Electros: higher voltage = longer life?

[skrunk]

I read somewhere recently (on a vintage parts suppliers site) that higher voltage electrolytics will have a longer life, as there is less strain on the cap.

Is there any truth in this?
Would say, a 50v rated cap last longer than a similar 16v rated cap, in a 9v pedal?

[teemuk]

That is seriously overlooking some way important factors, like power dissipation factor, ripple current and ambient temperature.

I'd be more inclined to believe opposite is true since too low voltage will cause electrolytic to dry out and in practice this means that sustained operation of, say, a 65V cap at only 20V will effectively drop the cap's tolarated working voltage to about 20 volts. Even the manufacturers often recommend not operating the capacitors at voltage levels that are less than about 33% of the capacitor's working voltage rating.

[.Mike]

Maybe.

Here's an interesting read: Deriving Life Multipliers for Electrolytic Capacitors.

It says three factors in determining electrolytic capacitor life are base life (expected life @ max temp T_m, and max voltage), temperature, and DC voltage.

It specifically states:

However, at higher stress levels, such as when the temperature is near T_m, the additional leakage current from operating near the maximum voltage rating Vr may cause enough electrochemical degradation and hydrogen gas evolution as to reduce the life of the capacitor. Therefore a reduction in the applied DC voltage may extend the life of the capacitor, especially at elevated temperatures in capacitors that are tightly sealed.

Mike

[DDD]

Longer life = proper DC voltage and proper ripple/DC voltage rate.

[amptramp]

If you want to get the wisdom of the ages for free, download MIL-HDBK-217F from this site:

http://quanterion.com/Publications/MIL-HDBK-217/index.asp

and go to page 251 of the pdf where aluminum electrolytic capacitor failure rates are discussed.  There is always a benefit from going to a higher voltage rating.  It is normal in high-reliability electronics to set derating factors which are fractional or percentage limits of the component rating that can be used.  For example, if you have a derating factor of 60% for electrolytic capacitors, a cap rated at 25 volts can be used up to 15 volts but no higher.  The MIL handbook covers almost all components used in electronic design and serves as a decent basis for comparing circuit reliability (but not really a good predictor of absolute reliability and in this case, does not cover the possibility of an increase in capacitance caused by operation at reduced voltage).  It is a 322-page bookandf almost 19 Meg download, but it is many millions of dollars of engineering effort available for free.

And this was post 1000 for me.

[MarcoMike]

in general, I would say: every capacitor will last longer if not working at full voltage.
this being said, you cannot compare a brand#1 cap with a brand#2 cap of different nominal voltage. the life expectancy of electrolytic capacitors depends a lot on the construction, operating temperature, type of electrolyte....

anyway, I guess they're all gonna survive enough.

[PRR]

Many of these de-rating factors are for HARD-working caps.

Welder power supplies and such, if not derated, can kill caps in months, reliably.

If you are not abusing the cap so bad that it is killed in months or years, most voltage power ripple ratings become insignificant. Now (after 5 or 10 years) you are at the mercy of electrolyte purity. Really pure chemistry can last 40+ years. A pinch of sodium in the vat causes cap-rot in 3 years. No matter what you did for voltage.

BTW: 36V-37V on "35V"-rated caps is bad, but I have a set in service since 1983.

[defaced]

Many of these de-rating factors are for HARD-working caps.

Welder power supplies and such, if not derated, can kill caps in months, reliably.

Im not really sure that's a fair comparison given the extreme differences in operating conditions between a stompbox and a welding power supply.
Welding power supplies operate in an entirely different world. They are very low impedance devices (hundredes of amp) operating at a hand full of volts. Open circuit on a typical stick welder is the ~80 volts and drops immediately to around 15 to 25 volts (depending on arc length and plasma composition) during operation. The highest running voltage ive ever seen come across is 40v on some MIG power supplies running "tornado" spray transfer. From the service guys I have spoken to, i get the impression that caps in welding power supplies are more likely to die from over heating due to current than over voltage. Now modern inverter supplies are a different animal all together. Id have to pop a hood on one at the shop to see what world they run in.

[Seljer]

Im not really sure that's a fair comparison given the extreme differences in operating conditions between a stompbox and a welding power supply.
Welding power supplies operate in an entirely different world. They are very low impedance devices (hundredes of amp) operating at a hand full of volts. Open circuit on a typical stick welder is the ~80 volts and drops immediately to around 15 to 25 volts (depending on arc length and plasma composition) during operation. The highest running voltage ive ever seen come across is 40v on some MIG power supplies running "tornado" spray transfer. From the service guys I have spoken to, i get the impression that caps in welding power supplies are more likely to die from over heating due to current than over voltage. Now modern inverter supplies are a different animal all together. Id have to pop a hood on one at the shop to see what world they run in.

To be fair, the ripple currents and such in anything with a switching power supply are going to be harder on a capacitors than our little stompboxes

[PRR]

> Im not really sure that's a fair comparison

I did not mean that it was a "fair comparison"! As you say, HARD-working caps are "an entirely different world". And in that world, derating is critical. Most audio caps, especially in pedal-world, get very much less abuse. The derating studies for HARD-working caps do not apply.

[amptramp]

The NASA derating rules are quite reasonable, even for commercial products.  The US Navy has NAVSEA derating rules that are more stringent.  The NASA ones are in this 3-page pdf:

http://engineer.jpl.nasa.gov/practices/1201.pdf

It is interesting to go through the derating documet and apply the values to the equations in MIL-HDBK-217F to see how much the failure rate (in failures per million hours) drops.

[greaser_au]

The theory is all very nice, & gives you an idea of operational characteristics, but from a practical servicing perspective, low-value high-voltage electros are often the first suspect... more times than I can count  I've 'shotgunned' the 1/2.2/3.3uf  50v caps on a board  & fixed the problem. To be honest, I have no idea why this is.

If you're using low-value electros, especially in the signal path, I'd suggest making an intelligent decision about the voltage across it- make sure they are rated appropriately, but not excessively...

david