Restoration glory of Keithley 2001 DMM

[ManateeMafia]

I never received a reply from Keithley on their FW upgrade release notes. There is only one mention on the website which is for version B16

http://www.keithley.com/data?asset=13096

It refers to the addition of an average-responding RMS converter. It also states that it is for serial numbers < 0936521. I wonder if the later (A) fw versions were for a different hw configuration like flash memory or a single prom?

I see that TiN's K1 has the A version and K2 has the B version. Maybe the A08 includes the additional features of B16? Perhaps this is why the calibration failed?

[cyr]

I thought "B" was the later revision, and that late units have the larger flash memory needed.

Is A08 really the latest, or only the latest of the A series?

Anyway, time to upgrade my B06 unit to B15.

[TiN]

Original FW on K2001 sent to cal was A05. They tried calibrate it, but could not do until updated FW to A08.
All my Keithley's are way below 09xxxxx S/N numbers, actually 054xxxx S/Ns

First unit: A06, Rev J digital board with single ROM chip, Rev L analog board
Second unit: B07, Rev K digital board with two ROM chips (Even and odd), Rev L analog board
Third unit: missing ROM chip , Rev H digital board with single ROM socket
Forth unit: A08, Rev H digital board with single ROM socket, Rev J analog board
Fifth unit: A08, Rev G digital board, single ROM
6th unit: same as 5th.
7th unit: A05, Rev G digital board, single ROM socket, Rev L analog board
Keithley 2002: A09, Rev K digital board with two ROM chips (even and odd).

[ManateeMafia]

TiN,

I would love to see the results with a LTZ1000, especially after seeing your reference boards. I am guessing you will be looking at replacing the stock metal foil resistors with lower tempco versions if all goes well to improve your resistance measurements?

I also noticed that they could have used a better AD687 variant for the TRMS converter.

I have currently fixed the self-test errors on my #3 meter, but I have to track down the cause of a 6.9V offset on ACV. All other functions appear to work. Either I put in a bad part from meter #2 or something else near that part has died.

[TiN]

Did you try running AC calibration? (it does not need external gear).

What errors you had in 2001. I have unit on my desk now with 304.x, 306.1, 309.1, 407.x, 408.6, 410.1, 411.1, 411.2 errors.

[ManateeMafia]

The errors that I had were 406.6,407.1,407.2,407.3,408.6,410.1,411.1,411.2.

Fortunately the service manual pointed me in the right direction. I had to follow the SELFTEST and SELFTESTEN outputs from U505 pins 5 and 6. I ended drawing out a basic schematic of the signal flow and found that the levels from U513:9&10 were not going high enough (1.1v) to change U513:8 to low which failed to enable U522:8.

Q518 had failed with a low resistance path to ground.

Do you usually step through each failed test and measure the input at AD IN on the converter board? It would be interesting to see what you are measuring at each of the failing test steps considering some of the failing steps are nearly the same as the ones that pass like 301.1 and 304.2.

I have not tried running AC calibration, but I will look at the cal manual and see if it is possible. The offset is such a gross amount and so close to one of the failing tests I fixed (407.1 unable to measure 7V), that I believe there could be more issues. It is possible someone tried to calibrate the AC ranges and saved a bad offset.

I am currently trying to layout a better diagram of the few signals I traced so that it can be uploaded to xDevs.

[TiN]

I was checking affected blocks by errors , trying to fix lowest number error first, but sometimes it's getting puzzling, as relation between different selfdiag steps not always straight forward.

Here's A/D input voltage with all error steps in manual mode

Passing tests:

201.2 - VREF check = +6.908974
304.1 - Ohm zero = -0.007313
304.2 - 0.98mA and 9.2mA = 0.745667 VDC (spec 0.7 to 0.86V)

Errors:

304.3 - 89uA and 0.98mA = 0.288570 VDC (spec 0.585 to 0.715V)
304.4 - 7uA and 89uA = -0.485134 VDC (spec 0.504 to 0.616V)
304.5 - 770nA and 7uA = raising from -0.2VDC (spec is 0.63 to 0.77 V)
304.6 - 70nA and 770nA = 6.0072++ VDC (spec is 0.63 to 0.77 V)
304.7 - 4.4nA and 770nA = 11.8315 VDC (spec is 1.6V to 2.8V)
306.1 - Ohm cal switch = -0.00975 VDC (spec is -0.001 to 0.001 V)
309.1 - Amps, 200uA = 0.03571 VDC (spec is 0.0801 to 0.0979 V)
407.1 - FE 2V = -0.00518 VDC (spec is 7V +-0.15)
407.2 - FE 200V = -0.00771 VDC (spec is 0.7+-0.2V)
407.3 - FE 750V = -0.00781 VDC (spec is 0.14+-0.04V)
408.6
409.6
410.1 - TRMS = 0.004572 VDC (spec is 7V +-0.18V)
411.1 - Filter TRMS = -0.0046 VDC (spec is 7V +-0.18V)
411.2 - same as 411.1

So obviously lots of stuff is wrong

[cyr]

Interesting enough, old ADC modules I had were Revision G, with date codes on IC's 1992, while new ones are Revision H and Revision J, with datecodes 1992 and 1994. There are few layout differences and extra OPA177GS (U813) near to integrating amp U816 (OPA602AP). I will update schematics soon, to see what was changed.

Did you ever notice any difference in performance between these revisions?

I'm thinking of modding the OPA177 circuit into my older revision ADC board...

[TiN]

Did not actually tested that point, as I did not had stable enough reference to test 2001-grade DMM till last few month.

I will do it as part of comparison with modified meter and stock calibrated 2001.

[ManateeMafia]

TiN,

Have you had to source any of the VN0605T Mosfets? I have borrowed a few from the one that I have been slowly repairing. I cannot find a reliable source for that part.

I have found the 2N7002E part that appears to have similar specs. I am not sure if it would make a suitable sub. Here is a link to the datasheets...

http://www.vishay.com/docs/70212/70212.pdf

http://www.vishay.com/docs/70860/70860.pdf

Do you think this will be a good fit? I am planning ahead to make sure I can attempt to fire up the board when I get the burnt areas fixed.

I did some quick checks on the TRMS circuit and found that the diagnostics work because the test signal on the TRMS input is always on in ACV mode. When the meter is switched back to normal use, the 0.69Vdc signal is still on the input to the rms converter. This is why it shows 6.9xxxx on the display. The input to the converter does change as I select the different manual tests.

I will have to step through the R1_STB and R2_STB registers and see if any of them are wrong.

[TiN]

Got some progress finally as of today.

Measured voltages on ohm current source resistor loads (common terminal of foil resistors R356,R358,R365,R366 and 100Meg R354) and control nodes
from registers (OHM_FA, OHM_FB, OHM FD, OHMCB, OHMCA, OHMCAL, /ACAL).
All these were OK. Mux U325,U332 work ok as well, also ~7VDC was present across resistors as per test points.



Find out that switching JFETs Q320,Q312,Q324 were dodgy. After replacing these from another donor 2001 board (from same locations) - no problem with ohms 304.x tests now.
JFET are TO92 package, PN4392 I believe, selected and remarked by Keithley as TG128.

Measurements with reference to pin4 common on ADC board connector.

Test	PASS unit	FAIL unit	After fix
304.1, U323.7	0.790 VDC	0.744 VDC	0.7854 VDC
304.2, U323.7	0.790 VDC	0.744 VDC	0.7854 VDC
304.3, U323.7	0.635 VDC	0.291 VDC	0.6301 VDC
304.4, U323.7	0.548 VDC	-0.523 VDC	0.5451 VDC
304.5, U323.7	0.736 VDC	-0.523 VDC	0.7331 VDC
304.6, U323.7	7.645 VDC	6.12+ VDC	7.6026 VDC
304.7, U323.7	0.482 VDC	5.807 VDC	0.4773 VDC

Now I have left 407.1, 407.2, 407.3, 408.6, 409.6 (sometimes this one pass), 410.1, 411.1, 411.2

[ManateeMafia]

Great work. I had Q518 almost shorted to ground. I measured ~0.5 at AD IN. If you can measure the 7v there, I would look at Q519.

[TiN]

Also for your question, I think 2N7002E should do fine, as it's just bring common potential in schematics.

[ManateeMafia]

Great work. I had Q518 almost shorted to ground. I measured ~0.5 at AD IN instead of 7v. I believe R233 connected to Q518 through U531:18

[ManateeMafia]

Sorry about the double post. My ipad is not cooperating (always blame the equipment). I don't think Q519 will apply and I tried to delete that info.

[TiN]

Changed Q518, but nothing improved
Will debug in more depth later on.

[ManateeMafia]

I found the issue with my 2K1 #3 with the offset of 6.9V on the AC ranges. K503 had an open winding. I am not sure how it was caused but when I removed it, one of the pins tied to the coil came free from the body. The meter still passes self test and now the AC range seems to be functional.

I can now get back to repairing the burnt area on 2K1 #2.

[TiN]

Good job. As of mine, i found that problem is somewhere before function mux U511, as opamp U520 seeing wrong voltages on its inputs when running tests 407.x.
Its nice to have another 2001 for cross checks.

So far Q518, Q519, Q507, Q509, current source Q543 (2 leg metal can near dual JFET) are all ok, and not cause of problems.

I can make backlit hires photos of affected area , if you think it can help ur repair.

[ManateeMafia]

Thanks for the offer. I already have some pictures with both boards when I removed the caps. The best pictures I have are from my ipad.

If you already have some pics, I would be happy to see them.

Your description of components tested are the same ones I checked. The big difference is that my op-amps had good inputs. Nothing seemed bad until I probed around Q547 and found the open coil.

[TiN]

Here are some photos in case people use this as reference in future.

Fix for ohm 304.x failing on analog board #4



Replace TG128 selected JFETs Q320,Q324 (near blue resistor network, on left side) and Q312 (near caddock network).
After doing so 304 errors are gone, resistance measurements fixed.

Now recap (remove old capacitors and replace with new ones). I used 1000uF 50V instead of 1000uF 35V for C116 and C117 positions. They are same diameter and pitch, just taller.



Remove analog board from chassis, prepare caps.



Remove all capacitors and clean board. Make sure no damaged traces or electrolyte corrosion anywhere.



Closeup on 4 caps which usually leak most on old 2001's.



Backlit photo to show traces in inner layers under capacitors area.



Backside backlit photo



Clean holes with toothpick, so you can put new capacitors in place.



Soldered caps on analog PCB.



1000uF 50V capacitor in mains switching circuit.

Also while trying fix 407.x errors and swapping components seems I got backwards and borked something else, as now selftest steps 402.1, 404.1 to 404.5, 405.2, 405.4, 405.6, 405.8, 406.6 are failing too!
Oh-oh, time to spend some time and recover schematics for all AC-measurement related circuit.

[macboy]

TiN,
Photos are not showing on your most recent post.
Photos work now.

[ManateeMafia]

Thanks for posting the pics. I will try and get a basic schematic of what extra information I found in the AC section. I can send it in an email in eagle format. Hopefully you can add it to your existing drawings if possible.

I found a source for replacement COTO relays. They are not cheap but I should have a NOS relay in a week.

[TiN]

Better in PDF format, I don't have eagle, sorry.

Here's what I got so far at schematic.

[benvei]

Hey there!

I've got one of these multimeters from an old lab which was working fine, but I noticed a klicking relay, so I took it apart, and looked for the klicking relay.
It was somewhere on the input voltage selector circuit, so I hooked up my Multimeter to the Capacitor on that circuit and noticed some ripple on the voltage.

To be sure that the ripple is related to the klicking relay i hooked up my oscilloscope to the bridge rectifier and the alligator clip to the negative side of the capacitor in the voltage selection circuit and ... yeah. I did watch Dave's Video about how to not blow up my oscilloscope, but I didn't do any measurements on that part. However, i did blew the voltage selection circuit. Luckily i didn't blew my oscilloscope (since i'm a sencondary school student, i can't really afford blowing up equipment), but i blew up the bridge rectifier and a trace to the capacitor.. I replaced it, and hooked it up to my bench power supply, and a mosfet started smoking.

However, i would declare the voltage selector as dead. I thought of connecting the transformer directly to the power switch, but I wanted to ask here if its safe to do it. I noticed that there is a optocoupler on the voltage selection circuit going into the multimeters circuit, what the reason for this? Also, what are the pins on the transformer cable for 230V mains?

I also don't really know how I shorted it out. Even if we talk about a mains earth referenced power supply, i hooked up my alligator clip to the negative side of the capacitor, so how could that happen?

Thanks!

[TiN]

Which capacitor you talking about? C108?