Classic Computing

I was presented with a PDP8 which was about to be junked from a local university.

The machine was in three main bits : the Power Supply and Chassis, the left hand memory card frame, and the right hand side processor frame :-

The three bits as received.


I would seem that it is many years since it was last used!

Labels on three sections :-

Aug 23rd 1998

After some very useful advice from the people on alt.sys.pdp8, I needed to gradually ease the power supply back into life. The idea of one of those huge electrolytic capacitors exploding is not one I relish!

The plan is therefore to gradually increase the mains supply over a week or so to hopefully give these guys chance to reform.

While this is going on I need to put the two racks back onto the chassis, and reconnect the ribbon cables between them, and connect all of those cables to the front panel. Yes, but how ??

Solution :- look at sunsite.unc.edu for a great series of detailed pictures (Courtesy of Paul Pierce prp@ssd.intel.com), and copy them - plus some help from the maintenance manual.

Sept 1st 1998

Progress on getting my PDP8 back into working life. Power supply voltages are ok, and ripple within tolerance. Also the machine didn’t disappear in a puff of smoke when I switch on with everything connected.

However, it’s not working. At least two problems at the moment :

1. When I hit the Cont key, relay R1 bounces. I think this implies that something is dragging the –15V rail down, and causing R1 to drop out to disconnect power from the core. Then once it drops out, I presume –15V returns to normal because the relay clicks in again – and so it repeats. This sounds bad so I am avoiding it at the moment!

2. Operating the Load Address switch will only OR the switches into the PC, it doesn’t clear PC first. This sounds more manageable, so I am following the Load Add logic to check the PC clear operation in SP1

Later that day ..

Turns out that one of MC voltage slide switches was open circuit, and there was no power to the PD rails on the processor side. Correcting that allows PC to be set, and DEP and EXAM now both increment the PC! Getting there !!

Still unable to set a value into core. Front panel lights indicate all is well while operating the DEP switch, but looking at the same addresses afterwards shows that nothing was actually done.

Sept 4th 1988

Can't seem to write to, or read, from core. Operating DEP or EXAM should transfer the PC to MA, during SP1. (The PC just having been set through the switches and operating Load Add.) Therefore if MA is not getting set, small wonder I can't modify addresses in core. MA seems to be stuck at 4066 and won't budge. IR is being set correctly - i.e. Set to DCA for DEP and to TAD from EXAM. MB clears each time - which it should in SP1, so I presume the actual contents of location 4406 is being set into AC, then transferred to MB. Possible that 4406 simply contains zeros?

January 1999

Time presses on, but no progress I am afraid. Seem to be trying to work myself to death at the moment!

February 2000

Moved home by now!   The Straight-8 is now back in my new study, and has been stripped down and all of the flip chip contacts cleaned.   One of the R211 boards has been clumsily repaired and needs attention.  (Actually the repair looks as if they were replacing a couple of transistors, the new metal can types have a DEC part number.  Looks a very rough repair if DEC actually did it!)

Looks like some more has died.   DEP and EXAM no longer set the IR lamps.   Tracing through with a scope reveals that the R107 at PB33 has died.   Swopped with another, and DEP and EXAM now work again.