Polymer Electrolytic Ripple Current Low Frequency

[KarlM]

I often use polymer electrolytic caps in power supply designs because of their high ripple current handling. I am looking at a new design with multiple frequency components in the current and noticed how the ripple current rating drops significantly as the frequency reduces. One for example is 4Arms at 100kHz but only 0.2A at 120Hz, 1/20 reduction.

Looking at normal electrolytics the ripple current rating at 100kHz is lower but the drop is much less as the frequency reduces. 2Arms at 100kHz and 1A at 120Hz, only 1/2 reduction.

Clearly the polymer caps are optimised for the higher frequencies in switch mode power supplies but was wondering if any one knows the mechanism within the capacitor that is behind this?

Karl

[T3sl4co1l]

Sure.  Polymer values are much smaller; have you checked what ripple voltage would be at that frequency?

This brings dielectric loss into play.  Dielectric loss looks like an R+C in parallel with the main C, with the extra C being small (about tan 𝛿 times main C).

Or if you stick with ESR (being equivalent series resistance), the ESR rises at low frequencies.

So the key is that the ripple voltage rises at low frequencies, while the Q factor stays roughly constant (Q ~= 1 / tan 𝛿).

Knowing Q factor, you can calculate loss.  Find the reactive power, then multiply by tan 𝛿 to get the real power (loss).  Reactive power is V*I (more or less), so once you know ripple and capacitance, you know this.

Tim

[David Hess]

Like T3sl4co1l says, ESR varies with frequency and is higher at lower frequencies.  In demanding applications if it is not specified at the operating frequency, then it should be verified through measurement under the expected operating conditions.