To calculate what value of a capacitor is suited for ripple suppression, we can use this formula:

Cmin = Load Current / (Ripple Voltage X Frequency)

It seems a higher value of the capacitor will get better performance for suppressing ripple.

Of course, a higher value for the capacitor will also be more expensive. If cost can be ignored, is a higher-value capacitor for ripple suppression always a good choice?

  • 7
    \$\begingroup\$ No it isn't always a good choice because where would this madness stop? \$\endgroup\$
    – Andy aka
    May 29, 2021 at 13:43
  • 7
    \$\begingroup\$ Also remember that big electrolytics are bad at high frequencies, so one big cap won't "solve" everything. \$\endgroup\$
    – Uncle Dino
    May 30, 2021 at 1:03
  • 3
    \$\begingroup\$ Well, cost can never be ignored. LOL. But even if it can, physical size will eventually become a problem. \$\endgroup\$
    – user57037
    May 30, 2021 at 6:39

4 Answers 4


A high value capacitor will cause a current surge when the device is first switched on. The capacitor needs to charge from 0 volts to whatever the circuit voltage is, in a very short time.

In extreme cases this could cause blown diodes or blown fuses in the power supply.

  • 22
    \$\begingroup\$ This is exactly right - when I was a know-it-all teenager I found some HUGE (44000uF) can capacitors. Also found a huge (25lb) step-down transformer and decided to build "the mother of all power supplies" in electronics class. Shop teacher let me do it too ... blew the CRAP out of the rectifier and regulators, and only THEN did he teach us about inrush current. The smart-ass electronics genius student ate some humble pie that day! \$\endgroup\$
    – akohlsmith
    May 29, 2021 at 19:20
  • 2
    \$\begingroup\$ PFC could be mentioned (related to the brief peak currents). \$\endgroup\$ May 30, 2021 at 11:44
  • \$\begingroup\$ I think that my electrolytic cap was only a couple of farads and in 1977 I made a computer and it's power supply. One evening, the cap blew up. Luckily, cases were also too expensive, so I had made a very sturdy case out of plywood, so it kept all of innards of the cap from blowing all over my room. \$\endgroup\$
    – NomadMaker
    May 30, 2021 at 17:59
  • 3
    \$\begingroup\$ @NomadMaker "only" a couple of farads? \$\endgroup\$ May 31, 2021 at 12:56
  • \$\begingroup\$ @htmlcoderexe Sorry, only a couple thousand microfarads. It was larger than a soda can. Made a really large boom. Then a friend and I redesigned the power supply properly. \$\endgroup\$
    – NomadMaker
    May 31, 2021 at 13:00

A large value capacitor needs to be charged. A very large capacitor will, at best, take a long time to charge assuming a safely current limited power source, and at worst reduce the life of the components from the source like switch contacts.

A large value capacitor is physically big and heavy. If size and weight matter, and they usually do, then going for as small a package, that is as low a value, as will allow the circuit to operate, is usually the thing to go for.


Ripple is a simple impedance ratio of an RLC filter with a noise source and a load noise for load regulation relative to the source impedance.

Other variables impacted are the surge energy to charge a bigger capacitor may be orders of magnitude bigger than the load and thus undesirable.

Cascaded filters outside the loop can reduce ripple with lower energy, but degrade load regulation from a higher source impedance. So there are many tradeoffs to the regulator and secondary filter response.

First you must define all the unexpected specifications that matter.


The larger the capacitor the worse the power factor and the higher the peak current in the rectifier.

This can be largely mitigated by adding a choke between the rectifier and the capacitors.

  • \$\begingroup\$ But at the expense of output level. It changes from nearly the peak value to 0.9 the RMS value. \$\endgroup\$
    – user207421
    Jun 1, 2021 at 1:21
  • \$\begingroup\$ yeah if you like ripple this is not a good solution. \$\endgroup\$ Jun 1, 2021 at 17:45

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.