Wild_Bill wrote:

Capacitor failures have been mentioned as very likely faults, even on
motherboards that are only a couple of years old.

FWIW, there are a number of reasons why present designs of capacitors fail
prematurely.
The number one reason is stress, whih can result in high ESR equivalent
series resistance, or excessive leakage.
High ESR and excessive leakage aren't conditions that a typical DMM (even
ones with a capacitance range) can test for, either in or out of circuit.
A dedicated ESR meter, and/or a capacitor analyzer are required to test for
these values.

Forgetting about the counterfeit capacitor electrolyte for a moment (or
forever now), SMPS switch mode power supplies and motherbord DC-DC voltage
converter circuits apply a lot of stress to today's miniaturized
electrolytic capacitors.

Older linear power supplies would typically apply 60 or 120Hz to the filter
capacitors.

Newer SMPS apply frequencies of tens of kilohertz, often at high peak
currents. These conditions stress the electrolyte and foil coatings in
capacitors.

We can see what a similar high frequency stress does to thin aluminum foil
when it's placed in an ultrasonic cleaner.
Likewise, surface changes take place on lead/acid battery plates when high
frequency, high peak currents are applied by desulfator circuits.

Newer, low ESR capacitors are very compact due to the miniaturization of
most equipment, and bare space on motherboards is becoming almost
non-existent.

Most higher grade capacitors are rated at 105 C degrees to be more tolerant
of heat, but poor circuit layout practices often place them where ambient
heat is problem. With a nearby heat source and high peak currents, many
capacitors are operating close to their breakdown specifications.


Unfortunately, at the current required by the CPU, they have no other
choice than to put the caps next to the CPU. if they were even two or
three inches away, the voltage drop would be excessive. The CPU can
consume over 100 watts at 1.8 volts. That is over 50 amps, so the I/R
losses have to be kept to a minimum. The older Pentium chips consume 50
watts at 2.8 - 3.3 volts. which is still a lot of current to push
through 1 oz. foil on an interior layer.



It's been mentioned that circuit/motherboards that are shelved when new, may
have faults in a couple of years, just from storage.
I've seen numerous "new" low ESR capacitors fail testing (badly, not just
borderline/questionable), when they're old stock of only a couple of years.
That was what started me to test new stock before installing them as repair
parts.

It seems that old stock is liquidated to sellers/brokers that package them
individually for retail sale and offer them to parts distributors, or just
sell them in bulk to surplus sellers. Much of the stock available from the
second or third transaction sources would no longer pass new specification
testing. They may function in a hobby circuit, but their reliability is
likely to be very low.

For important applications, fresh stock should be sourced from a distributor
with high volume stock turnover to ensure that the stock has been recently
replenished from the manufacturer. There probably isn't any added value in
buying a hundred pieces if they'll just sit in a drawer for a year or two.

Many quality electrolytic capacitors have 4 digit date codes on them which
are fairly easy to interpret, as week/year or year/week.

One of the best ESR testers by reputation and cost is the Bob Parker/Dick
Smith ESR Meter in circuit tester.
Other models costing many times more, are no more effective at measuring ESR
both in and out of circuit.
More costly capacitor and inductor analyzers generally offer more test
parameters for more thorough evaluations.

Desoldering thru-hole leads of electrolytic capacitors can be a bit
challenging on multi-layer motherboards, but easy enough to do with quality
desoldering equipment designed for the task.
Random replacement of electrolytics on motherboards may correct or prevent
some problems from failed capacitors (and replacement of all the voltage
converter and CPU voltage supply capacitors is wise), but I'm convinced that
testing capacitors before replacement is a requirement.

Similar to recent comments regarding rechargeable batteries, electrolytic
capacitors are also availble in various grades with particular
characteristics. The voltage rating and capacitance value shouldn't be the
only considerations for important applications requiring any degree of
reliability.


--
You can't have a sense of humor, if you have no sense!