11/34a problems continue

[Tony Duell]

> 
> 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