0
\$\begingroup\$

I have several supercapacitors in series, and they're quite old and probably drifted from their original spec by quite a bit. Completely draining them in series will draw at least one of them below 0v. How should I protect against this?

I thought to use a schottky diode, but the best one of those would still allow the caps to be drawn to -0.3v. Also, many ideal diode circuits need some power, but the goal here is to protect fully discharged caps.

The ideal behavior I'm looking for is something that acts like an open circuit above 0v, and looks like a dead short below (or just shy of) 0v.

\$\endgroup\$
  • \$\begingroup\$ What are you using the supercapacitors for, and why would you drain them completely? \$\endgroup\$ – Bruce Abbott Feb 3 '18 at 0:26
  • \$\begingroup\$ Put a string of equal series resistors across them to force the voltages across each to be equal. Large value resistors of course. \$\endgroup\$ – user207421 Feb 3 '18 at 0:29
  • \$\begingroup\$ Could you power the ideal diode circuit from the entire series string? \$\endgroup\$ – immibis Feb 4 '18 at 23:53
  • \$\begingroup\$ @BruceAbbott I suppose I wouldn't. It's just that these capacitors were expensive and I don't want to break them :) \$\endgroup\$ – Daffy Feb 5 '18 at 0:26
9
\$\begingroup\$

A -0.3 volts will not hurt any capacitor, even one rated for 5 VDC. Also Schottky diodes are available in very high current modules if needed.

A negative 3 volts may hurt them by confusing the chemicals in the dielectric, much the same as reverse charging a battery. But there the similarity ends. Capacitors can be drained to zero volts for a long time, yet are ready to use again if needed.

Supercaps have a higher leakage current when power is first applied but with time the leakage current drops way down.

There is no 'perfect' circuit to do want you want without a MPU to monitor and use MOSFET's to act as a bypass diode with near zero voltage drop.

If you can keep the reverse voltage under -1.0 volts the capacitors should be fine.

EDIT: Also something to consider is load balancing by placing a 1 Meg ohm resistors across each capacitor. This ensures that they charge up equally and maintain an equal voltage across each one. The penalty is a leakage current of 1 uA per volt of charge on each capacitor.

Addendum: Please see the following link for details of many types of super-capacitor. The small PCB mount type generally have a gold foil or carbon conductors, so do not worry about them going bad due to lack of use. Their only real enemy is over voltage or a high reverse voltage.

https://en.wikipedia.org/wiki/Supercapacitor

A statement from that article says:

As of 2015, a CDC supercapacitor offered a specific energy of 10.1 Wh/kg, 3,500 F capacitance and over one million charge-discharge cycles.

I do not think your supercaps are in danger from just sitting around, but watch out for lead corrosion due to moisture. Also, the PCB mounted type can charge quickly but may supply only 10 mA to a short circuit. It is due to how electrons are 'trapped' inside it and 'migrate' a long winding path to the pins. A cheap 1N5822 schottky diode is plenty good enough, unless you want one with a lower Vdrop.

For better charge-balancing you can use 10 K or even 1 K resistors. Add a 4PDT relay power by 3.3 or 5 VDC to engage the 1 K resistors when charging. When not charging the relay changes contacts to the 1 Meg ohm resistors for maximum storage time.

\$\endgroup\$
  • \$\begingroup\$ I intend to have an active system for charging, since I intend to charge them at upwards of 20 amps. That would overwhelm a 1 Mohm resistor. Glad to know -0.3v won't be a problem though. Guess I can use schottkys after all! Thanks! :) \$\endgroup\$ – Daffy Feb 5 '18 at 0:21
  • \$\begingroup\$ No harm to the 1 Megohm resistor, as it does not have this current flowing through it. These are just balancing/bleed resistors. They see only the caps voltage. \$\endgroup\$ – Sparky256 Feb 5 '18 at 0:26
  • \$\begingroup\$ What I mean is that it wouldn't work to balance. The 1 Mohm resistor would be draining at a few uA while the whole pack is charging at over a dozen amps. I don't see how it would help. Or is there something I'm missing? Also, it's worth pointing out that these caps have a leakage current of around 5mA already. \$\endgroup\$ – Daffy Feb 5 '18 at 0:28
  • \$\begingroup\$ The leakage spec is for capacitors being charged for the first time, or in a long time. The long-term leakage should be much lower. Feel free to use 100 K ohm or 10 K ohm balancing resistors. They do prevent an 'infinity' syndrome where one cap charges full and blocks the others from charging. \$\endgroup\$ – Sparky256 Feb 5 '18 at 0:34
  • \$\begingroup\$ The active system I have in mind would drop the current to around 400mA and put a resistor across the full cap. \$\endgroup\$ – Daffy Feb 5 '18 at 0:40

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.