# Electrolytic Cap in parallel with ceramic for Decoupling

I wanted to ask if adding a high value electrolytic cap in addition to the ceramic decoupling capacitor in a circuit would make the filtering of high frequency worse. Because electrolytic caps are bad at high frequencies. I wanted to ask in case i need to power this circuit from a linear supply (like a Transformer with rectifier) the high value electrolytic cap would be needed for filtering the lower frequencies.

• This is actually a common thing to do; the electrolytic can filter low frequencies and the ceramic handles the high frequencies. Jan 25, 2019 at 19:09
• What transformer and caps and power levels V,I The RLC parameters matter Jan 25, 2019 at 19:44
• 50VA Transformer, 2200uF cap and 20W Class D Amp Jan 25, 2019 at 19:47
• Under certain rare circumstances it can be a problem. The larger cap can have a fairly low SRF, meaning it turns inductive fairly early. The ceramic cap can still be capacitive, and can have an L-C resonance with the inductance of the larger cap. If you have transients at the resonant frequency you can have problems. Jan 25, 2019 at 19:54
• C must be large enough to filter rectified pulses at 2f rate and SRF must be high enough to notch at PWM class D rate so there is no guess work. Thus ALL datasheets and component specs are needed Jan 25, 2019 at 19:56

When you have two different capacitors in parallel you might run into problems with antiresonance. This phenomen is discussed more e.g. in here Antiresonance of multiple parallel decoupling capacitors: use same value or multiple values?

To put it shortly, you'll probably lose some filtering power on some frequencies right below the resonant frequency of the ceramic capacitor. Though, Electrolytic capacitors tend to have high ESR compared to ceramic caps, which reduces the effect of antiresonance.

You can also try to mitigate the problem by adding a small inductor or a ferrite bead between the capacitors, though their inductance might cause different problems depending on what you are doing.

The antiresonance is also not a problem if there isn't anything consuming current or causing noise at the antiresonant frequency. And if there is, you can just change the value of the ceramic capacitor to move the antiresonance to a frequency where it doesn't matter.

This is a bit of an XY question and irrelevant so unanswerable without more details on your specs and choices.

Always start an RLC filter design with a design spec for source impedance Z(f), Bandstop f , attenuation, passband f, passband loss or ripple.

There are far more parameters than meets the eye with load / source impedance ratios and phase response. Every passive part has at least 3 values ESR, ESL, C

A Class D amplifier presents noise current in the audio band, as well as the switching frequency and all harmonics. The power source pulses large caps thru the rectifier only for the duration of the rise in ripple voltage of about 10%. So the source current has 10x the harmonics and also double the line frequency.

This is not ideal but something like the filter you can get with high Q ultra-low ESR caps

This is guessing a poorer select of bulk cap and slightly higher ESR

Your question is irrelevant to class D Amp but very relevant to RF filters with ceramic series and parallel resonance effects from too low ESR.