PDP-11/10 repair started

[Henk Gooijen]

I have spent some time on the ill PDP-11/10 this weekend.
I have this machine already a few years, but I recently got an empty
H960 rack, which would be the perfect home for this PDP-11/10.
When I picked up this 11/10, the story with it was "not good".
"We put the boards in the backplane and switched on the machine.
When we saw smoke, we turned it off. We did not know that the boards
had to go in specific slots ..."!

This machine is an 11/10"S" with 16KW core in the 9-slot CPU backplane.
So, possibly the +20V core supply voltage fried ICs. However, a visual
inspection did not show any blown ICs.
After removing all boards (stored in ESD-safe bags), I first checked
the power supply. All voltages are present on the backplanes. DC LO
and AC LO are at 4.6V and have a ripple of 130 mV.
Seems acceptable to me.

Next, I placed the two CPU boards in the backplane. Of course in the
correct slots :-)  and connected the console ribbon cable. Power on...
The RUN LED stays on. Alas, it is not the infamous NPG problem. The
console is not responsive, although ENABLE/HALT and then START makes
DATA/ADRRS LED 0 go on. But that's all.  Seems like some fault finding
is needed in the CPU and console.
I downloaded "DEC-11-H05AA-A-D_1105um.pdf". In the back are some tests
that you can do. I hope to find the problem by elimination. Chapter 6.11
describes a few useful checks to verify the console stand alone. Great!
All 8 tests check out just fine, but ... the data pattern on the LEDs
is "inverted"! The manual says for test numbers 5, 6, 7, and 8 that the
LED pattern must be 052525, 031463, 007417, and 0003777 respectively.
I got exactly the opposite data: 125252, 146314, 170360, and 177400.
Looking at the schematic I could only explain that by a defective 7404
hex inverter, no longer inverting the input signal. But that seemed too
weird to be true, and I was not sure whether that would be a correct
assumption anyway.
Back home I downloaded "EK-KD11B-MM-001_Jan75.pdf" as a preparation to
some microstepping/debugging of the CPU. To my surprise, in the back is
a chapter 5.11 "Console maintenance".   It describes exactly the same
tests, but now the LED data is different ... it is what I am seeing!
So, I am happy to say that the console is eliminated as a problem source.

I guess that the first manual is not for the 11/10 "S" version, although
it surprises me that there would be a difference in the console hardware.

I dug up 3 double-width extender boards and the KM11 maintenance panels,
getting ready to do some microstepping as described in the manual(s).
Keeping both manuals side by side!

If you have some ideas how to proceed checking the CPU, I'm all ears!

greetz,
- Henk, PA8PDP