1
\$\begingroup\$

What type of capacitor should I use for a longer life (C8) aluminum electrolytic or aluminum polymer capacitor? It works at 400 kHz and must provide a current of 30 mA. Capacitor

\$\endgroup\$
4
  • \$\begingroup\$ Polymer caps don't dry out like electrolytic \$\endgroup\$
    – Aaron
    Mar 31 at 22:08
  • \$\begingroup\$ @Aaron Hybrid polymer can dry out I think, but solid/dry polymer can't. Hybrid is more resistant to overvoltage though. \$\endgroup\$
    – DKNguyen
    Apr 1 at 1:09
  • \$\begingroup\$ At this small capacitance value, I would contemplate a ceramic capacitor instead of both C8 and C9. But that depends on whether U6 is stable with low ESR-caps. If U6 needs the C8's ESR to work as a snubber and stabilize the output, then an aluminum polymer cap might not be an option and could lead to oscillations. \$\endgroup\$
    – tobalt
    Apr 1 at 5:33
  • \$\begingroup\$ @tobalt U6 appears to be a simple rectifier bridge (DB = Diode Bridge, I guess), so there is no stability issue here. The "U" prefix in its designator is indeed misleading. \$\endgroup\$ Apr 3 at 20:15

2 Answers 2

1
\$\begingroup\$

I would say: neither.

Use an X5R or X7R 10uF 50V capacitor in 1206 or 1210 SMT package. E.g. SAMSUNG CL31A106MBHNNNE, about a dime each qty 100. Compared to electrolytics, its life is effectively infinite.

Make sure that the rectifier U6 uses ultrafast diodes. You may end up needing discrete diodes instead. Even at mains line frequencies, fast switching of rectifiers is essential to keep the noise down. At 400kHz, the problems caused by slow rectifiers are 10,000 times worse, relatively speaking. Use diodes that are designed for ultrafast switching, and whose datasheets provide maximum switching time specifications. Otherwise, you'll end up with noisy outliers if only a typical switching time is guaranteed.

Ensure that the path from the rectifiers into the filter capacitors has low AC impedance: use wide copper areas to lay out this circuit. The "small" 30mA current is only the DC part of it, while the AC part at tens of MHz may have peaks 10-100x higher than 30mA. The low impedance of this current loop will ensure that the capacitors have a chance to shunt the majority of the switching noise energy away from the rest of the circuitry.

\$\endgroup\$
3
\$\begingroup\$

As you are a new member here in the community in order to receive better recommendations my suggestion is when you post a question first keep in mind this points:

1- show the community your researching about the topic you are asking about.

2- be explicit in your post (if you post schematics, tell us what's the schematic about, in this question also it's important to know what would be working temperature of the board for example)

Take a look at this table (https://www.we-online.com/web/media/07_electronic_components/news_1/blog/midcom_blog_photos/capacitor_series_august_2016/32_res18.png) having in mind what would be the working temperature of the board.

img1

Also take into account this points:

Pro's of aluminium electrolytic caps:

  • higher voltage ratings available (up to 600v)
  • way cheaper pricing ( for the same capacitance and voltage)
  • better leakage current behaviour than polymer

Pro's of polymer electrolytic capacitors:

  • lower esr/higher allowable ripple current
  • no dry out behaviour
  • higher expected lifetime/load life

When you calculate the expected load life of a capacitor take into account the temperature in which the board will work. At 105 degree celsius the aluminium electrolytic capacitor has longer life but if the temperature drops in a more normal range of 65 degree celsius the polymer capacitor has a longer life.

\$\endgroup\$
4
  • \$\begingroup\$ It is interesting that you tell that lower ESR was unequivocally a pro. Very often, when a schematic shows a parallel combination of a ~100nF ceramic cap and a larger polarized cap, the high ESR of the polarized cap on the order of ~1 Ohm is actually important to stabilize the rail against oscillations by working as a snubber. So at the least, emphasize that it critically depends on the specifics of U6. \$\endgroup\$
    – tobalt
    Apr 1 at 5:38
  • 1
    \$\begingroup\$ U6 is a simple 4-diode rectifier bridge. Can I give up the C8? \$\endgroup\$
    – Miau
    Apr 1 at 9:24
  • \$\begingroup\$ @Miau that depends on how much ripple voltage you can tolerate across R8 then. \$\endgroup\$
    – tobalt
    Apr 1 at 9:35
  • \$\begingroup\$ @Miau C8 is there to filter the output of the bridge. It must be there if you want small ripple. C9 isn't there to filter out ripple but to improve transient response if the load varies rapidly (C8 is slow in this respect, because of various reasons, ESR and also ESL). \$\endgroup\$ Apr 3 at 20:21

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.