6
\$\begingroup\$

I am trying to learn basic circuitry and I've been looking into capacitors and their uses in different areas.

When looking at capacitance several different sources say that circuits might malfunction or burn with higher capacity capacitors than designed with. Unfortunately, but none of those sources go into detail.

How can a capacitor cause malfunction if capacitance increases? Wouldn't the capacitor simply take longer to fully charge? Can high capacitance capacitor really cause any sort of "burn"? I mean it cannot store or produce higher current than what is given by power supply, right?

Am I missing some important detail?

\$\endgroup\$
  • 5
    \$\begingroup\$ There is no one-size fits all answer. But large capacitors can affect the stability of op-amps or switching regulators. And they can give rise to large inrush currents when power is first connected to a circuit. Even if the inrush can be accommodated, it may cause the power source Voltage to droop, and if other circuitry is attached to that power source, it may reset or malfunction due to the droop. Unfortunately, the only real answer to this question is, "You have to know what you are doing." \$\endgroup\$ – mkeith Oct 7 '15 at 16:27
  • 1
    \$\begingroup\$ @mkeith I realize that there's no universal best capacitor. I was just wondering what behavior a too big one actually displays and/or what effect it has on the current. The "know what you are doing" can only be achieved by learning and knowing at least some of the behaviors I can understand the topic easier without DIY capacitor explosions and burnt circuits. Also, thank you for an additional example. \$\endgroup\$ – Zero Oct 7 '15 at 17:13
  • \$\begingroup\$ @Zero, yes, I didn't mean to come off like I was squelching your curiosity. My apologies. I did list the problems I could think of quickly off the top of my head. Happy trails to you, and you are wise to try to learn from the mistakes of others. \$\endgroup\$ – mkeith Oct 7 '15 at 20:08
7
\$\begingroup\$

How can a capacitor cause malfunction if capacitance increases? Wouldn't the capacitor simply take longer to fully charge?

Capacitor is a charge reservoir. Switched-mode power supplies need to charge it first. Too large capacitors might make the internal power supply loop go unstable, which would create large voltage deviations across the capacitor and potentially burn it due to too large capacitor heating caused by its non-zero parasitic resistance called "ESR".

Can high capacitance capacitor really cause any sort of "burn"? I mean it cannot store or produce higher current than what is given by power supply, right?

Capacitor do burn quite often. Actually, the aluminum capacitor failure is the most common failure mechanism in large motor drives! Motor drives and other power electronics (solar inverter, wind inverter, car battery charger, ...) exhibit very large current ripples at various frequencies. These ripple currents cause capacitor heating (ESR), which degrades the capacitor capacitance and further increases ESR. It's like a positive feedback. Aluminum caps have limited lifetime measured in thousands of hours. Their lifetime also decreases with elevated temperatures.

The typical way to mitigate this issue is to use multiple parallel caps (splitting the ripple currents) or using higher quality capacitors. These methods, however, tend to increase cost of the final product. The electronics industry is very cutthroat nowadays, which gives rise to designing for full functionality and reasonable failure rate but not a bit more.

Other answers also list good examples of how not only the capacitor can burn but how the large capacitor can cause other components to burn.

\$\endgroup\$
  • \$\begingroup\$ Ya beat me to it. Also I'm not sure that enumerating all cases when higher capacitance can lead to a circuit not working as intended is a good, on-topic question: filters and what not... including any feedback loop really, not just SMPS: capacitance on opamp output etc. \$\endgroup\$ – Fizz Oct 7 '15 at 16:21
  • \$\begingroup\$ @RespawnedFluff I am a motor drive guy so please feel free to post another answer. I am curious myself! \$\endgroup\$ – SunnyBoyNY Oct 7 '15 at 16:31
  • \$\begingroup\$ I would say that the first case is just a design mistake, like many others. But i totally agree with the second statement. \$\endgroup\$ – Gregory Kornblum Oct 7 '15 at 16:46
8
\$\begingroup\$

I'll limit my answer/addition to "burn" type failures, because otherwise there's really a mile-long list of circuits that can malfunction (any high-pass filter, any feedback loop etc.) and depending what they control stuff might burn/blow up. The example that SunnyBoyNY gave with the SMPS output cap is actually an example of feedback loop. More modestly, any opamp can (and will) oscillate with a big/right capacitance on its output; and when it does so it will heat up. Unless it has thermal protection/shutdown, it may get damaged. In all these cases it's the reduction in phase margin by the [increased] cap that's causing a problem. I won't get into details on this, because LTI control theory is almost certainly too difficult to grasp [right away] by someone wondering about this too-big-capacitor-damage issue.

However, even in a linear power supply, a capacitor that's oversized relative to the rectifier (diode) can lead to failure of the rectifier via increased peak [charging] current. Below is a textbook illustration of the problem (for a half-wave rectifier with capacitive filter):

enter image description here

The smoother the filtered voltage gets (less ripple), the smoother the load current gets, but the same energy is transferred to the capacitor over a shorter period of time (conduction interval) so the diode sees a higher repetitive peak/surge current when the capacitor is increased. Actually this is just a repetitive version of the worst-case scenario, which is the inrush current when the capacitor is completely discharged (in which case the peak will actually be much higher), but it's worth noting that diodes have different/lower current ratings for repetitive stress vs "non-repetitive" events (which is a bit of a misnomer, such events can repeat just nowhere near the mains' frequency; this appnote is good read for diode-spec jargon.)

A less obvious variation [because it doesn't involve things sold as capacitors] of this inrush current issue is the gate capacitance of a MOSFET. Make it too big [say, by changing/upping the MOSFET model so it can switch a bigger load] and the gate inrush current can damage an IO pin of a microcontroller...; avoiding such a scenario is one of the [many] reasons why MOSFET driver circuitry can get complicated.

Finally, the problems mentioned are not insurmountable. One can compensate feedback loops (one way of doing that is actually by adding/increasing capacitor[s], but in a different part of the circuit). As for limiting inrush current (directly) there are numerous solutions, the most obvious being adding an inductor.

\$\endgroup\$
  • \$\begingroup\$ I realize there are many general circuit malfunctions that can happen, unfortunately, as i mentioned I couldn't find any specific info and I'm not getting into practical stuff just yet. Anyways, I assumed that if 1 capacitor is too big in an otherwise perfect circuit, the amount malfunctions that could show in current would be limited. SunnyBoyNY answered my question (as vague as it was) to the spot, but your answer is highly appreciated too. \$\endgroup\$ – Zero Oct 7 '15 at 17:05
  • \$\begingroup\$ That's perfectly fine with me. I've added a few more details, but don't take this as a request to change to your accept of SunnyBoyNY's answer. When I find a topic interesting I enjoy writing long answers for which I barely get one upvote (and sometimes none). \$\endgroup\$ – Fizz Oct 7 '15 at 19:07
0
\$\begingroup\$

Aluminum capacitors, which are normally cheap and have higher capacity, are also not reliable. They have working lifetime of thousands of hours. If you look at them, you can even see that they are prepared for explosion:

http://www.o-digital.com/uploads/2179/2184-1/Aluminum_Electrolytic_Capacitor_161.jpg

See those cuts? They will open wide and direct the explosion to the top, saving rest of the board.

\$\endgroup\$
  • 1
    \$\begingroup\$ Eh? What does this have to do with the question asked? \$\endgroup\$ – Fizz Oct 7 '15 at 16:30
  • \$\begingroup\$ I've noticed those, but this doesn't quite explain how too high capacitance affects the circuit. \$\endgroup\$ – Zero Oct 7 '15 at 16:30
  • \$\begingroup\$ Zero, high capacity itself doesn't affect a thing. There are supercaps of very high capacitance, and they work just fine. So leaving aside bad designs, the only problem is reliability. If your application must be reliable, aluminum and some other electrolytes are banned. By the way, in Servo drives you often can't use anything else due to hard competition- you must keep costs low. So you just put as much caps as you can, distribute load between them and hope for the best :) \$\endgroup\$ – Gregory Kornblum Oct 7 '15 at 16:43
0
\$\begingroup\$

Your posting question does not specify which part of your circuit the capacitor is in nor what supplies it. Take for example a fat car battery supplying your circuit with 12V through 1mm^2 copper wires. Suppose that the input of your circuit has a 100,000 microFarads of big fat capacitor there, what length of supply wire would be less than 0.05 ohms so go to 120 Amps when first connected?

Don't try that at home. To design a circuit which is safe and does have a huge input capacitance, you need to set up some extra switching near to the input end so that you can fill up the big input capacitor more gradually through a modest resistance and not draw power to your main circuit until you've charged the input C and bolted down the wires.

At the other end of your circuit, suppose that you have a HF transformer rectified into an output smoothing capacitor. After doing the safe start, that unit could have 12 Volts 10 Amps going into it. After ten seconds, that has received 1200 Joules, so if it has an output capacitor, it could have stored plenty enough to kill you or start a fire. Burns are an underestimate; it is possible to get much more current and much more voltage out of an unspecified circuit (would that, by any chance, contain a car ignition coil?) and a big industrial capacitor than the original car battery supplied.

\$\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.