Keithley 2420 Power Supply problem

[sgahr]

Hello, my Keithley 2420 3A SourceMeter is broken. The Fuse was blown, and replacement didnt work. After checking the board, an deffective "ICL15" resistor was easy to find - see attached pic.
After replacement, the board worked for a moment, but after some minutes the same happend. Fuse blown, ICL exploded. The only thing I noticed was a 50Hz flickering sound, but I wasnt to worried about that initially.

So I replaced all capacitors, checked all solder points, replaced the 2 Triacs, the three grey rectangular capacitors and the black rectangular thing next to the ICL -> again the same problem

Anyone who can help me fixing that issue? This would made my day/week or even month 

Greetings Stefan

[PedroDaGr8]

You have a dead short somewhere, the ICL likely stands for In-rush Current Limiter. So something is pulling WAY too much current and blowing the ICL. Start looking for shorted tantalum capacitors and bad electrolytics.

[sgahr]

@PedroDaGr8 Yes, you are right. There must me a short somewhere, but not easy to find. So all the electrolytic caps have been changed, and the rectangular-tantulum ones are tricky to get off, or at least I never did that. So I ended up measuring all the caps while being on the board and most show high resistivity, there are just 2 of the little yellow square caps that show around 220ohms (marked black).

The AC terminals resistance is 1Mohm, the power rails are all in the kohm region, measured on the testpoints. I could just identify one short, where the gate of the triac is shorted with the GND. I dont really understand how this works, but it seems to be shorted via a Thick-Film resistor with 0.15ohms, that was initally burned and changed. I included a picture that hopefully shows that (marked red).

Thanks for helping

[sgahr]

Still having the same issues. Actually I am about to give up, I have no more ideas. Please feel free to give me any advice.

Actually I can send it to have it repaired. Any suggestions?

[Ordinaryman1971]

To late for advice but I would try to use the thermal camera of some sort and see where you have something cooking...

[sgahr]

Thanks, Ordinaryman1971. That came to my mind, just I have no IR-camera. But I keep that in mind if I find no other solution.

[Zucca]

If no IR camera is available, there is a poor man solution:
IPA alchool evaporates very quickly on warm components.... 

PS: If you want to sell it then drop me a PM.

[sgahr]

Haha, ok 
So since my repair attempts, I powerded the board three times.
First time it was running for a minute... just a flickering sound.. I thought its back
The two attempts later wherent so promising. The ICL or later the 0,15ohm resistor next to it exploded within a second... faster than the fuse. Lightning, burning, smelling

So I am not sure if its smart to repeat this over and over again, to maybe have the chance to find a hot component. Its still an expensive pice of equippement.
But if a IR-cam would have been avaiable at the first try, this would have been a good opportunity.

I have an offer from a keithley dealer to repair it for 2k€..., I might do that if I dont find a safe option to repair it myself

[Zucca]

2K€.. Lord I am a sinner...

Wrong strategy in the last two attemps. 
Well I would try supply the power rail of the ICL and the other resitor with a Lab PSU and slowly increasing the voltage, better the current.
In 100mA (or less) steps I would check for hot thermal spot...

I don't know that board very well so don't know exactely if it would help.

[sgahr]

@zucca
The area of the primary AC/DC converter is actually quite densly populated and I wouldnt know where to connect my Lab-PSU and at wich voltage.

The 4Layer Board makes it very hard to find where the connections are going. But thanks for helping. Actually I considered replacing/bypassing the original PSU by another one, but couldnt figure out how.

Unfortunately there is not much information to find for the 2420 or 2430 board (wich should be identical).

[Zucca]

where to connect my Lab-PSU and at wich voltage.

You have to hunt the big DC filter bypass cap and put some VDC there (polarity is clear looking at the cap), if it's a short (99% it is) you don't need a lot of voltage but a controlled (low) current.
Anyway I wish I could be there in Austria with you to work on that beauty...

Yeah it's a lot of reverse engineering and time required....

EDIT: if the short it's on the AC side you could power some VDC (again with controlled current) on the main, if it's a SMPS it should start to work regardless. Anyway I am just guessing...
As long you are not pumping too much power (by controlling the current flowing) you can't do any further damages...

bah.... back to work now.