James said:
Has it been long enough to know yet?
We know not because of testing, but because we know what the problem
was. A Chinese capacitor manufacturer stole an electrolyte formula
from another Chinese capacitor manufacturer and got the wrong formula.
Now everyone is back to formulas that have been is service for many
years, so that particular problem is gone.
When I've ordered replacement caps I was somewhat shocked at the
lifetime ratings, many of them are only rated 2-3k hours, only
a couple times longer than the average life of a common
incandescent lightbulb.
First of all those ratings are for how long it will take for the
capacitance to drop below the rated capacitance, not how long
before it fails. Your circuit may work fine at 50% capacitance.
Second, the life time is the time the capacitor will perform
within the stated specification (usually ±20% of the initial
capacitance) at the maximum rated voltage ripple current and
temperature. When operating at temperatures below the maximum
rated temperature, the life expectancy of the component will
roughly double for every ten degrees C lower than the rated temperature.
For example:
2,000 hours at 105 degrees C
4,000 hours at 95 degrees C
8,000 hours at 85 degrees C
16,000 hours at 75 degrees C
32,000 hours at 65 degrees C
64,000 hours at 55 degrees C
128,000 hours at 45 degrees C (14 years+)
256,000 hours at 35 degrees C (28 years+)
Also, the life increases if the applied voltage is less than the
rated voltage, down to 60-79% of the rated voltage. A typical
figure is eight times the life at 60% voltage.
Having a low ripple current and not having voltage spikes helps,
as does avoiding long (years) periods at zero volts.