Understanding electrolytic capacitor ratings

Discussion in 'The Projects Forum' started by wayneh, Oct 6, 2014.

  1. wayneh

    Thread Starter Expert

    Sep 9, 2010
    12,083
    3,021
    I've just ordered a couple of these Panasonic EE electrolytics for a SMPS application. Panasonic refers to them as "high ripple at high current" and the specs for that seem impressive.

    Some manufacturers use the term "low ESR" or "low impedance", sometimes preceded by "ultra". Aren't these all basically conveying the same thing?

    Or to ask another way, is there a handy way to translate from ESR Ω to ripple current or to tan-delta or whatever else they use? It's darn confusing. Microfarads and voltage are sooooo much easier!
     
  2. kubeek

    AAC Fanatic!

    Sep 20, 2005
    4,669
    804
    ESR and maximum permissible ripple current are not the same thing, even though they are related. If you had two caps with the same ESR and other parameters, but one was physically much larger then the ripple current value would be higher, because larger cap will be able to dissipate more power than a smaller one.
     
  3. Lestraveled

    Well-Known Member

    May 19, 2014
    1,957
    1,215
    ESR and tan theta are very similar, but not identical, ways of expressing loss in a capacitor. Simple explanation:

    ESR - If you use a capacitor in a pulsed condition where the pulse is long enough for the inductive element to saturate, then you are left with capacitance with a resistor in series with it (ESR). In other words, ESR is a better parameter to use when the capacitor conditions are steady state.

    Tan theta - If you use a capacitor in an AC environment (sine waves), then the ESR and the inductance of the capacitor will sum (vector sum) together. Tan theta is the resultant phase angle difference from the phase angle that you would get with a perfect capacitance. The higher the ESR and inductance of a capacitor, the bigger the phase angle shift will be. Tan theta changes over frequency.
     
    Last edited: Oct 6, 2014
  4. Experimentonomen

    Member

    Feb 16, 2011
    331
    46
    High ripple current rated caps are usually used in good old iron core transformer power supplies where ripple current is high due to the low 50-60Hz line frequency.

    In a smps its basically death to both the capacitor and the psu to use a cap that does not have "low ESR" or atleast "designed for use in switchmode power supplies" in its spec sheet, basically the psu would appear to work normally for the time it takes for the cap to build up enough internal pressure to either vent(steam buildup is slow enough to allow the can to crack at its weakest point, sometimes blowing the can off leaving the roll intact) or in the rare case explode and spew its foil and paper over half the board due to hydrogen and oxygen gas buildup from the electrolyte breaking down into its main components.
     
  5. Lestraveled

    Well-Known Member

    May 19, 2014
    1,957
    1,215
    Hey guys, I would like to point out something, that is kind of cool, that concerns electrolytic capacitor life.

    Download the capacitor data sheet that Wayne posted (in post #1). In the first specification box, you will find "Endurance". They burned in the capacitor for 8000-10000 hours then measured the change in some parameters.
    - Leakage didn't change
    - Capacitance changed by 20%
    - Tan theta (ESR) changed by 200%

    When an electrolytic capacitor gets old, increasing ESR is what kills it.

    Yup, I'm a geek, and I'm proud of it.
     
  6. wayneh

    Thread Starter Expert

    Sep 9, 2010
    12,083
    3,021
    Uggh. So these EE series are not really a good choice for a SMPS application? They happened to be on sale at Newark.

    The capacitor I'm replacing is labeled HY CD288H and I haven't found a good description of their specs. 120µF and 200V. Maybe something here, which suggests they start with a lower tan-theta than the Panasonic ones. This site refers to them as "+105℃,High frequency,low impedance".
     
    Last edited: Oct 6, 2014
  7. Lestraveled

    Well-Known Member

    May 19, 2014
    1,957
    1,215
    Wayne, I think they will be fine. They have a very low tan theta and they have ripple current rating at 100KHz. I would not hesitate to use them in your project.

    Even the worst capacitor will not explode in a few minutes. The caps you bought are probably as good as the originals. How long did they last?? 5-6 years? Could you have gotten better caps? Yes, for a lot more money. What would that extra money have gotten you? Maybe another two or three years.
     
    Last edited: Oct 6, 2014
    wayneh likes this.
  8. wayneh

    Thread Starter Expert

    Sep 9, 2010
    12,083
    3,021
    Whew, thanks. That's welcome news.

    FWIW, the two I'm replacing measured at ~0.6Ω ESR in my DIY tester. (It uses 100kHz for testing and was calibrated against 1/4W carbon film resistors.) I don't think they're actually bad.

    Some of my other caps from this PCB have an ESR over 20 or even 30Ω, while identically rated caps from the same PCB measure ~5Ω. So I'm replacing them all anyway good or bad - they're very cheap once you're placing an order. Another few bucks for these big caps is a small price to pay.
     
  9. Lestraveled

    Well-Known Member

    May 19, 2014
    1,957
    1,215
    Using the formula, ESR = tan theta X Sqr. Rt. of (2 pi F C), The capacitors you bought have an ESR of 0.045 ohms at 120Hz.
     
    Last edited: Oct 6, 2014
  10. wayneh

    Thread Starter Expert

    Sep 9, 2010
    12,083
    3,021
    That sounds pretty good! I'll certainly test them before I install them, but my tester doesn't have much resolution below 0.5Ω.
     
  11. Lestraveled

    Well-Known Member

    May 19, 2014
    1,957
    1,215
    Wayne, I do not have much faith in your tester to measure very small resistances. The ESR of the capacitors in your tester would have to be lower than the ESR of the cap under test. That would take some special caps. Try testing them at 120Hz. The calculation in post #9 is based on data from the data sheet. The 0.045 number should hold true at 120Hz. Testing at 100KHz probably will not yield good data.
     
  12. wayneh

    Thread Starter Expert

    Sep 9, 2010
    12,083
    3,021
    No question. The ∆V between 0.25Ω and 0.33Ω (4 and 3 1Ω resistors in parallel) was only 30mV on a 20V scale. It's pretty good between 1 and 100Ω though.

    The only judgement I made based on my readings was that I might as well go ahead and replace all the little caps along with the big ones, since some of the little ones had much higher ESR than their mates. It's a $2 decision in favor of new caps.
     
  13. takao21203

    Distinguished Member

    Apr 28, 2012
    3,577
    463
    yes Panasonic is a good brand for electric parts.

    If you replace just all mind some of them are not loaded much so would last a long time.

    What kills capacitors is to use ones just at the margin + high temperature.
    And of course there are especially bad one's- cheaply constructed or even fake.
     
  14. The Electrician

    AAC Fanatic!

    Oct 9, 2007
    2,281
    326
    Dissipation factor is most often symbolized as tan delta, not tan theta.

    The formula you gave is incorrect; see here:

    http://www.radio-electronics.com/info/formulae/capacitance/esr-df-loss-tangent-q-tutorial-basics.php

    http://en.wikipedia.org/wiki/Dissipation_factor

    The correct formula is:

    tan\delta = DF = \frac{ESR}{Xc}

    This can be rearranged to give:

    ESR = tan\delta * Xc

    Finally:

    ESR = \frac{tan\delta}{2 \pi f C}

    So the capacitors in question would have a (maximum) ESR of 1.66 ohms at 120 Hz.
     
  15. Lestraveled

    Well-Known Member

    May 19, 2014
    1,957
    1,215
    OUCH!!! Electrician, you are right. My dumb math error. Thanks for the correction. Sorry wayneh.
     
    atferrari likes this.
Loading...