capacitor life time: what does it really mean?

[jarthel]

I was reading a datasheet for a Rubycon YXF. The capacitor is rated at 105C (temperature).

The "life time" as stated on the datasheet for the capacitor I'm looking at is 10,000hrs.

Does that mean the capacitor is good for up to 10,000hrs when run in an environment that is 105C without exceeding the rated voltage? If yes, should I get more out of the capacitor if the temperature is less than 105C? maybe even more if the capacitor is run at less than rated temperature and voltage?

Also, what is a capacitor's shelf life?

Thanks for the help

[Christer]

Originally posted by jarthel

Does that mean the capacitor is good for up to 10,000hrs when run in an environment that is 105C without exceeding the rated voltage? If yes, should I get more out of the capacitor if the temperature is less than 105C? maybe even more if the capacitor is run at less than rated temperature and voltage?

Also, what is a capacitor's shelf life?

Manufacturers define the life time differently, but for good brands it is usually defined as the end of life time when used at maximum temperature and maximumum ripple current, assuming proper mounting wrt. to cooling. The end of life usually means that certain parameters (typically capacitance, ESR and/or leakage current) are no longer within certain margins. How much they are allowed to change differ between manufacturers. Anyway, it is the heat that kills the caps, so if you compare two caps with the same life time but specified for 85 and 105 deg. respectively, the 105 deg. one should have a much longer life time at 85 deg. (Off the top of my head, I think the life time of a 105 deg cap is around ten times longer at 85 deg) and both will last even longer at en even lower temperature. Note however, that it is not the ambient temperature we are talking about, but the surface temperature of the cap (more precisely it is the hot spot temperature inside the cap, but that is compensated for in the figures we get).


Some further reading on this
http://www.pikpower.com/New%20Site/e...fe_factors.pdf

and on this and everything else you never even knew you even wanted to know about electrolytics
http://electrochem.cwru.edu/ed/encyc...4-appguide.pdf


Shelf life has to do with the oxide layer weakening during long idle periods. It seem not quite understood yet what causes this
http://www.faradnet.com/deeley/chapt_06.htm

A cap that has been idle for long time can often be revived by reormatting it. The most common procedure recommended for this is to slowly increase the voltage over it up to the max. rated voltage and then slowly decreasing it again. This is said to build up the oxide layer again. Just don't ask me for a more precise definition of "slowly".

[AndrewT]

Hi,
there is another thread quoting "life-times" running in the last 24hrs.

[mzzj]

Originally posted by Christer



A cap that has been idle for long time can often be revived by reormatting it. The most common procedure recommended for this is to slowly increase the voltage over it up to the max. rated voltage and then slowly decreasing it again. This is said to build up the oxide layer again. Just don't ask me for a more precise definition of "slowly".

ABB motor drive manual specifies 200mA maximum charging current for capacitor bank reformatting. Cap bank in these drives is huge, something like 40x6800uF 450V so scale current according to your capacitance.

Caps are kept at full voltage for 30mins- several hours before taken to servive, depending how long caps have been kept in storage.

[AKN]

I totally agree with Christer that heat kills caps.

By their nature of funktion they are often mounted close to voltage regulators and as we know those produce heat.
Manufactures tend to miniaturise and cut costs, resulting in demands for cheap, very good performance and small electrolytic caps.
Ok, caps today are very good but with small margins.

Take a small high performance cap that has to take a high ripple current at high temperature. Over time ESR rise due to drying resulting in even more heat in cap and we get an accelerated decreasing life time.

Of course this is if we tend to push caps to their limits.
I not so cost sensitive enviroments we can take good use of their good performance and get very long lifetime if not pushed close to limit.

An example of remarkable performance at least to me:

Rubycon ZLG series 2200uF 16V has ESR at 100KHz of 0.012Ohm and can take 2.9A ripple current at 105 degree celsius and all this in a tiny can of 12.5x25mm size.

Looking back 20 years this would be impossible.

[TwoSpoons]

Rule of Thumb : every 10C drop from rated temp will double the life of the capacitor. So a cap rated at 10,000 hrs @105C will live aproximately 40,000hrs @ 85C.

This is capacitor internal temp, not ambient temp. The best way to measure the internal temp is to connect a thermocouple to the positive terminal, IIRC.

Did all this when I was working on the design of the motor drive electronics for the Fisher & Paykel SmartDrive washing machine.