Connect with us

Life expectancy of Aluminum electrolytic capacitor

Discussion in 'Electronic Design' started by Obelix, Jan 15, 2004.

Scroll to continue with content
  1. Obelix

    Obelix Guest

    I've been looking for a formula describing the life expectancy of Al
    Electrolytic Capacitors. I found a couple of sources that descibe life
    expectancy based on an exponential relationship with ambient operating
    temperature and a linear relationship with operating voltage.

    Is there a good source for determining the life of an Al Capacitor?
    Does anyone know of a good white paper or similar? I would rather have
    a more reliable source than a suppliers website or at least know that
    the information is reliable enough.


  2. Robert Baer

    Robert Baer Guest

    For ordinary commercial useage at or near rating voltage, i would say
    the life expectancy would exceed 20 years.
    I base this on the fact that i have seen numerous *wet* aluminum
    electrolytics (like the ones used in the 20's and 30's) still fully
    functional after 10 years of use, and a good percentage fully functional
    after 20 years of use.
    The "older" ones failed due to evaporation of the liquid; but
    rejuvination was simple and quick via replacement of the "secret"
    Since aluminum electrlytics now do not use a liquid electrolyte, then
    there is nothing to evaporate.

    Keep in mind that this is based on ripple currents well within the
    rating of the capaciror, as well as the temperature of its environment.
  3. The ratings are more like 2,000 *hours* at rated temperature (4%
    Best regards,
    Spehro Pefhany
  4. Roy McCammon

    Roy McCammon Guest
  5. budgie

    budgie Guest

    The humble elector is probably the electronic component with the least
    consistent life expectancy across manufacturers, due mainly to their targetting
    different intercepts on the cost/performance curve.. Look at what happened
    recently with low voltage electros on PC motherboards, because someone with an
    "imcomplete" formula offered cheaper products to cent-sensitive board makers.

    Any time you see lifetime figures on a manufacturer's website for electros, it
    will not be applicable to the next manufacturer. I wouldn't even expect to be
    able to use the thermal expression from one on another's product.

    The life expectancy is probably directly proportional to the price you pay.
  6. Robert Baer

    Robert Baer Guest

    Yes, there is a (ahem) slight discrepancy of the 2,000 hour rating and
    the experience of seeing aluminum electrolytic capacitors operate for
    more than ten years.
    Granted that such caps used for 5 or so years and then equipment sit
    on shelf or outside for another 5 or so years may need re-forming to be
    used for another 5 or so years.
    I am speaking of equipment like scopes and other test equipment made
    by HP, Tektronix, EG&G etc back in the tube days (1960's or so)
    retrieved from scrap dealers.
    And they are still going strong...
  7. Genome

    Genome Guest


    Data, Data and..... more Data. Try page 16.

    That's BC components, now part of Vishay, Formerly Philips.

    Same applies to Siemens and what they sold off.

  8. That link is for caps with an organic semconductor dielectric, which
    behave differently to the regular electros.

    OS - Factor of 3.2 decrease in lifetime for each 10 degree increase in
    temperature. (The datasheet may quote this as factor of 10 for each
    10 degrees.)

    Normal - Factor of 2 decrease in lifetime for each 10 degree increase
    in temperature.

    If the temperature isn't too high, the organic semiconductor ones may
    have huge (expected) lifetimes.
    They will also have lower ESRs, which results in lower self heating
    (which also helps the lifetime), although in my experience, nearby
    semiconductors have more of an effect on the temperature than anything

  9. R.Legg

    R.Legg Guest

    Any time you see lifetime figures on a manufacturer's website for electros, it
    If a manufacturer doesn't publish their own data, beware.
    This would be a very simple way of improving component quality, if it were true.
    Why don't you patent it?

  10. Roy McCammon

    Roy McCammon Guest

    yup. I figure the OP can do his own fishing if I show him
    the pond.
  11. Yzordderex

    Yzordderex Guest

    (R.Legg) wrote in message


    Heheh. Must be the most difficult question to answer or calculation
    to perform. I for one would like to hear what your application is.

    In a nutshell it boils down to hottest temperature inside of cap. Not
    as difficult to predict if load and conditions are known.

    case 1 - Set top box for cable tv receiver. Ambient temperature is
    25c, load is well defined, and can assume some duty cycle. On at 6am,
    off at midnight if you wish. Line impedance is reasonablty well
    defined as most residential installations closer to 10kVA than 100kVA.

    case 2 - 3 phase AC motor speed control bus caps. Ambient temperature
    0-50c, load unknown, duty cycle unknown, phase imbalance unknown, line
    impedance unknown...yada yada yada..

    If case one then you measure inner capacitor temperature (or outer if
    you can then calculate inner temp from that) and use this to predict
    lifetime. Can also get frequency spectrum of current waveform and
    apply each harmonic to esr at that harmonic to figure out power
    dissapated. That and can geometry will tell you internal temperature
    and then get lifetime from there.

    If case two then it's basically pull a number out of your ass. In
    other words experience will be your guide. To attempt an analytical
    approach to case two would be an excerise in futility.

    Hope this helps. Again, what is application?

  12. "Robert Baer" <> skrev i en meddelelse

    Ususally, the 2000 Hours rating, which is typical, is at 80 degrees Celcius
    core temperature of the capacitor and at rated ripple load.

    Decreasing the core temperature will cause the life expectancy to grow
    exponentially - *how* exponentially is disclosed in the manufacturers
    datasheets, there are curves in there. Below 30 degrees Celcius, the
    lifetime can well be a decade or so at rated load.

    When we add a Gaussian distribution to the probability of failure, several
    decades for some capacitors in a given sample is probable.

    But, since we wouldn't like to rely on chance when going to Mars, we use the
    Manufacturers Datasheets instead of anecdotal evidence ;-)
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day