ADM-3A question

[Peter Coghlan]

Charles wrote:
> Today the replacement 'LS193 arrived, so I put it in the 
> previously-installed socket and the screen is now 24x80 again :)
> I'd been testing with the dip switch in half-duplex mode... For final test, 
> I put it in FDX, connected to my HP protocol analyzer, and what do you know, 
> no serial data out.
> The 1488 RS-232 driver was blown (TTL-level data going to it, but the RS-232 
> line was stuck in spacing). Not an uncommon failure with static discharge 
> and incorrect cable hookups...
> I destructively removed it, again installed a good dip socket, now waiting 
> for THAT chip (a variety of other line drivers in my drawer, but not 
> 1488/1489).

I'm glad to hear you got to the end of the chain of faults on your ADM-3A.

I am still struggling with my ADM-5.  The smoke I mentioned last time came
from a tantalum capacitor decoupling a -20V supply.  After removing it, I
got back to the monitor showing a cursor.  I tried sending and receiving data
and found neither operation worked.  While I was looking into this problem(s),
without any provocation the monitor went off, with the picture collapsing to a
vertical line and fading out with no nasty noises or any other signs of
distress.  From previous experience, I checked the HDRIVE signal to the
monitor and found it was absent.  (The usual beep at power on did not happen
after this either). Luckily, the ciruitry in this area seems to be identical
to the ADM-3A so I can use the ADM-3A maintenance manual to troubleshoot it,
however the component references are all different so it can still get very
tedious. (I also found some frequencies listed on the timing diagram don't
seem to be correct which doesn't help matters either).  I eventually found a
74LS161 which is used to divide the master clock signal was not behaving
correctly.  Not having a replacement, I "borrowed" one that seemed to be
working correctly from my other ADM-5 which has even more issues.  This made
the waveforms around the dot counter and character position counters look much
more reasonable, however, I still don't have the HDRIVE signal restored yet
so there must be another problem lurking and when I get to the bottom of that
one, I will probably still have the data transmission and reception issues...

While working on these latest issues, I remembered that the faults that led
me to replacing the two 74LS125 chips on a previous occasion were certain
received characters being incorrectly displayed due to stuck bits.  I also
noticed that I had to replace a 7805 voltage regulator during that
troubleshooting session.  I'm making lots of notes this time around so I won't
have these recall difficulties next time around.

This ADM-5 was stored in an attic space with poor temperature control for
many years but since then and before the most recently failures, it had been
in a less variable environment.  The 74LS161 that failed was clearly working
initially when the terminal came out of storage this time.  It didn't seem to
be a lead bonding issue either - the MSB of the count output seemed to be
following the clock input.

Every time I try to use my ADM-5 terminals, I seem to run into these kind of
successive failures.  I have some BBC Micro equipment from the same era which
also contains a fair amount of TTL but it doesn't give anything like the same
degree of trouble, although I did have to replace a single 74LS163
counter/latch in my oldest BBC micro.  It was wired to be a latch only but it
decided it's true destiny was to be a counter.  This failure could have been
a lead bonding issue.

Regards,
Peter Coghlan.