Cap failure prevention?

[vampares]

I have a number of self built and used systems that I maintain, both software and hardware. All of these use intel processors and most use intel chipsets. I have never knowingly had a capacitor related problem before (although I have p3 HP w/ asus mb that locks under load in about 5 mins but otherwise runs fine , 45+ days uptime).

I was wondering what, other than shoding manufacturing, causes a cap to fail or degrade. I assume that heat is a major factor. Would it be effective to disipate the heat with something like aluminum tape? I have seen circuit boards w/ hot glue on the top of the caps (psu's etc). Can a cap's life be prolonged by reinforcement, at ones own risk, of the lid (I assume that the white cross pattern across the top is for explosion prevention) or will the cap fail internally regardless of the cans condition? Do caps have lifespans or is failure more a function of circumstance and quality?

[linuxguru]

Re: Cap failure prevention?

The vent pattern is a safety device, similar to a fuse, to prevent any build-up of gas that could cause an explosion. The problem is not in the can, but rather in the composition of the electrolyte and the electrodes. It's a materials quality issue, and very little can be done to delay it, except perhaps slightly improved cooling and air-flow.

[PeteS in CA]

Re: Cap failure prevention?

Like lg said, usually the key issue is the electrolyte, with heat playing a role. The life of a good quality cap is a function of its operating hours, the magnitude and frequency of the ripple current, and the ambient temperature.

WRT glue/RTV, this can decrease the cap's ability to dissipate heat; OTOH, the purpose of the glue/RTV is to improve the cap's stability under vibration and mechanical shock.