Restoration glory of Keithley 2001 DMM

[TomThaMan]

Hi all,
Happy owner of 2 Keithley multimeters which I bought recently. 1 is a 2000 which is 100% functional, no self test errors. The second one is a 2001 which has substantial damage done by leaky caps. Somebody already did a repair attempt to it, but its definitely not 100%.

After some investigation this is where its at at this moment:
It does power on, let me take measurements, seems on the low voltages ranges its pretty accurate. Not really possible for me to check its accuracy anyway. It does jump up in voltage for no apparent reason once every few seconds for a very short time. Ohms range is not working well. Shorting the probes gives good readings, but anything above a few ohms sometimes give me an ok reading, but most of the time its al over the place.
Its missing its mains transformer. Probably it got a shorted turn, and got tossed away at some point in the past. Now it is running from an external transformer which is putting out the following voltages:
9.3v and 35.0v to analog board, and 9.5v to digital board.

I noticed that J1025 is not used atm. This is for the voltage selection. Do you guys think that this can give issues? Maybe the circuit is trying to sense and switch all the time causing issues?
I started to take some measurements but when I saw that the voltage at one of the avalanche diodes was at around 48v I stopped and thought it would be better to clean and replace all obvious problems first.

Failing self tests:
200.1  No A/D communication, noisy A/D
200.6  U808 Malfunctioning
200.7  U808 Malfunctioning
201.2  Cannot measure 7v at A/D
201.3  Cannot measure 1.75v at A/D
300.2   A/D Mux 1.75V Ref,  x5 gain
304.6  70nA Ohms source
304.7   5nA source

What I found inside:

Analog board:




A/D converter board, running rev A06




Power supply caps replaced (not by me). Notice the bodge wire running under it. If you look closely you can already see the horror that happened between those 2 caps.




Burnt hole in PCB, and some surgery already done (not by me)




Bodgewire running from U510 all the way to U320





Corrosion around c108




Wierd pin on U626, original or not?




Half eaten away resistors near C633




C623 and c624 leaky



So lots of rework needs to be done. First of al I'm planning to change all caps, also the ones that already have been replaced. Just so all caps are fresh, and the same brand/type (I ordered Panasonic 105c ). I also want to put in the original transformer. Quoted Tek via parts request on the website, but no response so far (about a week now). Called them, and they said it is available for 220 euro shipped to the Netherlands (where I live). If somebody is willing to sell one for less than that, I'm happy to hear.
Lastly, I want to thank TiN, and definitely all the other who contributed to this magnificent work!!!

Thanks,

Tom

***Having trouble getting the pics online, will fix that tonight***

[TiN]

You can upload pics on my ftp, i'll mirror them for you.

[TomThaMan]

I actually already uploaded them to your site this morning. Its in the folder with my username in it. Thanks!

[TiN]

Wow, pretty nasty
Remove all caps, all damaged components and let's see what we dealing with.

[TheSteve]

Wow, that one is going to need some serious TLC. It must have spent many hours powered on with the caps leaking away.

[TomThaMan]

Thanks TiN!!

Does anybody has a spare transformer and willing to sell it?

[ManateeMafia]

There is also a bracket that fits over the transformer. Don't forget to get one of those, it is used to mount the transformer to the case.

[TomThaMan]

thanks for the info, that one I already have.

[PedroDaGr8]

Lego master give us another gem. I can see you tried many things on many legos to find out these tricks.
Thanks for sharing, really. I have cracks in 2400's panel "ears" which made it barely holding as a whole.
And it's bit different than 200x one, so I can't replace it directly. So I'll try to buy that cement to give it a whirl. 

Finally, something I, as a chemist, can contribute instead of just reading in interest and awe. For plastic welding the three most common solvents are acetone, chloroform/methylene chloride or methyl ethyl ketone. Chloroform and methylene chloride are 100% interchangeable and perform the exact same, I will refer to them as methylene chloride from here on out.

Methylene chloride, also called dichloromethane, is commonly used for pvc welding and works on many rigid plastics. MEK is commonly used in certain other plastics that chloroform doesn't work on (such as ABS). I have used methylene chloride to plastic weld a stop pin back on the sliding door of an instrument. It maintained perfect rigidity. In fact, about a year later, someone broke the pin off again, and it broke in a different location indicating the strength of the weld. These are all VERY low viscosity liquids which will seep into any crack. They also evaporate really quickly. To weld the plastic together, coat both surfaces in the solvent and press together firmly for about 30sec. Be careful to not apply too much because you can actually substantially melt the plastic when they press together. Just a thin coat is all it takes. Then let sit and dry for a few hours to ensure complete evaporation of the solvent. You can increase evaporation by heating (methylene chloride is not at all flammable, MEK on the other hand is so heat accordingly).

Link of interest:
A list of common plastics and the solvents that can be used to weld them: http://nerfhaven.com/forums/topic/18527-intro-to-solvent-welding-plastic/

There is a pdf I am looking for similar to above but I cna't find it.

This will allow you to find the solvents much cheaper than the previously linked cement. As previously note: Acetone, Dichloromethane and MEK are the big ones to use. Personally, I haven't had much luck with acetone (not sure if I am doing it wrong) but the other two work AMAZING.

Sent from my LGLS992 using Tapatalk

[TomThaMan]

Nice info on plastic welding! I run my own independent car garage in the Netherlands, and what I usually do with thermoplastics (not thermosets, check that first!!) is weld them on the non visible side by melting the parts together with a nice fat soldering iron.  You can also add material, as long as its the same type. Maybe not the best solution for small stuff like this, but it works very well and the only thing you need basically is a soldering iron. Could help you out in some situations where the correct plastic welding solution is not available atm.

[girish]

I started out with a bunch of 400 series errors. After replacing all the caps and the chips around the capacitors I am down to just one error now.
412.1 ACA switch - AC amps switching error

I changed CR 305 and CR 309 but the error still won't go away. Before I changed the diodes it would pass all tests from a cold start, but now even that doesn't happen.

Any clues?

PS: Thanks Tin for the great post and the wealth of info on this multimeter. I am becoming a Keithley fan!

[girish]

Update: my 412.1 ACA switch error went away for while. I ran several tests after meter warmup and it was still passing all test. But after a few more hours the error is back. I am pulling my hair out. Someone please help!!  I am thinking its temperature related but I don't know how to troubleshoot without a thermal camera.

[TiN]

Perhaps you have leakage for current somewhere in U5xx area on board.
Also check voltage rails on LT1050 near TrueRMS. Make sure they are balanced against DMM LO.

[girish]

Perhaps you have leakage for current somewhere in U5xx area on board.
Also check voltage rails on LT1050 near TrueRMS. Make sure they are balanced against DMM LO.

I will give that section an alcohol bath and see if that helps. Otherwise, I just need to start replacing chips in that section. Or perhaps it could be a damaged track that acts up when the temperature increases? I  am wondering if I need to do the yellow bodge wire that Tom's meter has (Bodgewire running from U510 all the way to U320). Those chips are part of the Amps switching test circuit.

I'll check the LT1050 again. I replaced this earlier too so I doubt its busted. But for some reason the 4.3 zener near it shows 3.99V. I changed that too but the new one shows the same behavior.

[girish]

Bodgewire running from U510 all the way to U320

Hi Tom, what pin on U320 does the bodge wire go? pin #6?

[TiN]

I'd hold off bodging wires if no clear understanding of related circuitry purpose is achieved. Also try with clean IPA swipe first around. Putting whole board into bath likely cause more issues than currently you have. Such things are not something you usually do on 26+ bit analog instrument.

[girish]

I'd hold off bodging wires if no clear understanding of related circuitry purpose is achieved. Also try with clean IPA swipe first around. Putting whole board into bath likely cause more issues than currently you have. Such things are not something you usually do on 26+ bit analog instrument.

I agree. I was exaggerating when I said alcohol bath.  I have swiped the front and back of the PCB with 99.9% IPA without luck.

[TomThaMan]

I would probably do what TiN suggests, but to answer the question, yes, its connected to pin 6.

[TiN]

U320.6 is ADBUF+ node, common input of A/D buffer.
It's side of switch which routes 1.75V Reference voltage to ADBUF.
It shouldn't run to U510 which is switching in AC block.

[TomThaMan]

I have multiple 1.75v related errors, may be that's why... Or maybe someone tried to correct these errors with this bodge. I just ordered the transformer at Keithley, and new caps and diodes are in. Will remove all caps and chips around them which have oxidation on them and clean thoroughly. The transformer will take 3 weeks to arrive.

[TiN]

How much was for transformer. You got from official Tektronix support?

[TomThaMan]

The transformer costs 320(!!) euro shipped. awful lot of money for a simple transformer. Its from Tek. If somebody got a good one for less I'm still happy to hear!

[TiN]

Yah, that's inline what I heard from others. Yikes. Been there before with HP 3458A transformer... 

[girish]

U320.6 is ADBUF+ node, common input of A/D buffer.
It's side of switch which routes 1.75V Reference voltage to ADBUF.
It shouldn't run to U510 which is switching in AC block.

So, I traced the ~1.8V (3 diode drop) 412.1 test signal according to the repair manual to pin 6 of U510 and it should show up on pin 7 through the switch. When my 412.1 test fails pin 7 is always at 0V.
 I cleaned with ISO again around the U5xx chips, re-soldered  U510 and the error went away for a whole day. I ran the automatic test for 24 hrs continuously and it kept passing. Its a little warm here today (room temperature is at 27.56 C and the air coming out of the DMM vent is around 50 C) and guess what the error is back. Wondering if U510 is faulty or some other part after that is pulling the switch output to low.

Excerpt from the repair manual


Test 412.1 – AC amps switch
Cannot measure 1.7V ±0.3V at A/D IN.
The 9.2mA ohms source is turned on to dump current through switch U323 (ACAL
pulled low) to the amps protection diodes (CR305 and CR309). The three diode voltage
drop is routed through U317 (/200uA pulled low), U320 (/ACA pulled low), U510
(REL3 pulled low), U526 and Q516 to the input buffer (Q512 and U520). The buffered
signal is tied to AMP IN.
The signal at AMP IN is then routed to the variable gain amplifier (VGA) and, being
positive, follows the non-inverting x1 path that consists of R530, Q509, Q507, U519,
R531, U509, Q501, and U516 to pin 12 of multiplexer U511.
The output (OUT) of the multiplexer is routed through buffer U342 to ACV/A. The signal
at ACV/A is routed through U320 (/AC pulled low) and applied to the A/D buffer
(U322), which is configured for x1 gain. Measure 1.7V at A/D IN.

[TiN]

Try replacing LM339 which drive those switches. I had bad 339's in two out of 6 units before. Also could be your U510 is leaky/flaky.