Restoration glory of Keithley 2001 DMM

[__BriKs__]

PCB is well damaged by electrolit of capacitor :/
I try to clean and remove most of burned FR4 materials.
I hesitate remove a lot more of materials and rebuild trace ?
Following pictures of PCB before cleaning and zoom on damaged section after cleaning.

[w.v.s.]

I try to clean and remove most of burned FR4 materials.
I hesitate remove a lot more of materials and rebuild trace ?

I decided to remove everything, which is black/brown. Quite a bit of work with 4 layer PCB and vias. Feels a bit like a dentist with mask and drill (and epoxy) . But from my point of view, that is the proper way to fix it.

[w.v.s.]

I hesitate remove a lot more of materials and rebuild trace ?

This is what is ahead of you, if you go this way:

[mawyatt]

That's some pretty nasty burnt up areas, nice job cleaning this up

At lease the traces are wide and you have enough copper to work with.

Good luck with the restoration!!

Best,

[Kleinstein]

The brunt part is in the power supply / voltage regulation part. So it is not super critical with leakage. Somewhat brown may be acceptable in this area.

[w.v.s.]

The brunt part is in the power supply / voltage regulation part. So it is not super critical with leakage. Somewhat brown may be acceptable in this area.

I agree from the technical point of view. But we are talking about restoration and I tend to see what level of restoration i can do.

A nice weekend to everyone!

[__BriKs__]

Yes I think using conformal coating filling the hole in the pcb and stop to dig in.
My point is it's important to remove burned fr4 between trace and made  new insulation and mechanic strengh with conformal coating.

And also, a little bored digging in burned fr4 and don't want to retrace the traces with wires ...

[w.v.s.]

My point is it's important to remove burned fr4 between trace and made  new insulation and mechanic strengh with conformal coating.

And also, a little bored digging in burned fr4 and don't want to retrace the traces with wires ...

I was surprised, how deep the damage went under the surface, especially pads. In my case, there are only few traces which have to be restored and I use copper tape for HF-shielding for them and bury them again in glass and epoxy. But i did not see much conductivity in the burnt areas, so in the power-supply region a less invasive approach should do, too.

I guess you've read the thread and the cleaning recommendations?

Good luck!

[__BriKs__]

Yes indeed in the device I have burnt fr4 is deep, up to other side just under the capacitor.
In the CR514 area temperature must had been very high also... maybe with the fusion of CR514.
I didn't read 100% of topic, and I'm far way to be experienced in restauration of pcba (I'm used to work with brand new pcba in prototype step).
Conformal coating usage is because I can do it, one lab of my customer have conformal coating. I think if I have no easy access to conformal coating I will just don't do it.

[__BriKs__]

Hello,

I did all caps and diode relative to BS+/BS- psu.
The unit start! Screen stay hand on 'overflow VDC' just after fw revision.
And finally the transformer start to smoke aie aie aie  I shut off the unit...

So, I don't know if I take the quest to find a transformer or if I just give up...
I don't have the time to measure psu voltage rail.

[aronake]

Has anybody measured temperature under the plastic hood after K2001 has been properly warmed up in 23 degree (or close) ambient temperature?

In TIN's K2002 post he mentioned that the meter showed internal temp of -23.2 at startup (measured against internal temp at time of calibration) in ambient temperature of 24 degrees. This should indicate a warmed up under the hood temperature of 47.2. Sure this is a K2002, but reasonable to believe that K2001 have a quite similar temperature.

I live in the tropics and the K2001 is quite temperature sensitive as observed by many, so my idea is to make a PID based fan controller with temperature measurement under the hood and a bigger fan more powerful with a fan adaptor. This will obviously not be very meteorology grade, but much better than allowing a static speed fan try to deal with ambient temperatures swinging from 20 to 32 degree C.

[w.v.s.]

And finally the transformer start to smoke aie aie aie  I shut off the unit...

So, I don't know if I take the quest to find a transformer or if I just give up...
I don't have the time to measure psu voltage rail.

Hi,

thats bad news. Do you think the transformer is dead or just injured?
How will you proceed?

Did you compare the resisitance values at the regulator outputs against GND?

The were some values in this video:
https://youtu.be/S8okMaAKYiw?t=1503

The values I put down from it were:

for a working board:

+15V    U107     83 Ohm
-15V    U102    500 Ohm
+5V     U108    600 Ohm

for an older board in doubt:

+15V    U107     67 Ohm
-15V    U102    2.15k Ohm
+5V     U108    920 Ohm

for my board I measued
                with adc-board  without adc-board
+15V    U107    84.8 Ohm        86.2   
-15V    U102     523 Ohm        3.30k
+5V     U108     97.3 Ohm       613 ohm

[Macbeth]

Does your unit pass all self-test in continuous loop  for few hours? Sometimes there is random error due to bad optocouplers between analog/digital, that cause "locks/freezes" like you describe.
What if you gently smack unit on the side, do the readings change (freeze up/reset/etc)? Line sync is FREQ signal that goes into A/D board, so you can trace from there.

Pics would be nice to see, sounds like your meter rather recent.

Hi TiN!

It seems I can run self-test for several hours without any errors. I didn't think of "smacking" the unit, but after trying that there is definitely somethings going on: after enabling line sync, I get a reading every 10-15 seconds (conversion speed "normal" - default). Between the readings "---------" is shown in display. If I bump the meter (not too hard  ) I can unfreeze the meter to get a few readings before meter goes back to "--------" display. So my "average # of readings per minute" goes up when bumping the meter. Something electro-mechanical seems to be going on - just have to figure out exactly what's happening.

An observation: When changing to Line Sync, there is always a number of initial readings for 2-4s after pressing the TRIG button. It's almost like some capacitor was fully charged and then drained of energy, only to be recharged each 10-15s to get a few more readings, before being emptied of charge, again. This is a long shot, I know...

I have to disassemble to meter again to be able to probe the connectors to the A/D board - FREQ specifically.

Hi eplpwr,

Did you resolve this Line Sync issue? I have 2x 2001M with B16 firmware and they exhibit similar behaviour. (Except knocking them doesn't make any difference - the lockups are too long and sporadic to know really). The meters become completely unresponsive until they sort themselves out when finally getting a line sync.

Both had other completely unrelated faults when I bought them. I fixed them with the help of the built-in tests and they can run through all the built-in continuous tests for 24H no problem.

It's only now I want to do TiN's shorted input noise test with line sync on that I noticed this issue.

[Pete2]

Hi all. A new 2001M owner here with B16 firmware. Lots of good information in this thread and I already have some tips on how to approach the faults. This unit has reportedly the following faults:
- 407.1 Front End (2 V range)
- 410.1 TRMS
- 411.1 & 411.2 TRMS Filter

The unit measures at least 10 V DC and 10 mA DC fine (according to the seller's pictures, I have not turned the unit on). I have yet to do a more thorough diagnosis. No idea about leaky caps either. Luckily there are schematics and repair manual available.

Tips & tricks from @TiN would be greatly appreciated if you're still around! Thanks.

[Kleinstein]

Defintely check for leaking capacitors, ideally before turning it on.

The 20 V and 2 V range front ends are a bit separate. So it is possible to have a defect in the 200, 2 V and 200 mV range and the 20 V + 1000 V range still working. This would at least point to a limited area of the circuit.

The RMS part is anyway largely separate, but may also requite the 2 V range input part for the test.

[Pete2]

OK so I did a bit more thorough diagnosis. I turned it on since I saw no leaks (no guarantee, I know). It does have more errors:

- 407.1, 2 and 3. Front End 2V, 200V and 750V ranges.
- 408.6 200V Range Compensation Error
- 409.6 750V Range Compensation Error
- 410.1 TRMS Error
- 411.1 and 2. TRMS Filter Error.

I will start with the Front End since that's what the repair manual suggests. Initially I measured that the + and - 8VF is all over the place. Those voltage rails are used in the AC front end circuitry and are fed by the +/- BS rails (~40V) and regulated with a zener/transistor regulators. So far +/- BS seems fine.

This unit is surprisingly clean. No dust anywhere inside and the fan is clean too. Going to get a quieter fan as well but have to use a 40 x 40 x 10 mm thick fan since they've used a 13 mm fan and you can only get 10, 15 or 20 mm ones.

[w.v.s.]

I turned it on since I saw no leaks (no guarantee, I know). It does have more errors:

I would recommend to have a look at leastunder the 5 larger caps close to the transformer, soon. I've seen the damage the leaking caps can do in my unit. Good look with fixing your unit!

[Kleinstein]

The +-8VF are bootstraped supplies in the DC front end, not the AC part. They are relative to the input buffer and may thus move with the input, especially when open. So having them move around is a good sign. The voltage is also used for the 20 V range. So one can expect it to work OK.

[Pete2]

The +-8VF are bootstraped supplies in the DC front end, not the AC part. They are relative to the input buffer and may thus move with the input, especially when open. So having them move around is a good sign. The voltage is also used for the 20 V range. So one can expect it to work OK.

Thanks for this. How much will the rails swing? At one point I measured -2,6 V and another time +17 V on the rails. Seems like a lot for 8 V nominal which is supposed to be stable.

Going to replace all caps of course and then just trace the +7V (x1 / x10 divided) test signal that is expected at A/D input.

[Kleinstein]

The voltage is relative to the main DC input buffer and this covers the whole +-20 V range and a little over-rage. it looks like there are 22 V zeners as clamps. So it can more a lot. For testing one would need to apply a voltage to the input, like initially zero and later than maybe some 5 V or 10 V. For the start I would go with the non AZ mode, as this makes things easier with no constant switching.

[Pete2]

OK fault found. It's K503 that has failed and not actuating. The coil pins came off when I desoldered it. It wasn't even flush with the board which may explain why it failed eventually due to vibration.

I bodged in a standard relay for testing and now all tests pass without errors. The original relay is unobtanium (except this site, I wonder if it's legit: https://www.ecufixtool.com/products/relay-coto-8200-0031-reed-relays) and the pitch is so bizarre no direct replacement can be found either. Also the spec is 1 kV because of input isolation.

This relay is pretty close: https://www.digikey.fi/en/products/detail/standex-meder-electronics/SHV05-1A85-78L2K/6164037

The problem is the pitch which is close but not quite. I wonder could I make a small adapter PCB...

If anyone has a parts unit I could use the analog board top shield, since someone has stolen it from my unit. Also the rear input ferrite was taken... weird along with the other fan nut(!?). And finally the keyboard membrane is slightly broken (Range Down sheared off) so I will try to get one.

[w.v.s.]

I bodged in a standard relay for testing and now all tests pass without errors.

Congratulations!

[Pete2]

In my last post I bodged in a standard relay for testing. However the relay needs to have at least 1 kV isolation voltage between contacts since it's connecting the internal reference voltage to the input during self test.

The original reed relay is unobtanium so I found a replacement but the footprint is different. Pinout is the same but smaller pitch. So an adapter PCB was designed. The relay for it is Standex-Meder SHV05-1A85-78L2K (https://www.digikey.fi/en/products/detail/standex-meder-electronics/shv05-1a85-78l2k/6164037)

The attached pics show the relay mounted on the adapter PCB. Luckily the original relay's pin pitch was a multiple of the standard pin header so adding an angled header allowed the relay and the adapter to be mounted on various orientations, depending on location and clearance.

[CM]

I have a Keithley 2001 firmware B06 that threw a 363 error on the open circuit section of the comprehensive calibration.  Replacing the analog switches U317, U320 and U323 resolved that issue.  Thank you ddcc for that great slouthing to figure out that one!  I used a 4 wire measurement with 19 kohm resistance for the 20 kohm input and a 2 wire measurement with 1 Mohm resistance for the 1 Mohm range (as suggested by the calibration manual), and the comprehensive calibration ran through with no errors. Replacing the LM399 precision voltage reference also breathed new life into this unit.  I am going to proceed with replacing the other 6 analog switches to see if I can drive down the ppm errors I am seeing on the resistance measurements.

[w.v.s.]

Sounds good. Did you also take care about the electrolytic caps?