On Sun, 01 Dec 2019 14:39:19 -0500, Clare Snyder
wrote:

On Sun, 1 Dec 2019 10:27:39 -0500, Ralph Mowery
wrote:

In article ,
says...

The people who work for me fall into couple of groups on troubleshooting.

One is in the camp of: It's a brand XXX model 123 so it's always a faulty output relay. They're right a lot of the time and that reinforces the approach. When it's wrong they're lost.

Another is: Gather all the data, check everything, avoid a diagnosis until the last moment. These guys are slow but get it right. I've maybe got one like that.

The most common though is to make the diagnosis too soon and then exclude all data that doesn't confirm it. I try to keep these guys away from expensive equipment but there's no way to fix their approach. It's not that they ignore conflicting data, they can't actually see it.



I like to think that I fall in the middle of that.

When working on equipment, about 90% of the time it will be the same
part that fails. For example at work there was a piece of equipment
that had a relay in it that often failed. It just plugged in.
Instead of taking time to think about it, I would put in a new relay.
Takes about 10 seconds. If that failed to fix the problem, then out
comes the test equipment and circuit diagrams to do the trouble
shooting.


I do similar - but usually I "test" the relay (or whatever) to either
prove it IS the problem or it IS NOT the problem. This is because I
don't always have the replacement part available, and in most cases if
I get the part and it is NOT the problem it is not returnable - and
I'm stuck with it. If IDO have a "Known Good" part to swap in, it IS
(generally) the quickest way to prove the fault.

Goes a lot faster if you know what to test first. The best test is to
plug in the good one you have in your pocket. ;-)
I have worked on all sorts of machines and some you really needed to
diagnose (like a CPU with 1000 cards in it or a check sorter with
10,000 moving parts). Since the 90s, just replacing the bad part based
on a few symptoms, is not an unreasonable approach. When I left in
1996, IBM pretty much said take a scope with you, take two, we don't
need them anymore. (along with a trunk full of other test equipment).
They didn't need me either. Any dweeb can work on a box with a half
dozen FRUs, functionally packaged.