Yes, I bought another Keithley (196) to repair (FIXED!)

[staze]

So, just took delivery of a Keithley 196 I got for dirt cheap ($40). Which, for a 6.5 digit meter, can't be beat.

Initial diagnosis says it's probably something similar to what was wrong with my 199, which was the counts pathway between the analog and digital board (199 was the 74HCT04). Anyway, interesting that the 196 doesn't use optocouplers, but instead has pulse transformers (weird!).

First question is, does anyone have higher quality schematics for these units? the ones in the online pdf manuals are terrible quality.

Anyway, I'm working on physical repair of the unit at this point (shipping was un-kind to the unit), then I'll start (of course) with power rails, then I'm guessing TTL pathway, and I'm guessing the HC00's, or the LM311's that are in that same pathway. We shall see. Anyone else repaired one of these? =)

[lowimpedance]

 The only repair I have done was for one at work which was needed for a test setup (old software etc) and was powered 24/7.
The failure was bad mechanical mounting of the mains transformer leading to eventual dry joint fracture and then burning of the PCB around that area. So my only input would be at least make sure yours is sound in that area.
 I did not need the handbook at the time and the owner probably would be unlikely to put his hand on it any time soon unfortunately.
If you still need better scans I 'might' be able to find it !.

[staze]

The only repair I have done was for one at work which was needed for a test setup (old software etc) and was powered 24/7.
The failure was bad mechanical mounting of the mains transformer leading to eventual dry joint fracture and then burning of the PCB around that area. So my only input would be at least make sure yours is sound in that area.
 I did not need the handbook at the time and the owner probably would be unlikely to put his hand on it any time soon unfortunately.
If you still need better scans I 'might' be able to find it !.

If you have the ability, that would be awesome. Judging by the scans, they seem to have been done with a really crappy scanner, or someone that just didn't care. So anything would be great... but don't go out of your way. I think I can get it fixed with what I have, but it sure would be nice to have better quality schematics. =)

[rastro]

I also recently picked up a 196.  It seems to be functional and maybe 8-10 counts out on DC voltage measurements.  I still need to figure out accurate tests for ohms/amps/AC - certainly none are grossly out.  I sure like the large LED display.

I probably have the same documents you have from the internet.  My firmware is B15 and I think the main PCB is revision "C".  Let me know if you want me to cross check any test points or component values if that helps.

One thing I noticed is that the front panel seems slightly warm (90F with room temp around 78F).  I'm not sure if this is normal or there's a problem brewing.

Anyway good luck with your troubleshooting. 

[tsmith35]

Would anything like this be helpful?

[staze]

that's what I have that has terrible schematic scans... zoom in to see what I mean. =P

[tsmith35]

Yeah, pretty standard for scanned manuals, unfortunately. Back when the manuals were printed, the print quality wasn't that great either.

[staze]

Yeah, pretty standard for scanned manuals, unfortunately. Back when the manuals were printed, the print quality wasn't that great either.

yeah... guessing just really crappy photocopies or scale-downs from the original engineering drawings. I kind of want to just fix the whole thing in Illustrator/Photoshop, but I'm not sure I care THAT much. =/

[staze]

So, power supply rails all check out (without the analog board hooked up), and TTL signals all look good.

BUT, hooked the analog board up, and things go wonky again. Checked the analog rails, and the +5V and +15V are fine. -15V rail, is going between -15V and Gnd in a regular fashion, so something is pulling it up the ground.

Will post a picture once I have a chance, and need to figure out what is pulling it up. Guessing it's when the measurement circles around to something, maybe the 0V reading. So, will put it in diagnostic mode next and see if I can isolate what part of the cycle it's doing this in.

Wonder also if using my bench supply to drive the -15V rail would show much... wouldn't think so since if it was shorted, it would just sit at 0V rather than pulse back and forth. As a I said, with the analog board unhooked, the -15V rail is fine. And the +15V rail is fine either way, so it's not that the two are being shorted together.

Any thoughts? Real annoyance is how bad the schematic quality is... printed them out full size (11x17 (A3)), and they're unreadable as far as numbers, etc. Might have to re-do them. =(

[staze]

So I just noticed C53, a 68pf 500V polystyrene cap looks melted/blown.

I'm trying to find it in the schematics, but more importantly, I can't seem to find replacements through Digikey or Mouser. Can Mica's be used instead, or...?

Thanks!

[PedroDaGr8]

So I just noticed C53, a 68pf 500V polystyrene cap looks melted/blown.

I'm trying to find it in the schematics, but more importantly, I can't seem to find replacements through Digikey or Mouser. Can Mica's be used instead, or...?

Thanks!

I wouldn't; polystyrene have super low dissipation factors. You can source them online for a couple bucks. If I remember correctly those polystyrenes decouple directly to ground. So that could be the source of the failing -15V rail.

[staze]

Pedro,

Good to know... that should help me find it.

I happen to have a 68pF 500V mica cap in my bin... think I'd be safe replacing it temporarily with that? Then more permanently once I find a replacement polystyrene?

Got any links to where they might be? All I've found is an audio parts place that sells a big batch for $100 of all difference sizes. =/

[PedroDaGr8]

I think it should be OK! Just don't be surprised if things read a bit off. It would still be helpful for diagnosing the problems on your analog board.

I'll see if I can find my links. I do remember I had better luck with finding 630V polystyrenes than looking for 500V ones. Just like with other capacitors a higher voltage rating shouldn't affect things.


FRI just in case you didnt know. Polystyrene capacitors are by far the most heat sensitive of all of the capacitor types. They can be easily damaged by soldering if you aren't careful.

[staze]

I think it should be OK! Just don't be surprised if things read a bit off. It would still be helpful for diagnosing the problems on your analog board.

I'll see if I can find my links. I do remember I had better luck with finding 630V polystyrenes than looking for 500V ones. Just like with other capacitors a higher voltage rating shouldn't affect things.


FRI just in case you didnt know. Polystyrene capacitors are by far the most heat sensitive of all of the capacitor types. They can be easily damaged by soldering if you aren't careful.

yeah, noted that they're heat sensitive... it could be that this one was damaged by someone else trying to solder near it or something... they're a rather large chunk of it missing/melted.

anyway, will try swapping in the mica one tomorrow (assuming I have time). I don't doubt it'll be off, but at least it'll tell me if it's that cap, or something else. I'll also snap a shot of the -15V waveform with the board plugged in.

Thanks!

[tsmith35]

So I just noticed C53, a 68pf 500V polystyrene cap looks melted/blown.

I'm trying to find it in the schematics, [...]

It's on the schematic "Analog Board" (page 134 of the PDF I linked above). It's at location "A2" at the bottom right corner.

[staze]

Cool, thanks. And I do see a path between it and -15V. Hmm.

That said, I would think a 68pf cap, should it fail, would fail open rather than short... Since those plates aren't very close together. Will post picture of waveform.

[SeanB]

Polystyrene will almost invariably fail as a short circuit, just due to the construction, especially if it is hit with a soldering iron tip.

[staze]

Oh. Good to know! I would think those plates would be pretty insulated given the size of the unit, and how low a capacitance.

That said, I just checked it in circuit, and it shows 3.4k in one direction, and 5k in the other. So it doesn't seem shorted, unless they only short when there's more voltage/current than my 199 uses for resistance.

[staze]

Well, I just fired up the meter again, and that cap is making a hell of a racket now (high pitched buzzing). So it's obviously bad.

Will replace with mica and see what we get.

[staze]

So, replaced that bad PS cap with a mica one, and things are back to "normal" as in, what it was doing before the cap went sideways.

Something is bringing the -15V rail up to ground (or just above) in pulses. Obviously the ramping is due to the filter cap on that rail, but any thoughts? Guess I still need to put into diagnostic mode and see if I can figure out if a specific part of the sample cycle is the issue.

[edavid]

The problem is that polystyrene caps aren't made any more.  Usually the best substitute is polypropylene or NP0 ceramic.  I can't figure out exactly what that cap is doing, so it's hard to say if mica is good enough.  It looks like that cap is filtering the input, so I think they chose PS for low dielectric absorption, and I'd use polypropylene.

[staze]

Looks like it's just decoupling the signal ground. No sure that would matter that much, especially for testing.

[staze]

Diag mode stages made no difference, something is still causing the oscillation.

Guess there's always the chance the 7915 can't feed any real load...

Think I'd be safe injecting -15V into the analog board, and seeing if it stabilizes? Would imagine I'd just put the positive of my bench supply on the ground, and negative on the -15V rail... Or I guess I could be "safer" and inject it directly onto the pins for the 7915 (same basic idea though, negative lead to 7915 output, positive to analog ground.

Thoughts, or is this a bad idea?

[edavid]

Diag mode stages made no difference, something is still causing the oscillation.

Guess there's always the chance the 7915 can't feed any real load...

Think I'd be safe injecting -15V into the analog board, and seeing if it stabilizes? Would imagine I'd just put the positive of my bench supply on the ground, and negative on the -15V rail... Or I guess I could be "safer" and inject it directly onto the pins for the 7915 (same basic idea though, negative lead to 7915 output, positive to analog ground.

Thoughts, or is this a bad idea?

You would have to pull the 7915 first.

[staze]

Okay. Maybe I'll just try loading it down first and watching the voltage. Got a couple 8R 20w resistors that I can use to put about an amp out of it.

Out of curiousity, why would I have to pull the 7915 first? I know this is the practice... But why?