[VID ADDED] Keithley 2000 Multimeter - 101.2 test fails

[baltersice]

Hi!
I bought a broken Keithley 2000 and was hoping for an easy fix since the repair manual and full schematics are available now.

Problems: In lower DC voltage ranges the voltage reading drops below 0, slowly sinking up to ~-400 mV, Ohms always reads overload, Self-test 101.2 fails.

Things I already did: Previous owner told me about the 101.2 Fail. The meter basically compares the ADC output of a 0V Test (101.1) with the ADC output when converting the 7V reference voltage. So I ordered a new 7V reference and the multiplexer which is responsible for directing the VRef to the ADC and replaced them without further thinking about it. It didn't do anything though

My current hot scent leads me this part of the circuit:
If I touch anything of the red marked part with a high impedance finger or scope probe, the voltage reading gets much closer to 0 but as soon as I let go, it starts dropping again.

So what I'd really like to know: How are the gates of those 2 big Mosfets Q101 and Q102 ever going to get any voltage? Because I don't get anything on the scope and looking at the schematic I'm not surprised ...

But to be honest: Probably every piece of information whould help me out because I don't understand that part at all



To loosen your tongues I made a quick 4K vid here: ... Probably will take another moment to process tho ...

[baltersice]

Small Update:

The 2 big mosfets from my previous post are for input protection and are opened on overload condition. I can measure my imput voltage on pin 2 of Q102, so I don't care how that opto-isolated part of the circuit is able to provide voltage to their gates.



I poked around some more along the signal path and desoldered every one of the FETs (red) one by one. After every removal I checked whether the voltage would still drop, and it always did. Is it now safe to assume that the error is somewhere in the green section?

[baltersice]

 

The voltage slowly dropping is normal because of the high input impedance. Now I furiously soldered around in my K2000 when I actually just should have checked on test 101.2 fail 

https://www.eevblog.com/forum/testgear/keithley-2000-user-help-request/msg602783/

[baltersice]

Ok, now about that 101.2 Fail: In the previous test 101.1 it lets the adc convert 0V and memorizes the counts. Then in 101.2 it puts the reference voltage from the LM399 through to the adc input and subtracts the counts from 101.1 from this vref reading.



There seems to no way of displaying the result of that calculation, so troubleshooting is not easy. I hooked up another voltmeter to adc input and got 0V in test 101.1 and then 7.11V in 101.2 which is exactly VRef and it is within the stated breakdown voltage range of the LM399.

I need some way of seeing the adc counts which are expected to be 76275970 with 1800000 tolerance but since it isn't just a standard part I have no idea where to look

[picburner]

Select "STEP" for test mode and press "SHIFT" to see the expected value.
 

[baltersice]

Oh that sounds useful if you meant the measured value. Thanks a lot for this! Why dont they put it in the manual?

152523541 <- Test 100.1 gives me
153661550 <- Expected Value from Manual
    1200000 <- Limits

So that passes alright

152524749 <- Test 101.1 gives me
234298675 <- Test 101.2 gives me
  76275970 <- Expected Value (101.2) - (101.1)
    1800000 <- Limits
  81773926 <- Result of my (101.2) - (101.1)


But every measurement is perfectly accurate now, so I guess if you are unlucky enough to get an LM399 at the edge of its stated breakdown voltage, you'll have to live a 101.2 fail :\

[6thimage]

I wonder if they might have relaxed the limits in later firmware versions - from your video, its running a02, with the latest being a20. Unfortunately, I can't find a change log for the firmware versions, so updating the firmware might invalidate any calibration data the meter has.