[vcf-midatlantic] Repair Workshop Wrap-up 3/30 & 3/31

[David Gesswein]

On Sun, Mar 31, 2019 at 04:07:37PM -0400, Jeffrey Brace via vcf-midatlantic wrote:
> David Gesswein worked on the museum's PDP-8. He will send a report later.
>
Had the annoying two steps forward and one large step backward. The goal
for the weekend was to install the reader run modification to the teletype
which allows the PDP-8 to stop the tape reader and to fix the bulbs not
working right on the front panel and test using the tape reader. End result
is new faults developed so the machine isn't useable without more repairs.

I ran all the diagnostics when I started and all passed without error.

I had made the reader run board at home and tested it so it was supposed to
be easy. I got it installed without problem but when loading FOCAL programs
I got 11.35 errors which are input buffer overflow. In doing some research
this seems to be a design error in FOCAL. Normally when reading a tape it
prints everything received. It uses interrupts to read the tape into a small 
buffer which feeds the loading into memory and printing. The reading of the
tape does not use the reader run control to stop reading when the input
buffer is getting full. If the printing which is controlled by the transmit
clock and any software delays is slower than the reading which is controlled
by the mechanics of the teletype you will get an overrun. I need to check
further but I don't think the tape reading speed is adjustable. The transmit
clock had pots so I may be able to adjust it slightly faster as long as the
teletype is ok with it. I also made a tape with nulls to try to slow down
the reading but was unable to test due to faults that developed.

I was able to load the FOCAL interpreter from tape and it ran fine.

The teletype also had a capacitor go open that caused buzzing noise and
operation problems. One of the workshop attendees had a capacitor that was
usable to fix that problem.

I moved on to the front panel issues. I finally tracked down cause of a bulb
that would come on at varying brightness when it shouldn't to a diode on the
cards that have the wires that go to the front panel. I checked the rest of
those cards and found two more suspect diodes and replaced them also.

Now that everything was supposed to be working I tried to load FOCAL again
and it looked like it loaded fine but didn't run correctly. I tried to
load diagnostics and they wouldn't run. Ran the moving AC light demo program
and the bits disappeared when going into AC bit 3. Symptoms tended to change
some as I was troubleshooting but did find moving the AC3 card to another
AC bit moved the problem. I was unable to determine the failing component
on the card before the end of the day Sunday so the machine is currently
non functional unless the problem decides to go away on its own for a while.

I'll bring an AC card to VCF to replace it and hopefully that will get it
working again for VCF. I'll then take the card home and see if I can get it
to fail in my machine so I can fix it. It has 123 diodes, 12 transistors,
63 resistors, and 23 capacitors. The diodes are the normal fault but none
measured suspect so I will need to use the test equipment and schematic
to figure out what is going wrong after I can reliably get it to fail.