0
\$\begingroup\$

I choose an input capacitor for the LM2596 DC/DC converter. Since the device will be operated on the street (from -35 to +35°C), I decided to choose polymer capacitors, since they are not particularly dependent on temperature and there is no electrolyte.

Everything seems to be fine, but when compared with electric capacitors, I noticed that polymers have a very high leakage current.

Example:

  1. Electrolytic capacitor - 4uA
  2. Polymer capacitor - 2350uA

It bothers me. My circuit is protected by a residual current circuit breaker, which is of type A and is designed for a leakage current of 10mA. If there are a lot of capacitors and 10mA is typed, will the switch work? Or am I misunderstood?

\$\endgroup\$
10
  • 2
    \$\begingroup\$ RCD will have nothing to do with capacitor leakage. What’s specified in the datasheet? \$\endgroup\$
    – winny
    Commented Sep 18, 2019 at 11:15
  • \$\begingroup\$ LM2596 recommends low ESR capacitors in the datasheet. This indicates the use of electrolytic and tantalum capacitors. I do not use tantalum capacitors because of their instability to currents. Nothing is said about polymer capacitors at all, but I look in their direction because of durability and operating temperatures. \$\endgroup\$
    – Delta
    Commented Sep 18, 2019 at 11:20
  • 2
    \$\begingroup\$ Looking in a datasheet of AVX solid polymer caps: I do not see a capacitor with a DC leakage current coming even close to your 2350 uA. Please include a link to the datasheet from where you got this number. \$\endgroup\$
    – Bimpelrekkie
    Commented Sep 18, 2019 at 11:20
  • 1
    \$\begingroup\$ I don't see "endurance" as a feature of polymer capacitors. Maybe you can share a link to this fact/factoid? \$\endgroup\$
    – Andy aka
    Commented Sep 18, 2019 at 11:37
  • 1
    \$\begingroup\$ Compare the datasheet stated leakage current with your measured. Is it within spec? Forget about electrolyte! What is the rated capacitor endurance? \$\endgroup\$
    – winny
    Commented Sep 18, 2019 at 11:50

2 Answers 2

Reset to default
4
\$\begingroup\$

Your residual current circuit breaker is only measuring leakage currents from the supply (phase/hot/live) to ground (Earth). It doesn't measure any other leakage currents in your application.

So if you are putting suppression capacitors between the supply and ground, then leakage matters. If it's within the DC part of some controller electronics, then it doesn't matter to the circuit breaker.

\$\endgroup\$
2
  • \$\begingroup\$ Type A RCDs must also control the leakage of direct current (ripple) current. Should it be related to electronics? electrical-installation.org/enwiki/Types_of_RCDs \$\endgroup\$
    – Delta
    Commented Sep 18, 2019 at 12:39
  • 2
    \$\begingroup\$ @АлексГарисон It still only matters if the RCD can see that current. All it is measuring is the difference between the supply live current and the supply neutral current. If that difference is 10mA or more, it assumes current is leaking to ground and trips. It has no idea what is going on within an appliance. \$\endgroup\$
    – Simon B
    Commented Sep 18, 2019 at 13:32
-1
\$\begingroup\$

TRUE 5mA/ mF is typ for Alum Polymer Caps yet Ripple current rating can be 500x to 1K times this leakage rating. i.e. Rp/ESR ratio

These are intended for lower DC voltages rather than AC line voltage so how/WHY is an RCD breaker used here?

WHAT ARE YOUR DESIGN CONSTRAINTS? SHOW ALL IN/OUT SPECS OF CURRENT, POWER, AND RIPPLE.

Line filter Y caps use plastic film NOT electrolytic polymer caps. (That’s assuming you are looking at Al Polymer caps , which ARE electrolytic)

All design errors stem from bad assumptions or missing specs.

EDIT

will the switch work? Or am I misunderstood?

The switch will work as higher leakage is expected on high ripple designs. Again, this is return current and not earth-ground BALUN RF current, done only with plastic-caps. Do not confuse this earth leakage with bulk cap leakage that is realy defined by Dissipation factor.

Also the switch RdsOn , however must be as low as the ESR of the cap or lower for efficiency and thermal reasons, but there is a lot more tech. to make this "work".

\$\endgroup\$
1
  • \$\begingroup\$ the problem with undocumented downvotes is that it is like a @community troll who never returns to see the improved answer assuming it was lacking in the 1st place \$\endgroup\$
    – D.A.S.
    Commented Sep 28, 2019 at 9:53

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.