Partly Faulty Fluke 1900A

[wilhe_jo]

Hi again!

I'm just sitting here and debugging an old Fluke 1900A Counter.

Bought this some months ago because it was reasonably cheap, looked ok and was sold as "not working" because of low sensitivity.
(give it some cal and it will do I thought....)
Some weeks later I shot a Schlumberger dual counter with oven, gpib,...  really, really cheap (50euro) ...
so that little Fluke found its way to the cellar and waited for fixing...

Now I'm at home fore some weeks because a really dumb accident while relocating to my new flat and I needed something todo.
That's when I remembered that Fluke counter and searched for it in all that boxes.

After a day of debugging I found that that everything was in cal but someone just removed the batteries (that unit had that option).
The probelm: Fluke used  the NiCads for regulation (!) and buffering.

I quickly wired an 7805 up, added some more caps and what should I say, it works with sensitivity within spec (25mV spec, 15mv Typ, after some tweaking I got 12mV ).
Well so far so good, 2kHz to 75Mhz works (the unit goes up to 80 but I can only generate up to 75MHz) but below 2kHz something strange happens.
The 74LS112 seem to generate some spikes and glitches...
Since all that 74-logic was powered from 7.7V (no NiCad-Regulators) I suspect them at least partially faulty but where to start.
My guesses are the counters (74LS112+LS00 and 74LS90) but since none of them was socketed, I thought it's better to ask if someone has a better guess...

I plan to change them in the following order:

• LS112 (2- JK FF used as main decade counter in conjunction with the LS00)
• LS00 (nand)
• LS90 (ripple counter)
• LS196 (ripple counter)
• LS74 (d-ff)
• LS42 (decoder)
• 7447 (led driver)

Interestingly everything works if the LS00 and the LS112 are cooled down...
Thank's in advance for your educated guesses

Hans