Keithley 2010 Repair

[Cyclotron]

Does anybody know where Q156 is located? It appears to be another proprietary chip.

Q156: TRANS, DUAL N-CHAN JFET, SNJ450H99 Keithley Part #: TG-326

I can't find the bastard anywhere though. Maybe it doesn't exist on my meter, but I do wear glasses. 🤷

Thanks,
Josh

I'll take a look.

As a side note, I've printed the board layout from the service manual and I gave to the wife to play where's waldo when I can't find something because here close up vision is a lot better than mine. 

[KungFuJosh]

As a side note, I've printed the board layout from the service manual and I gave to the wife to play where's waldo when I can't find something because here close up vision is a lot better than mine. 

I usually lose a small screw, or something like that, on my office floor and my wife finds them for me. 

I took high-res photos of my board, and looked at it with a zoomed in digital microscope, and I couldn't find the bastard. Of course, I didn't take off the metal shield plate, so it could be in there.

[KungFuJosh]

the board layout from the service manual

I totally overlooked that since it wasn't searchable. 

You can zoom in on the PDF on computer screen. Why isn't the bastard searchable??

[KungFuJosh]

Here's a vector zoomable copy of just the main PCB layout.

[Cyclotron]

Here's a vector zoomable copy of just the main PCB layout.

I was just thinking about taking the image and running it through OCR to see if it would figure out the numbers.

I have found one part on the service manual board layout that isn't on my board. U196, which is depicted right at the rear of the front/rear switch near U153/U169.  There's no pads or any other hint that the part should be there.  I assume board rev differences.

Also, my wife finds the crap I drop, too. She'll hear me start cursing, and she pops in and says, "What did you lose?"

[Cyclotron]

Just realized there are board change notes on that image too. U196 is on that list.

[KungFuJosh]

I tried Acrobat's OCR, and it sucked. It worked on the obvious outside text, but ignored the component labels on the board.

ETA: I tried it again a couple times, including after rotating the page, and it did a little better. It still can't handle two text directions well though.

BTW- Q156 is a dual n-channel JFET, so we should expect at least 6 pins, right? So it should be even harder to hide. 

[Cyclotron]

I've also tried Acrobat and several tricks I've done before, and it really sucks.

I have another schematic I really want to OCR, so this is a side quest for me now or later.

I have my meter warming up. Been running for 3 hours, no problem. It will fail eventually. 
Side quest today is to plan my DC rail around my bench to supply LEDs and get rid of the flipping awful EMI I have in my lab.
I think I've settled on an LM317 current sink for each strip, since they vary in length and number of LEDs.

My guess is Q156 is a replacement part that may also appear in other board revs.  I can't find it on the images of my board or the service manual layout.
Maybe it was deleted from early boards?

[KungFuJosh]

The only thing near U177 that could possibly replace Q156 is maybe Q160. Q160 is TRANS, NPN PAIR, LM394, Keithley Part #: TG-142.

I dunno if this is true, or total bullshit from google's AI, but it said this:

Finding an LM394 labeled as Q160 confirms you have one of the original, early-generation production revisions of the Keithley 2010.

Here is exactly what is happening on your PCB and why Q156 is missing:
The LM394 / Q160 Swap
The LM394 is not a JFET; it is an ultra-matched Dual NPN Bipolar Transistor pair. In the very first design iterations of the Model 2010, Keithley engineered the low-noise composite pre-amplifier stage for U177 using this super-matched NPN pair (Q160) instead of a dual JFET.

...and looking at picburner's schematic, Q160 is tied to U177, so it's plausible.

[KungFuJosh]

To rule it out, I replaced U176. The problem in the 10V range persists. Not surprising.

Anyway, it's probably worse now. 

Before it would go to -11.9V in the 10V range stay there. Now it goes out of range. Of course, the problem only seems to exist in the 10V range, so switching ranges isn't helpful. Never mind. It's stopping at -11.9V again after warming up for a bit.

I suppose this verifies that I need to chase down some stupid switches.

Is there an simple way to test Q104, Q105, Q108, and Q113?

Does anybody know the correct replacement part numbers for them? The manual says SNJ132199. ETA: maybe MMBF4393LT1G?

One other question- is it possible I'm not cleaning the board well enough, and that's responsible for the leakage?

Thanks,
Josh

[Kleinstein]

The MMBF4393 is a reasonable choice. It is kind of the standard switch.

For checking the switches it is worth to look at the gate voltages. The LM339 comparators are known to be a little failure prone and this may cause some switches to be stuck open.
For testing the switching one look at the voltage after the buffer or main amplifier. AFAIR the issue was high hum for the DCV input. This would be Q104 not turning on.
Another test would be switching the meter to non AZ mode and have a signal generator with a low frequency (like 100 Hz) signal at the input and follow the signal path with the scope.


The schematics around U174 is definitely wrong: The circuit with TL431 and current source is at U190 - the plan somehow combined U190 and U174 to one. One can see the difference on the photos.

U174 should be wired as a buffer. A common block found in such meters is a buffer for feedback voltage (gate of Q193) as a kind of guard and maybe bootstrapping some protection / use for JFET gates as switches.

[KungFuJosh]

Here's a bunch of corrections. I did not test the source voltages yet on U190, I assume for now they're fine. U184 was really U185. Some device pin corrections, device ID corrections, etc. etc.

Thanks,
Josh

PS. Yes, I will share the updated schematic project after I've gone through some more corrections/additions.

[Kleinstein]

There should not be a link from Q184 to U174. That link is very likely toward U190 instead.
Instead there should be a resistor from the TL431 to ground.