Fluke 720a "Problem"

[muvideo]

Hi,
In the middle of covid lockdown boredom and despair I bought a Fluke 720a in" working state" from the usual auction site.
When on my bench I opened it to check if it would contain most of what is is supposed to contain, and superficially checked
fuctionality for major problems, fearing the successive big regret for spending a lot for some useless paperweight.
To my suprise (the final price was low, but delivery + customs added to total final cost too much) the unit seem to be mostly
fuctional, at least no big problems.
The test was simple and a bit naive: I connected a voltage source to the 1.1 input and my 6-digits multimeter to the output and
regulated the voltage source to have a nice round 10V reading to the output with the divider set to 1.000000
Then started moving the dials decade per decade to see if the reading was coherent with the ratio set.
To my awe the reading was stable to last digit plus or minus 1, had to compensate for some drift of voltage source and
multimiter a pair of times but the final divided reading was astonishingly stable... wow
That was going almost too well untill G dial, the least significan digit one, that was tricky to asses but reasonable
until I selected "3" as last digit... there the output became totally messy. x.xxxxx4 was ok x.xxxxx2 also but x.xxxxx3 was jumping all around.
My suspect went immediately to S7A resistors, so opened the lid and here is the last divider... notice anything strange?

Here is it:

Who knows what happened to this soldering joint... I don't want to think that Fluke would let the unit out of production with a missing
soldering joint but the only other explaination is that someone messed with the internals. Who knows.
Anyway, can I just solder it back with regular SnPb solder?

[Kleinstein]

This looks like a tin-free solder joint.

Using regular solder should be OK. It would help to have no clean flux that works well with oxidized parts.

[muvideo]

This looks like a tin-free solder joint.

Using regular solder should be OK. It would help to have no clean flux that works well with oxidized parts.

Okkk

[Conrad Hoffman]

It happens. I bought a very nice HP selective voltmeter at a hamfest. It seemed erratic. Come to find out, a big silver mica cap on one of the gold plated cards was never soldered on one lead. I bought a Crown audio amp where the mono/stereo switch on the rear had unsoldered wires. They were factory and were crimped around the switch lugs, but never soldered. I bought some Polk speakers where a wire wasn't soldered to the speaker terminal, just crimped around it. Everything else was soldered OK. It's amazing that things can get through final test and work for years before the problem rears its ugly head.

[dietert1]

Typical, you just don't know whether to cry or to laugh after finding the error. But that one is an easy fix.

In my experience test equipment is never as mature as consumer electronics. I mean they had some procedure and apparently it was made for a small number of devices, so this one slipped through. Recently i spent a week or so to fix a Keithley 213 four channel DAC (rebranded IOTECH DAC488) that would severely overheat after 30 minutes, a completely faulty construction. With overheating i mean LM7805s running continuously in thermal clipping mode. That is much more difficult to understand to me. I can't believe nobody noticed that at IOTECH and nobody noticed it at Keithley.

Regards, Dieter

[muvideo]

Reading your comments seem that a manufacturing defect can explain the problem,
this is reassuring, I would prefer that instead of someone tinkering with the internals.
I agree that this kind of equipment was assembled manually and piece by piece so small
inconsistencies are the norm.
Maybe the soldering was marginal (the selector pin is wetted) and the shipping treatment
has detached finally the wire from the pin.

[dietert1]

Yes this device is from the1960s. There was nothing like a cheap computer for automated production or testing. They would have needed something like a robot to turn the knobs and an expensive computer like a PDP11. So the quality check was manual, too.

Regards, Dieter

PS: Sorry the PDP11 was introduced in the 1970s, so it would have been an IBM mainframe.

[notfaded1]

My assembly language book in college (In the 90's) was for a PDP-11 even though we did it on a newer DEC VAX the coding is still done in the same way/syntax and hadn't changed really.  At least you found the problem @muvideo... that's the important thing.  I wouldn't hesitate to solder that back with some good 50/50 solder.  Despite all this anti lead stuff (also requires more heat) I still prefer good flowing Kester 50/50 and doing my solder joints quickly with a Metcal Soldering station.

Bill