Restoration glory of Keithley 2001 DMM

[gummy]

Thanks to everyone for the wealth of information on this thread, it has been invaluable on my road towards fixing my very first Keithley.

I bought 2 units off eBay which were listed as working. The units did not work. Both were showing a slew of errors 304.2-304.7 Ohms current sources, 307.2 cal divider ad mux/10, 308.1 4 digit mode ad signal path. I'll refer to them as unit 1 and unit 2 from now on.

Unit 1 worked and was displaying somewhat plausible readings straight out of the box. The self-tests above failed and I tracked it back to a faulty LT1007 in the AD buffer. I replaced the LT1007 and the meter works!

Unit 2 powered on but was displaying nonsense. With inputs shorted the DCV reading would slowly drift up from 4V all the way up to 12V. The ohms ranges were also displaying nonsense, half a gig ohm for a 500K resistor. I have been pulling my hair out all week, I replaced U319, U325, U332, and U322 before I decided to swap the AD board from Unit 1 into Unit 2. Unit 2 passed with flying colours with Unit 1s ADC. This leads me to believe the problem resides on the ADC board instead of the buffer. I haven't scoped data out but from the strange readings, there's definitely data going back to the CPU. I'm at a loss as to how to test the AD board. Any help will be greatly appreciated. Thanks!

I managed to source DG211 and HEF4052BP off eBay for a reasonable price. As far as I can tell they are legitimate components. The item numbers are 324013465152 (HEF4052) and 292514792305 (DG211)

[Kleinstein]

For the ADC board a first part to check is the supply and the reference. There is a 2nd zener based refrence on the ADC board and this is used for most of the ADC operation.  The +Vref and -Vref should be a some +-10 V.
An usefull point to check if the ADC is working overall is the integrator output voltage waveform (e.g. output  of U816 for version G,H,J). The question is especially if the signal reaches saturation at some points.
There seem to be different versions of the ADC board - the parts enumeration may be different.

[gummy]

I've let this problem marinate in my brain a little and I realized I've jumped the gun. I did end up testing the zener on the ADC and both reference voltages. The zener was 6.2V, not 6.4V as stated on the schematic (a little head-scratching involved here) and both V_ref voltages were about +10.2 and -10.5. I realized I never tried ADC2 in Unit 1. I just got home and changed about the ADCs and it's currently passing self-tests. I'm going to let it run for a while but it's looking like the ADC is fine. The problem unit only begins to fail self-tests after warming up for 5 minutes so I'll see if I can find an IR cam and spot the hot component. The bugger of a zener tested open when I lifted a leg but once I pulled it off the board it turns out it was behaving. I began binning all my 6.2V zeners looking for the perfect 6.4V one before I realized the thing was working...

The ohms errors popping up are really confusing me, with most problems pointing towards the buffer amp as that is the only place where the ohms sources and the 4 digit signal path come together (please correct me if I'm wrong, my brain needs a few extra wrinkles). Since I've replaced U319 already, the only other thing that jumps out at me is U317 being stuck in a partially on state.

I'm all out of ideas today. More might happen if I let it marinate for a bit longer.

[Kleinstein]

Zener diodes have quite some tolerances. The 6.2 V for the reference are likely OK. That reference zener on the ADC board is a special low noise type, not one of the more normal ones.
The asymmetry in the reference is a little strange, though it should not be a problem for the ADC.

I would not count too much on having a single error to effect both the ohms mode and fast 4.5 digit mode. This could well be 2 different problems. The 4.5 digit problem could be related to the ADC - as most of the other parts would be the same with slower conversions.

The Ohms ranges may use a different input path for a ~ 10 V range, more similar to the 2 V range than the 20 V range. AFAIK that 10 V range is not available as a normal vortage range - it would be nice to have it, as this could be a good range (low noise, good stabilitym though maybe slightly limited linearity).
The ohms part is rather confusing and there are quite a lot of parts that can go wrong there. This includes the input protection and an opto-coupler.
It could make sense to do a few more tests in the ohms range on how far it works.

[Gridstop]

Anyone ever find a replacement/substitute for the 2000/2001/2015 banana jacks?

I have a very nice 2000 I'm repairing (just needs electrolytic, thankfully no damage to the PCB at all yet). Everything is great except the front jacks are pretty beat up. I was looking at trying to swap the front and back panel jacks but there doesn't seem to be a way to remove them without destroying them. I tried drilling a brace for the pin end and slowly jacking it out with a c-clamp, but it just started to collapse. Obviously I started with one of the rear panel ones I will never actually use.

The rest of my front panel is in perfect condition so I'm not eager to pay a lot to replace it with another used one just for slightly less used jacks.

[mzzj]

Yet another 2001 here: partly burnt pcb under one of the capacitors and shorted secondary windings on transformer. I'm assuming shorted pcb cooked the transfomer.

PCB is fixed to somewhat working condition to verify that there is no major underlying problems. (needed 2x12v transformer and  2 lab power supplies..)
Only one error on self-test  so I'm planning to rewind the shorted transformer secondaries.

Washing the pcb? I was planning to go with my usual: warm soapy water, after that RO-water, compressed air, IPA and compressed air again.
What do you do with the Rear-Front switch? Remove it or keep the area out of liquid?

Washed the pcb, repaired the corroded traces and burnt FR4, rewound transformer secondaries.
Biggest trouble was dissassembling the transformer, getting the first piece of core steel out took serious persuasion.

Appears othervise operational except Test 412.1 – AC amps switch 
DC calibration seem to be off by about 70ppm compared to my trusty 34401a

[maxwell3e10]

I tried to get a replacement U808 from Keithley, but of course they wouldn't sell it,  nor would they sell a whole A/D card.
So at this point my only hope for fixing the meter was to pray to the T&M fairy. And... the fairy appeared in the form of TiN himself with a replacement chip!  After popping in the chip the meter works great with no diagnostic errors. The picture was of course timed during warm up, but generally it seems quite close to specs.

[IanJ]

Appears othervise operational except Test 412.1 – AC amps switch 
DC calibration seem to be off by about 70ppm compared to my trusty 34401a

My 2001 has same error (after repairing a bunch of others).......if you get to bottom of it let me know.......i probed around for a bit and couldnt see much wrong per service guide.

Ian.

[IanJ]

Appears othervise operational except Test 412.1 – AC amps switch 
DC calibration seem to be off by about 70ppm compared to my trusty 34401a

My 2001 has same error (after repairing a bunch of others).......if you get to bottom of it let me know.......i probed around for a bit and couldnt see much wrong per service guide.

Ian.

Update, I have fixed error on test 412.1 - AC amps switch.

Broken track between Pin 6 U320 (ACA) and pin 6 U510.
A temporary link and now I get no errors but I guess I will have to nail the exact location down or at least make sure C408 is still connected properly at U510 end.

PS. I had many 4xx.x errors prior to this last error, but fixed them the other day, a broken DATA connection to U530 was stopping the dual DAC from working.

No errors on this K2001 now....whoop whoop!

Ian.

[maxwell3e10]

It's important to keep the internal black plastic cover over the analog section closed when testing. Otherwise there is a lot of drift.

[mapleLC]

Is there a way to plug in a Keithley 2001 so that I can see if it works beyond the screen being broken? 

I don't have a replacement screen yet and I don't want to invest unless the rest of the meter is functional.

Is there a way to do this?

[macboy]

Is there a way to plug in a Keithley 2001 so that I can see if it works beyond the screen being broken? 

I don't have a replacement screen yet and I don't want to invest unless the rest of the meter is functional.

Is there a way to do this?

It has GPIB and will respond to basic queries like IDN? Other commands are documented in the manual. It has no serial port; the DE9 is not serial but a digital GPIO.

[mapleLC]

It has GPIB and will respond to basic queries like IDN? Other commands are documented in the manual. It has no serial port; the DE9 is not serial but a digital GPIO.

What about some kind of generic screen that shows what would have been shown if the broken screen worked?  Or at least capture the raw signal data that would says "yes, if a screen was here, it would be getting the right data"

Its a stretch to ask, but who knows

[voltsandjolts]

It has GPIB and will respond to basic queries like IDN? Other commands are documented in the manual. It has no serial port; the DE9 is not serial but a digital GPIO.

What about some kind of generic screen that shows what would have been shown if the broken screen worked?  Or at least capture the raw signal data that would says "yes, if a screen was here, it would be getting the right data"

Its a stretch to ask, but who knows

Well, here is one way but I doubt you'll like it, as it is quite some work.
In this <thread> the OP has linked a decoder he wrote for sigrok/PulseView.
It shows a pictorial view of an lcd using data decoded from the logic analyser.
I think that is very cool and it could be fun to do something similar for the K2001 display protocol.
But my idea of fun and most other people's idea of fun are likely different

[voltsandjolts]

https://www.eevblog.com/forum/projects/keithley-2001-display-substitute/

Some display info...
https://xdevs.com/fix/kei_dd51/

And worth reading...
https://xdevs.com/fix/kei2001/

[mapleLC]

Thanks, I read that display info on xdevs.

I thought there might be a simple way to verify it works before investing in the screen.  Can anyone think of something easy that at least indicates its still works otherwise.

[fma]

Hi!

I found a dead Keithley 2001 DMM in a trash, with a burned transformer.

Before I start to buy capacitors, diodes... I would like to know if I can find a transformer; does anyone have a used one to sell?

Thanks,

[fma]

Is there a way to plug in a Keithley 2001 so that I can see if it works beyond the screen being broken? 

I don't have a replacement screen yet and I don't want to invest unless the rest of the meter is functional.

Is there a way to do this?

You can try to communicate through GPIB or RS-232 lines...

[ps]

Is there a way to plug in a Keithley 2001 so that I can see if it works beyond the screen being broken?

The main board communicates with the display board via a serial interface. Connect your Saleae protocol analyzer, run the self-test, and decode the captured byte stream. Assuming ASCII characters are used, you should be able to read "All tests complete" at some point in the decoded data stream.

[mapleLC]

The main board communicates with the display board via a serial interface. Connect your Saleae protocol analyzer, run the self-test, and decode the captured byte stream. Assuming ASCII characters are used, you should be able to read "All tests complete" at some point in the decoded data stream.

What is a Saleae protocol analyzer?

[AVGresponding]

The main board communicates with the display board via a serial interface. Connect your Saleae protocol analyzer, run the self-test, and decode the captured byte stream. Assuming ASCII characters are used, you should be able to read "All tests complete" at some point in the decoded data stream.

What is a Saleae protocol analyzer?

These:  https://www.saleae.com/

[mapleLC]

The main board communicates with the display board via a serial interface. Connect your Saleae protocol analyzer, run the self-test, and decode the captured byte stream. Assuming ASCII characters are used, you should be able to read "All tests complete" at some point in the decoded data stream.

What is a Saleae protocol analyzer?

These:  https://www.saleae.com/

Too modern.  I like old junk lab gear.  Do the old ones do the same thing like a scope does or is this next gen?

[fma]

Hi!

I removed the caps. They all leaked, but only one (C117) made a mess. A big mess! The PCB has been damaged on both sides...

The cap continuity is ok (still connected to the diode), but there seems to be a via between the cap legs.

Does onyone know where it should go?

I guess this is a 4 layer PCB?

[mapleLC]

Has anyone cleaned up a really dirty face plate?  This one seems a bit pitted up, like would happen to a headlight.

Has anyone tried those headlight treatments on the faceplate or is there a common approach to this?

[mapleLC]

Has anyone cleaned up a really dirty face plate?  This one seems a bit pitted up, like would happen to a headlight.

Has anyone tried those headlight treatments on the faceplate or is there a common approach to this?

Also, forgot to ask.  Is there a Keithley logo out there to replace the faceplate one that is damaged?