I'm tyring to build my first step down component using TPS562219 to get 5V output from a 12V input (instead of 1.05V output shown in the example). But as I'm new to this electronic stuff, I have some problems of understanding.

TPS562219 circuit

  1. Input capacitors

In this example circuit there are three capacitors. Why should I use three of them and how do I know the values of these. In the text of the datasheet (page 15, section they are talking of two capacitors: "A ceramic capacitor over 10 µF is recommended for the decoupling capacitor. An additional 0.1 µF capacitor (C3) from pin 3 to ground is optional to provide additional high frequency filtering."

  1. Output capacitors

For the output capacitors table 2 on page 15 gives a range of 20-68µF. How do I choose the correct value? And why should I use three of them?

In another circuit which is using the same TPS562219 they are using just two input capacitors and one output capacitor - with different values:

enter image description here

So I would be very thankful for some explanation, so I can understand that and build some own things. How many capacitors do I need and which values are recommended?

  • \$\begingroup\$ On some eval kits and similar "versatile" applications its convenient to have multiple footprints because you might op to populate them or not. Say a case where the load is lighter or more ripple is acceptable. \$\endgroup\$
    – Wesley Lee
    Sep 15, 2018 at 5:35

3 Answers 3


Input capacitors

Ideally, the input to the regulator has 0 impedance. That's not possible, of course, especially if there are long wires between the regulator input and wherever the input voltage is coming from. The capacitors guarantee a low impedance at the frequencies where this really matters.

Since 0 impedance is ideal, infinite capacitance would be great. The datasheet gives you guidance of how much capacitance is good enough for what the chip needs.

The extra 100 nF is to provide a lower impedance at even higher frequencies than the larger caps are effective at. The extra 100 nF by itself is inconsequential when added to 20 µF.

Output capacitors

The output of this switcher is pulses of current. That requires capacitance to smooth out to a reasonably steady voltage. In this case, the capacitance does two things

  1. It smoothes the pulses to make a flat-enough output voltage.

  2. It is part of the overall system the controller in the regulator is working with. Both too little and too much capacitance can cause control instability. Stick to what the datasheet says.

Values towards the higher end of the range will result in less output voltage ripple, but also allow less additional capacitance to be distributed among the consumers of the output voltage.

Multiple capacitors

Often multiple identical capacitors are used in parallel on both the input and output of such switching regulators. The purpose is to get more capacitance than a single part of the right type can provide. These capacitances need to be low ESR (equivalent series resistance) to the point that they pretty much need to be ceramic with today's technology. Ceramic caps can only be so big before mechanical problems make them unfeasible. Multiple caps in parallel allow for higher capacitance at the same voltage.

If a new technology came along that could fit the combined capacitance into a single part without mechanical problems or higher ESR, then it would be fine to use such caps. The technology tradeoff keeps changing, within the lifetime of many datasheets. Look around. Perhaps you can get the total capacitance with the right specifications in a single package today.

  • \$\begingroup\$ I've calculated a RMS of 0.2756A for the output capacitor. I don't know anything about the ESR. So I don't know how to find out if there is a single package existing today. I thought of using JMK212BBJ226MGHT 22µF capacitors but don't know if these are fine. Therefore I don't know if there is a 66µF cap, which has the right specification. \$\endgroup\$ Sep 14, 2018 at 17:02

Output capacitors: -

You'll probably find that using 3 x 22 uF (66 uf in total) has a lower parallel ESR and higher self-resonant frequency (SRF) than just using (say) a bulkier 68 uF.

Input capacitors: -

This can be problematic if the source impedance of the power feed is sloppy / too high or has resonant impedances involved. Add to this a cable feeding the regulator and you need to use belt and braces to cover the likely eventualities. 100 nF will have a reasonably high SRF and ensures the incoming supply impedance is "made" fairly low above a MHz or so and anything added in parallel is there to bulk out the capacitance should there be the possibility of drop-out or poor transient behaviour under load.


They are using multiple capacitors to save money.

MLCCs above a certain size get are only available as stacks of smaller parts and these stacks are crazy expensive

the two 10s make up a 20uF capacitor

on the output they use 3 22s in parallel as a single capacitor. as the switch is the other end of the inductor there's no need for a RF filtering capacitor here.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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