Restoration glory of Keithley 2001 DMM

[SKE]

Sorry, my bad 201.2 test. I measure it again...

U317 after 201.1 test (201.1 on display)
14 -0.018mV
15 6.917V
16 4.9423V

U317 after 201.2 test (201.2 on display)
14 6.917V
15 6.917V
16 0.0253V

[TiN]

OK, mux working fine.
Your output from U322 (which goes to AD_IN thru 1K resistor R284 on bottom of analog board) with 201.2 on display should be same 6.917V.
If it's not then gain switching made up with U318,U319 and resistors around them might be sick, or U322 itself could be sick.

[SKE]

Thank you for the advice.

Output voltage on the U332 pin 6 seems to be wrong.
U318 and U319 has been replaced before. New parts are DG211DY from mouser.

[TiN]

Mm, if you have reference voltage on 201 test, then result should be pass. Let me measure voltages for you, seems not 100% right.

[TiN]

OK, here are results for your reference:

200.1 - AD_IN = 0.044 mV, Pin.2 of U322 = -0.012mV, REF (pin.1 AD board) = -10.2893V
all othe 200.x tests about same voltages, so it's mostly internal test on A/D board using U80x registers and CPLD.
Check you voltage supplies (+15,-15,+5) to make sure they are not noisy. If I artifically cause shift on VREF even 0.1mV - test will fail.

201.1 =  0.137mV, SCOM = 3.491mV, pin2 u322 = -0.020mV, REF = -10.28967V, pin.6 U322 = -8.467mV
201.2 = 6.88303V, SCOM = 3.484mV, pin2 = 6.88364V, REF = -10.2897V, pin.6 U322 = +7.994V
201.3 = 1.72111V, SCOM = 3.489mV, 1.72094V, -10.2897V, pin.6 U322 = +1.9922V

All measurements taken relative to black wire LO terminal. I have about 8mV pk-pk noise on +15V, 3mV pk-pk on -15V, 12mV pk-pk on +5V going to A/D board, measured by other 2001 using VAC PEAK mode.

[SKE]

Measured with keithley #1 7.5d high-accuracy, no filter
x - mean unstable digits

201.1
SCOM  from -0.2 to 0.3uV (slowly change, it's difficult to measure)
U322-2   -0.034xx mV
U322-6   -8.550xx mV
-VREF     -10.316239 V

201.2
SCOM     0.012xx mV
U322-2   6.91693x V
U322-3   6.91693x V
U322-6   9.03565x V
AD_IN    6.91701x V
-VREF    -10.31625x V

201.3
SCOM      0.006xx mV
U322-2   1.729238x V
U322-3   1.729205x V
U322-6    2.002529x V
AD_IN     1.729306x V
-VREF     -10.31625x V

200.1
SCOM     0.005xx mV
U322-2   0.06xx mV
U322-3   0.06xx mV
U322-6   -8.548xx mV
AD_IN     0.036xx mV
-VREF   -10.31627x V

Oscilloscope in peak detect mode, passive probe 1:1

200.1 measure with keithley 2001 #1 VAC in peak-peak, speed-normal, no filter
I don't think this reading is giving me any usable data

+15V 11.1mV p-p
-15V 6.3mV p-p
+5V 12.6mV 9-p

Oscilloscope in peak-detect mode
Noise floor mean short end of ground pigtail to probe tip.
Length of ground pigtail is cca. 15cm (I now, it's too much)

[SKE]

After my last post, I started thinking that it was the ripple in +15 and -15V which was causing of the problem.
Therefore I measured the ripple on pins +-15V voltage regulators and I was surprised with what I found.
With no signs of 100Hz ripple in power rails I tried to connect GND pins on C101 and C104 together (see photo).
The ripple on ADC and +-15V testpoints looked the same as in working kei2001 #1.

In the previous measurement I had to use the long GDN wire as I had nothing else to use at that time. Sorry I forgot to mention it in my last post.

Then I disconnected  AD_IN pin from buffer and connected it to S_GND. When I ran 200.x test, the test failed, showing the same errors as before (200.1, 200.6, 200.7) whereas on the working kei2001 #1 the test passed with no errors.
In that case the problem should not be in the voltage offset on AD_IN from buffer.
Is it possible that these errors could be caused by wrong calibration data except power supply for ADC?

(The analog board is revision K, firmware version of digital board is A5.)
Would it be possible to connect digital board from working kei2001 #1 with FW version B07 to this analog board?

Which is the latest FW version  I could use for this analog board revision?

Do you think I should update FW in faulty kei #2 and in working #1 too?

Many thanks for your help

[Le_Bassiste]

I recently acquired a used 2001 that was previously restored by TiN.  It is now spuriously throwing 201.x errors, starting at 201.1, when cold started, with voltage display drifting up to OVERVOLTAGE message. After about 10 mins of warm-up, a power-cycle somtimes brings it into error-free status and it then functions normally (that's at least what i guess, as voltage readings against a known good voltmeter are matching pretty well, with a tad of offset in the 200V range).
somewhere else in this thread, a similar fault was described, which was cured by swapping out the ADC board.
So, my question here: could it be the CPLD gone wild with corrupted config memory? a quick check with scope on pin 23 of the module shows signal stuck at either low or high, when fault occurs. did anybody manage to read out the cpld (i know, probably read protected, don't have programmer to look into it)? And, if all else fails, is somebody willing to sell me a spare working ADC module? any help very much appreciated!
 the VCD is pretty close to leaving the building, although all pixels still give some light. I'm planning to do a VCD replacement board with either TLCD or GLCD and an FPGA on it. Any interest in such a replacement board?
(reason for this project is keithley asking 430 EUR in germany for front panel assembly. Insert bomb emoticon here)

[TiN]

Magnum unit travels around, I see. You may want check your mains (I know it sounds wierd), and primary earth-reference circuits and optocouplers. I saw these errors one other unit when running 50Hz mains. here some test results with AC source. Even with those errors I was not able to see any difference in operation.

I believe SKE tried another A/D and it was still failing, while one from failed unit work fine in another unit.

As of VFD I recall Magnum's unit had bright screen, so not sure now if we talking same meter then. You can always grab 7001 for cheap and swap the glass, if you have steady hand on soldering

[Le_Bassiste]

Thanks 1e6 for the quick reply. Yep, it's the one from magnum. good to have a starting point for further investigations now! will be posting results here when i have time to dig into it, as i'm also travelling atm. i had suspected the AC mains side first, but couldn't find any evidence for malfunction there, didn't check the optos, though. These are quite prone to ageing of their LEDs, so there is  indeed a probability that their current gain has dropped below specs. Avago has a white paper about opto ageing, but i don't have the name from top of my hat. otoh, i'm also suspecting the 5V supply rail on the ADC, because its at 4.8V, which is just 50mV above VDDmin for the CPLD.
will keep you posted!
as for the 7001 display, that paricular unit you are mentioning looks like a taker, but customs into germany is simply forbidding. Thus,  i'd prefer to stay with the bad VFD and save the bucks for a nice ISO calibration, rather than sinking more bucks into parts of unknown condition.

[Magnum]

Sorry to hear that you are having problems with the unit. I had it running without any problems, I sent you a PM.
I swaped the screen with my other 2001, but that shouldn't cause the problem.

[Le_Bassiste]

error 201.1 on cold start seem to be gone, could trace it back to ribbon cable on P1027:3 not always making contact, so DTA_IN on A/D board was missing.
the reading on 200 mVDC with HI-LO shorted (HI ACCURACY, SYNC AUTOZERO) has some offset, something around +2.0 µV after 2 hrs. warm-up, with extremes going between 1.8...2.2 µV.
can somebody please confirm as to whether this is normal behavior due to thermal emf on the inputs?

[Magnum]

I checked with my 2001, same settings:
Min/Max: -1,404uV / -0,868uV
Avg: -1,166uV
StdDeviation: 0,11uV (0,556PPM)

[Magnum]

When playing around with a small tool I wrote to get the values for Le_Bassiste, I found a problem with my 2001:
When measuring the voltage in 20V range and only with 10PLC the voltage first drops and then rises again. Other ranges don't have this problem.
I attached a screenshot. The graph on top is the measured voltage, first 200 samples with 0.1 PLC, then 1 PLC and after that 10PLC. The graph below shows the difference between Min and Max and the standard deviation, both in PPM.
Someone having an idea what might be wrong?

[Kleinstein]

@Magnum: the drop in measured voltage looks strange. Is it reproducable or just a singel event.

It could be a problem with the auto zero function. The AZ reading may not use just one zero reading for each measurement, but could use the average of several past zero readings too.  This somewhat makes sense, as the zero is not changing very fast.

So switching just the NPLC setting could upset the meter und mixing readings for different settings or get one poor conversion to upset more than just one value. So it might be a good idea not to change NPLC setting one the fly, but stop / restart after that.

[plesa]

ebay alert

http://www.ebay.com/itm/Keithley-2001-Digital-Multimeter-7-5-Digit-Parts-Repair-/231797719292?hash=item35f83880fc:g:C-EAAOSw3KFWgtw9#shpCntId

Too late, $500 seems to be fair, repair based on picture will be piece of cake (find 7001 as VFD donor and solder new VFD).

[plesa]

The unit in auction was not the best for your type of modding, it looks like new. So I will put there new VFD and resell it.
BTW check the http://www.go-dove.com/en/index
About month ago there were huge auction where was about 20x K2001 ($100-$150 not working and $300 operational).

[TheSteve]

Hopefully there is nothing else wrong with it - changing the caps should be done no matter what of course.
Values on the 2001 seem good, probably a good time to sell mine.

[michaeliv]

Hi

I recently acquired a broken 2001 that won't power up at all.
After a bit of testing I think the transformer is blown, specifically the red windings on the secondary.
Would someone be able to please confirm going by the resistances ? (Also note the exact model / revision of your transformer ).
I've looked around but can't seem to find where to buy a replacement one.
Would it be possible to replace it with a different generic one if I match the output voltages ( assuming someone can post the AC out voltages ).

I've looked through this thread but can't find the schematic for the mains input circuit ( The one with the relays K101 / K100, Q528, CR100 .. ). Has that not been documented so far or have I missed something ?
One other glaring issue is that R123 & R250 (bottom side, close to power resistor & MOSFET heatsinked to the case) reach 130C within 1 minute of connecting power. I'm assuming this is not normal ?

Thanks!

[ManateeMafia]

The transformer was measured here...

https://www.eevblog.com/forum/testgear/anyone-got-a-keithley-2001-and-want-to-check-something-for-me/

The schematics are listed on page 1 and TiN has a webpage documentating the same repair here...

https://xdevs.com/fix/kei2001/

The part still might be available through Tektronix parts ordering. I have ordered Keithley parts from them over the last 12 months. Shipping is a little high but you know the part will work out of the box.

[Kleinstein]

The resistance values look resonable. So the transformer is likely not blown. There is a small chance to have a shorted turn - but this can not be measured from the ohms readings. Here a test with the secondery not connected can help. In case of a shortet turn, the no load consumption would be way to high.

The transformer has a PE connected shield winding, to reduce capacitive coupled current leakage. So a normal transformer would not give the same performance. Also there are 3 secondary winding so it would be rather hard to find a suitabel transformer.

Getting a MOSFET run so hot, there is likely something very wrong like a short or blown chip to draw way to much current.

[Jay_Diddy_B]

Hi,

Here is the partial schematic of the power supply:




I would expect the resistance red to red/wht and the other red to red/white to be similar. It is a centre tapped winding.

Regards,

Jay_Diddy_B

[macboy]

The resistance values look resonable. So the transformer is likely not blown. There is a small chance to have a shorted turn - but this can not be measured from the ohms readings. Here a test with the secondery not connected can help. In case of a shortet turn, the no load consumption would be way to high.

The transformer has a PE connected shield winding, to reduce capacitive coupled current leakage. So a normal transformer would not give the same performance. Also there are 3 secondary winding so it would be rather hard to find a suitabel transformer.

Getting a MOSFET run so hot, there is likely something very wrong like a short or blown chip to draw way to much current.

Actually it looks like the red wired center-tapped secondary is shorted, his measures 0.75 ohms or less, and the example in the other thread measures about 3 ohms from CT to either side or 6 ohms across.

I 100% agree that a normal transformer will not work here, it must be the special shielded type. Otherwise there will be much too high capacitance to the primary and/or ground, compromising the "floating" nature of the instrument.

[michaeliv]

The transformer was measured here...

https://www.eevblog.com/forum/testgear/anyone-got-a-keithley-2001-and-want-to-check-something-for-me/

The exact model of the transformer was not posted there so I was still holding out some hope .

The schematics are listed on page 1 and TiN has a webpage documentating the same repair here...

https://xdevs.com/fix/kei2001/

The part still might be available through Tektronix parts ordering. I have ordered Keithley parts from them over the last 12 months. Shipping is a little high but you know the part will work out of the box.

Found the input schematic thanks. In the URL you provided I can't find anything relating to troubleshooting the power supply (wrong link ?)

Getting a MOSFET run so hot, there is likely something very wrong like a short or blown chip to draw way to much current.

It is not the MOSFET that is getting hot, it is 2xSMD resistors that are placed in parallel ( 2x20k for a total of 10k ). See attached.

Actually it looks like the red wired center-tapped secondary is shorted, his measures 0.75 ohms or less, and the example in the other thread measures about 3 ohms from CT to either side or 6 ohms across.

That's what I'm afraid of ... I think I'll connect 120V mains to the transformer directly (1-3 / BLK-WHT / 20VAC + 90VAC) and see what happens on the secondary .. would this be a bad idea ?

I 100% agree that a normal transformer will not work here, it must be the special shielded type. Otherwise there will be much too high capacitance to the primary and/or ground, compromising the "floating" nature of the instrument.

So I'll get a noisy meter or something more sinister ? Is there a cheaper device that shares the same transformer with the 2001 ?

[ManateeMafia]

The manual with part numbers is here ...
https://xdevs.com/doc/Keithley/2001/2001-902-01.pdf

The transformer part number in the manual matches what is printed on the side (TR-280). I have had a few issues tracking down new part numbers with Tektronix as some have changed. For example a letter (A,B,C,etc.) has been added to the part.
The manual also lists test voltages. I don't think any real t/s procedure has been written. The current method of repair is to remove all the electrolytic caps around the supplies and inspect for leakage. I have done this for three 2001 and one 2002. At least two had pcb damage to traces.