My pile of 7075's

[wiss]

In may I scored 6 Solartron 7075 cheap of ebay-UK 

Got the a while later and the first one I tried powered up but large part of the display did not come on, the driver-chip (IC11 and 12) was toast, get a donor out and all segments light up! 

After running for about an hour (to check drift and stuff) there was a very loud bang from the hobby-room...
The mains-filter blew up! Leaving lots of residue in the power-supply...
Getting two more filters out, from other 7075's, different model/make, still they went fzzz after 10-15 minutes!
That one I just replaced with a plain connector and it is running as "Unit 0" now.
There were lots of noise and low level drift, one of the AC-reed-relays seems to have been dirty or cracked, and after amputating it it got more stable, still some ppm-level noise, more on that later.

[Vgkid]

I hate you, good luck with the repairs.

[wiss]

I disassembled the other two and ran most of the boards through the dish-washer (not the power-supply).
I reassembled "Unit 1" and tried to get it running properly. The Ohms-selftest failed with a reading of ~0. Found R604 to have failed open, replace from a donor and resistance agreed to calibrate on 10k, 1meg and 10meg range but 1k was pretty off  . R604 and R602 are a matched set, you have to replace both! And the Ohms-zero took me a while to figure out, there are two current-sources, one on the positive side and one on the negative, Ohms-zero set their balance and you have to connect source+, source- and +in, floating from -in.

One of the AC-reed-relays failed to make contact, replaced.

Still the same ppm-noise as Unit 0 
The ref and zero-self-test are very stable (+- 1ppm) so the AD seems to be ok, so it's in the input-amp...
Started probing around in the chopper-part just out of curiosity, hum, the copper-O/P seems rather noisy...
Is the last integrator op-amp (AD545) crappy? Found a data-sheet with some help from in here, not super-good, maybe a op07 might be better? So I tested a few in a circuit with 1meg//1uF feedback and +in grounded, CA3140, OP07, AD820, AD545 and LTC1150: 3140 was worst and 1150 best, 820 was better than 545 and 07 but not by much. So ad545 might be pretty ok.

Next I probed the stage before the integrator, IC405, LM301 with a 3.3n compensation. When the jfet-demodulator goes open (and I guess the load on IC405 gets capacitive) it starts to oscillate! The same happens on both Unit 0 and Unit 1! 
Dropping in a 33-100n-copensation fixed that!

[wiss]

To calibrate the ranges I built a divider that when feed 13 V will not induce a range-switch but when feed 10 V will.
A ratio of about 1:8.5 will do this. Resistors are 100k and 13.5k and I note the exact ratio at 13 V and recalculate the output for 10 V in for cal.

The pin-side fits into the supply and the DMM is connected on the hole-side.

[Vgkid]

can you post a picture of the R602/4, it is strange that those 2 would be fried. They are 1/2 watt  wirewounds.
here is the schematic of the section:

[wiss]

I do not think that it is fried, it does not look damaged, just old age, or vibrations...

602 and 604 in the picture.

[wn1fju]

I would be interested in the long-term drift of your units.  I have a Guildline 9577 which is the same as the Solartron 7075.  Mine would work perfectly for several hours, with minimal drift, and then almost suddenly would go off on a wild goose chase.  I tried replacing many of the op-amps, but never really got to the bottom of it.  My not so pretty solution was to cut away some of the plastic side panels and mount a small fan in the unit to blow across the middle board.  Then it was stable all day long.  But I sure would like to know the real solution.

[Vgkid]

@ wn1fju: How is the input impedance behaving, do you notice the drift on all ranges, or some. How stable is the resistance reading as well?

[wiss]

I would be interested in the long-term drift of your units.  I have a Guildline 9577 which is the same as the Solartron 7075.

At the current weather-conditions (massive heatwave here, 28-30 degC... longing for february and -15) and my crappy connectors I get some 10-20 ppm/day. Unit 1 with the DIY fisher-male is worse than Unit 0 with a metal DIN-5.
Noise is very low, Non-linearity + noise is less than 1 ppm in 6-digit mode.

  Mine would work perfectly for several hours, with minimal drift, and then almost suddenly would go off on a wild goose chase.  I tried replacing many of the op-amps, but never really got to the bottom of it.  My not so pretty solution was to cut away some of the plastic side panels and mount a small fan in the unit to blow across the middle board.  Then it was stable all day long.  But I sure would like to know the real solution.

I would check potential oscillation in IC405, measure output of it and/or solder a 33n cap to the underside of the board piggybacking on C415 (?).

[wn1fju]

To Vgkid:  If I recall (it was a few years ago), I simply shorted the test leads and set the DMM to DCV, 1 sec integration (6 digits).  The nominal 0.000000 reading
would be stable for a few hours (with a little noise on the last digit) then start changing in at least the last three digits, if not more.  I never bothered to check
the input impedance or other ranges/settings as you suggested.  Probably a mistake on my part.

To wiss:  I will try adding a cap on IC405.  Until I open the unit up and take a look, I don't remember if I tried replacing that IC originally.  First, I have to disable
my fan "solution" and see if I can reproduce the original fault....   A good project for a rainy day!

[wiss]

To wiss:  I will try adding a cap on IC405.  Until I open the unit up and take a look, I don't remember if I tried replacing that IC originally.  First, I have to disable
my fan "solution" and see if I can reproduce the original fault....   A good project for a rainy day!

You will have kHz ringing on the output of that opamp if that is the problem, all accessible from the underside of the lower board, no need to lift the upper board out

[Vgkid]

@ wn1fju
I whould have said input current.

[wiss]

I would be interested in the long-term drift of your units.  I have a Guildline 9577 which is the same as the Solartron 7075.  Mine would work perfectly for several hours, with minimal drift, and then almost suddenly would go off on a wild goose chase. 

Last night My Unit-0 read 10.000040 and my prema 6001 10.00004, this morning the solartron had dropped to 9.999982 and the prema 10 10.00002, so -6 ppm to the ref and -4 ppm to the LM399-based prema.
Temperature was about 26 degC last night and 25 this morning.
The solartrons were more unstable when the temperature was higher (~28 degC).

Maybe a reference heater problem?

[wiss]

Went to check on the reference and noticed something odd.
When in Self-Test:
 in reference-, zero-, ac- and display-step the reference holds the same voltage.
 in resistance-step in increases ~13 ppm.
(measured on TR301 emitter, the driver for the positive reference)

Measuring output of IC301 (driving TR301) there is no corresponding increase/decrease in voltage but rather higher voltage (1 mV) for Res, AC and Zero compared to display- and reference-steps

what?

edit:

The negative reference moves about the same amount in the positive direction as the positive reference, both shift relative 0.

edit 2:

Nevermind, used wrong ground-ref...

[wiss]

The schematics specify AD510L for the reference opamps, in my instruments are PMI OP07EZ, the AD have open-loop gain min 10^6, OP07 2*10^5, would that difference matter?

[wiss]

I had Unit 0 running over night in ratio-mode. Same 10 V in both on reference-inputs and front-inputs.
It dropped 5-6 ppm until breakfast, but then dropped another 20 ppm during my shower!
Moisture-sensitive? Problem with Trimmers?

The schematics specify AD510L for the reference opamps, in my instruments are PMI OP07EZ, the AD have open-loop gain min 10^6, OP07 2*10^5, would that difference matter?

No, it won't. The change in output voltage from the op varies by a few mV, which for OP07 would be a change in input voltage of a few 10 nV. Insignificant.

[wiss]

For Unit 0
Maybe due to morning, it was still 30 ppm low when I got home from work, but not in ratio-mode, maybe I switched that off before I woke up properly...

Still sitting on a solid -30 ppm, no change during the day, typical spec says +- 29 uV in 10 second mode while measuring 10 V, well within that at 1s integration today Maybe just need a very long worm-up time...

Will check Unit 1 the coming night...

[Vgkid]

For your unit, were the OP07's installef in place of the ad506, or did you swap them for testing?

[wiss]

No, I did not replace any opamps in the two assembled units, I did take one AD545 from the donor. I would expect that an upgrade of the reference opamps would improve performance. OP07 is not bad, LTC1150 would be better, but...

The whole instrument seems a bit temperature sensitive, I will from now use a minimum of 4 h of warm-up. Unit 0 has been running for 28 h now and are stable within +-2 ppm measuring my ref. Both DMM and ref have wider spec than 2 ppm

[Vgkid]

I have a 7065 myself, and it does take quite a while to become stable. I have noticed that the analog section is sensitive to air movement around the instrument.

[wiss]

The 7065 is far more modern, CPU and stuff in there
I have one at work, I don't use it much but it seems to stabilize within 2 h, as far as I remember. (Mine is 10 ppm high )
But that is controlled climate and not possible to open the windows in my work-workshop...

Unit 0 has been upgraded to have a DIN5 connector on the front. Found it in some random pile of leftovers somewhere, does not look like the cheapest one can find. It does not really degrade the instrument, as far as I can tell.

Graduation photo for Unit 0 (Top of the shelf, Volt ok, all I care about), Unit 1 (on the bench, all ranges work) and the HP (on top of Unit 1)

[Vgkid]

That 10ppm is only 2x off of the 24h specs 

[wiss]

Will check Unit 1 the coming night...

It was stable!   

Had both of them turned off during the day today and now, after 3 h they are back where they were this morning/yesterday.

[wiss]

My DE5000 arrived today
First thing to measure was the original compensation-cap from Unit 1, it came out 20% low.
Did it drift over time?

Edit: After a CAL of the DER it was even lower, 2.6 or 2.5 nF

[wiss]

A few external mugshots of #3.

The compensation-cap on IC405 in this one was a disc-ceramic and it did not oscillate.

There were a slow drifting up and down, very regular, me thinking beating between 50Hz
and chopper 200 Hz: re-soldered the optical choppers (joints looked suspicious), solved the problem!