On Sunday, September 2, 2018 at 4:45:44 AM UTC-7, Stephen Wolstenholme wrote:
On Sun, 2 Sep 2018 06:03:04 -0500, Fox's Mercantile
wrote:

[about electrolytic capacitors]

Their installed and at the end of assembly line, full power is applied.
They either work or they don't.

I remember when part of my job was fixing power supplies. Some that
failed the automatic part of the test last thing on one day would work
the day after.

I never found out why that happened but I assumed it was an
electrolytic so I changed the lot.

Electrolytics have to be formed (kept under bias for a period) before they
develop the dielectric (oxide) layer, AND that layer is continuously
renewed when they're under bias. On the shelf, however, it degrades.
Some switchmode power supplies (the old Apple II types) rely on
timing capacitors for startup sequencing, and the power filter electrolytics are
NOT the prime suspects when such a power supply doesn't work.

You can often fix those power supplies by leaving 'em turned on overnight.
The most plausible reason for this is that the timing capacitors (not the big
filter caps) have to be re-formed. This syndrome can't be improved with
new filter capacitors (I've tried). Once they work, they're reliable again.