in a conventional feedback loop to control the cavity length based on the usual mode balance of two adjacent orhtogonally polarized longitudinal modes with warmup and time-to-lock taking about 10 minutes. The output of the HeNe laser tube feeds the mode detector optics inside the dark gray thing with the photodiodes mounted on the angled circuit board. One of the two polarized modes is blocked by a filter at its output. The remaining mode, oriented at a 45 degree angle, enters the AOM, which is a birefringent crystal slab. This splits the angled beam into two parts. With no RF applied to the AOM, all of the power that exits the laser comes from the ordinary ray, which is horizontally polarized. The output power in the vertically polarized beam shifted 20 MHz in optical frequency from the horizontally polarized beam comes from the extra-ordinary ray and is proportional to the RF power applied to the AOM. This also subracts some power from the ordinary ray, so the total output power from the laser is lower with the AOM driven, by about 25 percent when the H and V components are equal.
The beams then pass through a prism, reflect off the two bounce mirrors, and then through a mystery optic - possibly a spatial filter/selector to pass only the horizontally polarized beam and the single desired 20 MHz offset vertically polarized beam. (Older versions of the 7701/2 lack this optic though its mounting holes may be present in some cases) so it's apparently not totally essential but was prehaps added to remove vestiges of the unused sideband of the vertical mode as well as hormonics resulting from the squarewave drive of the AOM.) The output is expanded/collimated to a diameter of either 3 or 6 mm depending on the specific version. The output shutter provides for the full diameter beam, blocked, or a small diameter beam for alignment.
Left-Side View of Interior of Zygo 7702 Stabilized HeNe Laser and Right-Side View of Interior of Zygo 7702 Stabilized HeNe Laser show the substantially similar Zygo 7702. The main difference between the 7701 and 7702 seems to be that the 7702 has added a fiber-optic output connector for the 20 MHz reference while the older 7701 provides only a differential ECL electrical signal. Internally, the 7702 has the AOM RF driver built into the Control PCB rather than a separate module.
The 7702s (or at least all those I've seen including the one in the photos) have a totally redesigned digital Control PCB using a microcontroller for initialization and warmup, and only switching to analog feedback for actual stabilization feedback. This is actually a rather sophisticated system, much more so than the HP/Agilent lasers, at least those with the Analog Control PCB. (And even those with a digital Control PCB have no externally accessible communication ports.) The status may be interrogated via an RS232 port, and with some firmware revisions (or maybe there's a jumper to select this), the the controller will report step-by-step what it's doing and any errors encountered. There must also be a way of setting parameters like temperature calibration and output mode balance (AOM power), but this information is very tightly controlled by Zygo.
The firmware starts out by measuring the tube heater resistance two or three times to determine its temperature and whether it's in the steady state (e.g., hasn't just been powered off as would be indicated by a changing temperature). If so, it waits a bit to gather more information. Then, it computes the number of mode slews (what everyone else calls mode cycles) to wait and turns the tube heater on full. Since a single mode slew corresponds to a known temperature change of the tube, this allows the firmware to set the temperature of the tube to a presumably optimal point. Whether all this actually results in a laser with better performance compared to HP/Agilent, or even a Coherent 200, might be debatable though. :)
There is also a fundamental shortcoming with respect to status in all of these Zygo lasers: The OK LED and presense of the REF signal only means that the system has locked successfully, NOT that both orthogonal polarized modes are present or correct. In fact, from the AOM to the output aperture, there is no monitoring of any kind! Thus, the AOM can be weak or dead, or have had its drive cable unplugged, and the OK LED will still be lit, and the REF signals will still be present. There could be an actual bug in the beam path resulting in no output power at all and the system would not know. :) The HP/Agilent lasers at least derive the REF signal from the output beam and require a minimum level to turn on READY solid. This is nearly always a sufficient indication of correct operation.
Unfortunately, while it's possible to swap controller PCBs and usually get a combination that will lock, the mode balance is typically way off (set by the AOM power) and the temperature calibration is also messed up (which may or may not really matter). But, one can't claim these meet Zygo specs like the HP/Agilent which have no adjustments except of the temperature set-point.
Another similar issue may be related to laser tube power and/or beam sampler photodiode sensitivity. To check the electrical signal from the photodiodes, there is a test point on the controller PCB (it's location varies depending on version). This is a negative voltage that ranges from 0 V (no light) to 10 V (amplifier clipping). It would appear that the useful range is from about 5 V to 9 V. If the level goes above 10 V, the laser will probably not lock, or lose lock after the tube warms up. Where a combination of laser tube/beam sampler and controller produces too high a signal level, it is possible to put a neutral density filter inside the beam sampler housing just before the photodiode ("Light Valve") PCB. An angled plate would be best to minimize back-reflections, but I simply stuffed a couple pieces of thin plastic from some packaging in there, resulting in a 15 or 20 percent reduction in the light hitting the photodiodes, and that worked fine. I don't know how low a signal it will tolerate, nor whether there is any sort of automatic gain control implemented in firmware. It may be that as long as the signal varies by some minimum percentage, the exact level doesn't matter as long as it never exceed 10 V. (The 7701s with analog Control PCBs will still lock reliably with less than 100 µW total power.)
So, a major disadvantage of the digital Control PCB for service by someone without access to Zygo documentation and software is that there is no way to adjust any parameters (there are no pots!) if they are swapped. While I have swapped Control PCBs in 7702s, it isn't possible to get the mode balance to meet Zygo spec for the 7702 with electronic adjustments (less than 5 percent difference between H and V modes, but see below), or to change the temperature calibration to accommodate slightly different tube heater resistances.
Of course, it may be that the mode balance is actually set in some other way, by the precise orientation of the laser tube, for example. Swapping AOMs doesn't seem to change it much, nor even swapping PCBs. It seems to be associated with the tube and chassis. Where the vertical mode is greater than the horizontal mode, it may be possible to add a small resistor in series with the RF to the AOM, or a voltage divider before the RF driver. Or, in either case, adjusting the alignment of the AOM very slightly may be able to balance the modes without affecting the symmetry of the beam profile or total output power noticeably, if at all.
One note about adjusting the AOM: I don't know if it's possible to get the alignment so far off such that a sideband is passed to the output as the horizontal mode rather than the vertical mode. Assuming both polarizationed modes were still present, the result would be to change the sign in the measurement of position/velocity. There is no easy way to check for such an "oops" without testing in an interferometer. Without one, this would require comparing the optical frequencies to a reference laser.
And, speaking of swapping parts, my recommendation until more info magically becomes available would be to avoid disturbing the HeNe laser tube, prism following the AOM, mystery optic, and beam expander/collimator. Very *small* adjustments can be made to the turning mirrors to peak output power. Everything else can be removed and replaced only requiring straightforward realignment at most. OK, so there isn't much else - the beam sampler assembly and AOM! :)
I had a 7701C/E that was very weak - about 40 µW out the end. It still locked reliably but this output power is well below minimum spec. So, what I did was to transplant its 7701C/e Control PCB and AOM into a 7702C/E that had a good tube but bad Control PCB. This turned out to be quite straightforward with the only tricky part being aligning the AOM to maximize the vertical polarized mode. Everything including the bolt holes are the same so the result is a like-new 7701C/E.
There were six 7702C/Es (all except ID #7) and that weak 7701C/E (ID #7) as follows:
Power ID Condition/Diagnosis Installed/Modified Status Output --------------------------------------------------------------------- 1 Functional OK 600 µW 2 Controller failed POST PCB from #4 OK 720 µW 3 Mialigned AOM, weak BS BS from #7 OK 720 µW 4 Would not lock PCB, AOM from #7 OK 650 µW 5 Weak AOM drive AOM unplugged Locks 780 µW 6 Weak AOM drive AOM unplugged Locks 720 µW 7 Very weak tube #2 PCB, #4 AOM Parts 30 µW
ID #4 was a 7702C/E, converted to a 7701C/E with organ transplants. IDs #5 and #6 aren't useful for measurement, but are decent stabilized lasers. If the AOM and mystery optic were totally removed, the output power would be between 1.5 and 2 mW in a single mode. Since ID #7 is not useful in its present condition, I may attempt to install a good Zygo tube just for kicks. IDs #1 to #4 all have excellent output power and acceptable mode balance. They appear to meet Zygo specifications, though I do not have enough information to be absolutely positive.
The following is from the operation manuals for the 7701 and 7702. Everything on the DB25F is the same:
Pin Function Description/Comments ----------------------------------------------------------------------- 1 NC 2 NC 3 NC 4 NC 5 TXD RS232 Transmit 6 RXD RS232 Receive 7 GND 8 GND 9 GND 10 ECL REF SHLD 11 NC 12 -15 VDC 13 -15 VDC 14 +15 VDC 15 NC 16 NC 17 SERVICE~ Laser is unable to lock (active low) 18 UNSTABLE~ Laser is in the process of locking (active low) 19 LTO~ ??? 20 LTO ??? 21 +15 VDC 22 ECL REF Reference (beat) frequency output 23 ECL REF~ Complement of above 24 REF SHLD 25 MAIN SHLD
To run the laser only requires the +/-15 VDC and GND connections.
Settings: 9600 baud, 1 start bit, 1 stop bit, no parity, XON/XOF, half duplex.
All commands consist of 3 decimal digits terminated by a Carriage Return (CR) character. Any other characters or improper format will return an E01 "Unrecognized Command" error.
There are 7 commands documented in the Zygo 7702 operation manual:
Command 100 - Report System Status
This command returns a code (usually 2 digits) that indicates the state of the firmware during system initialization until locked, or if an error is encountered. The valid codes are:
It's not clear from this list of states, when the "Service" LED is turned on. However, from monitoring states continuously with firmware that constantly sends status to the RS232 port, it seems that going to state 0 does not turn on the "Service" LED even though it's going to be trapped there and never re-acquire lock.
Command 101 - Report All Errors
This command reports the contents of the error log.
As errors are encountered, the firmware enters them into an error log. The possible error codes are shown below:
A "G" follows the last error sent, and the error log is cleared. Duplicate consecutive errors are recorded only once. If there are no errors, the response is only a "G".
The following RS232 port (user) errors are NOT recorded in the error log:
Command 102 - Report Laser Head Serial Number
This command returns a text string corresponding to the laser head serial printed on the nameplate. So, in essence, this is really a Control PCB serial number. :)
Command 103 - Report PCB Firmware Version
This command returns a text string corresponding to the laser head firmware revision.
Command 104 - Report Laser Tube Total Hours
This command reports the total laser tube power on hours. Presumably, if a tube is replaced, this can be reset to 0.
Command 105 - Report Heater Control Status
This command reports the heater control state as follows:
Note: the values for heater on and off are interchanged in the manual.
112 - Report Output Power
This command reports the laser head output power in microwatts (µW).
Well, sort of. :) The integer value *is* proportional to the output power but the typical values returned are closer to 1/2 the output power, and even then, tend to be low by 30 percent or more.
Case#-> 1 3 2 5 6 Code Function PCB#-> 1 3 4 5 6 Comments ------------------------------------------------------------------------------- - Output Power 580 µW 715 µW 730 µW 620 µW 680 µW Measured 100 Laser State 45 45 45 45 45 Locked 101 Error Log G G G G G No Errors 102 Serial Number G20XX G8XX G13XX G19XX G20XX PCB S/N * 103 Firmware Rev V01.17b V01.14 V01.15a V01.17b V01.17b PCB FW 104 Head PCB Hrs 11832 32947 24226 12330 11576 PCB Hours 105 Heater Status 3 3 3 3 3 Analog Ctrl 106 2518 2480 2738 2505 2718 107 77 81 68 76 67 108 AOM RF Power? 176 174 193 72 77 IDs 5,6 Low 109 1.00 1.00 1.00 1.00 1.00 110 0 0 0 0 0 111 1.00 1.00 1.00 1.00 1.00 112 Output Power 221 219 204 190 212 -1 Warmup 113 Output Power 221 219 204 190 212 Always 114 128 128 128 128 128 115 2398 2316 2371 1847 2189 116 1.00 1.00 1.00 1.00 1.00 117 127.00 126.36 126.83 122.03 127.59 -1 Warmup 118 127 127 132 129 130 255 Warmup 119 127 125 122 125 124 0 Warmup 120 127 126 128 127 127 255,0 Warmup 121 22.60 19.50 22.10 23.20 21.80 " 12.29 12.15 12.11 12.15 12.16 122 -1 -1 -1 -1 -1 123 255 145 140 255 255 " 54 54 56 54 53 " 249 249 250 248 247 124 E03 E03 E03 E03 E03 Needs Value 125 11.10 11.10 11.10 11.10 11.10 " 17 19 20 22 20 126 198 204 200 205 204 127 FW Ckecksum? 002DD722 002D0C4A 002D8DD0 002DD722 002DD722 002DD722 002D0C4A 002D8DD0 002DD722 002DD722 128 180 180 180 180 180 129 -1 -1 -1 -1 -1 130 116 113 118 116 116 131 3.16 2.67 3.16 3.14 3.12 " 68.95 58.24 68.95 68.52 68.10 132 8.58 6.97 8.50 8.23 8.35 0.00 Warmup 133 10.00 8.00 9.00 12.00 10.00 0.00 Warmup 137 500 E01 500 500 500 144 EEprom Locked E01 E01 Unlocked Locked How unlock?
Note Code 124: It returns E03, which means that a parameter is missing. I tried a select number of possible values. It seems to want a 4 digit number so as not to generate an error. 4 digit numbers returned mostly a 0, but sometimes some other 1 or 2 Hex number. But never anything particularly enlightening. Also note that one controller has its EEPROM unlocked, whatever that means.
One firmware revision, V01.15a, produces a running commentary (see the next section) on exactly what it is doing, and counting off mode slews. I didn't see anything different on the PCB with this firmware compared with the others, including no obvious jumper settings, so I don't know if it's simply a feature (or bug) associated with V01.15a. I did try changing jumpers with no positive results. Removing the jumpers that were already present (1 at a time) only resulted in the controller not responding to the RS232 port at all, and probably not doing anything else either. Then there was that one that did really bad things. (More below.)
Here is another older laser, probably one of the first with the Digital Control PCB:
Code Function PCB#-> 7 Comments ------------------------------------------------------------------------------- - Output Power 610 µW Measured 100 Laser State 45 Locked 101 Error Log G No Errors 102 Serial Number G17X S/N * 103 Firmware Rev V01.10 PCB FW 104 Head PCB Hrs 51775 PCB Hours 105 Heater Status 3 Analog Ctrl 106 2708 107 70 108 AOM RF Power? 190 109 3.96 110 0 111 1.00 112 Output Power 630 -1 Warmup 113 Output Power 630 Always 114 128 115 2548 116 1.00 117 127.00 -1 Warmup 118 128 255 Warmup 119 124 0 Warmup 120 127 255,0 Warmup 121 21.80 " 12.35 122 -1 123 89 " 48 " 239 124 E03 Needs Value 125 11.10 " 23 126 15.30 " 206 127 FW Ckecksum? 002A9A5B 002A9A5B(All inquiries from 128 through 150 return E01 for this early firmware.)
If that "Head PCB Hrs" of 51775 applies to the original tube as well, that would be impressive as the tube behaves nearly like new.
CAUTION: Don't try installing jumpers at random on Zygo Digital Control PCBs. There is a set of posts that *looks* like a location for a 2 pin jumper, but is actually a pair of test points for power (+5 VDC and +15 VDC) with no current limiting resistors! Guess what happens if you install a jumper there? :( :) Don't ask how I found out! At least it was a unit that had other problems and has now been put out of its misery, but it still hurts. I wonder how may Control PCBs have been ruined by Zygo field service techs accidentally installing a jumper in the wrong place! Stupid PCB layout. Or perhaps intentional? ;-)
Power on from cold start
Zygo Laser Head Starting up. Firmware Version V01.15a Entering Preheat Startup Entering Preheat Wait for A/D Entering Preheat First Read Coil Res #1 Counts = 42 Ohms = 11.6133 Temp = 11.5798 Entering Preheat First Delay Entering Preheat Second Read Coil Res #1 Counts = 43 Ohms = 11.6367 Temp = 12.0741 Entering Preheat Second Delay Entering Preheat Third Read Coil Res #1 Counts = 44 Ohms = 11.6600 Temp = 12.5683 Temperature is stable, Heater ON, Tracking. Slews = 126 Entering Preheat Track Wave Target 125 Elapsed Time 32.00 Target 124 Elapsed Time 34.00 Target 123 Elapsed Time 36.00 Target 122 Elapsed Time 37.00 Target 121 Elapsed Time 39.00 Target 120 Elapsed Time 41.00 Target 119 Elapsed Time 42.00 Target 118 Elapsed Time 44.00 Target 117 Elapsed Time 46.00 Target 116 Elapsed Time 48.00 Target 115 Elapsed Time 49.00 Target 114 Elapsed Time 51.00 Target 113 Elapsed Time 53.00 Target 112 Elapsed Time 55.00 Target 111 Elapsed Time 57.00 Target 110 Elapsed Time 58.00 Target 109 Elapsed Time 60.00 Target 108 Elapsed Time 62.00 Target 107 Elapsed Time 64.00 Target 106 Elapsed Time 66.00 Target 105 Elapsed Time 68.00 Target 104 Elapsed Time 70.00 Target 103 Elapsed Time 72.00 Target 102 Elapsed Time 74.00 Target 101 Elapsed Time 76.00 Target 100 Elapsed Time 78.00 Target 99 Elapsed Time 80.00 Target 98 Elapsed Time 82.00 Target 97 Elapsed Time 84.00 Target 96 Elapsed Time 87.00 Target 95 Elapsed Time 89.00 Target 94 Elapsed Time 91.00 Target 93 Elapsed Time 93.00 Target 92 Elapsed Time 95.00 Target 91 Elapsed Time 98.00 Target 90 Elapsed Time 100.00 Target 89 Elapsed Time 102.00 Target 88 Elapsed Time 105.00 Target 87 Elapsed Time 107.00 Target 86 Elapsed Time 109.00 Target 85 Elapsed Time 112.00 Target 84 Elapsed Time 114.00 Target 83 Elapsed Time 117.00 Target 82 Elapsed Time 119.00 Target 81 Elapsed Time 122.00 Target 80 Elapsed Time 124.00 Target 79 Elapsed Time 127.00 Target 78 Elapsed Time 129.00 Target 77 Elapsed Time 132.00 Target 76 Elapsed Time 135.00 Target 75 Elapsed Time 137.00 Target 74 Elapsed Time 140.00 Target 73 Elapsed Time 143.00 Target 72 Elapsed Time 146.00 Target 71 Elapsed Time 148.00 Target 70 Elapsed Time 151.00 Target 69 Elapsed Time 154.00 Target 68 Elapsed Time 157.00 Target 67 Elapsed Time 160.00 Target 66 Elapsed Time 163.00 Target 65 Elapsed Time 166.00 Target 64 Elapsed Time 169.00 Target 63 Elapsed Time 172.00 Target 62 Elapsed Time 175.00 Target 61 Elapsed Time 179.00 Target 60 Elapsed Time 182.00 Target 59 Elapsed Time 185.00 Target 58 Elapsed Time 188.00 Target 57 Elapsed Time 192.00 Target 56 Elapsed Time 195.00 Target 55 Elapsed Time 199.00 Target 54 Elapsed Time 202.00 Target 53 Elapsed Time 206.00 Target 52 Elapsed Time 209.00 Target 51 Elapsed Time 213.00 Target 50 Elapsed Time 217.00 Target 49 Elapsed Time 220.00 Target 48 Elapsed Time 224.00 Target 47 Elapsed Time 228.00 Target 46 Elapsed Time 232.00 Target 45 Elapsed Time 236.00 Target 44 Elapsed Time 240.00 Target 43 Elapsed Time 244.00 Target 42 Elapsed Time 248.00 Target 41 Elapsed Time 252.00 Target 40 Elapsed Time 257.00 Target 39 Elapsed Time 261.00 Target 38 Elapsed Time 265.00 Target 37 Elapsed Time 270.00 Target 36 Elapsed Time 275.00 Target 35 Elapsed Time 279.00 Target 34 Elapsed Time 284.00 Target 33 Elapsed Time 289.00 Target 32 Elapsed Time 294.00 Target 31 Elapsed Time 299.00 Target 30 Elapsed Time 304.00 Target 29 Elapsed Time 309.00 Target 28 Elapsed Time 314.00 Target 27 Elapsed Time 319.00 Target 26 Elapsed Time 325.00 Target 25 Elapsed Time 330.00 Target 24 Elapsed Time 335.00 Target 23 Elapsed Time 341.00 Target 22 Elapsed Time 347.00 Target 21 Elapsed Time 353.00 Target 20 Elapsed Time 358.00 Target 19 Elapsed Time 365.00 Target 18 Elapsed Time 371.00 Target 17 Elapsed Time 377.00 Target 16 Elapsed Time 383.00 Target 15 Elapsed Time 390.00 Target 14 Elapsed Time 396.00 Target 13 Elapsed Time 403.00 Target 12 Elapsed Time 410.00 Target 11 Elapsed Time 417.00 Target 10 Elapsed Time 425.00 Target 9 Elapsed Time 432.00 Target 8 Elapsed Time 440.00 Target 7 Elapsed Time 447.00 Target 6 Elapsed Time 455.00 Target 5 Elapsed Time 464.00 Target 4 Elapsed Time 472.00 Target 3 Elapsed Time 481.00 Target 2 Elapsed Time 490.00 Target 1 Elapsed Time 499.00 Target 0 Elapsed Time 508.00 Temperature is SET, Heater ANALOG, Monitoring. Unit is STABLE. Heater Temp = 116.85 Lock time(mins) = 8.47 Time between Mode Slews 1 and 0 9.00 Entering Preheat Monitor Wait Entering Preheat Monitor Wave 1/4 HOUR. ET 0l 1/4 HOUR. ET 0l
Laser powered off for about 1 minute and then powered on
Zygo Laser Head Starting up. Firmware Version V01.15a Entering Preheat Startup Entering Preheat Wait for A/D Entering Preheat First Read Coil Res #1 Counts = 199 Ohms = 15.2767 Temp = 89.1750 Entering Preheat First Delay Entering Preheat Second Read Coil Res #1 Counts = 194 Ohms = 15.1600 Temp = 86.7038 Entering Preheat Second Delay Entering Preheat Third Read Coil Res #1 Counts = 190 Ohms = 15.0667 Temp = 84.7269 Temperature changing, wait for 2 minutes. Entering Preheat Wait Long Coil Res #1 Counts = 170 Ohms = 14.6000 Temp = 74.8421 Temperature is stable, Heater ON, Tracking. Slews = 51 Entering Preheat Track Wave Target 50 Elapsed Time 154.00 Target 49 Elapsed Time 158.00 Target 48 Elapsed Time 162.00 Target 47 Elapsed Time 165.00 Target 46 Elapsed Time 169.00 Target 45 Elapsed Time 173.00 Target 44 Elapsed Time 177.00 Target 43 Elapsed Time 182.00 Target 42 Elapsed Time 186.00 Target 41 Elapsed Time 190.00 Target 40 Elapsed Time 195.00 Target 39 Elapsed Time 199.00 Target 38 Elapsed Time 204.00 Target 37 Elapsed Time 209.00 Target 36 Elapsed Time 213.00 Target 35 Elapsed Time 218.00 Target 34 Elapsed Time 223.00 Target 33 Elapsed Time 228.00 Target 32 Elapsed Time 233.00 Target 31 Elapsed Time 239.00 Target 30 Elapsed Time 244.00 Target 29 Elapsed Time 250.00 Target 28 Elapsed Time 255.00 Target 27 Elapsed Time 261.00 Target 26 Elapsed Time 267.00 Target 25 Elapsed Time 274.00 Target 24 Elapsed Time 280.00 Target 23 Elapsed Time 286.00 Target 22 Elapsed Time 293.00 Target 21 Elapsed Time 300.00 Target 20 Elapsed Time 307.00 Target 19 Elapsed Time 314.00 Target 18 Elapsed Time 322.00 Target 17 Elapsed Time 330.00 Target 16 Elapsed Time 337.00 Target 15 Elapsed Time 346.00 Target 14 Elapsed Time 354.00 Target 13 Elapsed Time 363.00 Target 12 Elapsed Time 371.00 Target 11 Elapsed Time 381.00 Target 10 Elapsed Time 390.00 Target 9 Elapsed Time 400.00 Target 8 Elapsed Time 410.00 Target 7 Elapsed Time 420.00 Target 6 Elapsed Time 431.00 Target 5 Elapsed Time 442.00 Target 4 Elapsed Time 454.00 Target 3 Elapsed Time 466.00 Target 2 Elapsed Time 479.00 Target 1 Elapsed Time 492.00 Target 0 Elapsed Time 505.00 Temperature is SET, Heater ANALOG, Monitoring. Unit is STABLE. Heater Temp = 116.85 Lock time(mins) = 8.42 Time between Mode Slews 1 and 0 13.00 Entering Preheat Monitor Wait Entering Preheat Monitor Wave 1/4 HOUR. ET 0l 1/4 HOUR. ET 0l 1/4 HOUR. ET 0l
The 7712/14 is in a much fancier case than the 7702 with ball joints for mounting - change every 50K miles. :) It may be better environmentally sealed than the 7702. The DB25 connector, the cable power and signal wiring is compatible with that of the 7701/02. I have not seen the inside of a 7712/14 as yet though. It has the same fiber connector for the REF signal as on the 7702, as well as additional fiber connectors labeled "Sync", purpose unknown. And there are the water connections which appear to be 1/8" NPT female fittings. A montage photo can be found in Zygo 7712 HeNe Stabilized Laser.
Code Function Value Comments --------------------------------------------- --- Output Power 1.3 mW Measured 100 Laser State 50 Locked 101 Error Log G No Errors 102 Serial Number XXXXX Laser S/N * 103 Firmware Rev V01.7b 104 Laser Hrs 48740 105 Heater Status 3 Analog Ctrl 106 0 107 7 " 0.13 109 E01 110 0 111 E01 112 E01 113 47 " 1.23 114 5 " 2.504 " 2.315 115 E01 116 E01 117 120 118 134 119 1 " 2 " 4 120 130 121 E01 122 -1 123 255 " 54 " 249 124 E03 125 11.10 " 17 126 198 127 FW Ckecksum? 0046D024 " 0046D024 128 180 129 50.61 130 55 131 1.31 " 8.29 132 4.63 133 E01 134 570 135 40 " 0.75 136 23 " 1.60 " 1.95 137 500 138 1800 139 128 " 2.409 140 Zygo P/N 8070-0159-02 141 413 142 28 " 0.53 143 100 " 4.292 " 0.527
Note that the "Laser State", Code 100, is slightly different for the 7712 compared to the 7702. On the 7702, it is 45 for the locked condition. On the 7712, it is 50. The intermediate codes also differ.
If the drive is strong, then either the AOM crystal (or the electrical connection to it) is bad (this is unlikely), or it is misaligned. Loosen the two hex screws holding the AOM in place and carefully move it back and forth and adjust its orientation while monitoring the vertically polarized output (or preferably, both outputs). If there is a position where the H and V output power can be made approximately equal, someone before you may have been inside this laser! The alignment generally won't change on its own.
These are among the hardest faults to track down, and may have no solution other than a replacement tube or laser. They tend to occur on high-mileage tubes but not aways. And, the output power and all other characteristics may be perfectly acceptable.
Now to elaborate further on one set of behaviors that seems most common:
As noted on some lasers, the output power will drop (probably to zero) for a few milliseconds and then recover, possibly resulting in a mode flip. If there is no mode flip, then the output power will return to exactly where it was before the dropout. If there is a mode flip, the stabilization feedback will result in a slow (over a few seconds) reduction in output power and then recovery as it relocks the modes.
These events may not cause a change in the the Status LEDs on the back of the laser or even be detected and flagged as errors by the firmware, but they should be caught by the Zygo data processor. In principle, any measurement/motion that was in progress could be aborted and redone without position error, though I don't know if that's what is done.
The time from one dropout to the next can be many hours so the only real way to detect this behavior is to run the laser on a data acquisition system that monitors the laser output power. Such a run is shown in Random Dropouts and Mode-Flips in a High Mileage Zygo HeNe Laser. This shows one polarized mode of a Zygo 7701 laser tube powered using an external HeNe laser lab power supply with nothing either active, so there is no stabilization and thus there is normal mode sweep. The dropouts have a duration of less than 1/30th of a second, though exactly how short they are is unclear due to the data acquisition sampling rate and input filtering. My guess is a few milliseconds. When a dropout occurs, there is a 50/50 chance that it will induce a mode-flip. Without stabilization, this results in a sudden change in output power. When the laser is locked with approximately equal power in both modes, there is no sudded change in power (or only a very small one), but then the laser relocks by going through a 180 degree change in the modes as shown in Zygo Laser Random Mode-Flip Event. This is basically the bottom half of the normal mode sweep that would be observed during warmup as in the previous plot - a dip of 20 or 30 percent of the locked power with a duration of 10 or 20 seconds. Both polarized modes will do the same thing.
In order to narrow down the cause of this behavior wtih one such laser, I plugged the +/-15 VDC power supply into a Sola constant voltage transformer on a surge suppressor with no noticeable change. I also ran the laser tube using a separate HeNe laser power supply with nothing else in the laser powered (no stabilization) to eliminate the internal one and other circuitry in the laser as a cause and still got dropouts. The AC power was also monitored with no evidence of corresponding spike, surges, brownouts, or other unsightly blemishes. So, unless someone installed an MRI machine next door while I wasn't looking, The cause of the dropouts is almost certainly related to the HeNe laser tube or its immediate wiring (ballast resistor, etc.) I'm now leaning toward a bad ballast resistor as being at least one of the possible causes. These lasers have two ballast resistors. One is near the tube and another is further back along the HV wiring. Both are enclosed in a plactic cylinder and potted. There is almost always evidence of stress in at least one of the resistors in used Zygo lasers - a rough and/or bulging outer surface. I found the ballast resistance of one 7702 laser that was having these spasms to be around 140K ohms, which seems rather high. The ballast in a healthy laser measured 75K ohms, which is what would be expected. Another one that had problems read 90K ohms. Now, 90K ohms might be normal as I've found that some Zygo tubes do prefer a somewhat higher value, but 140K seems to be off the charts. And I later found two lasers with ballast resistances of over 300K which is absolute nonsense. So, these resistors degrade over time. They probably have the most deviation from a normal resistance when cold, but run with a resistance that is reasonable most of the time. However, the resistance spasms for want of a better term resulting in a tube dropout. Unfortunately, there is no real way to dissect the potted ballast resistors. So, even where the total resistance measures under 100K, the only test will be to replace both resistors and see if the problem disappears. And it may be necessary to run the laser for several days on a continuously monitored setup that will catch even very short glitches to be sure. The "Service" LED is generally not turned on by these events.
However, even the new ballast resistors doesn't seem to cure all lasers. At this point, I am fairly convinced that there must be internal tube problems with some of them. My wild guess would be that it has to do with contact between the cathode-end mirror mount and the thin film cathode deposited on the glass envelope.
More to come.
PCB PCB Wire Connector Cable Wire Pin Color Pin Color Function ------------------------------------------------------------ 1 Black 1 Gray Ground 2 Green 2 Orange V- (-8 VDC) 3 Blue 4 White V+ (+8 VDC) 4 Brown 5 Red MEAS 5 Orange 6 Black ~MEAS 6 Gray 7 Bare MEAS Shield 3 Blue NC
I don't know how the LEMO connector is actually numbered so I have labeled assuming the diagram below (viewed facing the female connector on the receiver):
Key |_| 1 o o 6 2 o o o 5 Pin 7 is in the center. 3 o o 4
The cable had been cut so I have no idea if this is a standard Zygo cable and thus the cable wire colors may be meaningless.
In addition to the LEMO, there is a test point, presumably for signal strength monitoring, like on the HP/Agilent optical receivers.
According to the Zygo operation manual, V+ is +8 VDC +/-0.5 V at 0.3 A max and V- is -8 VDC +/-0.4 V at 0.75 A max.
V- may be used only to provide negative bias to -Vs of an AD9696 ultrafast The positive supply (at least) needs to be well regulated - when using one with a bit of ripple, operation was erratic and the presence of the output signal correlated with the power line frequency.
The optical input for testing was a red LED driven through a bridge rectifier made of 1N4148s (to double the pulse rate) from a function generator. As expected, the output was a differential signal with a response starting at about 150 kHz on the input. I haven't tested the high end yet, only to around 8 MHz, the (doubled) limit of the function generator. But the response appears very sensitive to signal level, much more so than with the HP/Agilent optical receivers. In fact, depending on signal level and offset (relative intensities of the positive and negative half-cycles for the LED), the output frequency sometimes equaled the input pulse rate and at other times was half of it.
However, when tested with a Zygo 7701 laser head, the output was always 10 Mhz, the split frequency divided by 2, which is what it should be. So, I guess it really expects a roughly sinusoidal waveform for the optical signal.
For examples of home-built stabilized HeNe lasers, see the sectiions starting with Inexpensive Home-Built Frequency or Intensity Stabilized HeNe Laser.
Although the label on the HeNe laser tube says Carl Zeiss, there is no doubt that it is made by Siemens: The model is LGR-7621, which is a Siemens model, and the remainder of the label is Siemens style. So, what about this is not clear? :) It is about 10 inches in length with anode-end output and a rated output power of 2 mW. (Its label also has the same irrelevant reference to Patlex patent number 4,704,583 as most HP/Agilent tubes discussed elsewhere in this chapter - and it is equally irrelevant for this one.) The normal dual mode thermal control technique is used with feedback provided by a pair of reflections from the uncoated surfaces of the roof prism - the one from the front, and the other through and back off the opposite surface. Polarizers oriented at 90 degrees select the two modes for the the photodiodes.
The entire controller is on the little PCB. Heat transfer to the environment is aided by the large heat-sink. Note the extra 1,000 uF capacitor, possibly added when it was determined that the HeNe laser power supply, which runs on 24 VDC, resulted in excessive droop on the DC power input when starting.
The locking sequence is rather interesting). Initially, it runs the heater at a power that switches between about 12 and 19 V on the heater apparently depending on whether the output mode is low or high, respectively. This results in about 35 complete mode sweeps, probably based on the mode sweep period exceeding some value like 15 or 20 seconds. Then it switches to optical feedback and locks with the heater voltage at about 6 V. This entire process takes about 5 minutes. Since 35 mode sweeps represents a relatively modest temperature rise - half of the 70 mode sweeps typical of other stabilized lasers - it loses lock after 2 or 3 minutes when the internal heating from the tube current increases the temperature such that the heater would have to generate negative heat to maintain a stable mode position. Once it's detected that the modes changed enough to be unlocked, it turns the heater back on doing it flipping thing for another 7 or 8 mode sweeps and again switches back to optical feedback, this time with the heater voltage at about 8 V. A similar sequence happens again after anywhere from 10 to 50 minutes if the laser cover is in place (but possibly never if it is not and convection cooling is more effective). Eventually, it settles in at a position which can be maintained by the heater. Maybe. :) Although I haven't checked, it's possible that if it were unable to lock with maximum heater power meaning the lock point was at too high a temperature relative to ambient, it would do the reverse and allow the tube to cool for a few mode sweeps and then switch back to optical feedback. This approach does make sense in a convoluted sort of way so I don't think the particular laser I have is broken, just a bit smarter than the average stabilized HeNe despite the apparent simplicity of the controller PCB! :) On the other hand, maybe it's just misadjusted or broken. While I can identify the mode balance pot, there is another one whose function is not known. It looks like an afterthought being globbed in hot melt glue without direct connections to the PCB, and reverse engineering the circuit will be difficult for that reason.
The output from the laser is a single polarized mode passed by a Polarizing Beam-Splitter (PBS) cube visible at the bottom-right of the photo. There may be an optic missing - or at least an option that isn't installed. A short focal length lens after the PBS cube results in a divergent beam that is never collimated. So it becomes a blob larger than the diamond aperture at the far left, and much of the power is lost. There is also a pair of suspiciously empty mounting holes in about the right position for a collimating lens, a few inches to the left of the PBS cube. But a similar laser installed in the Bomem spectrometer also lacks a collimating lens, so perhaps a large area beam is more important than a lot of power. And the holes look pristine.
The output power is rather pathetic. By the time the beam reaches the output, not much is left. About 2/3rds is lost in the prism, PBS cube, and expanding lens, though half of this is because only a single polarized mode is sent through the PBS cube. But 3/4ths of the remainder is blocked by that diamond-shaped output aperture! So, starting with almost 2.2 Mw out of the tube, only about 0.25 mW makes it out of the laser!
Specifications for frequency and/or intensity stability of this laser are not known. However, since this is for a spectrometer, really precise frequency stability probably isn't essential.
The entire laser assembly is about 19x9x2 inches overall. For some mysterious reason, it is much larger in person than it appears in the photo, above. :)