Restoration glory of Keithley 2001 DMM

[kada]

SKE encountered with the same errors:

Self test reported these error codes:
200.1   A/D Zero
200.6   x0.02 Line Cycle Integration
200.7   x0.01 Line Cycle Integration
201.2   7V Reference

Did you find a solution?

I'm sorry for delayed report, the solution (TiN, you was right  ) was to update firmware to A08 and I don't have these errors anymore.

[AF6LJ]

This has been a cool project to follow   

[connor]

My 2001 have this problem ,I donot know how to fix it???

[Signal32]

My 2001 have this problem ,I donot know how to fix it???

The err indicator is on ... is the front panel working ? Navigate to the list of errors and tell us what they are.

[connor]

My 2001 have this problem ,I donot know how to fix it???

The err indicator is on ... is the front panel working ? Navigate to the list of errors and tell us what they are.

Error ID=513  AC calibration data lost.

[Signal32]

Is that the only error or are there more ? Use <  > to switch between them.

[connor]

Is that the only error or are there more ? Use <  > to switch between them.

This error is not the built-in test error.when the power on ,it will show this error and then turn to the display page like the first picture. This problem can not run built-in test.

[TiN]

Just run AC calibration. It does not need any external equipment.
Do you have any other errors from self-test?

[connor]

Just run AC calibration. It does not need any external equipment.
Do you have any other errors from self-test?

Thank you ,I will run AC calibration later .but now the 2001 have a big problem ,the panel display  "-------" ,like my first picture,when i run self-test the 2001 will crash, i must reboot it . 
the +5V ,+15V ,-15V ,+38V ,-38V is OK.

[TiN]

?rash? What you mean by crash?
If its turns off or hangs, turn the unit immediately off and read this thread from the first post

------- is default state if measurements system is not armed and waiting for trigger.

[connor]

This 2001 made in 1993 or 1994.
I have read your blog,i have replace all electrolytic capacitors ,except the 15000UF/16v.
I replace the LM311,NE555 and fix the power problem.
The crash means the display have no changes when i press any keys. Always display -------,How to jump out this status? Thanks.

[TiN]

Even MENU button?
Is there any electrolyte on the boards or trace damage?
What is digital 5V voltage on digital board?

[connor]

The MENU button is ok.
There is a little  electrolyte ,i have clean it ,no trace damage.
digital board 5V is OK.

[TiN]

How about self-test result then?

[connor]

Sometimes when the power on,the "ARM" will show on the top ,i can run self-test ,the result is ERR 200.2 ,200.3.
 sometimes i reboot it ,the "ARM" will not show there,the display always "-------",the self-test can not run finished ,the display will stop at any page,then all the buttons are not work.

[TiN]

Check your digital board or/and transformer and mains circuitry. Processor feeling sick, hence the hangs and communication errors (200.x)

[connor]

Thanks ,i will check it .
Is it necessary to replace the 15000uf/16v on the digital board ?

[McBryce]

This may sound like a very stupid suggestion, but is there any chance that the trigger (or some other) button on the front panel is jammed on? I had a very similar "facepalm inducing" repair a few months back (not on a Keithley though).

McBryce.

[connor]

This may sound like a very stupid suggestion, but is there any chance that the trigger (or some other) button on the front panel is jammed on? I had a very similar "facepalm inducing" repair a few months back (not on a Keithley though).

McBryce.

Thanks,I check it already when this problem happened.

[Le_Bassiste]

Sometimes when the power on,the "ARM" will show on the top ,i can run self-test ,the result is ERR 200.2 ,200.3.
 sometimes i reboot it ,the "ARM" will not show there,the display always "-------",the self-test can not run finished ,the display will stop at any page,then all the buttons are not work.

i have one KEI2001 that had similar symptoms like yours: on cold-start (meaning that it was _cold_ on start-up, left off for a day or so), it came up with standard boot screen but then repeatedly ran into and out of DC OVERVOLTAGE, although range was set to AUTO and inputs shorted. on top of that, REAR annunciator was blinking with slow and irregular patterns. of course, self tests failed with similar results like yours did. these symptoms went away after about 10 mins of warm-up, and then the meter worked fine. 
ribbon cable that connects digital and analog board had flakey contact inside the connector on analog board. replaced connector with 0.100" 20pin two-row male header and made completely new cable. all good now.

[TiN]

Good point from Le_Bassiste! Could be very possible, as that cable is the only interface between digital brains and analog world. Optocouplers near cable are another component worth checking.

[lukier]

Finally, I managed to repair my 2001, I think.

My meter had most of 200.x and 200.1 errors. Leaky old caps on the digital board (replaced) and somebody replaced the caps on the analog board around 2004-2007 (replaced those as well and the fan too). All the supplies seemed ok, but the errors were still there.

I found that resistors through which +-15V is supplied to the NE5534 on the ADC board are broken. One burned badly, leaving tar on the PCB, second one failed on high resistance/open. Replaced these resistors as well.

I've tried running comprehensive calibration, but it failed during open inputs step with error 354 = "200 mV offset out of spec". Then I've started chasing red herrings, replaced diodes in the power section, replaced the shift registers, some of the DG switches, NE5534 on the ADC board, desoldered the optoisolators and tested them on a protoboard - nothing helped. After all this messing around things got even worse and a dozens of new errors appeared.

But with TiNs help and some legal tender I got a new ADC board. With that I was able to run the calibration procedure with some poor artifacts (linear psu, 20k and 1M from a DIY decade). Things improved massively and I was only left with the errors that kada and SKE had: 200.1, 200.6, 200.7, 201.2 and turns out these can be fixed with upgrade to A08 (from A06). So once I got an UVC lamp (only had UVB) and a special PLCC44 adapter (the one I got with TL866 didn't work with 27C4096) I've upgraded the firmware and self tests passed.

I've started measuring some resistors that TiN kindly gave me, along with measurements from his 3458A, and I've noticed completely bonkers measurements on the megaohm ranges.

Now I wanted to improve things a bit, peroxide treatment (but the 25 year old paint on the display lens didn't survive kapton tape :/) worked nicely and while I was going to order DS1245 for MEM2 option I've ordered slightly better calibration artifacts, because I don't have anything top notch yet (multi-LM399 or LTZ1000) from Farnell: Y006220K0000T9L (20K 0.01 1 vishay-ppm/C), PTF651M0000BZEK (1M 5ppm/C), two ADR01BRZ to put in series and LT5400BCMS8E-8 9K/1K divider to divide ~20V to 2V.

And with these the calibration failed! Error 385, 20 MOhm 2W gain out of spec. High ohm resistors are a nuisance so I figured out that my soldering spree panic on the analog board left a mess affecting ohms current sources. After cleaning everything under a microscope, every guard trace, resistor cases, with dozens of cotton swabs and isopropanol the calibration worked!  I could even measure 1GOhm resistor.



Important lesson here is that self-test depends heavily on the calibration constants (which may be wrong or corrupted) and even these elaborate self-tests don't test everything that comprehensive calibration procedure tests. I still wonder what killed the resistors on the ADC board and worry that there might be something more wrong (a lot of the stuff is connected to +-15V rails).

I calibrated K2001 against my 3457A, as it seems most stable and extremely well aged (this one was closest to the resistance measurements done by TiN). Also, I really like this meter as it has most functions quickly available from the front panel, has offset compensation, 3 GOhm range, some autocalibration, no VFD but then I don't have to worry and it often runs 24/7 (also no fan) and does 7 digits in math mode - averaging. Unfortunately, it is not otherwise volt-nut friendly due to 3V front end arrangement. Transfer accuracy from a multimeter to a multimeter using some random voltage references wasn't that great, but none of my meters are calibrated anyway, it was more of a sanity check. Funny thing, when the GPIB cable is connected I get 104.1 error (GPIB handshake) but GPIB seems to work otherwise (well just *IDN? from ibtest, need to test more, luckily I have some NAT9914 chips at hand).

Now the meter occupied the last available space on my workbench and I want to build a more stable reference (multi-LM399 or maybe LTZ1000) and check how stable the meter is, drift, tempco maybe, etc. I also need to test some other functions (AC, current) more in depth to double check if there are no surprises that self-tests don't catch. Performance verification, even basic one, of such meter is definitely not an easy task.



Then if it turns out it is roughly OK I'll have to probably calibrate (low level calibration) this meter. I don't feel like spending more on the calibration than on the meter, also Tektronix has been annoying me for a long time, either once I wanted to buy 2000 benchkit parts and they gave me silly lead times, or other times I had to create an account there, as I was downloading TLA stuff or I wanted to download a datasheet for an SMD oscilloscope probe that I got off ebay. It was enough to trigger their marketing department and I started getting annoying phone calls. Now they stopped calling, but I still get emails telling me to buy TBS2000 or other crap - they seem really desperate. Anyway I couldn't even find UKAS Schedule of Accreditation for Tektronix (not that I need certificates, I was just curious on the DC 10V specs, Keysight UK has 0.8ppm). I'll ask RS Calibration for a quote. On their price list it says 62 GBP for a bench meter. They might not do 7.5 digit, but at least they could transfer whatever they have currently in their Fluke calibrator - better than nothing. I hope they have GPIB scripts for K2001.

Overall, I quite like this meter. Maybe not as much as 3457A, as I have to go through multiple menus and the average display from readings stored in a buffer are using engineering notation with only 4 digits after the decimal point, but on the positive side it has big VFD, sensible ranges with GOhm impedance up to 20V, thermocouples and similarly to 3457A offset compensation as well as scanner card option.

I would like to thank the forum members for helpful posts here and TiN in particular for providing me the ADC board, as well as putting a lot of effort on writing articles about his repair adventures - it probably takes as much time to write and take pictures as it takes to repair some things.

[TheSteve]

How about credit where credit is due? I mentioned this technique in TiN's K2000 thread here. I came up with this crazy idea one day, decided to try it on a Keithley 199, and it worked like a charm. I am certain that I am not the first to do this but I hadn't heard of anyone doing it.

Received an Agilent device today with an extreme dent in the front face - very extreme! I've used heat before to form plastic and it has worked so/so. Thought I'd give it a go tonight when I remembered this thread.
I think it turned out very good. The heat was probably a little bit high from my heatgun but there was never a chance to make it look 100% original. The main goal was to get it back into shape so that when a rubber boot was put on it was completely hidden. With the original dent it was very obvious and looked terrible even with a rubber boot in place.

[Zucca]

Thanks lukier! Congrats!

One day I will write a post like you, my K2001 is still under my bench  .

A couple of little questions...

the one I got with TL866 didn't work with 27C4096

Why? Not fitting well... made in China crap?

but the 25 year old paint on the display lens didn't survive kapton tape :/

So the kapton tape peeled the paint off?

[lukier]

Why? Not fitting well... made in China crap?

No, different routing of pins and AFAIR TL866 software doesn't support 27C4096 in PLCC44 packages, just DIP40. So to mimick DIP40 one needs adapter like this:

http://www.ebay.co.uk/itm/111553438268

So the kapton tape peeled the paint off?

Yup. Maybe I should've ignored that and just apply peroxide without worrying about the paint. I think the display lens can be removed entirely as well.