Keithley 2010 Repair

[KungFuJosh]

The pins appear to be 2.5mm, so maybe one of these would work depending on the wire gauge:
https://mou.sr/3QaIAE7
https://mou.sr/4xmaCgq

[Cyclotron]

The pins appear to be 2.5mm, so maybe one of these would work depending on the wire gauge:
https://mou.sr/3QaIAE7
https://mou.sr/4xmaCgq

Those look nice.

I think I would go with one of these. These are both in stock at digikey.
LAPP 11197100
Phoenix Contact 1674833

[KungFuJosh]

My first link is the 1674833. I don't know what the wire gauge is though. Not sure which is better.

I emailed one of my contacts at Tek to see if they sell those sockets or not. Let's guess- $4 at Mouser, so $50 at Tek?

[Cyclotron]

My first link is the 1674833. I don't know what the wire gauge is though. Not sure which is better.

I emailed one of my contacts at Tek to see if they sell those sockets or not. Let's guess- $4 at Mouser, so $50 at Tek?

Apologies, I wasn't clear, I only posted to say that one of your selections and one I found were in stock at digi. 
I'm interested to hear if you get a price from Tek.

[KungFuJosh]

Apologies, I wasn't clear, I only posted to say that one of your selections and one I found were in stock at digi. 
I'm interested to hear if you get a price from Tek.

Ah, gotcha. No worries. You prefer digikey? Mouser annoys me sometimes, but usually less than other suppliers.

Digikey has been fine in the past, but they always seemed slower to me, and I'm not always that patient.

Tek will either tell me a (probably crazy) price, or that they don't have them available for individual sale. Either way, I'll probably buy a couple TF-245 chips, and the jack current input piece (white fuse holder, part number 2001-312D).

[Cyclotron]

Mouser was my primary till they ticked me off to no end and basically told me it was my fault for believing their statement that orders by a certain time would ship that day. I will still order from them if required but that was one support person's ability to lose the company whatever it is that I contribute to their revenue.

I have been preferring digiky and using FedEx as UPS from them was wishywashy for a while.  I also do not suffer and overabundance of patience.
Just like my attempts to support Mr Carlson's lab recently. 

One of my favorite rules of CyberSecurity is "Say what you do, do what you say". When I find people or companies not just doing what they say they do, I lose faith rapidly.

[KungFuJosh]

Mouser was my primary till they ticked me off to no end and basically told me it was my fault for believing their statement that orders by a certain time would ship that day. I will still order from them if required but that was one support person's ability to lose the company whatever it is that I contribute to their revenue.

I have been preferring digiky and using FedEx as UPS from them was wishywashy for a while.  I also do not suffer and overabundance of patience.
Just like my attempts to support Mr Carlson's lab recently. 

One of my favorite rules of CyberSecurity is "Say what you do, do what you say". When I find people or companies not just doing what they say they do, I lose faith rapidly.

I agree with all of that. Mouser has shipped my last 5 orders the next day, despite only 1 of them being even close to the cutoff time. On the bright side, they did upgrade the most recent order to next-day air, so I still get it when expected. While the delay does annoy me, other companies, including digikey, may take even 5 days before they bother to ship an order.

Mr Carlson's Lab annoyed me too. I signed up for his premium bullshit years ago. I looked around, I was disappointed, and didn't feel it offered anything worthwhile above the free stuff. So I canceled my subscription. He emailed me all pissed off and threated to ban me. Like I care- I canceled because I didn't see the value. 

I hear you on companies keeping their word. I give a little leeway though because I don't expect anybody to be perfect all the time, and delays happen.

[KungFuJosh]

So now I'm down to only one fault remaining.  305.2 BIAS CURRENT

Did you resolve this?

[Cyclotron]

So now I'm down to only one fault remaining.  305.2 BIAS CURRENT

Did you resolve this?

No, I was working on that issue when the primary issue ( I assume the reason it was sold) finally caused it to stop communicating with the ADC. I just got home from travel this week so I should be digging back in this weekend on this ADC issue. I'm hoping that also solves the 305.2 BIAS Current issue. In the thread where I've been documenting my repairs I mentioned that my meter came with A08 and that revision doesn't have a 305.2 test or any 305 tests. I have A08, A12, and A15 chips to load and there's A09 and A10 out there I think.  Relevance unknown...

It would be super cool to find a way to determine which is the latest per hardware revision as there appears to be at least 3 I believe. My meter was built/shipped around 1996/97.

In other news, the lens came in and its fantastic. I really like the look.  Thanks for your efforts on that. Great job!

[KungFuJosh]

Per TeKeithley rep:

They said there are two hardware versions (A & B). Final A firmware is A15, and final B firmware is B05 (should be obvious that they are not interchangeable).

Only two hw versions, A15 being the most recent relevant option.

305.2 came back on mine again after running for a long time. I wonder if my R359 resistors being heat shrink wrapped are getting hotter than they should maybe. Just a guess, I dunno. 🤷

I just found your thread.

Thanks,
Josh

[KungFuJosh]

One could do a little better with the OPA177 (e.g. OPA207 or OPA205).

Would the OPA205 be a drop-in replacement, or will some other modification be required? I read something about potential oscillation issues, but it might have been slop nonsense.

Thanks,
Josh

[Kleinstein]

Th OPA205 is considerably faster than the OPA177. So it is not a direct drop in replacement. It can still work in many places. The OPA207 is still a little faster, but not by much and in most places should work.

There are also different output stages so that the OPA205/207 have near rail to rail out and in rare cases this could be an issue.
Another point is that the OPA177 has offset trim functionality - though rarely used, the new chips don't have this anymore.

For the K2010 integrator the higher speed of the OPA205 should be OK and can even be an advantage. If for some reason there is too much ringing (measurable at the OPA205 output) one can compensate for the higher speed by changing a resistor (e.g. reduce R217). I would not expect it - the 1:1 divider looks about right for the 3.6 and 16 MHz GBW for the OP-amps. 

... I wonder if my R359 resistors being heat shrink wrapped are getting hotter than they should maybe. Just a guess, I dunno. 🤷

The heat at R359 not should not be an issue. That would be some 6 V and 3 mA and thus 20 mW range. The original R359 would have a similar power.

[KungFuJosh]

Th OPA205 is considerably faster than the OPA177. So it is not a direct drop in replacement. It can still work in many places. The OPA207 is still a little faster, but not by much and in most places should work.

There are also different output stages so that the OPA205/207 have near rail to rail out and in rare cases this could be an issue.
Another point is that the OPA177 has offset trim functionality - though rarely used, the new chips don't have this anymore.

For the K2010 integrator the higher speed of the OPA205 should be OK and can even be an advantage. If for some reason there is too much ringing (measurable at the OPA205 output) one can compensate for the higher speed by changing a resistor (e.g. reduce R217). I would not expect it - the 1:1 divider looks about right for the 3.6 and 16 MHz GBW for the OP-amps. 

At least according to the schematic we have, pins 1,5, and 8 are not connected. So the offset trim is not being used anyway.

Sounds like I should install an OPA205 at U138, measure the output, and if there's no substantial ringing, it's all good?

Thanks,
Josh

[Kleinstein]

The replacement should work OK. Testing for high ringing is a good measure, but I would not expect this to be an issue. If it is a fix should be relatively easy. The question is a bit if it would be worth it. If U138 is broken, changing to a more modern part would be a small step.

The expected benefit would be a little lower noise, but it is not so clear how significant. The specs for the current noise of the OPA177 are a bit unclear. They only show noise with 2x200 K input resistance and it is not clear how much of the current noise is correlated. There is also some uncertainty about the relevant frequencies.

For the error messages the problem is that it is not so clear what the tests actually measure.

[w.v.s.]

I was inspired to try the built-in test with my 2010 (cold start). In the first run, I got a 400.2 and a 500.1. The 400.2 was gone in the second round, but the 500.1 stayed and showed values in the 7V range instead of the 100nV allowed maximum. 4-wire short resistance is measured around minus 20kOhms. The short is O.K.. So I think there is something wrong, which wasn't before.
 

[w.v.s.]

Update: It's fine at the rear input. So probably only a bad contact/switch.

[KungFuJosh]

Try cleaning the contacts, and push the switch a bunch of times and see if that helps.

[KungFuJosh]

The pins appear to be 2.5mm, so maybe one of these would work depending on the wire gauge:
https://mou.sr/3QaIAE7
https://mou.sr/4xmaCgq

Those look nice.

I think I would go with one of these. These are both in stock at digikey.
LAPP 11197100
Phoenix Contact 1674833

This is kinda funny given our conversations on the topic of Mouser/DigiKey.

Mouser doesn't stock 1674833, so I ordered it from DigiKey along with some OPA205ADR. DigiKey's site said that 1674833 needed an extra day before shipping...but they shipped my order 2 hours after I placed it. Should I be mad at them for not keeping their word also? 

[KungFuJosh]

Here's an update. I replaced U138 with an OPA205ADR. Error 305.2 is not present either way with a cold start. I'm letting the meter warm up for a while to see if that changes anything.

I do still have an intermittent 400.2 fail.
Fails:
5.229872
5.149884

Pass:
5.098455
4.996471
5.005132
5.050364
5.117431
5.113144
5.095101

That's not all the tests, but a good sampling. Out of over 25 tests it might have failed 3 or 4 times. It seems something related is a little unstable for whatever reason. I will check this again after a longer warmup.

I tested the output of U138, and there was a little noise if I zoomed in a lot, but otherwise the attached is what's seen when the 305.2 test runs.


Repairs performed so far:
All electrolytic caps replaced.
2 heatsinks added.
R271 removed, tested, cleaned (ultrasonic), reinstalled, and 2 bodge wires added.
U139 replaced with LT1124AMPS8#PBF, and 1 bodge wire added.
VR122 replaced with 1N4578A.
R359 tuned to 4.059kΩ for VR122 stability.
U138 replaced with OPA205ADR.

Thanks,
Josh

[Kleinstein]

Without knowing what the self tests actually measure, it is hard to tell what a fail / pass or the value means.

Maybe do some normal measurements on the 4 wire short and check what value one gets in the different ranges (manual ranging for the lower 2 or 3 ranges). Similar one test a few known good resitors.

[KungFuJosh]

I saw one of your other posts about drift, and the meter does do weird open drift behavior when it's first powered on, and then continues less dramatically.

For example, when first powered on in DCV, it may show -5V or 5V, and then drop down closer to 0. Right now, once it reaches close to 0 (maybe 0.03V), it takes about 4 seconds to drop to -1.17V, and then back to 0.

400.2 and 305.2 (which is back after a warmup), are both performed open.

305.2 is bias current. My guess is that the state of the drift caused by 305.2 is causing the intermittent 400.2.

Thanks,
Josh

[KungFuJosh]

I do still have an intermittent 400.2 fail.
Fails:
5.229872
5.149884

Pass:
5.098455
4.996471
5.005132
5.050364
5.117431
5.113144
5.095101

Test 400.2 — NON INV PATH
Bank: VAC
Inputs: Open
Expected Value: 5.6 volts
Limits: 0.6 volts
Fault Message NON INV

If the expected value is 5.6V, and the displayed values during the tests that passed are the farthest from the target, how does that make any sense??

I'm guessing this fail is the product of current leakage on the front end messing things up that is likely caused by and/or causing 305.2.

Thanks,
Josh

[Kleinstein]

The test limit may be relative to another test done earlier. I would expect something like looking for the same gain for positive and negative side pulses in the AC signal. The description is still a bit confusing.

What is a bit odd is that the readings in the tests are scattering quite a lot.

[KungFuJosh]

The test limit may be relative to another test done earlier. I would expect something like looking for the same gain for positive and negative side pulses in the AC signal. The description is still a bit confusing.

What is a bit odd is that the readings in the tests are scattering quite a lot.

That's true for the 400.2 fail. Unfortunately, we have no documentation for 305.2. If we assume it's the same situation, I can try 305.1 and see if it shows a value. 🤔

[KungFuJosh]

I ran tests again from a cold start, and continued repeating them until they failed. It didn't take that long.

400.1: ---------- (always)
400.2:
4.864007 Pass
4.862374 Pass
5.196737 Fail
5.021211 Pass
5.205027 Fail
400.3: always matches 400.2; has not failed.

305.1: 0.059605
305.2: 0.010496 Pass
305.1: 0.020624
305.2: -0.017864 Pass
305.1: 0.002562
305.2: -0.033800 Pass
305.1: 0.038397
305.2: OVERLOAD Fail

Thanks,
Josh