On Mon, Sep 02, 2024 at 01:45:22PM -0400, Mike Rieker via vcf-midatlantic wrote:
I'm restoring a PDP-8/L. I got it mostly working, had to replace a couple chips.
Not bad for what people have reported for restoring other 8/L's.
I have checked the supply voltages and there is a timing diagram in the PDP-8/L maint volume 1 book that seems ok. This one seems a little slow (like 600nS instead of 500nS), but the stuff seems to be in place, relative to each other. The manual shows some current waveforms to check but I don't have a current probe (one is on order). Anyone have experience getting this stuff to work?
Haven't had much problems with my 8/I or 8/L memory but have spent time with straight 8 memory. For 8/L people have reported bad diodes on the core modules. If bad people normally cut the body out and solder a new one to the leads. I think the diodes can be tested with meter diode check function.
Hi Dave, ok I can pound through them
On these core modules when they broke a wire making them they would twist the wire together then put some coating over it. Some people have reported those connections going high resistance. If suspect you can check that with ohm meter comparing to other wires.
ok that's a good idea. I think we'll bail on it if the core itself is shot. or maybe cobble up an sram card to replace it.
There is a strobe alignment to make sure you are sampling the core in the middle of the pulse and adjustment for memory voltage/current.
yes I read that in the book, I'm waiting for my current probe.
If you think failures change with temperature check the thermistor.
ok good idea but I think temperature has been pretty steady past few days
What current probe did you order? I have an old Tektronix one since that
supposedly the same tektronix 6019 they recommended in the maint manual
was the only reasonable cost option to have the 10 MHz bandwidth needed when I was looking.
I was having issues where I would adjust the straight 8 memory to pass diags but would stop working later. With the current probe I could see that the current varied with address. I found bad steering diodes and high resistance backplane contacts. After fixing that the memory has been pretty stable.
ok maybe that is what is happening here
Would also be good to see what memory locations are good and bad and see if you can find a pattern in address or data bit. That may point to
I've toggled around a little and haven't noticed anything but will keep playing with it. I am waiting for a circuit board I designed to plug in place of the front panel that I can use to scan the memory via a raspi. Will see what it comes up with. Thanks for all the great suggestions. Mike
particular areas to check. If you figure out suspect addresses you can try swapping driver cards and see if problem moves.