Keithley 2000 repair, only works in Hz mode

[Per Hansson]

This meter threw me off a bit: I bought a bunch of stuff on Mouser so I'd be sure to be able to repair it.
I started with the electrolytic capacitors: to my relief there was no leakage and the old ones tested ok.
After firing up the meter it worked, and kept on working, and working, and working for over a month with no issues!
I was not really thinking that was the fix, but I was surprised all the same!

But then a week or so ago it decided to fail in the same manner again.
So I replaced U139 with the high-end LT1124AMPS8 because I chickened out with the test
And it has been fine now for a couple days.

There is still the bodged through hole zener diode VR109 that I want to take care of.
And in my Mouser cart I put a 18v linear regulator: MC78L18ABPG
I just want to check what you think about that.
Is it better to just stay with the original zener circuit that bakes itself?
I did get both 17v and 18v Zener diodes and a 1K resistor to replace R168 if I wanna go that route...

P.S: I also measured C219 & C220 while U139 was out of circuit and it was 300pF at 100Hz, 190pF at 1Khz and 120pF at 100Khz
That all does not make any sense so I left C219 & C220 alone.

These where the replacement electrolytic caps I used, I tried to stay as close to the originals as I could:
UVR1E682MHD
UVR1J101MPD
UVZ1C222MHD
UVZ1H102MHD

[Kleinstein]

Measuring C219/C220 in circuit, even with U139 removed is not expected to be accurate. There are still other paths thround the resistos.
I would also leave C219 /C220 as is.   Changing the capacitors and moving them to a different point would be a thing if U139 would be changed to a LT1013.

For the zener / transistor solution for the 18 V, I would probably use a 18 V regulator if at hand, but go for 18 V zener and 1 K if I would have to order the regulator.

[Per Hansson]

For the zener / transistor solution for the 18 V, I would probably use a 18 V regulator if at hand, but go for 18 V zener and 1 K if I would have to order the regulator.

Thanks!
Yea I already bought the 18v linear regulator.
I'm at a loss how I thought about connecting it up though, the circuit as shown confuses me greatly.
As shown VR109 turns on the base of Q118 via the current limiting resistor R168.
But then Q118 has a separate +20v unregulated supply from collector to emitter.
Why does this result in a regulated input to U123?
Maybe this question would be better suited in the beginners section lol but here I am

[Kleinstein]

The base voltage gets stabilized to the zener vottage and the transistor works as a emitter follower to produce some 0.6 V less than the zener votlage. So it is not actually regulated (as with a feedback loop), just stable enough.

A voltage regulator would be in place of the transistor with the 20 V as input, and the GND pin wired to GND with a short at the zener and R168 removed.
Chances are one can bend the wires on a TO92 case regutlator the right way.  With a smd case one may have to be lucky.

[Per Hansson]

Thanks for the explanation, I'm glad I asked because I really had a hard time grasping this, now it is clear!
Yes the MC78L18ABPG that I bought is a TO92 so what you describe was what I had in mind.

P.S: Would anyone happen to know the part number or replacement for the amps jack?
It is broken in mine and I'm anyway placing an order on Mouser so if they have it in stock...
But I guess it might be a custom Keithley part?
Late edit: The partno is 2001-312D and Mouser does have it but it's not available to order from Sweden

[Per Hansson]

The 18v linear regulator has been installed in place of VR109 and it runs at ca 30°C compared to 105°C of VR109 before
The ripple U123 sees is now the same on Pin 4 and Pin 8, compare to post #22 how it was before.
The instrument still passes the built in self test
Thanks so much for all the help!