DEC PDP-8
Restoration

2026

-----------------------------------------------------------------------------------------------------------------------

4-JAN-25

Grant wrote a program that will accept text from the console and then punch the paper tape 

11-JAN-25

We loaded and ran MAINDEC-801-3A PDP-8 Instruction Test (EAE_Type_182)-Part 3A. This diagnostic tests the MQL, MQA, SHL, LSR, NMI, SCA, SCL instructions. After about 30 seconds it printed "3A" on the console an restarted the test. We let it run for 30 minutes and declared victory.

20-Aug-25

Grant's Bannerrama program would not run. We toggled in a serial console loopback program and saw garbage when we typed in characters. We toggled in a program to increment the AC and load the data into the serial port. We looked at the outputs of the flipflop on pins E & J on the R205 in slot ME25. We didn't see any changes to the flipflop outputs. We looked at the input on pin H and the latch signal on pin M. The BAC data looked OK, and the latch signal was about 100 ns wide.

The incremental printing program is stuck on the TSF instruction waiting for the character to finish being sent to the serial console.

We also found some memory addressing problems were addresses in the 0004 range didn't work.

27-Aug-25

We loaded and ran MAINDEC-801-1-D PDP-8 Instruction Test Part -1 to see if the basic instructions and interrupts work. The diagnostic worked as expected.

We loaded and ran Grant's Banerrama program. That also ran OK.

We will do more experimenting on Saturday.

10-Sep-25

The PDP-8 is behaving strangely. The console UART will not work when the system is first powered on, but will after fiddling with it for a while. We finally traced the issue to the R405 FlipChip that generates the 7.05 kHz signal that is divided to make the TTI and TTO clocks. We cleaned the fingers on the R405, but that didn't help. Tapping the crystal will start the clock running and allow the UART to work. 

6-Dec-25

We moved the PDP-8 into the new RICM Lab space. You can EXAM memory, but you can't DEP memory. During a DEP the Memory Buffer contains the switch data, but it doesn't get written to memory.

Diagnostic Tests to run:

• MAINDEC-801 PDP-8 Instruction Test Part 1 (RIM Format), 11/23/24

• MAINDEC-801-3A PDP-8 Instruction Test (EAE_Type_182)-Part 3A, SA=0200, SW=5000, 1/11/25

• MAINDEC-801-3B PDP-8 Instruction Test (EAE_Type_182)-Part 3B, no copies of this diagnostic found so far

• MAINDEC-08-D01A-D PDP-8 Instruction Test Part 2A, SA=1200, SW=7777, 11/27/24

• MAINDEC-08-D02B-D PDP-8 Instruction Test Part 2B, SA=0200/2000, SW=0400, 11/27/24

• MAINDEC-08-D03A-D Basic JMP-JMS Test, SA=600, 11/27/24

• MAINDEC-08-D04B-D Random JMP Test, SA=0200, 11/27/24

• MAINDEC-08-D05B-D Random JMP-JMS Test, SA=0200, 11/27/24

• MAINDEC-08-D07B-D Random ISZ Test, SA=0037, SW=4004, 1/11/25

• MAINDEC-08-D0AA-D PDP-8 Instruction Test Part 3A EAE, SA=0200, SW=5000, Uses the SCA instruction, which the Classic PDP-8 does not implement, so it halted at 5406.

• MAINDEC-08-D0BA-D PDP-8 Instruction Test Part 3B EAE, SA=0200, SW=5000, 12/21/24

• MAINDEC-08-D0UA-D Family Of 8 Add Random Test, SA=0200

• MAINDEC-08-D1B1=D Memory Address Test (Low)

• MAINDEC-08-D1B2=D Memory Address Test (High), 11/30/24

• MAINDEC-08-D1L1-D Basic Memory Checkerboard (Low)

• MAINDEC-08-D1L2-D Basic Memory Checkerboard (High), 11/30/24

• MAINDEC-08-D1MA-D Memory Address Test (Low)

• MAINDEC-08-D1MA-D Memory Address Test  (High), 11/30/24

• MAINDEC-08-D2AA-D Teletype Reader Test

• MAINDEC-08-D2BA-D Teletype Paper Tape Reader Test

• MAINDEC-08-D2CA-D Teletype Paper Tape Punch Test

• MAINDEC-08-D2DA-D Teleprinter Test

• MAINDEC-08-D2PE-D Family-of-8 ASR 33/35 Teletype Tests Part 1, SA=0200, SR=0000

• MAINDEC-08-D2QD-D Family-of-8 ASR33/35 Teletype Tests, Part 2, SA=0200, SR=0000

• MAINDEC-08-D5BB-D DF32 Diskless Logic Test

• MAINDEC-08-D5BC-D DF32 Diskless Logic Test, SA=0100

• MAINDEC-08-D5CG-D DF32/DF32D Disk Data Minidisk, Interface Address, Data Test, SA=0100, SW=0000

• MAINDEC-08-D5DB-D DF32 Multi Disk

Addressing Modes Tested:

• Indirect 9-Oct-24

• Zero Page 9-Oct-24

• Local Page 9-Oct-24

• Auto Increment 9-Oct-24

Instructions Tested:

• AND 0xxx 2-Oct-24

• TAD 1xxx 2-Oct-24

• ISZ 2xxx 14-Sep-24

• DCA 3xxx 9-Oct-24

• JMS 4xxx 9-Oct-24 (Broken again)

• JMP 5xxx 14-Sep-24

• IOT 6xxx 2-Oct-24

• ION 6001 2-Oct-24

• IOF 6002 2-Oct-24

• KSF 6031 2-Oct-24

• KCC 6032 2-Oct-24

• KRS 6034

• KRB 6036 2-Oct-24

• TSF 6041 2-Oct-24

• TCF 6042

• TPC 6044

• TLS 6046 2-Oct-24

• NOP 7000 2-Oct-24

• IAC 7001 14-Sep-24

• RAL 7004 18-Sep-24

• RTL 7006 18-Sep-24

• RAR 7010 18-Sep-24

• RTR 7012 18-Sep-24

• CML 7020 18-Sep-24

• CMA 7040 18-Sep-24

• CIA 7041 2-Oct-24

• CLL 7100 18-Sep-24

• STL 7120 2-Oct-24

• CLA 7200 18-Sep-24

• GLK 7204 2-Oct-24

• HLT 7402 18-Sep-24

• OSR 7404 18-Sep-24

• MUY* 7405 9-Oct-24

• DVI* 7407 9-Oct-24

• SKP 7410 18-Sep-24

• NMI* 7411 9-Oct-24

• SHL* 7413 9-Oct-24

• ASR* 7415 9-Oct-24

• LSR* 7417 9-Oct-24

• SNL 7420 18-Sep-24

• MQL* 7421 2-Oct-24

• SZL 7430 18-Sep-24

• SZA 7440 18-Sep-24

• SCA * 7441 9-Oct-24

• SNA 7450 18-Sep-24

• SMA 7500 18-Sep-24

• MQA* 7501 2-Oct-24

• SPA 7510 18-Sep-24

• CLA 7600 2-Oct-24

• CLA* 7601 2-Oct-24

• CLA 7610 2-Oct-24

• LAS 7604 2-Oct-24

• CAM 7621 2-Oct-24

*These are EAE instructions.

Bad Behavior to Fix:

• The STOP switch does not stop the processor when it is running. Dirty switch. Was also caused by the S107 in slot PD33. (Done 1-Aug-24)

• The SING INST. switch does not stop the processor when it is running. (Done 11-Aug-24)

• The DEP switch does not change the contents of memory. (Done 14-Sep-24)

Restoration To Do:

• Glue the acrylic covers back together. We need to find a glue specifically for acrylic plastic, not just epoxy. Add reinforcing to the covers so they don't crack again.

• Replace missing screw on the front left fan.

• Replace finger guards on left fans

• Oil the right front and right rear fans.

• Document the white clock wires on the Memory and Processor sides of the chassis.

• Install the oldest DF32 from the warehouse in the short cabinet.

• Inventory the FlipChips in the Processor side of the chassis. (Done 24-May-24)

• Determine which front panel paddle switch pivots are broken. (Done 24-May-24)

• Reform the capacitors in the power supply. (Done 1-Jun-24)

• Disconnect the power cables from the Memory and Processor chassis and test the power supply. (Done 1-Jun-24)

• Repair broken wire on the Panel Lock switch. (Done 5-Jun-24)

• Get a short DEC cabinet from the warehouse, and set the PDP-8 on top of it. (Done 5-Jun-24)

• 3D print Doug's pivot pin drill fixture, and drill out the broken pivot pins in the front panel switch paddles. Replace the broken pivot pins with 1/16" or 0.0625" acrylic rod, and reinstall the repaired switch paddles. (Done 1-Jun-24 through 27-Jul-24)

Repaired FlipChips:

• G603 in slot MP1, replaced the damaged FlipChip with a spare, 14-Sep-24

• G603 in slot MP2, replaced the defective FlipChip with a spare, 17-Oct-24
  The original G603 has two loose transformers with open coils

• G603 in slot MP5, replaced the missing FlipChip with a spare, 14-Sep-24

• G603 in slot MP6, replaced the damaged FlipChip with a spare, 9-Oct-24

• G603 in slot MP8, replaced the damaged FlipChip with a spare, 14-Sep-24

• R211 in slot PC/D10, replaced D-662 D37 with a NOS D-662, 14-Aug-24

• S107 in slot PD33, replaced D-662 D1 with a NOS D-662, 11-Aug-24

• S111 in slot PB32, replaced Q1 DEC-3639 with a 2N3639, 1-Aug-24

• S111 in slot PC31, replaced D10 a D-664 with a NOS D-664, 9-Oct-24

Other Electrical Repairs:

•  Q2, a 2N3236 in the R/W power supply with a correct NOS 2N3715, 11-Sep-24

Notes for Warren's RS-232/20mA Current Loop adapter

• RTS is driven high and DTR is driven low from the PC to provide power to the Current Loop 

• CTS is driven high from the Current Loop adapter
  CTS is controlled by the Reader Run current loop from the PDP-8
  It looks like the Reader-Run control on the adapter is not working

• The RXD(2), TXD(3), DTR(4), GND(5), RTS(7), and CTS(8) are wired

• With Warren's terminal emulator CTS flow control needs to be enabled