The University of Iowa's DEC PDP-8

Restoration Log

Part of the UI-8 pages
by Douglas W. Jones
THE UNIVERSITY OF IOWA Department of Computer Science

Introduction

This is a chronological log of the progress restoring the University of Iowa's PDP-8 computer. Entries are added at the end as work progresses. Click on any thumbnail image to see full-sized image.

Mar 19, 2019, Reverse Engineer G007 Sense Amplifier

Bug 64: We reverse engineered the G007 sense amplifier board, on the theory (advanced in the log entry for Dec 6, 2018) that one of these boards might be pulling the Memory Strobe signal to ground. The parts are labeled with the part numbers given in the G007 schematic in ght Feb. 1966 edition of the maintenance manual. Our board is a G007C, while the board in the manual is a G007D. The difference between these outlined in the reverse engineered board with an orange square outlining a DEC D664 diode that was later replaced by R29 in Rev D.

This does not agree with the G007C schematic. on Vince Slyngstad's web site, where a resistor is also shown. Evidently, DEC experimented with different parts while continuing to use the same revision of the board etch.

May 17, 2019, Reassemble TTY, fix G007 diode

The Teletype
Bug 26: In preparation for the University of Iowa Tech Forum on June 6, where artifacts from the Retrocomputer Lab will be on exhibit, we put the cover on the Teletype for the first time since we began work on it. Looks good!


The Teletype

Bad diode replaced
Bug 64: We used an ohm meter to check all the diodes on the G107 sense-amplifier boards, and found one of them that was bad, measuring as an open circuit in both directions. The diode in question was D15 on the G007 schematic in the Maintenance Manual. DEC's part number for this diode is D668, and their semiconductor substitution instructions in the Maintenance Manual (Table 10-1, page 10-5) say to use two D-664 (1N3606) diodes in series.


Bad diode replaced

As can be seen in the photo, showing the repaired board (bottom) and a board with a good diode (top), the pair of diodes in series is a bit bulky. We painted the transparent glass body of the diode with black paint because the forward frop of this diode is used as a voltage regulator, and the forward drop of glass-encased silicon diodes is notoriously light sensitive.

June 4, 2019, Get BRPE punch, straighten door trim

The BRPE punch Problem plugs
When we recovered the PDP-8 from its brief stay at university surplus, we did not get the high-speed paper-tape punch that the psychology department had used with the machine. That was apparently a Tally punch, while the PDP-8 Users Handbook (page 64, 1966 edition) says that DEC supported Teletype's BRPE reader. When a pair of old BRPE readers became available, we got them. The machines were used by Jackson Typesetting, in Jackson, Michigan, acquired through Dan Foust.


The BRPE punch	Problem plugs

These machines pose some problems, because they are configured for 7/8-inch paper tape, suitable for punching with 6-bit codes, but with two punches, there is hope that we can combine parts from two punches to make one punch that works with 1-inch tape and 8-bits per character.

A second problem is finding the plugs to connect to the machine. The 3-pin twist-lock connector for power may be harder to source than the 24-conductor ribbon connector since this is mechanically (but not electrically) identical to what later came to be known as a Centronics connector.

BRPE, is an example of a naming scheme Teletype corporation used. The root, RP, stands for reperforator, because any punch that was electrically driven was assumed to be creating a duplicate of a paper tape that had been originally perforated elsewhere and then read. The final E seems to have been used on all high-speed punches, perhaps indicating an electrical interface. The first letter was assigned sequentially; according to Nick England, the first in the series was the ARPE a magnet controlled reperforator. The CRPE may not have made it to market, but Jim Haynes says that there was a DRPE tuned reed punch.

Door trim
before and after
In preparing to exhibit the PDP-8 at the University of Iowa Tech Forum, we noticed that the bottom trim strip on the left front door of the PDP-8 was bent. Removing the door and banging on the back of the trim strip was sufficient to largely remove the bend, but after doing this, it was clear that whatever had bent the trim also bent the metal door panel itself. We left that alone for the time being. In the before and after photo to the right, the slight bend in the door panel is only apparent after the trim strip was straightened, and only when viewed looking down from an extreme angle.


Door trim before and after

June 6, 2019, Retrocomputer Museum Exhibit

The exhibit

An appreciative audience
photos ©2019 Michael J. Jenn, U. of Iowa ITS-Enterprise Services
The whole point of restoring old computer equipment is to educate, so when the organizers of the University of Iowa Tech Forum asked if we had anything worth exhibiting, I volunteered. Later, as it turned out, we were asked to abandon the space in the old Communications Center (the former home of the School of Journalism), so we ended up coordinating the move out of the old space into new space in Jessup Hall with the Tech Forum. We exhibited

The exhibit was very well received. We left the Teletype turned on with the paper-tape punch on so that people could type (and punch) souveniers, and we left the PDP-8 on and running (unfortunately, with all memory reading as zero, the program was just a sequence of AND instructions.)

The exhibit


An appreciative audience
photos ©2019 Michael J. Jenn, U. of Iowa ITS-Enterprise Services