Keithley 2015 - Optocoupler- need help

[letron77]

Good morning,

I continue my attempt to repair my keithley 2015.
And I come back to ask for your help
When switched on it displays "no comm link"

I have read all the posts on this forum dealing with this issue.
All the comments are excellent and I learned a lot about the operation of these devices (k2000, K2015)

After having eliminated the problems of contact braid, capacitor, rom ... I have come to the conclusion that the problem may have come from the optocouplers of the analog/digital part.

But if I understood the operation of a simple optocoupler (with 4 legs) I have much more difficulty understanding the operation of an optocoupler with 8 legs (I am a small amateur in electronics),

Also, knowing that the k2000 and k2015 share the same main board, I bought a k2000 (in perfect working order) so I could compare various voltage measurements.

It was indeed the only way I saw to try to find the failure of the 2015.

I compared the various voltages at the legs of the optocouplers (a2601, a2611) of the two devices and found the following differences:

      - A2601 - leg 6: k2000: 1.62 v k2015: 255 mv
      - A2611 - leg 6: k2000: 351 mv k2015: 5V

I have respected the masses AGND and DGND

I am attaching a drawing of the tensions observed at each leg.

Do these observed voltage differences suggest that the optocouplers are faulty?

Is there another way to test them?

thank you for any help

[Kleinstein]

If the meter is in fault mode it may send something different than in normal operation.
To be more sure on the optocoupler one would need to also check the control signal (pin3, pin 2 should be tied high). Normally the voltages at the pins 3 and 6 relative to there grounds should be similar.

If the comunication from the output side µC to the ADC is not working, chances are that the ADC is not sending data. So the difference at the 2601 may be just a consequence of the other direction not working.

If the schematics is corret (and the resistor not moved to the other side of the LED)it should be OK to connect pin 3 to ground   (if it's not already there) and than check pin 6.

p.s. : I relaized a direct link to ground would send to much current through the LED - it would need a resistor like 150 or 220 Ohm to limit the current.

[letron77]

Thank you for your answer

On A2601 I put pin 3 to ground (AGND)
Leg 6 (DGND) then comes out at 5 Volts
what should I deduce from this?

Another note: I measured at the same time legs 1 and 4 of A2611 (DGND) which starts at 2.7 V and gradually decreases to 217 mv

[Kleinstein]

Getting 5 V out with 0 V in would indicate a bad optocoupler.

AFAIK the pins 1 and 4 are not connected so drift there should not cause trouble.

[letron77]

to be precise, here are the measurements I get with pin 3 connected to ground (AGND)

Can you confirm that the optocoupler is faulty?

[Kleinstein]

The opto coupler is defintely bad. Chances are high it was bad before (from the intial data) and there is a chance that the direct link of pin 3 to GND could have additionall damaged it from too much LED current if was not bad before.

The more common equivalent opto-coupler type is the 6N137. Another name is HCPL2601.

[letron77]

Thanks for confirming this.

Two small questions:

    - can I take this reference?:
     
      https://www.mouser.fr/ProductDetail/Vishay-Semiconductors/6N137?qs=xCMk%252BIHWTZMrQz4FyDXhMg%3D%3D

    - I would have now wanted to remove the defective optocoupler and instead solder a support in order to be able to plug in the new one (without having to resolder it) upon receipt.
      Is it a good idea ?

[Kleinstein]

Having a socket is a good idea.

The 6N137 from mouser looks OK and should be a suitable replacement. Chances are it could also work for the A2611, at least for a 2st test. The A2611 is a speciall version for increased tolerance to fast common mode transistents. It is a bit odd to have 2 different types here, as both sides would see the same transients.

[letron77]

Thanks a lot

I'll get it all and keep you informed.
friendly