10 Feb
2009
10 Feb
'09
7:41 p.m.
Hi,
Apologies for the constant stream of questions at the moment, I hope I'm
not consuming too much list bandwidth at the moment. I've brought my
whole collection out of storage and I'm bringing them all back to
working condition and going through a steep old learning curve with my
Unibus machines!
So here are the latest round of questions:
1. I have started work on a second 11/05 I have. At the moment it has a
processor, an 8K core set, NPR cards and a pair of unibus terminators.
At present when powered up the machine has its run light constantly on,
shows what looks like a random set of lights on its address/data LEDs
and doesn't respond on its console.
I have tried swapping the CPU boards and core cards for known working
cards with no success. I have also hooked up a scope and confirmed the
power supply is supplying correct voltages.
PLEASE stop board-swapping and actually try to trace the fault logically.
It's not hard, it's generally a lot quicker in the end, and you learn
something about how these fine machines work. An 11/05 CPU board set is
under 200 ICs, and that can be easily understood.
3. I have an 11/34A (in 5 1/4" box) which is
experiencing a similar
constant run light problem to the 11/05. The machine has CPU cards,
terminators, console card and a MOS memory board in it. With both
terminators in, run light stays on. If I halt it, the bus error light is
illuminated.
If I remove the end terminator and power on, the console is responsive
and I can deposit values into memory and read them back.
What terminators are you using n the 2 cases? The little M930 ones (just
resistors or resistor packs) or the larger M9302 one with some ICs on it?
I would guess the latter, at least in the 11/34 machine.
This terminator has a feauture that if a grant gets all the way to the
temrinator (that is, if no device on the bus intercepts the grant and
doesn't pass it on), then the terminator will assert SACK. At this point
he processor (or more precisely the arbitor part of the processor) should
deassert the grant lines, and the Unibus carries on working as normal.
The problem comes if one of the grapts is 'open' either due to a missing
continuity card, a defective device, a defective backplane, whatever. The
grants are aobut the only Unibus signals that are active-high. So an open
one appears to be asseted, the terminator asserts SACK, the processor
can't do anything (it was never really asserting that grant line in the
first place) and you get a bus error.
I would guess you hav an open grant somewhere. Maybe a missing NPG jumper
(between pins CA1 and CB1 of each SPC slot.
-tony