Author Topic: Repair of a Keithley 6517A; drifting V-Offset and I-Bias-Offset  (Read 1207 times)

0 Members and 1 Guest are viewing this topic.

Offline voltampere

  • Regular Contributor
  • *
  • Posts: 58
  • Country: de
Repair of a Keithley 6517A; drifting V-Offset and I-Bias-Offset
« on: January 22, 2019, 04:56:55 pm »
Dear community,

i am struggeling with my Keithley 6517A that i bought on ebay.

The main problem is offset drift in V-Offset and I-Bias-Offset.

The unit will pass startup tests without reporting faults.
On startup it shows main firmware REV C05 and display firmware REV A02.
The highest date code inside that i could find is 07/45 so the unit should be manufactured in 2007.
The last calibration according to CAL constant was Feb/2008.

The Voltage Source seems to be correct. I checked some cardinal values from -1000V to +1000V and the output is stable and pretty accurate.

My troubleshooting so far:

I checked all power supplies.

All values according table 5-3 of the service manual:

Check point                  Required value          Measured Value

Reference to Common K
Q305, pin 1                  +5V, +- 5%         +5,03V
U301, pin 1                  +12V, +-5%         +12,12V
U302, pin 3                  -12V, +-5%         -12,00V
+120VK (VR301 anode)            110V to 130V         +124,0V
-120VK (VR302 cathode)            -110V to -130V         -125,1V
+1200VK (VR303 anode)            1100V to 1300V      N.A.
-1200VK (VR304 cathode)            -1100V to -1300V      N.A

Reference to Common B
+5VB (U408, pin 3)               +5V, +-5%         +4,97V
-5VB (Q401, pin 3)               -5V, +-5%         -5,06V

Reference to Common S
+25VS (C461 +)               22V to 28V         +30V
-25VS (C460 -)                  -22 to -28V         -30V
+5VS (U207, pin3)               +5V, +-5%         +4,94V
+220VS (C401 +)               200V to 240V         +250V
-220VS (C430 -)               -200 to -240         -249V
+15VS (U206 pin 3)               +15V, +-5%         +14,90V
-15VS (U205 pin 3)               -15V, +-5%         -15,23V

These seem to be all good.
The +-25VS and +-220VS are a bit high. But these are unregulated and i saw two other repair projects ( TIN and another guy, i forgot) where these Voltages are similar high.
I also checked the PS of the VFD on the board after the transformer. Above 60 Volts and looks clean on the scope.

This is how the failure looks like:
When i apply a voltage to the input and release Z-CHK ( zero check ) then the displayed value will ramp op ( or down ) slowly with a rate of around 1 mV per second.
If i let it happen then it will settle after a couple of minutes.
The same happens when i short the input with a triax-short. It will not always settle at the exact same value and it somehow depends how often and fast i „trained“ the relays by switching Z-CHK on and off in short sequence.
Because of this behaviour i cannot perform the V-Offset calibration.
It will usually stop after some time with Error 413 -> „Voltage offset not converging“.
Very seldom it will stop with Error 421 -> „Voltage Offset out of spec“.
I managed only one time to succeed with V-Offset but the next step after that is always I-Bias-Offset measurement and that failed.
I-Bias is almost the same but a bit more unstable and it takes longer to settle ( if it settles !).
It usually ends with Error 414 ->  „Current Offset not converging“.
I dont think i made it to Error 422 -> „Current Offset out of spec“ ( but not sure ).

Sometimes it is not possible to measure in V mode. The display just shows ——— on the left side.
I found out that it helps to rotate the unit 90 degrees left ro right ( and sometimes knock it a little).
After that i will get displayed measurements again.

Sometimes it is not possible to switch to V mode. (Seems like the relays are not engaging ). Usually here it also helps to rotate the unit left or right.

These problems are only apparent with the Volt mode.

Could it be that the unit was damaged during shipping?
It was very poor packaged, only thin cartboard, no cushioning, no extra foam, nothing!
In fact the rear plastic frame broke completely during the shipping.

I have checked a lot of information repairwise on the internet.

I found TIN´s Video and „teardown“ of the reassembly of his 6517A.
teardown ->
video ->

Thread about replacing leaky reed relays in the Keithley 617 ->
The Keithley 617 is an older model but seems to be very similar in the analog part.
I think this guy, Tom Mathews also responded in one of the numerous K617 repair threads.
He exchanged some ( or all? ) of the reed relays because they were „leaky“.
I think this could be my problem too, but i never saw him showing how the failure looked like.
( and i dont want to exchange all the relays and find out that it does not help ).

I am not an electronic engineer, but my skills are getting better fast.
I learned a lot already ( also from many mistakes ! ).
This time i want to take a good systematic approach to this repair.
I dont want to just guess and poke around and change some random parts.
First because the unit is in good condition besides the failure.
Very nice cosmetic condition ( some dust inside on the edges of the analog board) and there is really no sign of someone „troubleshooting“
the board before.
And then i also really want to understand how the input stage works.
Unfortunately there are no schematics in the service manual.
But i could maybe use the schematics of the K617 and try to „reverse engineer“ the input stage.

That all said i will be very thankful for ANY advise!
I dont expect you guys to talk me through this.
If you give me some pointers i will do my „homework“ and try to find the solution myself.

If you need pictures or a video of the ramping offset just tell me.

Thank you very much for helping!


Edit : changed link from youtube-video to thread about replacing reed relays in K617
« Last Edit: January 22, 2019, 11:04:24 pm by voltampere »

Offline Brad O

  • Regular Contributor
  • *
  • Posts: 85
  • Country: us
  • Keithley Apps Engineer
    • Keithley homepage at Tektronix
Re: Repair of a Keithley 6517A; drifting V-Offset and I-Bias-Offset
« Reply #1 on: January 22, 2019, 11:31:04 pm »
You're using verified Triax shorts and shielded caps when performing V offset and I offset right?  You should expect the voltage to gradually drift up after disabling Z-CHK if nothing is connected, the unit will just be pulling charge out of the air the same way a DMM in high Z mode would.  If it's ramping when you have a short then that's a different matter.

The most common reason for not seeing a V reading and only getting "-----" is noise, either preventing auto ranging from picking a range or preventing settling.  Try manual ranging, then connect a scope to the Preamp Output to check for line frequency noise. 

It does sound like there could be some damage based on your other comments though, particularly about rotating the instrument changes behavior,
The following users thanked this post: TiN

Offline TiN

  • Super Contributor
  • ***
  • Posts: 4245
  • Country: us
  • - 24/7 lab feed
Re: Repair of a Keithley 6517A; drifting V-Offset and I-Bias-Offset
« Reply #2 on: January 23, 2019, 04:57:13 am »

I worked on restoring 6517A (as replacement of the analog board cannot be called repair anyway), with photos here instead.
Good working unit should settle with femtoamp-level noise on lowest 2pA range with triax cap. Digikey has some suitable. If you don't see that, then there could be leakage/contaminated surfaces in signal path.

What does the full diagnostics shows? You need to enable secret mode to see diagnostics in menu.

Issue with orientation and random behavior might be as well related to 26-wire cable between digital and analog boards. Try to wiggle that one a little to see if it helps/changing. I had to replace the cable on 6517A before to resolve random hangs/failures to detect ADC. A/D board is same as used in Keithley 2001, and if there is problem between digital CPU and ADC comms you will get overflow or ----- on screen. On few K2001's I had to replace bad TLP opto too, similar symptoms.  :)

Usual warning note - be careful working on the unit, it have lethal voltages present on analog board even if voltage source is not enabled at the front panel.
YouTube | Metrology IRC Chat room | Live-cam | Share T&M documentation? Upload! No MB limit, firmwares, photos.
The following users thanked this post: Brad O

Offline voltampere

  • Regular Contributor
  • *
  • Posts: 58
  • Country: de
Re: Repair of a Keithley 6517A; drifting V-Offset and I-Bias-Offset
« Reply #3 on: January 23, 2019, 11:55:25 am »
 I dont have a genuine "Triax-Short". I am using a Triax-to-BNC connector ( Keithley 7078-TRX-BNC ) and a BNC-Short.
Since i have Guard-Off configured for the V-measurement it should be no problem right?
For the I-Bias measurement i am using a genuine Triax-Cap ( no shorting ).

Thank you for joining here.
And also "BIG THANK YOU" for your efforts with!
Really very helpful Place with loads of ggof information and archive of Manual and Teradown pics!!!
Helped me a lot so far!

Of course i saw your video of K6517 "repair" ( okay i did a little fast forwarding at some places :-) )
I already found the "secret menu" and did the full test suite.
As far as i remember it failed only on the V-Offset and I-Bias-Offest tests.
Will repeat it when unit is assembled, right now its in seperate peaces!  8)
I am about to do some cleaning. Especially the edges of the analog board do have a lot of dust!

Problem is i dont have the super-highgrade-metrology-boardwash. I am using the simple IPA stuff from Kontakt-Chemie.
Will it do more harm than good?

Searching for proper reed relay replacement parts now.
As sson as i have those i might do some testing with the relay that is used for the Zero-Check ( should be K416 ).
I think Tom Mathews desoldered one leg on his K617 and had a good and stable measurement with one leg lifted an that relay.
If that does not bring me forward i have to dig deeper into the preamp circuit!  :phew:

Thank you for security advice regarding high voltage inside the unit even with source selected off!
I am very careful when probing around and i will hopefully not forget it!  :P

About the C version firmware.
Should i consider downgrading it?
Few people mentioned that it was never released officially.

Thanks for all so far

« Last Edit: January 23, 2019, 11:59:21 am by voltampere »

Offline voltampere

  • Regular Contributor
  • *
  • Posts: 58
  • Country: de
Re: Repair of a Keithley 6517A; drifting V-Offset and I-Bias-Offset
« Reply #4 on: January 23, 2019, 10:22:40 pm »
I did a bit of cleaning on the analog board. But not in the shielded area of the preamp.
After assembly it looked very good initially.
Lowest Amp range almost zero. Only last digit moving a bit.
Volts also close to zero and stable.
But after warmup back to the old problem.

I did the full test suite again with following faults:

400.1  VOS -> Volts input VOS x 5
400.2  VOS -> Volts input VOS x .5
602.1  VS 10mA Limit -> Vsource 10mA Limit
603.1  VS 1mA Limit -> Vsource 1mA Limit
603.2 VS -1mA Limit -> Vsource -1mA Limit

I also did a quick check with current from a stable current source.
With 10 mA the 6517 needs more than 20 minutes to reach the end value.
Its 9,83xx mA after one minute.
After 20 minutes 9,9758 mA.
Looks like a logarithmic function.

Does this really look like a leaking relay?

Maybe the opamp has a problem?
Maybe i should check the feedback resistor, or the relay configuration for the feedback?

Suggestions welcome


Edit : typo "relay"
« Last Edit: January 25, 2019, 11:36:27 am by voltampere »

Offline Brad O

  • Regular Contributor
  • *
  • Posts: 85
  • Country: us
  • Keithley Apps Engineer
    • Keithley homepage at Tektronix
Re: Repair of a Keithley 6517A; drifting V-Offset and I-Bias-Offset
« Reply #5 on: January 24, 2019, 09:28:10 pm »
About the firmware, C05 did receive a proper release and is the most current version so you shouldn't need to worry about it.  I think it was released after the 6517B came out, so likely most units that have it were serviced at one point.

Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo