0
\$\begingroup\$

While working on compensation networks or ripple calculations, it helps to know the ESR for my output capacitors. In high voltage SMPS systems, this seems to be difficult for the following reasons:

  1. ESR if quoted is at 120Hz. - Commonly people will explain to use Tan(\$ \delta\$) to calculate ESR for a given frequency beyond 120Hz... See issue #2.
  2. Tan(\$ \delta\$) seems to not be consistent across frequency (source p. 7) So unless you are close to the often quoted 120Hz, this is not useful.
  3. Does Voltage bias affect ESR? I assume not, but I have not found any info supporting either yes or no.

I have made crude setups to test ESR of capacitors I am planning on using, but I would like to perform it more analytically if possible... Also depending on if ESR is affected by voltage bias, that has a huge affect on the complexity of testing ESR, since this is typically for ~400V output supplies.

Any advice or experts on the 3 points above? I'm guessing the consensus will be to just estimate and test...


Below are a datasheet for reference and an image of the plots from page 7 of the Nichicon capacitor manual I reference above.

Datasheet Img

                                       Tan with Freq

\$\endgroup\$
6
  • \$\begingroup\$ Is not ESR quoted in the data sheet? Can you link to it? Have you looked for capacitors that do have the rating you want? \$\endgroup\$
    – Andy aka
    Commented Dec 5, 2020 at 17:24
  • \$\begingroup\$ You say lab tests show close to 5 ohms, but how are you performing said tests? Measuring a capacitor's ESR at a specific frequency is not a trivial task. \$\endgroup\$
    – Hearth
    Commented Dec 5, 2020 at 17:34
  • 1
    \$\begingroup\$ Also, if your application is sensitive to ESR, it might be a good idea to look into ceramic and film capacitors instead of electrolytic. It may cost a bit more, though. A common thing to do is to put a smaller ceramic cap in parallel with a larger electrolytic, for the ceramic cap's lower ESR at high frequencies. \$\endgroup\$
    – Hearth
    Commented Dec 5, 2020 at 17:39
  • \$\begingroup\$ @Andyaka Most ~10uF 450V capacitors are designed for offline applications, thus all ESR is quoted at 120Hz, if quoted. \$\endgroup\$
    – MadHatter
    Commented Dec 5, 2020 at 17:41
  • \$\begingroup\$ @Hearth Good Question, I have used a signal generator, coupled to a biased capacitor with a resistor to ground, then on a scope i examine the difference of voltage at a given freq. across the capacitor. GND---RES--v2--CAP--v1--SIGGEN, Zc=dV/i... This is at-least a high level explanation, i agree it is difficult to do and requires a decent amount of setup, hence why i do not want to have to do it all the time haha. I can say my measurements have closely analytically matched calculations for ripple voltage when using measured ESR in calculations, so i think my methods have been "decent" \$\endgroup\$
    – MadHatter
    Commented Dec 5, 2020 at 17:43

0

Reset to default

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.