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

[PedroDaGr8]

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?

I think because otherwise the vreg will fight the voltage from the power supply.

Sent from my EVO using Tapatalk

[staze]

Bam. So, loaded down the existing 7915 output with 16R (disconnecting the analog board), and the voltage looked pretty similar. So, unsoldered the 7915, and wired in my bench supply for -15V, and fired up the unit. Viola, it worked! No random fluctuation at all, and the -15V rail was only drawing 70mA.

So, going to have to look through my parts for a 7915, or just go pick one up. The PS cap though... I'll just have to order one... only supplier seems to be in Canada, unless someone knows a good way to find one in a scrap bin, or if the Silver Mica one would work just fine, I'll leave it in.

Shame I didn't get to use the beautiful schematics scanned by lowimpedance! I will make sure to share them with the world though!

[SeanB]

Silver mica is not as good as polystyrene, but for this low value it will do, the 500v rating will keep unwanted DA and soakage to a very low level. Leave it in and get the regulator changed out and then check if the cal is still in spec.

[staze]

well, damn. it wasn't the 7915... which means it must be the filter "caps" on the bridge rectifier side of the ±15V regulators.

Only issue is, the schematics show two 680uF caps, but the unit uses a single bi-polar 1500uF cap. I know you can make a bipolar by hooking two back to back, but I'm unsure how to do that in this case. Would I go negative to negative and just replace the bi-polar, or should I test first by just hooking a 680uF 25V cap directly to the 7915 and see if that stabilizes things? Other issue is, I don't have any 1200uF+ 25V caps, so making a bipolar is going to require a trip to the store.

Any input would be great.

[staze]

Correction... I was looking in the wrong place. There are two 35v 680uF caps. The one for the negative rail is completely toast. Pulled it, and it shows 7nF. Awesome. ESR is seemingly infinite (Blue ESR just shows a dash).

So... I'll go pick up a replacement for both of those caps tomorrow, as all I have are 25v 680's.

So, guessing that's fixed.

[lowimpedance]

 Was about to post some thoughts on the said 680uF and bipolar stuff but you have already found the culprit, excellent.
Another nice old meter saved. I have two 195a 's (5.5 DIGIT) which wont win any beauty contests but are good solid performers and best of all nice LED displays like the 196.
Hopefully the two 196's at work will become redundant before me .
 BTW that scope shot is what you get with filter caps that have 'dried out' losing capacitance.
With the 'new scanned' schematics you can zoom in and actually read the part details!. I printed one on A3 and its clean with no fuzziness as well as easy to read at this paper size.
 

[staze]

Was about to post some thoughts on the said 680uF and bipolar stuff but you have already found the culprit, excellent.
Another nice old meter saved. I have two 195a 's (5.5 DIGIT) which wont win any beauty contests but are good solid performers and best of all nice LED displays like the 196.
Hopefully the two 196's at work will become redundant before me .
 BTW that scope shot is what you get with filter caps that have 'dried out' losing capacitance.
With the 'new scanned' schematics you can zoom in and actually read the part details!. I printed one on A3 and its clean with no fuzziness as well as easy to read at this paper size.

Yes, pretty amazing. And good to know about that. I thought it was the VR, but I guess the filter cap makes more sense. So really the VR is only getting enough current when the BR is peaking. Will get those ordered, and also replace the ancient 1500uF cap on the 5V analog rail... the gigantic cap for the 5V digital rail appears much newer than 1986 (when that meter appears to have been built).

[staze]

Ordered some Nichicon HE (10000HR 105C) caps to replace the two on the ±15V rails, and the +5V analog rail. Those should last well past my lifetime (since I don't plan on leaving the unit on at all times).

I don't plan to sell the unit, but I'm curious... anyone know what they actual go for? I see working ones on eBay for like $460, but I can't imagine they actually go that high.

[staze]

Was about to post some thoughts on the said 680uF and bipolar stuff but you have already found the culprit, excellent.

Though, please do post about the bipolar stuff. I know I'm curious.

[staze]

Re-capped the power supply (replaced both 680uF 35V caps on the ±15V rails, and the 1500uF 25V cap on the 5V analog rail). Also put the original -15V regulator back (figured it worked, why replace it).

Now I'm just waiting for the Polypropylene cap. I really hope the calibration is still good... though I guess it wouldn't be to bad to have it calibrated considering how inexpensive it was.

[staze]

So, all fixed.

I dropped a polypropylene cap in (https://www.eevblog.com/forum/buysellwanted/wtb-couple-small-value-polystyrene-caps/) in place of the polystyrene 68pF that was bad, and it both fit perfectly, and I would guess should behave very close to the PS cap.

Calibration seems very close, but I don't have a meter that makes a good transfer "standard"... it just compares well to my Keithley 199 that I had previously "calibrated" to my bench power supply. At some point, I'll send one of them off for calibration.

Thanks for the input all!