40 year old PDP-8
dpi
my Straight 8 running again. For the last couple of months it has
been one thing after another going wrong. Over the weekend we were
running a program we wrote to diagnose a JMS problem that was driving
us crazy. Sometimes, it would set the return address to the address
of the instruction after the one it was supposed to return to. The
failure was time sensitive with JMS instructions in close succession
always working but if you spaced them out too much it would fail. So
we were running this program and suddenly the front panel light
display changed drastically. The stop key didn't work, it was really
peculiar looking. Turned out the +10V power supply was now reading
-2.25 volts. For anyone who has never looked at these the power
supplies are really simple things. Big transformer, a couple of big
diodes, three large can type electrolytic caps wired in parallel. I
looked at the diodes and they were fine. I looked at the voltage
across the caps and 10 volts on the one closest to the rectifiers.
Also 10 volts across the second cap. But -2.25 volts across the third
cap. WHAT? Turned off the power and wiggled the wires and I saw
sparks as the charged caps discharged into the load. So now I had
found the problem and was thinking it was a loss of spring tension and
some oxidation or something like that. I was certainly surprised when
the wire fell out of the crimp when I tugged on it. There was a tiny
burn on the edge of wire and a tiny burn on the edge of the crimp. It
was clear that when it was manufactured 40 years ago someone didn't
get the wire placed before it was crimped. I don't know what caused
it to finally quit working but I am guessing the only thing holding
the wire in contact was the red heat shrink tubing that they used for
insulation. Total time to diagnose and correct was about 10 minutes.
This was the easiest thing to fix so far.
We ended up correcting five other problems over the last weekend
including the JMS issue which went away when I replaced a shorted
signal diode and one that was reading 0.235V of forward drop. Typical
for the DEC 664 diodes is around 0.59V.
Although I am not certain I believe that this particular machine was
built in the month of October in the year 1969 making this the month
of its 40th birthday. The Hobbs meter shows 15034.5 hours of
operation so far.
Doug Ingraham
Rapid City, SD
P.S. I am beginning to hate the R211 board which is one bit of the
PC, MA, and MB registers each. I've had to replace parts on three of
the 12 so far.
David Gesswein
dpi <doug.i...@gmail.com> wrote:
Your other recent posting reminded me of this one I was going to reply to.
>I was certainly surprised when
>the wire fell out of the crimp when I tugged on it.
>
I had a wire pull out of the crimp on my 8/E without much force so bad crimps
were still present on later machines.
I had to remove and clean the Faston connections on my 8/I to prevent it
from saying the 5V was too low.
I have two 799 power supplies and now both of them have had an AC oil filled
cap fail. They are Mallory 2uF 660V made in August 1969. The one in the 8/I
went bang when I was getting it ready for a VCF east and the other on my 8/E
for the DECtape failed just before I was going to shut it down at the
end of this years VCF. It didn't go boom but is bulging. How much trouble
have people had with these?
Tim Radde
I can't say I have had any problems wiht crimped on connectors. My 8i
did have a
bad cap in the power supply, but it did not explode. It just wasn't
holding well.
So I replaced it. I happened to have a few extras that I tested to
make sure they
were good. I'd hate to have one of those big suckers explode. I was
lucky that
my TU56 had had the big caps replaced before I got it.