8 Restoration progress
dpi
desire to talk about the progress we are making on getting it going
again.
This last week we made pretty good progress. By we I mean Warren
Stearns, who also is a lover of old dusty computers, and myself. We
both used this particular computer in college back in the 1970's. I
obtained the machine in the mid 1980's. It is an 8k straight 8 with
the power fail interupt and EAE. For peripherals it has a two platter
DF-32 and three TU-55 Dectape drives. It has a high speed paper tape
reader/punch and the 12 bit A/D converter with the 64 input mux. At
this point the processor seems to be working again as a 4k machine.
The next big thing is to get the second bank of memory working
properly and increase the time it will run the memory checkerboard.
At the moment it runs about 2-3 minutes before it picks a bit and
alters the program. At one point I had it running over 9 minutes
before it picked a bit but something is degrading. Earlier in the
week we started running diagnostics and ran into an annoying problem
with the front panel. Sometimes when you would press the load address
key the machine would start running and when it did this the stop key
wouldn't work. After looking at it for a while we discovered that it
actually would do this with deposit and examine also. The problem
turned out to be with stopping the machine during the pause portion of
an IOT instruction. I found the transistor responsible and replaced
it. Now the switches are working better than I ever remember.
Thinking back I suspect the machine had this problem even when we were
actively using it for classwork in the early 1970's. Someone had put
some capacitors across some of the switches in an attempt to cure this
but nothing they ever tried worked. I have removed all of those
additions now as well. After that success I decided to fix the few
broken lamps. The IOT instruction indicator, bit 1 of the MQ and the
MSB of the Instruction Field and Data Field indicators were non
functional. I had only one spare lamp so I decided to make a couple
of LED replacements and put those in the upper bits of the Field
indicators since they are not normally used and reuse those bulbs
other places. I had taken the panel apart once around 1985 and after
I got it apart this time I remembered just how awful it is to put back
together. In the process of trying I managed to damage 3 more bulbs.
The bulbs used have no support and the wires are probably 31 or 32
gauge, about like a coarse human hair, so are very fragile and unless
you are extremely careful they will tip over and not go into the hole
in the front panel and you stand a really good chance of breaking
wires right where they go into the glass. And that is what happened.
It looks like the bulbs are available so I may just buy a box of 100
and that should be a lifetime supply. My LED replacements are about
the right brightness but there is a problem I didn't anticipate when
using them. The come on when they shouldn't because of noise and
crosstalk in the wire bundle running to the front panel. The bright
white LED's I used come on at 2.64 volts. Across the bulbs there is
13.9 volts when they are turned on. If you want a milliamp of current
that means (13.9-2.64)/0.001 = 11260 ohms in series with the LED. I
have 10k ohm resistors so the current is actually 1.1ma. After
looking at it they could stand to be a little brighter but they are
not too bad. The real bulbs pull about 30ma each. So if you could
have all 88 bulbs lit at the same time it would draw 2.6amps from the
-15 supply. An LED replacement at even 2ma would be only 0.176 amps,
a considerable difference. 39 watts vs 2.64 watts. Like I said, they
glow when they shouldn't when the machine is running code. The lamp
drivers are spiking up to between 4 and 6 volts which is enough to
make the LED glow dimly. These spikes occur every 1.5us and the time
when goes above the threshold is perhaps 100ns. I suspect I can solve
this by placing a small cap across the leads of the LED which will
keep the voltage below the threshold. It appears these LED's glow
visibly on 140ua, 100ns pulses. The light they put out also looks
very blue compared to the yellow/orange lamps. But it is a very
pleasing color that I could live with if it meant never taking the
front panel apart again.
We have been keeping a log book of all the work we are doing and some
day I should put up a web page and chronicle the effort with photos
and perhaps even a sound recording of the fans running. They are
louder than you might think. There is also a distinctive smell in the
air because of the filtercoat on the air filters.
If you object horribly or find this posting interesting let me know.
I will weigh the two and decide if I should post more or less often.
Doug Ingraham
Rapid City, SD
Roger Ivie
> If you object horribly or find this posting interesting let me know.
One vote for interesting, and looking forward to pictures.
--
roger ivie
ri...@ridgenet.net
van...@vsta.org
> If you object horribly or find this posting interesting let me know.
> I will weigh the two and decide if I should post more or less often.
This is excellent to hear about, and I hope you will continue to
post updates. Good luck!
--
Andy Valencia
Home page: http://www.vsta.org/andy/
To contact me: http://www.vsta.org/contact/andy.html
JackRubin
inherently interesting, but by posting it is now in searchable form to
help someone else in the future. Likewise with pictures.
Thanks,
Jack
Tim Radde
I vote to keep up the good work. It's always good to hear of someone
else restoring an 8.
David Gesswein
dpi <doug.i...@gmail.com> wrote:
>For peripherals it has a two platter DF-32 and three TU-55 Dectape drives.
>
in how yours goes. Looks like your making good progress.
FYI: I made a board to allow use of the DF-32 on my 8/I without needing to spin
up the platter. Right now the design is messy but works. On
the todo pile is cleaning it up more and reprogramming it to be useable
as a timing track writer.
http://www.pdp8online.com/shows/vcfe08/pics/dscf0065.shtml?small
Tim Radde
> In article <10a95745-1ba4-40f1-88b9-a3189b669...@j19g2000yqk.googlegroups.com>,dpi <doug.ingra...@gmail.com> wrote:
> >For peripherals it has a two platter DF-32 and three TU-55 Dectape drives.
>
> I may be involved in a straight 8 restoration (not mine) so am interested
> in how yours goes. Looks like your making good progress.
>
> FYI: I made a board to allow use of the DF-32 on my 8/I without needing to spin
> up the platter. Right now the design is messy but works. On
> the todo pile is cleaning it up more and reprogramming it to be useable
> as a timing track writer.
>
> http://www.pdp8online.com/shows/vcfe08/pics/dscf0065.shtml?small
I am even more interested in that as I have two DF-32s that I have
never tried to spin up knowing the head
to disk interference. So you still use the DF-32 electronics but this
plugs in place of the drive? Or
this replace the entire DF-32? Curious to know. VInce and I talked
about a board to plug into the
DF-32 so you could still use the electronics, but not spin up the disk
itself.
David Gesswein
The board looks DS-32 expanders so you're using most of the electronics
of the master drive. It plugs into the connectors you would plug DS-32s
into. I though about making it plug in place of the platter connector
but decided on the other method so I could make only one board but
emulate multiple drives so you have a reasonable amount of space. Since
the cable to the platter on my master drive went bad I have it hardcoded
to emulate 4 DS-32 so the master drive unit select is set to off. I put a
switch on the board I was going to use to switch between two images (the
memory was big enough to store 8? DF32 copies. It was aso going to be used
to select how many DS-32 it emulated so it could run with the master
drive enabled. That logic isn't implemented. I either need to come up with
more efficent equations or put a bigger part on the board to finish that.
I also have created a version of my dump and restore program for the DF32
to send images over the console port to a PC that it looks like I haven't
packaged back up and put online yet.
Mike Roach
<doug.i...@gmail.com> wrote:
>I don't know if there are people interested or not but I find I have a
>desire to talk about the progress we are making on getting it going
>again.
Oh, I'm definitely interested! If you're up to it you could document
your activities with word and pictures using canned software and sites
like Blogger or Word Press. There are other packages available.
--
The plot was designed in a light vein that somehow became varicose.
-- David Lardner
Rich Alderson
> TIME Magazine Person of the Year for 2006 dpi
> <doug.i...@gmail.com> wrote:
>> I don't know if there are people interested or not but I find I have a
>> desire to talk about the progress we are making on getting it going
>> again.
> Oh, I'm definitely interested! If you're up to it you could document
> your activities with word and pictures using canned software and sites
> like Blogger or Word Press. There are other packages available.
The only problem with blogging sites, canned or otherwise, is that readers
cannot save off copies of messages for themselves. This is one of the great
features of Usenet news, IMAO.
--
Rich Alderson "You get what anybody gets. You get a lifetime."
ne...@alderson.users.panix.com --Death, of the Endless
Ty Cobb
"JackRubin" <jack....@ameritech.net> wrote in message
news:a3ad0e3c-0fb7-4117...@l2g2000yqd.googlegroups.com...