3
\$\begingroup\$

I am making a buck power supply and I need on the input and the output 4x 22uF @ >5V ceramic capacitor.

Even at 10V I find plenty of cheap X5R MLCC in the sizes 0603, 0805 or 1206. By looking at this datasheet, I don't see any noticeable difference between these models.

I guess the smaller is the better, am I right?

\$\endgroup\$
2
  • 4
    \$\begingroup\$ I guess the smaller is the better, am I right? No, smaller size (same capacitance, voltage rating etc.) means thinner dielectric. Thinner dielectric is more vulnerable. So larger capacitors are more robust. Watch here: youtube.com/watch?v=VwdnGbI5ls8 what can happen in a power supply with a small and delicate capacitor. Realize that for a switching supply the ESR might be important so check the datasheets of both switcher IC and capacitors. \$\endgroup\$ Aug 9, 2021 at 15:58
  • \$\begingroup\$ bigger package with same ratings also are more stable with DC voltage typically. \$\endgroup\$
    – tobalt
    Aug 9, 2021 at 16:25

2 Answers 2

3
\$\begingroup\$
  • the size is less sensitive to ESR in the AVX source range you linked but the voltage rating and bias has a significant effect on reducing C value.
  • In order to make the voltage rating greater the size must be increased so bigger can be better
  • the chip with the greater Vmax rating in any family will have a greater C value used at 50% vmax. with a higher Vmax rating. Choices are slim with 16 or 10V @ 22uF in MLCC
  • Also supplier choice makes a big difference and in this category, I would suggest TDK’s X6S parts with 10 mOhm ESR over AVX parts with 100 mohm for ripple reduction.
\$\endgroup\$
6
  • \$\begingroup\$ Do you have a source for ESR differences between manufacturers ? I find that ESR is not a prominent (i.e. hard to find) spec for ceramic caps. \$\endgroup\$
    – tobalt
    Aug 9, 2021 at 16:27
  • \$\begingroup\$ I agree it was hard to find \$\endgroup\$ Aug 9, 2021 at 16:31
  • \$\begingroup\$ If you search for low ESL caps in your range of values, you will find they also have lower ESR and probably higher price tag and lower aspect ratio, (thicker , wider) so you could choose more cheap ones in parallel. \$\endgroup\$ Aug 9, 2021 at 16:45
  • \$\begingroup\$ The Vmax rating and effect of DC bias is pretty tenuous on modern small package MLCCs. Looking at Murata simsurfing now and sorting by effective capacitance at 5v bias, I see lots of examples of 0603 and 0402s where the 6.3v rated part has less change under bias than the 10v rated part and lots of examples where the 6.3, 10, 25 and 35v rated parts have identical effective capacitance at 5v. I suspect manufacturers have figured out some way to independently optimize voltage and effective capacitance. \$\endgroup\$ Aug 9, 2021 at 16:58
  • 1
    \$\begingroup\$ If you can't find ESR for a capacitor, you usually find the dissipation factor (tan delta). You can convert from tan delta to ESR by ESR = (tan delta) / (2 π C f) where f is the frequency at which tan delta and ESR are measured. \$\endgroup\$ Aug 10, 2021 at 6:52
2
\$\begingroup\$

Beware, the very highest capacitance in the smallest package are often defined as "mobile only" capacitors. With a service life of just a few years. They trade reliability for size.

Other than that, DC bias derating is the biggest factor. Your 0603 22uF will likely have a real capacitance of less than 8uF at 5V. 0805 or 1206 will give you higher actual capacitance.

All such information is of course available at the manufacturer webpage.

\$\endgroup\$

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.