33
\$\begingroup\$

Usually you either combine capacitors in parallel because you want to increase the total capacitance while fitting the components in a certain shape/position, or you just combine capacitors by buying a single capacitor of a larger value.

Combining capacitors in series reduces the total capacitance, and isn't very common, but what are some possible uses for it? It shouldn't be used to increase the voltage rating, for instance, since you can't guarantee that the middle will be at half the DC voltage of the total, without using bleeder resistors.

\$\endgroup\$

10 Answers 10

Reset to default
18
\$\begingroup\$

I have only seen it done to increase voltage. On some power supply front-ends (AC/DC conversion) with a voltage doubler the capacitors are in parallel at low voltage and in series at high voltage. This works out well since for a constant power out the current is double at the lower voltage.

As you mention balancing resistors are required.

\$\endgroup\$
20
\$\begingroup\$

Another reason when done in production designs is to reduce your bill of materials (BOM). If your design has loads of 100 nF caps but needs one ~50 nF, it is often cheaper to use two 100 nF's in series due to the quantity you're buying the 100 nF's in, and also reduces pick/place setup time.

\$\endgroup\$
18
\$\begingroup\$

In an automotive application I've seen two ceramic capacitors in series to increase safety against shorts. In the extreme case a short could start a fire, and I heard that had happened at least once.

\$\endgroup\$
4
  • 6
    \$\begingroup\$ I had to do something similar when getting a product UL-listed. Their 'failure of components' test included seeing what would happen if an internal capacitor failed to a dead short. (Completely unrealistic test, but UL is a stickler for rules, and bless them for it. So the 50,000 amps source was shorted directly across two 100A diodes. Yes, the flame escaped the box.) In order to pass the test without requiring expensive external fusing, we put two caps in series. That way, if one shorts, the DC bus itself isn't shorted. Of course, another way around this is to just keep the flame in the box. \$\endgroup\$
    – Stephen Collings
    Commented Aug 18, 2013 at 2:25
  • \$\begingroup\$ @StephenCollings I'm curious as to why you say that it's an unrealistic test. Is that because of your specific circuit or in general? I've seen a ceramic bypass capacitor fail to a short. It was nearly impossible to find it without a thermal camera. The current was limited by the PS so the cap itself didn't go up in flames. \$\endgroup\$
    – horta
    Commented Aug 31, 2014 at 18:41
  • \$\begingroup\$ In my case it was self healing film caps. Ceramics are more likely to short, for sure. \$\endgroup\$
    – Stephen Collings
    Commented Aug 31, 2014 at 20:16
  • \$\begingroup\$ "Where voltage is not an issue, two capacitors in series may allow one to be safely shorted – a technique useful when adding a noise-reducing bypass capacitor between primary and secondary grounds." ieee.li/pdf/essay/… \$\endgroup\$
    – endolith
    Commented Oct 28, 2014 at 16:19
9
\$\begingroup\$

You sometimes see electrolytics connected in series, with opposite polarization directions. In other words, one cap will always be forward biased, no matter what the externally applied voltage. This is, I believe, how one arrives at the situation of having a 'non-polarized electrolytic' capacitor.

follow up -

turns out that what might LOOK like two ordinary electrolytics are not, in fact, two ordinary electrolytics. Despite the uncanny resemblance to exactly that, it is highly probable that the devices I saw had other properties as well.

So the moral of the story is, if you see what looks like two electrolytics stuck together back to back, it is high probability an 'NP' electrolytic, but don't try to make one on your own with regular electrolytics. (Kind of like "you can't make a BJT from two diodes")

live and learn, right? that's why we all love our stackexchanges. thanks out to Kortuk for the eye opener.

\$\endgroup\$
11
  • 1
    \$\begingroup\$ No, do not do this. It will act as a capacitor also, but once you pass a few volts it will blow out the insulator. \$\endgroup\$
    – Kortuk
    Commented Jun 18, 2010 at 3:38
  • 1
    \$\begingroup\$ @kortuk: i think it's crazy, because if one is always forward biased, one is also always reverse biased, but i tell you i have seen this sort of thing sold commercially - maybe something else was going on there. fwiw it was in an audio crossover circuit. \$\endgroup\$
    – JustJeff
    Commented Jun 18, 2010 at 10:02
  • 1
    \$\begingroup\$ You could do it if you used a diode in parallel with each cap. As long as the diode kept it below the blow through voltage. Still very risky. \$\endgroup\$
    – Kortuk
    Commented Jun 18, 2010 at 15:42
  • 4
    \$\begingroup\$ "Special capacitors designed for AC operation are available, usually referred to as "non-polarized" or "NP" types. In these, full-thickness oxide layers are formed on both the aluminum foil strips prior to assembly. On the alternate halves of the AC cycles, one or the other of the foil strips acts as a blocking diode, preventing reverse current from damaging the electrolyte of the other one. Essentially, a 10 microfarad AC capacitor behaves like two 20 microfarad DC capacitors in inverse series." \$\endgroup\$
    – endolith
    Commented Jun 18, 2010 at 22:10
  • 4
    \$\begingroup\$ and yes you can do this: electronics.stackexchange.com/questions/21928/… \$\endgroup\$
    – endolith
    Commented Feb 7, 2014 at 17:28
9
\$\begingroup\$

Kortuk's comments here are the first time I've heard that putting two identical electrolytic capacitors back-to-back is "very risky".

The following references seem to contradict Kortuk:

  • "electrolytic capacitors ... Non-polar (or bi-polar) devices can be made by using two anodes instead of an anode and a cathode, or one could connect the positives or negatives of two identical device together, then the other two terminals would form a non-polar device." -- http://electrochem.cwru.edu/encycl/art-c04-electr-cap.htm

  • "you can substitute two regular electrolytics in series ... with their negative ends joined in the middle. ... in fact, that's what you'd find if you opened a real nonpolarized capacitor." -- Charles Platt. "MAKE: Electronics: Learning Through Discovery". O'Reilly Media, Inc., 2009. p. 249. (excerpts available on http://books.google.com/ ).

  • "A common argument is over whether or not you can make a non-polarized capacitor by putting two polarized electrolytics back-to-back. People have been doing this for years, with no problems to my knowledge." http://my.execpc.com/~endlr/electrolytic.html

Of course, the capacitor will blow up (or not) no matter which way the majority votes. Sometimes the underdog is right.

\$\endgroup\$
3
  • 1
    \$\begingroup\$ I think this is similar to the LED as a diode problem, as long as you stay at 5V or lower you are fine, at my job we use 24V(industrial) and my coworker did not listen. Those 8 LED boards cost a lot to replace if they are certified for medical applications. \$\endgroup\$
    – Kortuk
    Commented Mar 3, 2011 at 22:58
  • 2
    \$\begingroup\$ LEDs work fine as normal diodes at 220V if in series with a 100k or so. They do break down around 100V but the the current is too low to do any serious damage (at least not in several days). \$\endgroup\$
    – jpc
    Commented Mar 29, 2011 at 21:35
  • 2
    \$\begingroup\$ this led to electronics.stackexchange.com/questions/21928/… \$\endgroup\$
    – endolith
    Commented Oct 28, 2014 at 16:21
8
\$\begingroup\$

I have only ever done it to increase voltage rating, and we were using large super-capacitors. They were rated to 2.7 V and we wanted 5V, so we connected in series. We purchased a nice charging controller, which did the job of ensuring they both had the same charge, charging them in parallel.

It decreased our Capacitance to ~25 Farads I believe, but the ESR was <.1 ohms.

\$\endgroup\$
1
  • \$\begingroup\$ What is the total ESR of a series connection of (super-)caps? ESR1+ESR2? Does your last sentence imply that the series connection improved ESR? \$\endgroup\$
    – alexei
    Commented Feb 9, 2017 at 17:33
6
\$\begingroup\$

I've used ten 3.3F supercaps rated at 2.7V in series to build a buffer capacitor for a digitally controlled locomotive of a garden model railroad. This makes it run much better over dirty tracks or switches. The nominal voltage is 24V.

For the first version I used a schematic proposed by somebody else, which doesn't balance voltage. There is a 3k3 resistor anyway to discharge it over some minutes, to prevent surprises.

The second one will use a 1% 330 resistor per capacitor to balance voltage, we'll see if there is a difference in longevity.

\$\endgroup\$
1
  • \$\begingroup\$ this is basically the same as Kortuk's answer \$\endgroup\$
    – endolith
    Commented Feb 7, 2014 at 17:26
3
\$\begingroup\$

To make a railgun =)

Connecting them in series increases the voltage capability (add voltage limits of all caps in series).

\$\endgroup\$
2
  • \$\begingroup\$ Doesn't a railgun want them all in parallel for high current? \$\endgroup\$
    – endolith
    Commented May 30, 2011 at 0:11
  • \$\begingroup\$ I guess not? "The capacitors are assembled in 8 sub banks wired in series, each bank containing 4 capacitors in parallel, for a total rating of 3200V nominal, 3.6kV peak charge and 3088.3uF (measured) capacitance." \$\endgroup\$
    – endolith
    Commented Aug 12, 2015 at 20:21
1
\$\begingroup\$

To have robustness against short circuit specially ceramic capacitors that are connected to power lines. If capacitor shorts, it can burnt PCB trace or worst it may cause fire.

\$\endgroup\$
1
\$\begingroup\$

Well, maybe people rarely see this configuration; however, this trick could be used to create high-voltage bipolar capacitors. If you series-connect two equal value capacitors in series, cathode-to-cathode and use only the positive lead of each cap to connect to other part of the circuits.

This trick are very often seen in audio equipments.

My two cents.

\$\endgroup\$
1
  • 3
    \$\begingroup\$ This is mentioned in previous answers. \$\endgroup\$
    – JRE
    Commented Oct 24, 2018 at 10:59

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.