Alan, if you want some detailed info regarding my 5135A repair, let me
know. I still have to get my head around what's going on in there, and
organize my documents and notes.
I did some experimenting last night, and tacked some wires to the output
latch in this section, to bring the results out to a header connector. I
made some quick scope measurements to see if these are just
combinational logic decoding, or are sequential, and part of the
counting process. They appear to be simply for decoding, but made little
sense, and I decided I'd better sleep on it for a while.
Today I set it up again with a counter, and planned to record the 8-bit
logic pattern for every 1 kHz step from 0 to 99, then compare the values
where the frequencies were correct, and where the 8 kHz error shows up.
To my surprise, the 8 kHz problem had disappeared! All the steps are
right now, and the few patterns that I noted last night look just the
same and as weird as for the steps today. The digits at this point
appear to be only partially or specially decoded - not at all a nice
clean two-BCD to 7-bit binary conversion that I expected.
Somehow, moving the instrument to the scope setup, or attaching the
wires, had "fixed" it, but I didn't know it until looking with the
counter today. I have moved and flipped the unit many times while
working it over, and that DB1 board has been pulled and remounted dozens
of times for inspection, shotgunning half a dozen ICs, and touching up
maybe a couple hundred solder joints - all with zero effect. Now,
tacking wires onto the LS373 that I had already replaced, seems to be
what did it.
I hate it when I don't find the real cause of a problem for sure. I'm
assuming something was intermittent due to the questionable condition of
the board metallization (hence all the resoldering I did - another form
of shotgunning), and I finally hit the real problem spot one more time,
which was sufficient. The only other thing I can think of is a solder
glob or a wire clipping was stuck somewhere, and finally fell out. I'll
look it over again, especially in this area, but I think I'll be
proclaiming it a done deal shortly.
If anyone has or wants more info pertaining to this project, please
contact me off-list - no need to continue here.
Ed
_______________________________________________
time-nuts mailing list -- [email protected]
To unsubscribe, go to
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.
