DATRON 1071 - Klickety klack, AC board RL2 is set each measurement

[DC1MC]

Hello everybody, I'm struggling to repair a DATRON 1071 got from a rather well known German seller of used equipment, with the comment similar with "...I've got tired to replace the bloody tantalums, besides that, all is perfect...", yeah sure  .
OK, the display was nice and bright, the device arrived and besides a lot of dead capacitors, both tantalums and elkos, it was having the dreaded corrosion pattern, some chips were having some kind of talcum like dust on pins, the wires were strongly corroded (I had to replace the wires to the 5V regulator, among others) and the front plastic mask was having the glue dried and the orange filter curbed.
Alright, replaced the caps, redid the wiring (with similar teflon isolated wire), reglued the front panel mask, did some amateurish  basic calibration and let it run to test the long term stability.

Now comes the strange part: after 8-10 hours, I hear some relay switching ca. 2 times/s !!!, everything seems to work OK, I open again the lids and is RL2 on the AC voltage board, run tests, test run OK, switch to AC, now TWO relays (one on the analogue board (the small one) are clicking, strangely the measurements seem to work OK, autotest procedure completes without errors !!!, just the bloody relay is clacking annoyingly.

I thought that the optocouplers went bad, I've measured them in and out (painful ) perfect defined signals, tried shuffling them, no change, WTF, look at the buffers and flipflops, nothing, all signals are perfect.

The issue is clear because of the digital command that comes during the acquisition procedure, if I put the device on hold and do manual sampling I got exactly the same behavior (one click-clack per conversion)

Yesterday the defect was gone when the device was cold, today is there from the beginning, I'm completely baffled, I was hitting the boards with a pencil, seem to not be a loose contact, what should I try next ?

 Many thanks,
 DC1MC

 Thanks,
 DC1MC

 

[coromonadalix]

try theses :   

https://doc.xdevs.com/doc/Datron/1071/  schematics begin at page 121 ...

and a eevblog thread :  https://www.eevblog.com/forum/metrology/datron-1071-repair/

[DC1MC]

Thanks for the answer, I've already got the schematic for the previous work I've described in the beginning of my post, it is most useful, or else the whole device would have been a lost cause.
The eevblog thread is very inspirational, although not related to my issue AFAICS.

The issue seem to be some kind of strange timing trouble, happening with the temperature (this is confirmed), that the AC board (sch pg. 163) operation selector flip-flop M5 seem to switch on and off the RL2 via some kind of very strange delay or integration circuit (?!?!?) involving a 0.1uF tantalum and some huge 1.5Mohm biasing resistors, as these tantalums are the spawn of the  , I  will replace it with a ceramic, but I was wondering if anybody could help me by explaining what this strange contraption there was supposed to do. Analog guys designing digital stuff are really creative  . 

In the end I will try to see if there's a difference in the waveform/timing with the temperature, but due to the huge resistors involved, the scope 1Mohm probe adds a significant influence .

 Cheers,
 DC1MC

 

[DC1MC]

Post again on  this topic just to update the status of the device and tell you that not only the bloody tantalums are evil, the other  parts are the opto-couplers, after changing even more tantalums (everything on the power lines and most of the signal ones) the instrument become more stable but the loooong term drift was terrible ( a 470R was shifting almost 100miliohms !!!!! in few hours from startup and the relay oscillation was there again) I've decided to bite the bullet and for a ridiculous small amount I've got 10 double and 10 single NOS ones and replaced the whole optocoupler gang there.

 WHAT  A GLORIOUS IMPROVEMENT   , after 3 hours from cold start the drift is ca.  8miliohms and (so far) no clickety-clack yet.  Only for the drift and it's  worth replacing them all, bloody hell Datron with their stupid colored paint dots on  digital parts, really one has to optically isolate  any analog designer from the digital part.

 Moral of the story, no matter how good an cool they look, if over 20yrs. old, and it's for the device you want to keep, all elcos and optocpulers have to be replaced, or suffer later.

 Cheers,
 DC1MC

[DC1MC]

So it clacked again and this time is done for good, here are the few easy steps in case someone is interested:

- Replace the random exploding tantalums and dried elcos: fun level 2,5/10, get a reasonable stable device and a nice bright display (the HiVoltage elcos were in nanos range, with 100ohm ESR !!!)

- Clean the strange white oxidation, replace the teflon isolated but corroded wires, replace the dried fragile plastic screws (not any more) holding the voltage regulators and power transistors to the heat sinks, get a clean and stable voltage and no more burning smell from the overheated regulators and power transistors that were floating in the air.

- Get rid of all the miserable optocouplers used, get a very low drift and quick temperature stabilization with the closed lid ( 20min instead of 3hrs+).

- Be puzzled about why the AC board still starts switching the test relay after the device heats, at the start at every conversion  !!!
Observe the strange delay shit made in the digital domain by analogue engineers (never let them design digital stuff), around M16/M17, use a cooling spray to determine the area where the trouble is, use a 0xx Metcal tip to heat each chip individually, replace a CD4071 and mumble something sweet  in their direction . Get a real rock stable device.

- Find a nice forum member around Mannheim in driving/ICE distance with a nice V/A/R calibrator and lots of patience to help me calibrate <<<-- I'M HERE.

- Finally enjoy the warm glow of the display 


 So yes, if anyone want to help me advance and has a calibrator available arroung Mannheim, as above, please let me know.

 Cheers,
 DC1MC

[Johnny10]

How did you troubleshoot AC board?

[DC1MC]

How did you troubleshoot AC board?

That was "easy", there you have two types of tantalums: the ones that fail short on power lanes and blow up and the ones that fail shot on signal/reference voltages and produce realy strange effects including oscillations  !!!
I was kind of lazy a$$ and only replaced the power lanes ones, and payed the price, one relay coil filtering bastard took the drivig transistor with it  .

If you want to be 100% sure replace ALL of them if you've had even ONE failing, if you want to be 169% sure, replace al of them anyways, they will faill in the most inconvenient moment, eventully after  you've payed and had a nice calibration.

 Cheers,
 DC1MC