Maker Pro
Maker Pro

Successful Stanford Research SR530 lock-in amplifier repair

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
Woo hoo!

IMG_20180616_201039.jpg

Here it is working. My signal generator's sync is apparently 90 degrees out of phase with the signal.

The signal is just under 1mV at 48.7kHz.

It's not completely obvious, but the max sensitivity is 100nV full scale, but you can then do a X10 on that too get 10nV

The current input is 106 V/A, so 0.5μA corresponds to the 500mV minimum sensitivity. I think that means the highest sensitivity is 10fA FSD, but I'd have to check the specs again.

Anyway, it worked, so I'll put the shielding back on and consider it a job well done!
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
Whoever designed the fastenings for that shield either had incredibly small fingers or a very evil sense of humour.
 

hevans1944

Hop - AC8NS
Jun 21, 2012
4,878
Joined
Jun 21, 2012
Messages
4,878
Anyway, it worked, so I'll put the shielding back on and consider it a job well done!
Congratulations! You may now be prepared to verify and repeat an "experiment" I did in the mid-1970s to detect the presence of people (warm bodies) using a single uncooled pyroelectric detector, a simple two-bladed light chopper, and a pricey PAR lock-in amplifier. Piece of cake to do this today with cheap PIR sensors, but "back in the day" it did require some expensive "test equipment," although the pyroelectric element was dirt-cheap by any standards.

Harshaw Chemicals gave me a free sample to "play" with. I set up on a bench in our electronics lab and pointed the detector element more or less toward a hallway door on the opposite side of the room, maybe fifty feet away, with the chopper blades right in front of the detector. No optics involved, no salt lenses, no plastic Fresnel lens like modern PIR sensors use. IIRC, the chopper interrupted incoming radiation at about thirty hertz or so (variable speed with a photocell interrupter to sync the lock-in), and the PAR integrating time was about one minute. Not exactly fast responding, but my experiment was just a "proof of concept" demonstration to show my boss that I understood how those new-fangled pyroelectric detectors worked. People were touting incredible D* (D star) sensitivity ratings for an infrared detector operating at room-temperature, potentially rivaling liquid nitrogen-cooled PbSnTe photoconductive detectors, the "gold standard" then for IR detectors. If that's all Greek to anyone, here is a nice little explanation from Hamamatsu on how infrared detectors work.

Well, it was was all hoo-hoo and marketing for this new "solution" looking for a "problem" to solve. About that same time, the basic operating principle (a permanently polarized, stressed, dielectric causing differential charge separation) was also being used to make electret microphones. Many engineers working with high-impedance piezoelectric strain gauges and accelerometers were already familiar with triboelectric effects, because special graphite-impregnated shielded coaxial cable was necessary to avoid triboelectric voltages, created by relative motion between the inner coaxial conductor insulation and the outer copper braid, which would overwhelm data acquisition from the high-impedance sensors mounted on vibrating structures.. like an aircraft wing. It was a real PITA attaching that special coax to Microdot connectors while avoiding shorting out the connector by smearing the graphite in the wrong place. So most of us didn't get real excited by this "new" development of temperature-stressed and sound-wave stressed dielectric transducers. Just shows-to-go-ya how short-sighted I was. The idea of using two identical pyroelectric detectors, with their outputs differential connected, and their active areas exposed to two slightly different fields-of-view (FOV), to detect motion was absolutely brilliant. Sure wish I had thought of it.

So after playing around with the Harshaw Chemicals pyroelectric device for a few days (weeks?) we finally got back to work trying to figure out how to build a laser weapon system to shoot down ICBMs during their re-entry to atmosphere phase. Well, not me, actually. I never did believe that would work, photons having practically zero momentum and easily reflected by mirrored surfaces, but I was happy to help the scientists build their experiments for dog and pony shows. That type of work was good for several more decades of funding, but I graduated in 1978 and went to work in a different field. AFAIK, the electromagnetic rail gun (which we were also trying to develop) eventually became a successful weapon, but I had no direct involvement with that either. <sigh> Us warmongers are having a tough time, what with peace and prosperity threatening to break out all over the world... good thing that I am finally "retired" from all that.
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
the chopper interrupted incoming radiation at about thirty hertz or so (variable speed with a photocell interrupter to sync the lock-in)

I'm planning to make a variable speed chopper as my next construction project. I've got a bldc motor because I can drive it at a known speed very easily. They are also designed to connect to a propeller, and the same mounting can be used to fix it to a chopper disc.
 

hevans1944

Hop - AC8NS
Jun 21, 2012
4,878
Joined
Jun 21, 2012
Messages
4,878
They are also designed to connect to a propeller, and the same mounting can be used to fix it to a chopper disc.
Excellent idea.

I haven't done any work with optical choppers since the 1970s, but do recall they came in several variations and were quite expensive, considering that "all" they did was spin a disk. A DIY optical chopper should be eminently practical to build today using commercial-off-the-shelf (COTS) components. Most of the choppers we used had motors whose speed could be phase-locked to an external reference frequency source, or they would generate a once-per-revolution timing pulse that could be used to synchronize an external oscillator. When used with either a lock-in amplifier (the usual application) or a boxcar integrator, a constant chopping speed was essential for good performance. Maybe the motor construction required to minimize "wobble" contributed to the high cost.

Please keep us informed on how you are going to use your professionally repaired lock-in amplifier, and whether or not the second one has arrived and is in serviceable condition. Do you plan to keep them or sell one of them to cover the cost of the first one you repaired? Just curious.
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
My rationale in using what is essentially a synchronous motor is that the speed should be very tightly controlled by the frequency.

At high enough speeds, the disc will act like a flywheel and reduce jitter.

At low speeds I am looking at using pwm to drive the motor from what equates to a low frequency sin wave to smooth out the lumps.

At this stage the speed will be mostly stabilised by the fact that the controller is clicked by a crystal oscillator. I've not considered phase locking it to an external signal at this stage.
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
Here's a question for you guys...

I'm starting to give dipped tantalum caps quite a bit of side-eye.

I've had a number of failures in a row caused by them now. My initial thought was to replace them with low-ESR electrolytic caps, but I wonder if that's a good idea.

My next thought is to replace them with MLCC caps. These have an even lower ESR (lower than tantalum), but they do tend to suffer from a reduction in capacitance with increasing voltage. However, since they're used (mostly) for bypass, the actual capacitance is probably not too important.

Whilst MLCC caps have both short and open failure modes, these failures seem to be related to stress. If I'm using leaded caps, that should minimize the stress placed on them.

They're moderately priced (in quantity). Not that I'll be buying 2000 of each value :-o

Should I start an MLCC-based tantalum eradication program?
 

hevans1944

Hop - AC8NS
Jun 21, 2012
4,878
Joined
Jun 21, 2012
Messages
4,878
MLCC caps for power-rail by-pass applications sounds like a good alternative to tantalum caps, maybe "backed up" with an aluminum electrolytic or two if this is a retrofit and there is room available.

I have in my "junque box" somewhere a few dozen or so tantalum capacitors of uncertain voltage rating and a few microfarads capacitance. I pop them onto a solderless prototyping board to by-pass the power supply rails when "playing" with a new circuit. Almost never, at my age, do I go on to make a "for real" PCB and stuff it with new components. If I did, and in light of your experience, I would certainly avoid tantalum capacitors. Not much experience with using MLCC caps either, but I have watched in astonishment how much they have diminished in size, especially the surface-mount devices (SMDs). SMD aluminum electrolytics have also appeared, but I have no idea how they compare with MLCC SMD.

As you (and others here, @Arouse1973 IIRC) have mentioned, MLCC caps have a pretty severe voltage-coefficient of capacitance that is tolerable for bypass applications but probably unacceptable for RC or RLC timing/tuning applications.
 

Arouse1973

Adam
Dec 18, 2013
5,178
Joined
Dec 18, 2013
Messages
5,178
As long as you have a stable supply voltage and choose the correct cap value versus voltage you will be fine. You should be able to find this information on the data sheet. One thing you do have to watch with low ESR caps is if you are suppling power to them using long wires at switch on you can get ringing which could have voltage spike that are greater than the voltage rating of the capacitor. Over time this can cause the capacitor to fail. A ferrite bead normally resolves this issue.
Adam
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
Now I have successfully repaired the second lock-in amplifier, is there anything I can do with two of them that I can't do with one?

I'm still waiting for the special triple half H-bridge controller (a fancy gate driver for 6 MOSFETs) before I build my chopper.
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
electromagnetic rail gun (which we were also trying to develop) eventually became a successful weapon,

I believe the latest US aircraft carriers are using these to launch their manned weapons platforms (aka aircraft). I also believe they're not having a good time of it. Old fashioned stream is simple and well understood, but I hope they can get them working because retrofitting conventional catapaults would be a real embarrassment.
 
Top