0
\$\begingroup\$

Implementing a LM2677 SMPS, the reference recommends 4 identical 10uF input caps for my design. I understand some of the reasons you may want to use multiple capacitors instead of a single one, and the datasheet has this note:

In some cases multiple capacitors are required either to reduce the ESR of the output capacitor, to minimize output ripple (a ripple voltage of 1% of Vout or less is the assumed performance condition), or to increase the output capacitance to reduce the closed loop unity gain bandwidth (to less than 40 kHz).

However I'm wondering if it's just an old reference design and newer capacitors can achieve these same goals with a single capacitor. I'd like to preserve space with a single cap and I've had some problems with tantalums failing in the past.

For instance, why not use a single low ESR high ripple aluminum polymer like this one: 25SVPF47M

\$\endgroup\$
7
  • 1
    \$\begingroup\$ Lower inductance? Check the datasheets and see if there is any difference. \$\endgroup\$
    – winny
    Jun 8 '20 at 12:55
  • \$\begingroup\$ ripple rating is another reason \$\endgroup\$
    – JonRB
    Jun 8 '20 at 13:07
  • 2
    \$\begingroup\$ Tantalums cannot tolerate voltage reversal, even a single microsecond event. They short out easily. As long as they are always forward-biased, they work well. If ESR is a concern, tantalum is not the answer today. Once, they had "good" ESR but we have MLCC ceramic and hybrid/solid options today. \$\endgroup\$
    – rdtsc
    Jun 8 '20 at 13:27
  • 2
    \$\begingroup\$ The extract is about output capacitors and not input capacitors as stated in your title. Be clear. \$\endgroup\$
    – Andy aka
    Jun 8 '20 at 13:31
  • 2
    \$\begingroup\$ Buck converters have high input current ripple. If a single cap is rated to handle the ripple current, AND the capacitance, ESR and ESL are such that you are OK with the resulting ripple voltage on your input supply, (don't forget load transients) you should be good with one input cap. \$\endgroup\$
    – John D
    Jun 8 '20 at 17:33
1
\$\begingroup\$

As far as I can tell a single aluminum polymer cap is better and cheaper than the 4 tantalums specified by the datasheet. ESL for the aluminum polymer I cited is zero at the 260K frequency this SMPS operates at.

Ripple Current rating is 2.8A versus 4 x 0.63A = 2.52A

Voltage Rating is 10V lower, but doesn't require as much derating as the tantalum, and I'm using overvoltage protection at 18V anyway for my application.

At 47uF it's a little higher capacitance than the 4x10uF specified in the datasheet

As far I can tell there's no reason to prefer the capacitors the datasheet is calling for. I did notice the datasheet was authored in 2004 and last revised in 2016, so hopefully it's just a matter of outdated docs not keeping with newer capacitor technology.

\$\endgroup\$

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.