Re: Darned capacitors again!


Miroslav Pokorni
 

I priced a new control module at over $375. GE subcontracted this part
to RobertShaw, a controls company. Being unwilling to pay what I refer
to as 'appliance repair extortion money' I got the controller on the
bench and started looking at it. I was able to power up the unit on the
bench and reproduce the failure. Lacking a schematic, I was not able to
pinpoint where the problem was.
You do not know what 'repair extortion' is, unless you had an electrical
repair on the car. It was pretty bad before, but with advent of engine
processor control, everything is 'you see this small microchip, that is the
brain of the car. I have to replace it, because I do not know any better'.
For mere $1000 that box gets swapped out and your car runs with its
defective parts, until memory in module builds up history and thereby works
itself out of wiggle room and your car is ready for another 'brain
transplant'.


I then decided to inspect the PCB for a cold solder joint Using a
jeweler's loupe and a probe, I started pushing on components while
looking at both sides of the PCB. I happened to touch the base of a
small electrolytic that left a stringy, green gel on the end of my
probe. This was a dime part at most.
Now, let us not get silly and be like popular media: 'shuttle explodes
because of failure of a 50 cent part. The O ring leaks ..'. Would those poor
people who died in the accident be any less dead if O ring did cost $50k or
50 millions?
Most parts on those controller boards are 'dime parts' and they are not
necessarily bad quality, at least not so until GMs and Fords of this world
start twisting manufacturers tail to lower price 'or we will go elsewhere,
now that we are 80% of your business'.


Tracing out the circuit revealed
that this particular cap was responsible for decoupling the sensing
bridge circuit, of which the oven sensor was a part. The cap was
gradually failing, leaking, shifting the reference voltage and
generating noise in the sensor circuitry. Previously I had to bias the
temperature range about 40 degrees high, as the oven was too cold (the
controller firmware had a feature for doing this). I replaced the
offending cap (which was a 4.7uF 16V 105 deg C unit) with a junkbox part
I had laying around. All the problems vanished, the oven now calibrates
nominally with no bias offset necessary.
I have an old Fairchild alarm clock; yes, at one time, in mid seventies,
Fairchild did alarm clocks, too. After 26 years of service, the clock did
start to change time, quite erratically. I opened it up, found it a rat's
nest (labor must have been quite cheep where Fairchild assembled those
clocks), moved wires and components around, looking for bad solder joint or
broken wire and bingo, there was that electrolytic hanging on one wire. You
can imagine what it is like to change a component on a consumer grade PCB
(low trace adhesion, probably easily cracked feed throughs), crowded with
wires and components. I got out that electrolytic, both parts, cap proper
and broken off wire. It looked like electrolyte seeped out and etched
through wire clean. However, on an afterthought, I decided that wire must
have been just marginal and it was my pocking around that broke it, so my
labor was counterproductive. After playing with clock controls for quite a
while, I realized that an intermittent switch caused problems. The switch
was practically free because it was part of the mold for clock housing and I
certainly will not get into correlation between price and performance.


Regards

Miroslav Pokorni

Join TekScopes@groups.io to automatically receive all group messages.