# When can Self Resonance of a capacitor be a bad thing in bypass and decoupling applications?

Specifically in decoupling and bypass applications, can the Self Resonance frequency cause problems (compared to ideal)? If so, what kind of problems can it cause?

• Apart frm the answers below, you can find how to calculate the self resonance of common surface mount devices at electronics.stackexchange.com/questions/193608/… – Peter Smith Jun 26 '16 at 16:34
• A number of white papers / application notes have been published on this subject over the years. I recommend searching the Interwebs using the keywords "bypass capacitor resonant", for example. (And optionally add the keyword "pdf" to the search keywords to help constrain the search results to PDF files.) – Jim Fischer Jun 26 '16 at 21:18

An ideal capacitor has an impedance that falls with increasing frequency, which is good for decoupling high-frequency noise.

However, real capacitors have some amount of parasitic inductance, which appears in series with the capacitance, forming a series-resonant circuit.

Such a circuit has a minimum impedance at its resonant frequency, and at frequencies higher than that, the impedance starts rising again, which is less useful for decoupling.

That's why it's sometimes useful to use a number of different capacitors to decouple wide-bandwidth applications; each one provides the low impedance required for a particular band of frequencies.

But beware of strange cross-resonant effects! Sometimes the capacitance of one capacitor will interact with the inductance of another capacitor to create a parallel-resonant circuit, which has a very high impdeance at its resonant frequency. Verify your implementation with a wide-band network analyzer.

• Rememeber to separate the different values with a small track inductance so the capacitors do not appear lumped :) – Peter Smith Jun 26 '16 at 16:39
• I would only use multiple parallel values for analog ICs, and only after analyzing their working frequency and tuning the decoupling for that. For digital, several identical parallel capacitors is always better. – Jacob Jun 26 '16 at 16:46
• I think using multiple values is bunk. If you are using 0402 caps, use the highest capacitance available with the dielectric you want. It will have lower impedance at all frequencies than smaller value capacitors of the same size, even if the SRF is lower. At least this has been the case every time I looked into it. – mkeith Jun 26 '16 at 16:51

A bypass capacitor is never in isolation, even when there is only "one" on a board. There are internal (i.e on chip) bypass caps, formed from gate capacitance or MIM/PIP (Metal Insulator Metal or Poly Insulator Poly), there is the bond wire capacitance, the stray capacitance with other pins, capacitance of traces to ground. When the whole of these are considered then ...

Absolutely, this resonance would be tickled/sparked or excited by frequency content present in the current surges that are meant to be suppressed by the by-pass caps. In that case not only is there no energy absorption of unwanted signal but there may also be amplification or additional energy in very distinct frequencies being impressed upon those pins.

In most designs the the higher the frequency the harder it is to have good PSRR (Power Supply Rejection Ratio) and thus this resonant frequency likely would be noticeable).

Take a look at the results from here, a snip from a salient plot is shown below: (please the article for context)

• The self-resonance of a single capacitor is not going to cause a rise in voltage levels. It actually represents the minimum impedance that the capacitor can achieve. – Dave Tweed Jun 26 '16 at 16:36
• Responding to your edit: I'll remove my downvote, but the resonance effects you're talking about still have nothing to do with the self-resonant frequency of the capacitor, other than the fact that they occur above the self-resonant frequency of at least one capacitor, where it has inductive characteristics. – Dave Tweed Jun 27 '16 at 10:12