9 Mar
2006
9 Mar
'06
12:07 a.m.
If something "blew", there must be a major 'problem' somewhere
else! Did you find out why the board "blew"?
Tip on power supplies (in the 11/34): check the rectifier bridge.
Often (for what's "often" worth) a diode inside is open circuit,
thus the full-wave sine is used just for one half.
I agree with Tony. You must do several simple voltage measurements
at the power supply, distribution panel, connectors, sockets, etc.
Basically everything a connection "meets" from the PSU to the
backplane. Don't forget the wire itself!
- Henk.
What do you mean 'the board blew'? Did the
magic smoke come
out (if so, out of what?). Was there a hole burnt through the
board, melted traces?
FWIW, unless the board was actually burnt, I think it would
be repairable. I once did something very silly to a H754
brick with the result that 4 transsitors blew apart, one
ressitor totally burnt up, etc.
And that brick is working again now.
And the +15V board is actually very simple. It's a linear regulator...
This message and attachment(s) are intended solely for the use of the addressee and may
contain information that is privileged, confidential or otherwise exempt from disclosure
under applicable law.
If you are not the intended recipient or agent thereof responsible for delivering this
message to the intended recipient, you are hereby notified that any dissemination,
distribution, or copying of this communication is strictly prohibited.
If you have received this communication in error, please notify the sender immediately by
telephone and with a "reply" message.
Thank you for your cooperation.
deal with and it came back to life - sans the
H745 coming up.
I will ask you once more. Have you done any actual electrical
tests on this PSU?
-tony 
9 Mar
9 Mar
1:09 p.m.
If something "blew", there must be a major 'problem' somewhere
else! Did you find out why the board "blew"?
I've not even started my normal rant about board-swapping (rather than
repair) yet....
Anyway, I do feel that it is somewhat foolish to replace a part -- any
part -- withoug a clear idea of what the problem is. You might be looking
in totally the wrong area,. or the part might be being damaged by a fault
elsewhere...
The initial fault, IIRC, was that the lamp on the -15V regulator brick
wasn't lit. Now that doesn't even mean that the -15V rail is missing (the
lampe might have blown). I can think of many reasons why that lamp might
not light, including :
The lamp itself has burnt out
The -15V rail is missing due to a fault in the reuglator brick
The -15V rail is missing due to a shrot-circuit in on of the loads on
that rail (e.g. a serial card)
The input to the -15V regulator is missing due to a bad connection,
open-circuit transformer secondary winding, whatever
The +15V supply to the -15V regualtor is missing due to a fault on that
regulator board
The +15V input is missing due to a bad connection/broken wire between the
+15V board and the -15V brick
The +15V rail is missng due to a short in one of the loads on that rail
(e.g. a serial card)
The +15V rail is missing because the input to that regulator is missing
(brocken connection, open-circuit transformer secondary, etc)
It is pointless, IMHO, to go any further unless you know where to look,
and that means doing tests to eliminate some or all of those
posibilities. When you know where the fault is (e.g. you know the +15V
rail is misisng, you know the inputs to that board are OK, and the load
doesn't seem to be a dead short), then you can look at the next level
down and find, e.g. an open-circuit power transistor.
The best advice I was ever given for faultfinding (and it applies to
everything, not just classic computers) was 'Measure, think, then
repair'. Make measrurements and gather evidence. Think about what those
measurements are telling you (it may then be necessary to make more
measurements..). Then, and only them, start replacing components.
This may not be the modern way to repair things, but it's not let me down
yet (which is more than I can say about some other 'methods').
-tony
4:06 p.m.
Well, I can tell you I am throughly confused when it comes to measuring this
particular power supply system with a multimeter. I *did* do some measuring
a few months ago on this, and could not for the life of me get a reading off
of the backplane, nor the power distribution board. However, the only
response I got was the light going out on the H7441 when I tried to measure
the +15V line that was, AFAIK, working properly.
I even traced down which black was the ground for what, if it mattered, and
made sure I used that to measure.
I'll repeat the measuring process this weekend, and write up a very detailed
status report here of my findings.
-----Original Message-----
From: cctalk-bounces at classiccmp.org
[mailto:cctalk-bounces at classiccmp.org] On Behalf Of Tony Duell
Sent: Thursday, March 09, 2006 3:09 PM
To: cctalk at classiccmp.org
Subject: Re: 11/34a problems continue
If something "blew", there must be a major 'problem'
somewhere
else!
Did you find out why the board "blew"?
I've not even started my normal rant about board-swapping (rather than
repair) yet....
Anyway, I do feel that it is somewhat foolish to replace a
part -- any part -- withoug a clear idea of what the problem
is. You might be looking in totally the wrong area,. or the
part might be being damaged by a fault elsewhere...
The initial fault, IIRC, was that the lamp on the -15V
regulator brick wasn't lit. Now that doesn't even mean that
the -15V rail is missing (the lampe might have blown). I can
think of many reasons why that lamp might not light, including :
The lamp itself has burnt out
The -15V rail is missing due to a fault in the reuglator brick
The -15V rail is missing due to a shrot-circuit in on of the
loads on that rail (e.g. a serial card)
The input to the -15V regulator is missing due to a bad
connection, open-circuit transformer secondary winding, whatever
The +15V supply to the -15V regualtor is missing due to a
fault on that regulator board
The +15V input is missing due to a bad connection/broken wire
between the
+15V board and the -15V brick
The +15V rail is missng due to a short in one of the loads on
that rail (e.g. a serial card)
The +15V rail is missing because the input to that regulator
is missing (brocken connection, open-circuit transformer
secondary, etc)
It is pointless, IMHO, to go any further unless you know
where to look, and that means doing tests to eliminate some
or all of those posibilities. When you know where the fault
is (e.g. you know the +15V rail is misisng, you know the
inputs to that board are OK, and the load doesn't seem to be
a dead short), then you can look at the next level down and
find, e.g. an open-circuit power transistor.
The best advice I was ever given for faultfinding (and it
applies to everything, not just classic computers) was
'Measure, think, then repair'. Make measrurements and gather
evidence. Think about what those measurements are telling you
(it may then be necessary to make more measurements..). Then,
and only them, start replacing components.
This may not be the modern way to repair things, but it's not
let me down yet (which is more than I can say about some
other 'methods').
-tony
10 Mar
10 Mar
4:54 p.m.
Well, I can tell you I am throughly confused when it comes to measuring this
particular power supply system with a multimeter. I *did* do some measuring
a few months ago on this, and could not for the life of me get a reading off
of the backplane, nor the power distribution board. However, the only
Right... Alas I am a lot better at diagnosing electronic faults than
mindreading :-)...
Let me again start with a general comment. If you get what seem to be
crazy readings, then it's time to investigate. Maybe you've made a
mistake. There's nothing wrong in making mistakes, I do it all the time
(as does every other engineer, hacker, whatever that I know). By findout
out what you're doing wrong, you learn a lot.
Or maybe there's a real fault that needs to be traced and cured. In which
case those seemingly odd readings will help in finding it.
response I got was the light going out on the H7441
when I tried to measure
the +15V line that was, AFAIK, working properly.
Hang on a sec...
Do you mean H754 (the -15V regulator) here?
I assume you do mean the +15V line. Now, IIRC the +15V regulator is a
very simple circuit without much overload protection other than a fuse on
the regulator PCB. If you managed to short the +15V lien to ground, the
most likely result would be to blow that fuse. That would kill the +15V
line, and with it the -15V line.
One very easy way to do that is to accidentally plug your meter leads
into the current sockets on the multimeter. Some meters (including my
Fluke) beep if oyu set them to a voltage range with the leads in the
current soockets to warn you about this. Most don't.
I even traced down which black was the ground for what, if it mattered, and
made sure I used that to measure.
The ground is commnn throught the entire system, it's linked between the
backplanes. The other power rails are not linked -- there might be, for
example,. several +5V lines in the system (there are 2 in just the 11/34
CPU box) that are not linked. They both should be at about +5V wrt the
common ground, but there might be slight differences. This should all be
shown in the power distribution PCB schematics.
-tony
6861
days inactive
6863
days old
3 comments
3 participants
participants (3)
-
ard＠p850ug1.demon.co.uk -
fireflyst＠earthlink.net -
henk.gooijen＠oce.com