8 restoration update 20091123

[dpi]

Over the weekend we attempted to align the core using the setup
procedure. It didn't work out all that well and I am not sure why.
We ended up putting all the settings back to the values that we had
measured when it was working and now it will run the memory
checkerboard test for as long as I am willing to run it. At least a
couple of hours. I am going to revisit this at some point but not
right away. The next thing on the list was to get the 184 memory
expansion operating again. This provides an additional 4k of memory
and is pretty much essential if you want to run much of the software.
We looked at it and the only problem found was that all the addresses
ending in 33 octal could not be written to. Actually all we could
tell was that they always read as zeros. This is pretty
straightforward and is either an X or Y line selector problem. One of
the things I found out was that these cards which are double sided and
plug into the right hand side of the core memory have alignment
problems. You can find references to filing the edges of the cards so
they can be seated correctly. I have found cards where the front and
back fingers dont really line up too well. These are the G603 (memory
selector matrix) cards and I would recommend that with a light shining
in from the right, you re seat all these cards. You can see when the
gold contacts line up by the reflections off of the card fingers and
you can align them so that they go in correctly. Many of the ones in
my machine are sloppy enough that the fingers on the card will short
adjacent contacts on the edge connector. Once seated correctly they
don't seem to have much tendency to move off center but I would guess
that eventually they will and I will need to re seat them at that
time. These cards have 64 diodes and 16 nearly irreplaceable
transformers on them (I will pay good money for a spare card or even
some of the transformers.) It looks like they had manufacturing
problems with them early on because on a lot of the boards they have a
number of these transformers replaced by ones from a different
vendor. I feel fortunate that so far I haven't had to worry about
making or finding some sort of replacement. Anyway after re seating
the first time the problem moved to the addresses ending in 12 (was
ending in 33) so all I did was change the problem. It was at this
time I carefully re seated the cards and wonder of wonders the field 1
memory now passes the checkerboard test as well.

We ended up the evening reading in a few more paper tapes from my
library. Warren is collecting all the paper tape images he can find
on the internet and is trying to regularize them with consistent
leader and trailer lengths so that he can deduplicate them. At the
moment he has something like 830 tape images. I would be surprised if
70% of these are not duplicates. I will put these on my web site at
some point. I would like to associate these tapes with pdf's of the
associated manuals also.

Next on the list is the EAE.

Doug Ingraham
Rapid City, SD