Can a capacitor let DC current through?

My thought about this is... No, capacitors do not let DC current through, nor AC currents either.

Capacitors block DC voltage. AC signals appear to pass through capacitors, but they don't actually do that.

Flame away.

The *signal* definitely passes through the capacitor. If it "appears" to pass, then it does; that's the nature of a "signal".

Whether or not the current "passes" is largely a question of semantics. Certainly, the electrons themselves don't pass, but at any moment in time, current flows into one lead and the same amount of current simultaneously flows out of the other.

Correct.

It's simple electrostatics.

An electron arriving on one plate of (assumed vacuum dielectric) will repel another from the opposite plate.

An electron pulled off one plate will attract an electron onto the other.

Pulling and pushing in this way is the essence of AC.

The current doesn't actually flow through, but because the net effect on both sides is the same, the current appears to flow through!

But it's not a matter of semantics - it is a matter of what is actually happening with AC signals. Practically speaking, it doesn't usually make much difference, however.

Yes!

No, it really does flow through. The problem here is that while we typically imagine current to be a stream of electrons, all nicely charging down a path (perhaps in single file, and all in step), real "electrical current" only follows this model in such thing as the beam of a CRT.

Current is the effective flow of electrical ENERGY; it involves the motion of charge carriers, and its units are defined in terms of charges-per-second, but is not directly equivalent with the motion of a particular charge carrier or group except under very specific circumstances. For example, what is the "speed" of electrical currents through a conductor? For almost all practical concerns, it makes sense to look at only the propagation speed of the energy or signal in question - even though the individual electrons are moving FAR slower than this.

Bob M.

Oh Hum.

Nope.

(snip)

I have a motor that runs on 60 Hz AC. I put a big honkin' capacitor in series with the line. The motor keeps running. I put an ammeter also in series, first on one side of the cap and then on the other, and get the same reading both times. So what isn't "flowing through," in your opinion?

Or are you thinking that the very electrons that are passing through my desk lamp at this very moment actually made a visit to the nearest power generating station in recent memory?

Bob M.

--
No, _charge_ flows into one lead and out of the other. ;)

--
The flame of truth burns eternal. :-)

And I am its bearer. Shazam!

Well, you imagine wrong. It's "pulsating DC", yes, but it's composed of many AC components riding on a DC reference. The way it gets through the capacitor is that at the rising and falling edge of the waveform, the capacitor suddenly finds itself with a voltage difference across it, and charge wants to flow so as to minimize that potential.

So, essentially, for this application, the square wave can be considered AC.

In fact, with the right instruments, when you first apply DC to one terminal of a cap, that change in voltage is coupled to the other plate by the capacitance, and the capacitor charges according to T=RC (google "time constant".)

Hope This Helps! Rich

Unfortunately, so do the sparks of stupidity. :(

--
Service to my country? Been there, Done that, and I\'ve got my DD214 to
prove it.
Member of DAV #85.

Michael A. Terrell
Central Florida

You can say the same thing about a length of wire. The electrons which flow into one end aren't the same ones which flow out the other end, unless the current is DC and you are prepared to wait a long time.

When it comes to AC, the question of whether a current flows *through* a component is the same whether that component is a capacitor or a resistor.

An alternating current either flows "through" both, or it flows through neither, or you are arbitrarily changing the definition of "through".

Thank you all for your response.

But a few more (perhaps dumb) questions :

1) What is the output from a microcontroller's pwm that is generating square waves considered (+5v -> 0v -> +5v -> 0v...etc)? Is it considered an AC or DC signal or a combination of both? After reading all the replies, I'm left with the impression that "pulsed DC" is considered as AC by some folks here even though it does not reverse direction (go negative) when it goes low (0v). 2) Should not the output from a capacitor be HIGH if it encouters a steady DC current? I know you've said that capacitors 'block' steady DC currents but why? Mentally, I imagine when the dc current first hits the capacitor's plate like a tsunami, it charges up the plate and pushes electrons on the opposite plate away. I imagine those (pushed away) electrons then go "racing away" from the capacitor perhaps towards a load which should be driven HIGH by those electrons so long as the capacitor's opposite plate is charged (which is what a DC current should keep doing). Why then is this not the case ? Somehow the mental model just does not fit. I can imagine how AC passes after reading the descriptions you guys have provided but why DC does not generate a contiuous high.

If only I could watch cartoons of what the electrons were doing, it would all be clear to me.

DC has two common meanings. One is unidirectional current or voltage, regardless of how it varies with time, and the other is steady current or voltage. I think most circuit designers (who are familiar with Fourier and LaPlace analysis) tend to think in the frequency domain, at least part of the time, and are more likely to think of DC as a steady current or voltage, essentially a zero frequency signal.

By that frequency domain way of thinking, that pulsing PWM unidirectional voltage has some DC component (the average voltage, and a whole series of AC components with various magnitudes and phases, relative to the pulse timing, that are all harmonics of the pulse frequency. Filters with various frequency responses will pass varying amounts of all those components.

Very. Infinite, if you wait an infinitely long time. The only way to get current through a capacitor is to have the voltage across it change at some rate. The formula that relates current to rate of change of voltage is :

I=C*(dv/dt) with I in amperes, C in farads, and dv/dt in volts per second.

So, the only way to get 1 ampere of DC to pass through an ideal 1 uF capacitor is to have the voltage across the capacitor climb at 1 million volts per second, and keep climbing at that rate for as long as the current must occur. Since few of us have voltage sources, that can climb toward infinity, sitting around, we generally think of capacitors as devices that cannot pass DC.

That's good. But that sudden onset of voltage is not DC in the frequency domain sense, but DC plus an infinity of AC frequencies all added together. The sudden high rate of change of voltage across the capacitor accounts for the high, momentary current.

Sounds pretty good to me. The output DC will be maintained, if there is no load current. So static DC voltage is possible through a capacitor. This is the same as the DC on rubbed balloons. But you can't get a steady current from this effect, only steady, insulated voltage.

AC supplies repeated (and alternating) rate of change of voltage, so alternating current can be passed.

It is a shame we can't put on special glasses and watch the little buggers. All this would be a lot less abstract.

"AC" and "DC" are extremely vague terms, and a capacitor doesn't care about terminology; most capacitors are made in foreign countries and don't even understand English. Whether a signal is one or the other depends on the time frame over which it's observed. A 1-cycle-per-year sine wave sure looks like DC if you observe it for an hour.

For practical purposes, in this case, look at the signal as having a longterm average value and call that the DC component. If you subtract that from the waveform, what's left is the AC component. So a general signal, like your square wave, has both AC and DC components. A capacitive coupling circuit, if the capacitor is the right value, will block the longterm average ("DC") and pass through the short-term wiggles (the "AC" component).

If the square wave is slow enough, the capacitor will get confused and think that the high and low parts of the wave are actually DC, so will lose interest in passing them, so the coupled waveform will start to droop. Capacitors aren't terrible bright, and have short attention spans.

John

A resistor has no dielectric barrier. A capacitor does.

Why do you believe "An alternating current either flows "through" both, or it flows through neither...?"

In a sense, though, there's no real difference between the two from the "dielectric barrier" angle. To examine this further, let's for the moment ignore the "resistor" question and just look at the difference between AC conduction through a capacitor vs. a plain conductor.

Electrical energy passes through any "conductive" path by virtue of the interaction of the fields of charged particles. In a conductor, this occurs at the atomic/subatomic level; in a capacitor, the interaction-through-fields clearly happens on a physically gross level, through the dielectric, and individual charge carriers cannot pass through the dieletric. But that really doesn't matter - to pass electrical energy or an electrical signal, it is the motion of carriers "downstream," induced by a similar motion "upstream," that matters - not that a particular carrier physically passes through the entire length of the conductive path.

In the case of AC, what goes on within a conductor in terms of the charge carrier motion is very interesting. Imagine a perfect conductor as being a frictionless pipe filled with ping-pong balls which just fit inside the pipe. If you push in a ball at one end, a ball pops out the other end (with the time between these two events governed by the physical attributes - elasticity and such - of the balls). There has been a transfer of energy, even though the ball you put in at the one end really didn't get very far and is certainly not the same ball that popped out at the far end. If we model AC this way, then you take one ball at the near end and alternately push it in and somehow suck it out AT THAT END. And at the far end, we have a ball which is behaving in exactly the same way, alternately popping out and being sucked back in. We can again transfer energy (or information) through this process, even though the ball we're pushing/pulling on at the near end NEVER makes it beyond that point!

Going back to actual electricity, let's further note that if we have a capacitor of sufficient size, there is no way at all to distinguish the capacitor-in-series case from the "straight conductor" case, if all we have to look at is the situation "downstream" of the capacitor. The only way the two cases could be distinguished in any event is through the capacitor's effect on the phase relationship between current and voltage, which, for a sufficiently large capacitance, becomes negligible. (The only other means you could use to distinguish these cases at all, given access to any information you want, would be to somehow "tag" individual electrons at the "upstream" side, and then wait a sufficiently long time on the "downstream" side to see if those particular carriers are coming through. But you'd have to wait a very long time to be certain...)

Another way to say this is that an infinite capacitance is indistinguishable from a "short circuit" (a "perfect" conductive path), again unless you have the ability to tag individual charge carriers. This makes sense because a truly infinite capacitance would always have the ability to make the corresponding change in charge on the "downstream" plate as ANY amount of charge enters or leaves the "upstream" plate. You can envision an "infinite" capacitance as either possessing plates of infinite area (if you can ignore concerns re the propagation times across the plate itself) or (possibly better) as having an infinitely thin dielectric (which equates to saying we have a zero-thickness "magic barrier" inserted between two conductors, such that individual carriers cannot pass through but still have an effect on the carriers on the other side, as if the barrier were not there).

In the real world, of course, we can't have infinite capacitances (or at the very least, you can't easily go down to Radio Shack and buy one...:-)), so we have to rely on the frequency of the AC, relative to the capacitance, to make the effects of the capacitor effectively "drop out" of the circuit. In more familiar terms, we would say that for a sufficiently low capacitive reactance has no significant effect when inserted in series into an AC circuit; there is no way to discern its presence by looking at the conditions "downstream" of the capacitor. In any practical sense of the words, then, we would have to say that yes, a capacitor DOES "pass AC."

Bringing resistance into the picture, as opposed to a perfect conductor, is only relevant if we want to compare the effects of resistance to a comparable capacitive reactance. And in this case, we would fall back to the fundamental difference between these two forms of impedance: resistances dissipate energy, while reactances merely store it and return it to the circuit later in time (which results in the voltage/current phase effects in a reactive circuit). But there's still nothing going on here that would cause us to say that the resistor is actually "passing AC," while the capacitor (reactance) is not.

Bob M.