From: ard at p850ug1.demon.co.uk
---snip---
However, they seem to be a lot less use when it's a one-off unit which
you need to repair no matter what (like a classic computer). You're going
to have to do real fault diagnosis sometimes anyway, it doesn't make it
much more difficult to do it properly every time.
Hi
Once many years ago while working for Intel, I was
called in to help fix a number of what we used to call
yellow dot boards ( failed but not repaired after third
rework and had a yellow dot placed on them ). The
yellow dot inventory had a sudden jump so someone
needed some additional help.
The boards had a lot of analog circuits on them because
they had a discrete component PLL used for the disk
controller.
I had the fellow from system test with me to take notes
on how to diagnose complicated failures on these boards.
As it turned out, about 90% were caused by a mix up
in the parts loaded on the boards. A low voltage zener
was placed in a number of locations that expected
signal diodes. This especially effected the balanced
mixer used in the phase detector.
After some time of carefully showing the system test
fellow a number of scope signals and carefully explaining
the reasons I was looking at particular locations,
I happened to glance at his notes. It went something
like this:
D3, D5, D8, D20, D21
There were no symptoms and nothing of what I showed him
about scope information. I soon realized that he was
just interested in a list of parts that I indicated as
causing problems and wasn't even paying attention to
anything else I was doing.
The worst part was that since this was a one time
parts mix up, it probably would never happen again
so that his list was virtually worthless. I told
him this but that didn't seem to ring a bell with
him.
Dwight