Teamwork, and PDP-1 status (was Re: TU56 restoration)

[Tom Uban]

At 03:10 PM 3/16/2005 -0800, Eric Smith wrote:

>Ashley wrote:
> > Yes, it is more fun (and educational!) if more than one person is working
> > on the same project.  Sharing knowledge, successes, and failures is a
> > good thing.
>
>Definitely!  I've found the team work on the PDP-1 project to be very
>rewarding.  Previously Mike and I tried to get a PDP-11/70 working on
>our own, and although we did make a fair amount of progress (before
>his house burned down), it was very slow going.  The PDP-1 is in many
>regards a more difficult machine to work on, yet the combination of
>skills and interests of the team make the project work much better.
>We still have sometimes gotten stuck on "grunt work", like the months
>we spent on reforming all the capacitors, but at times we've made
>very rapid progress.
>
>Latest PDP-1 status:
>
>We are now trying to repair a memory fault.  When we first ran the
>memory diagnostics, all 12K worked.  Now one 4K bank has problems.
>We think we've identified the faulty component, so next week we'll
>work on it.  I'm probably going to have to write a different memory
>diagnostic; the one DEC supplied has to be loaded into the same 4K
>bank it is testing, which was required for it to work on 4K machines
>but is rather suboptimal otherwise.
>
>The Spacewar control boxes Ken built (with some minor help and/or
>hindrance from me) mostly worked OK last week, but the hyperspace
>buttons had a strange interaction with other buttons on both
>control boxes.  The hyperspace inputs on the original cable from
>the PDP-1 were wired using diodes at the PDP-1 end so that the
>hyperspace buttons would activate both the rotate left and rotate
>right bits, which is how the software detects it (rather than using
>a distinct bit position).  It appears that those diodes weren't
>working well (!), so Ken put 1N4148 diodes in the control boxes
>to serve the same function, and left the dedicated hyperspace control
>line open, and that seems to work better.

Do you have an original set of Spacewar control boxes? I'm curious
if there was originally a dedicated hyperspace button or not. The
control boxes on my Imlac never had a dedicated button and relied
on the player to push both the turn buttons at the same time. I
would think that you would want the controls to be historically
accurate...

>There was some controversy over whether the control boxes should
>be wired to prevent manually pressing the rotate left and right
>buttons simultaneously from activating hyperspace.  Some players
>don't like that because they do it accidentally when not meaning
>to enter hyperspace; I like it as a deliberate way to do so.  At
>Peter Samson's suggestion, Ken rewired the rotates so that the CCW
>rotate switch is in series with the normally closed contact of the
>CW rotate, so that if you press both, CW takes precedence.
>
>We found a really nice display demo called "SNOWFLAKE", which
>produces hypnotic kaleidoscope-like images with sixfold symmetry
>(hence the name).
>
>Our initial attempts to use the light pen were unsuccessful.  It
>is possible that this was just due to the sensitivity adjustment
>(a knob on the underside of the Type 30 CRT housing) being too low;
>at the time we didn't think to try to adjust it.  We also need to
>test the programmatic intensity control, and the character generator.
>We don't have any diagnositics for the character generator, so I'll
>write something for that.  Possibly I might use that as the status
>display for the new memory diagnostic.
>
>Next week we'll also continue trying to get the typewriter working
>again.  There were mechanical faults which we believe are fixed, but
>then electrical issues cropped up.
>
>We expect the new belt for the BRPE paper tape punch to arrive next
>month, and it will be nice to get that going.
>
>I guess it's amazing that any of this 40-year-old stuff works at all,
>and not too surprising that various components are now failing.  We're
>at the far end of the bathtub curve.  We're hoping that the condition
>stabilizes soon.  We have not yet run the system on voltage margins,
>which should help identify components that are likely to fail soon.

Do you have any records which indicate the reliability of the machine
from back in the day? I've always been under the impression that machines
of that era were not terribly reliable and so perhaps expecting them to
be so now is unrealistic.

The progress that you all have made on the PDP-1 restoration is very
impressive! I am totally jealous that I don't get a chance (due to
geography) to visit and see the fun more frequently. I was out for
VCF last year and did get to see the machine, which was quite a thrill.

--tom

>Eric