Electrolytic capacitor lifetime and rated voltage?

[treez]

Hi
This web page states that electrolytic capacitor lifetime decreases with increasing capacitor voltage. Whereas that is the case for film capacitors, i am sure its not the case for electrolytics. In fact, believe electrolytics last longer if they are kept with say 90% of rated voltage on them....

What Influences Electrolytic Capacitor Lifespan?

If you speak with a bunch of design engineers, you might quickly believe that the electrolytic capacitor has a particularly dubious reputation. This view certainly hasn’t been helped by the so-called “capacitor plague” that occurred in the first few years of the new millennium.

resources.altium.com

You can achieve a decent lifespan increase when the operating voltage is lower than the maximum rated voltage. The most conservative estimate is that lifespan doubles when the component is operated at 50% of the rated voltage

 

[Easy peasy]

in simple terms , voltage is a stressor, if, for it's whole life, you run an electro at 50% voltage, the chances of failure are lower than running at 90%

 

[wwfeldman]

after a while, whatever the nameplate says, the "rating" of the electrolytic drops
essentially to the voltage it is used at.

to use it at a higher voltage (but less than then nameplate rating),
or if it has sat on the shelf for a long time, you have to "re-form" the
capacitor by applying a dc voltage starting at 0 V and bringing the
voltage up slowly to the rated voltage.

 

[treez]

Thanks, If two 10uF , 35V Panasonic FC-V series Electrolytic caps were used as the startup capacitance for an offline flyback, then how long would they last before drying out and becoming useless? (they have a voltage of 13V on them all the time).
These caps have essentially just a few mA of ripple current in them. If the product was permanently connected to the mains 24/7, and with day temperatures of 40degc, then how long would they last?
Surely that electrolyte is going to dry out sooner or later?

Panasonic FC-V electrolytic capacitor
**broken link removed**

 

[FvM]

The important wear factor is temperature not voltage. An electrolytic capacitor operated at ambient or slightly elevated temperature rarely dries out, even in 50 years.

 

[Easy peasy]

the "drying out" factor is a function of temp per the above - which is itself a function of internal heating and ambient - the just as great a determinant is the quality of the seal between the - usually - rubber bung at the bottom and the case and the leads.

If the seal is not made to 6-sigma manufacturing standards then there will be variations in parts - also if there is power & temp cycling inside the cap - this can "age" the seal faster than constant (low) power / constant ambient temp.

So there are many affecting factors - as always you generally get what you pay for ...

 

[c_mitra]

These caps have essentially just a few mA of ripple current in them. If the product was permanently connected to the mains 24/7, and with day temperatures of 40degc, then how long would they last?
Surely that electrolyte is going to dry out sooner or later?

It is the ripple current that the cap is delivering is the culprit. Well, the current along with the electrolyte series resistance, to be accurate.

Large changes in the applied voltage causes stress in the electrical double layer (the real dielectric; Helmholtz double layer) and capacitor failure is related with the failure of the dielectric.

The real dielectric is a paste of exotic chemicals (means I do not know) that does not like heat. ESR and ripple current produces heat and the inside pressure increases (gas formation) and that causes loss of capacitance. Tan delta simply measures this loss (for all circuits; not specific to capacitors).

Surely the electrolyte is going to dry out but capacitors are sealed and chances of drying out all by itself is remote. Surely you have seen a failed electrolyte capacitor? If it has leaked (along with the exotic chemicals), it has left its mark. If it has just vented, only steam is lost. In both cases the culprit is the heat (more accurately the internal temp causes of losses).

So coming to your first question, the caps are expected to last design life of the device; they are not the first ones going to fail.

 

[treez]

If the seal is not made to 6-sigma manufacturing standards then there will be variations in parts

Thanks, that sounds right, but as we know, the seal must allow gas to escape, since over time, electrolytics gradually produce gas anyway, and it will cause too much pressure if not allowed to seep out........so the seal must be "leaky", and here lies our problem.....it is the situation with wet electrolyte...it simply leaks gas over time as it drys out, this gas leaks out of the seal as it must do......so isnt this the problem with all wet electrolytes?

You just cant stop them drying out...its what they do?

 

[FvM]

I think it's time to get the discussion back to earth, what's your lifetime requirement for the specified capacitor application?

These caps have essentially just a few mA of ripple current in them. If the product was permanently connected to the mains 24/7, and with day temperatures of 40degc, then how long would they last?

 

[treez]

Thanks, more than 10 years at the use described...ie very little ripple current, and ambient daytime temperature up to 40degC, and product ON all the time, 24/7

 

[FvM]

Capacitor application notes usually refer to lifetime expectancy calculation with doubling every 10 K. Respectively a 1000h/105°C rated capacitor achieves about 10 years at 40°C with effective no self heating (actual ripple current < 10% of rated current). The calculation is somehow inaccurate due to the large extrapolation factor. If you want larger margin, you can use a capacitor with 5000h/105°C specification.

 

[Easy peasy]

Any gases produced can and do react with the Aluminium, hence limiting the pressure and the need to "vent" - thus fully sealed and long life caps can exist ...

 

[treez]

Any gases produced can and do react with the Aluminium, hence limiting the pressure and the need to "vent" - thus fully sealed and long life caps can exist ...

Thanks, i see your point, but its a shame that the Aluminium electrolytic capacitors' datasheets dont state whether or not this "non venting operation" has been facilitated.

 

[FvM]

Venting is a safety mechanism in case of failure, not normal operation.

 

[treez]

Thanks, i will try and find it, the doc which states that very minor "gas leakage" out of an electro is not only normal, but generally needed...to prevent the gradual build up of what eventually becomes big pressure. I am not convinved that the gas reaction with aluminium is complete, i am certain i have read that some gas is always and continuously produced, and after a while it would build up to damaging pressure if it were not gradually allowed to vent

 

[c_mitra]

Capacitor application notes usually refer to lifetime expectancy calculation with doubling every 10 K. Respectively a 1000h/105°C rated capacitor achieves about 10 years at 40°C with effective no self heating (actual ripple current < 10% of rated current). The calculation is somehow inaccurate due to the large extrapolation factor. If you want larger margin, you can use a capacitor with 5000h/105°C specification.

The doubling of rate (given by the Arrhenius equation) for every 10C increase in temp is a very crude approx for processes that take place at a reasonable rate around the room temp. It is messy to explain here and I make the reference to https://chemistry.stackexchange.com...ure-increase-doubles-the-rate-constant-k-when

Further, if my memory serves me right, these numbers (lifetime expectations at a given temp) were obtained by statistical analysis of relatively small samples and have large error margins. For example, if you use 100 capacitors at 105C for 1000 hours, perhaps more than 30% of the capacitors will survive. In other words, a single capacitor at 105C has a probability of 70% of failure. Again, the model used to estimate the average life is excluding infant mortality.

--- Updated Apr 18, 2021 ---

Thanks, i will try and find it, the doc which states that very minor "gas leakage" out of an electro is not only normal,

I doubt it; possibly it was true in the good old days but modern electrolytic capacitors do not produce gas unless you go significantly above the rate voltage for some extended time (means short voltage spikes won't cause any indigestion). I strongly suggest that you keep a 10% headroom on the working voltage.

 

[FvM]

The right point is that failure rates are based on statistical analysis. But your example numbers are totally misleading. The expected failure rate of an individual capacitor exposed to the specified conditions, e.g. 1000h/105°C is rather in a ppm than a 10 % order of magnitude.

10K doubling rule is used for lifetime calculations in manufacturer datasheets. I'd suggest to refer to it, unless you have better data.