8,192 words of core memory with an 18-bit word length
1.0 usec cycle time
Real-Time Clock
Operator's Console
300 cps Paper Tape Reader
50 cps Paper Tape Punch
10 cps Console Teleprinter, Model KSR-33
The processor cost $35,000 in 1968.
The TC59 tape controller cost $10,000.
The TU20 tape drive cost $12,000.
The cost in 2013 dollars of this system would be $315,421.32.
 This is the I/O controller serial number tag.  Click on the image for a larger view. A listing from DEC showing the original system configuration.
The original configuration of the PDP-9 according to DEC. There is a date of January 17, 1968 on many of the CPU fans. That is about right for the manufacturing date of this system. The official production run was 445 PDP-9 systems. Additional systems were built on special order after production ended. The PC0 paper tape reader/punch is S/N 863. The TC02 DECtape controller is S/N 5048.
The PDP-9 was a follow on to the highly successful DEC PDP-7 series of computers. While there were some differences, one of the PDP-9 system utilities would convert PDP-7 programs to run on the PDP-9. Input was by paper or magnetic tape.
This system was purchased by Max Levy Autograph in October 1968 and shipped to Concord Controls who designed and built the controller for the 60" plotter. This was one of the largest such devices in the world. It was used to prepare glass for etching as filters and measuring templates. The plotter rested on a five foot thick granite slab, with supports set in bedrock. The PDP-9 and plotter was then shipped to of Philadelphia, PA and ran 24/7 from 1968 to the summer of 1999. When the memory limitations of the PDP-9 caused problems with programming they bought a PDP-11/23 and connected it to the PDP-9 through a special I/O controller that was added to the plotter controller. We were contacted by the person who operated this system for 30 years. He filled us in on much of the history of the system and will come to the RICM for a visit. The notes that are taped to the front of the system are his. This PDP-9 was located for the RICM by Kevin Stumpf, and donated by Max Levy Autograph of Philadelphia, PA. A total of 445 PDP-9 systems were produced, and so far, only nine are known to have survived. The current locations of these nine include a PDP-9 and a PDP-9/L in California, one in Australia, one in Germany, one in Great Britain, two in France, two in Sweden and a new one #203, and of course our own, seen below. Parts of #209 that was delivered to Australia Iron & Steel were recently found. There are lots of details on the developers of the PDP-9 here. Apparently the design of the ROP Control Memory in PDP-9 systems caused lots of problems. Lichen Wang from SLAC developed an improved design that proved more reliable. The description is here.  Click on the image for a larger view.
Front view of the TC59 Tape Controller for the TU20.
 Rear view of the TC59 Tape Controller for the TU20.
Click on the image for a larger view.
The image below is of the TU20, 45 IPS, 7-track tape drive that was attached to the PDP-9. It supports 200, 556, and 800 BPI tape densities. From the label on the rear of the transport it appears to be an HP 7975A made for DEC using IBM's tape velocity fluctuation damper patents.     Click on the image for a larger view. The capstans and read-write head on the TU20 Tape Drive.
Mattis Lind made an Arduino interface for his DEC PC04 paper tape reader. He read some of his PDP-9 diagnostic paper tapes and emailed the paper tape images to us. We sent the paper tape image through the serial port on a PDP-8/S to a paper tape punch to recreate the paper tapes. A video of the paper tape reader in action is here.
|  The PDP-9 Console.
The PC09 Paper Tape Reader/Punch located above the control panel. The PC0, S/N 863, contains a PC02 Paper Tape Reader and a PC03 Paper Tape Punch. The cabinet to the right contains the TC02 DECtape and the TC59 Magnetic Tape controlers. The magnetic tape drive to the left is a TU20 7-track drive.
 Click on the image for a larger view. If the system was run 8 hours per day, this shows 20 years of run time.
 Click on the image for a larger view. The PDP-9 system console.
 Click on the image for a larger view. A rear view of the system console.
 Click on the image for a larger view. Front view of the system chassis.The missing boards in the upper memory chassis are for the Parity option. The missing boards in the middle processor chassis are for the EAE option. The missing boards in the bottom I/O chassis are for the KF09-A Automatic Priority Interrupt,
the KP09-A Power Failure Detection, and the 34H-0 display options.
 Yellow wires are factory installed. Blue ones are design changes.
The green ones could be design changes or additions.  Click on the image for a larger view. This is the board that holds the microcode for the processor. A pulse is sent through one of 64 wires and microcode bits are coupled through the transformers.  Click on the image for a larger view. The DEC 709 Ferroresonant Power Supply.
It makes 3A of +20VDC, 5A of +10VDC, 20A of -15VDC, and 10A of -30VDC.
You can see some evidence of a past diode block failure.
 We found this FlipChip chassis in the collection quite a while ago.
We assumed that it was part of a PDP-8 system and set it aside.
During a recent inspection we noticed that it has the nylon block that is used to secure a PDP-9 or PDP-10 I/O cable.
We now think that this is probably the interface that connected the PDP-11/23 to the PDP-9.  Charles Lasner donated a TC02 DECtape controller for 18-bit DEC systems. When we get this working, we can run a real operating system on the PDP-9. One of the TU55 DECtape drives in the PDP-8/I is an older design, contemporary with the PDP-9. We could borrow it, and one of the slightly newer drives for the PDP-9.  Click on the image for a larger view. We installed the TC02 DECtape controller above the TC59 1/2" Magnetic Tape controller.  Click on the image for a larger view. This TC02 display panel is very close to the TC01 display panel on the TC01 in the PDP-8/I, but the DATA BUFFER will display 18 bits, instead of 12 bits on the TC01. The larger DATA BUFFER display required moving other indicators to make room. |
Home > Large Systems Collection >
