I have a
"VAXserver 3100" which afaik is the same hardware
as a Microvax 3100. This unit has a bad power supply.
The power supply is labled: MODEL H7822-00
Anyone have schematics or other technical information?
I've never worked on this supply (and don't have schematics), but it
can't be that odd, can it???
It's complex enough that I'd feel better going at it with a set
of schematics. If nothing turns up I'll draw it out myself, but
if someone has technical data it could save considerable time.
The second
supply generates +5 and +12v which is applied
only to the second disk drive power connector. This supply
constantly starts and stops - with or without a load. I
can see the output voltages coming on, then going off
over and over.
Wouldn't it be best to find out why it's doing this? My _guess_ is that
either one output is rising too high, tripping the crowbar, and thus
shutting the supply down, or, less likely, the supply to the control
circutiry has failed (the reason I mention the latter is that that fualt,
caused by a dried-up capacitor on the _primary_ side of the supply, once
led me a merry dance in a Tektronix X-terminal). I'd be inclined to start
by checking _all_ the electrolytics on that second PSU board with an ESR
meter. An high ESR capactior on the output side can put some
'interesting' high-ish voltage spikes on the output.
Agreed completely - I guess I wasn't clear. I haven't done any debugging
yet, just a brief characterization of the symptoms - I received it with
the statement "it doesn't work on this supply, I think the 12v or something
is dead, but it works on this one borrowed from another unit" - With a light
bulb limiter and multimeter I did a quick verification on the rails and
discovered that they were all present, and followed far enough to give the
observations stated previously. Thats when I decided to see if technical
data was available before digging much deeper. The problem with "just going
at it" is that sometimes you make what seems like reasonable assumptions
that one look at the schematic will tell you are completely wrong. The
description of symptoms was just in case this is a common/known DEC supply
failure mode.
I too suspect a lack of regulation and crowbar trip, and yes in the absense
of technical data, I'll probably "shotgun" the electrolitics in the hope
that I'll get lucky.
It is at least fortunate that the working board and the non-working board
are VERY similar. Once I draw it out I may be able to do some comparitive
debugging.
The ?1 and ?2
signals appeare to be control signals.
All power rails look good. With a supply borrowed
from a Vaxstation 3100 (smaller supply without the
extra drive connector/supply, but same main pinout),
I see 4+ volts on ?2 and some low but non-zero voltage
on ?1 - with the bad supply, both ?1 and ?2 are at
0v.
Can anyone tell me the names and description of
function for the ?1 and ?2 signals - I'd like to
know exactly what these are supposed to do as the
supply comes up.
Being a DEC supply, these might well be open-collector signals, pulled up
on the mainboard. At least one is likely to be power-OK. The other might
be a second power-OK (some DEC machines have separate ACLO and DCLO
signals), it might be a line-frequency [1] clock interrupt signal. Have
you tried a 'scope on the signals from the good supply?
They are open collector signals (or at least they appear to be). If no
tech data surfaces, I'll spend some time with the scope and provide
more detailed information sometime in the next week or so.
[1] After Will's flame last night, I won't
call it 'mains frequency', for
all that's what wee normally call it over here. 'Line' to us is the
horizontal deflection on a raster CRT display, so 'line frequency' over
here is what you'd call 'horizontal frequency'....
Funny - I don't see any problem with "mains frequency" - I've seen
household
AC power referred to as "mains" many times.
Dave
--
dave06a (at) Dave Dunfield
dunfield (dot) Firmware development services & tools:
www.dunfield.com
com Collector of vintage computing equipment:
http://www.classiccmp.org/dunfield/index.html