Keithley 2010 Repair Attempt - Ideas Welcome...

[danoki]

Hi,

Firstly, hope everyone is doing OK, considering world events.

Secondly, first time here, so please be gentle and I did search the forums, but coudln't found much on my particular issue (even several evenings of googling let me down).

Situation: I scored a Keithley 2010 DMM from the "not worth repairing pile" at work and thought it'd be fun to get it working and have a nice 7.5 digit DMM for the lab (better than the old Solartron 7045 I've had for years!). I did scour the forum and found the partial schematics and user guide/service manual, but no many other posts on this particular model or similar issues.

Problem: Boots fine, but goes into "overflow" on all ranges and input types. Opening it up showed no visible magic smoke release, borken caps or anything fishy... Damn. There was a rare case where I saw some random digits, when I wasn't ESD grounded and poking some of the input wires.... Soon as I got the ESD mat out, it was overflow all the time.

So far:
- I measured all the voltages which generally seemed OK. 5V digital rock solid, 5V analogue was a little high and very hot. Hmmm. +/-15V rails seemed OK at 15.6V, but the supplies to the split rails, namely +/-24V and +/-18V are a bit on the high side...
- Thought the 5V analogue rail was a bit too warm.
- Confirmed that the front panel/main board digital section is working fine.
- Traced the comms from the digital section to the OKI chip in the analogue section. Can see data going back and forth, which changes when I chose different ranges/modes.
- Traced the tracks for the shift register (from the partial schematic) and confimed that all the clk/data/strobe lines are there.
- Traced the shift register data through all the chips and found it wasn'ty getting to the last two chips in the chain. Hmmmm interesting...
- I then confirmed that the ULN2003 (for driving the relays) connects to one of the suspect shift registers, so no wonder it's not switching!
- I checked the power pins of the suspect shift registers and low and behold, there's a short on the 5V analogue plane!? 
- I checked other parts of the 5VA plane, yep shorting; however some of the +5V/0V points are 50R 5V->GND ?!   

Solution: In progress... 

So, before I start systematically pulling each 5V crossing component to find the one that's shorting, anyone seen this sort of thing before and can suggest any cuplrits? The dreaded multi-element switch or something more sneaky?

Cheers and thanks in advance,

danoki

[picburner]

Welcome to the forum.
Have you tried to see if and what errors the diagnostic test shows?
(Switch on the instrument by holding "open" to access the diagnostic test, then shift - test).

[danoki]

Hi,

Thanks for that. There was no mention of the extra button to get the self-test function in the manual... 

Right, let's see... BRB.

Thanks.

P.S.: The transformer is buzzing more than I would expect, but with a 5V short, it's probably no wonder.

[picburner]

Thanks for that. There was no mention of the extra button to get the self-test function in the manual...

Yes, it's a hidden menu...
If the instrument is old, do not blindly trust appearances, it often happens that there is a leak under the electrolytic capacitors without the electrolyte coming out externally.

[danoki]

Fair point. I did try measuring for shorted caps and all the electrolytics did seem to charge up with the ol'e multimeter across them trick (in both polarities).

Errors were numerous in the 100/101/200/301/302/202/204/400/401/600 ranges, most .1/.2.

Will take a look at the service manual... OF course there's no obvious list in the 2010 model manual!

Think there's a 2000 series list somewhere...

[Nuno_pt]

Here you go the service manual from TiN's website < https://doc.xdevs.com/article/keithley-instruments/#2010 >, and lot more check out the website xdevs.

[picburner]

Errors were numerous in the 100/101/200/301/302/202/204/400/401/600 ranges, most .1/.2.

mmmh, check all the electrolytic capacitors in the analog section of the power supply....
Even my K2010 had all those errors and I had to replace all the capacitors, clean the pcb tracks, and replace some ICs.

[danoki]

Hi,

Sorry, meant to say the service manual earlier; I did come across the xdevs page previously. Thanks anyway!

So many errors, it's got to be some systemic power issue.

Quick check of vregs/eletrolytic cap temps...
U184 = 52.5 degC
U180 = 70 deg C
U179 = 81
C146 = 44
C281 = 32.5
C284 = 46
C277 = 36
C276 = 41
C131 = 39
C148 = 33.5

OK, I'll pull all those caps and see what they're like and if the 5VA shorts clear...

[danoki]

All those caps seem OK and PCB is clean underneath each of them. Still get the 50R across the outputs of the 5VA reg and other 5V plane areas are still shorted...

Sarcastic yay. Time to start pulling the 5Vreg and then other parts on that plane after all! 

P.S.: It's a 2005 PCB.

[picburner]

For the test list you have to rely on the "K2000 repair manual", the "K2010 repair manual" cannot be found on the net.
My pcb is from 2001, older than yours.
I am the author of the K2010 partial schematics that you can find on the net, if someone wants to finish them, is welcome of course

[danoki]

Thanks for the 2000 series manual with all the fault codes. Should be useful for others in the future.

Murphies law, the 7805 Vreg is fine... Typical.

Right, time to go crazy diagnosing a plane fault!   

[danoki]

@picburner

Out of curiosity, which ICs did you end up replacing on your one?

[picburner]

Out of curiosity, which ICs did you end up replacing on your one?

It's been a long time but I remember these: U139(LT1124), R271(TF-245 resistor network) and some CD4094(don't remember now the label) all damaged by the electrolyte of the capacitors.

[Kleinstein]

The large Al electrolytic caps usually don't fail short. It is more like tantalum caps that can fail with a short, especially the old drop form.

Also SMD ceramic caps (especially the larger ones) can fail short. This is often from mechanical stress.

[danoki]

Did notice some of the ceramic caps had a few little marks on top which looked a bit suspicious; but it turns out the cap under C106 was the main culprit. Swapped it out for a 220nF (what I have to hand).

Now I can switch ranges and see some values!:

Quick checks:
D Cell -> 1.58V
Shorted test probes (DC VIN) -> Fluctuating around 0.003 mV
Shorted test probes (2 wire method) -> Fluctuating around 0.125 
Shorted test probes (temp) -> 22.89 deg C
Continuity test with shorted probes - > Beep, beep!
Quick checks on AC ranges with shorted probes gave good low numbers and the period count sort of worked when tapping the 1.5V D Cell.
 

This calls for a cup of tea in celebration! Next (another day) is to work out if the remaining fault codes are just the unit being out of cal (last done 2017)! 

Grazie mille! / Vielen dank!

[danoki]

Took a quick pic of the bench so you can see what I'm dealing with (please excuse the muddle of cables)...

So aside from the obvious "which digital scope to get for home"... The big question is what level of DIY test standards do I need to build to check the Keithley is in spec?

P.S.: Just found out the Keithley is a 7.5 digit meter which current goes for about £3.5K! Not bad for a bin find and to be repaired job!

[Kleinstein]

Many of the tests to really check the DMM need quite special instruments - e.g. a good calibrator way more expensive than the meter.
Without many extra instruments one can do a few sanity checks, also to learn about the meter:
1) Check the noise with short at the input.
2) Check the input current (with a resistor or capacitor)
3) A crude linearity check with a turn over test: check an external reference with both polarities.
4) Watch the warm-up phase.
 

[martinr33]

One thing you might try is tracking the input through to the ADC. Set it up with, say, a 9.3V input, and see if you can follow it through to  the ADC area. The front end should be similar to the 2001, more so than the K2000. The CMOS switches are good things to check.


9.3V lets you identify a signal as opposed to a 10V lined used for other purposes.