PDP-8 core memory problems.

[Mattis Lind]

2016-09-07 17:33 GMT+02:00 Doug Ingraham <dpi at dustyoldcomputers.com>:

> The most likely cause of what you are seeing is a broken wire when the
> plane was originally assembled.  The wire was pulled back a few cores and
> the end stripped.  New wire was soldered to old, insulated and then they
> continued threading in that wire.  Over the years the solder joint has
> degraded or the wire broke at the stress riser found at one end of the
> solder joint and now you have an open circuit. I've not heard of this kind
> of problem on the Straight 8 but that may be due to the rarity of the
> processors.  It is apparently a fairly common failure on the 8I core
> planes.
>
> As was stated you have nothing to lose in attempting a repair as the core
> is useless as is.  A steady hand, good desoldering tools, lots of photos
> and you should be able to take it apart, effect the repair and
> re-assemble.  Keep in mind that the core beads themselves are extremely
> fragile so take precautions that nothing gets dropped on it.  Broken core
> beads are pretty much a death sentence to the memory.  Replacements are
> unobtanium and if you decided to make the beads you would have trouble
> matching the originals well enough to tune the core to work with both new
> and old.  You would end up making a whole new core assembly consisting of
> 49152 beads.  You would need to be really determined to attempt that.
>
> I did come up with an idea that is simply too dangerous to try.  Connect a
> power supply to the ends of the wire and ramp up the voltage until it just
> starts to conduct.  This could be several hundred to several thousand
> volts.  As soon as it starts to conduct the broken ends of the wire will
> start to heat and the moment the current starts to shoot up (the resistance
> drops) you need to cut power.  You will have welded the broken ends of the
> wire together.  The problem is that if anything goes wrong you are in worse
> shape than now and you really only get once shot at it.  And the assumption
> is that the broken ends are in close proximity.
>
> Here is wishing you a steady hand and lots of luck!
>
>
Thanks Doug and Anders for the encouragement. I have decided to put the
core in a safe place while I am gathering tools (stereoscopic microscope
and better tweezers). The core repair project will be on hold until some
vacation day when I have plenty of time and there are no disturbances
around. Possibly in the mean time I will look into the instruction decoding
problem that affects the front panel operation sometimes.

I prepared a small video on the non-operating machine:
https://youtu.be/8WYiUz4SNi8

/Mattis Lind
PDP-8 s/n 351 (non-operating)



> --
> Doug Ingraham
> PDP-8 SN 1175
>