20

This question will most likely come down to semantics. A friend and I were discussing a point that his EE instructor made, that if you consider the output from a solar panel over more than a few days, it effectively is a cyclic wave, and thus is AC at that time scale. I disagreed with that point and said that it is still DC, albeit with variable output.

Now, while I agree that it's useful to consider the longer measurement period and explain to students how one might conclude it is an 11.6 µHz signal with a DC offset or something, I feel it is misleading to label it AC.

The output voltage is always positive or zero -- unless you measure that signal with reference to some arbitrary point other than one of the panel's terminals. In order to get a negative voltage one would have to measure with reference to a midpoint (e.g. between two serially-connected panels).

I liken the sub-1-Hz "signal" example to an analogy, and maintain that it is somewhat useful, but only as much as, say, water analogies are to explaining electrical phenomenon.

In any case, if I measure a signal where no polarity reversal occurs (at any time scale), should it be considered AC? If so, is "alternating" merely describing the change in amplitude/voltage? Again, if so, does that make current reversal due to polarity reversal a non-requirement to be considered AC?

(A related question asks if a square wave still considered DC. Perhaps a short version of my question is: "Is a square or sine wave with no zero crossing still considered AC?")

JYelton
  • 32,302
  • 33
  • 134
  • 249
  • 5
    *the output from a solar panel over more than a few days, it effectively is a cyclic wave, and thus is AC at that time scale.* I would say that it is neither AC nor DC but actually **both**. In my definition, DC is the **average** current over all days. This DC current is **constant** over the considered time period. Then on top of that there are variations which is the AC part. – Bimpelrekkie Jun 21 '21 at 19:49
  • 1
    @Bimpelrekkie Ah the old AC/DC duality function, sister to that of wave/particle duality. :) – JYelton Jun 21 '21 at 19:50
  • 9
    IMO, it's "AC with a DC bias". You can't really justify NOT calling it "AC" because (unlike DC), it will pass signals through a capacitor. Only AC can do that. – Kyle B Jun 21 '21 at 20:48
  • 4
    Well using that analogy, the voltage of a phone battery is also AC, as i fluctuates up and down over the course of a couple of days... – Linkyyy Jun 21 '21 at 20:51
  • 3
    A capacitor passes DC current during the response to transient conditions. It is only by custom and convention that we say the capacitor passes AC current and blocks DC current. Technically the capacitor takes DC charge until its potential reaches the potential imposed by the surroundings and then it stops taking any more DC current. – SystemTheory Jun 21 '21 at 21:39
  • 1
    Since there is no absolute definition for this, would all answers be opinion-based? – Justme Jun 21 '21 at 23:05
  • 1
    @Kyle but **every** signal or potential or current can be said to be AC with a DC bias.... – Stian Jun 22 '21 at 04:50
  • 1
    @StianYttervik Pretty much yes, or it wouldn't be a "signal". It'd be a DC power supply ;) – Kyle B Jun 22 '21 at 05:30
  • 1
    I'll add... When we do the 'short caps and open inductor' to do what everybody agrees is an "AC analysis" in a circuit, it doesn't matter if the signal itself ever actually goes negative. So if we ALL call that AC analysis, how can ANYBODY make a case that's actually DC???? – Kyle B Jun 22 '21 at 05:32
  • 1
    @Linkyyy perhaps we should consider the magnitude of the 2 components. My camping solar panel is 12V nominal, and presumably 0V in the middle of the night, so 12Vac@11.6µHz+6Vdc. At the right angle is could be close to sinusoidal. My phone battery will be more like 4.2V full and 3.2V if I run it right down over the course of a day, so 1Vac (same 11.6µHz)+3.7Vdc. That will also be an asymmetric sawtooth with a flat top, if we assume usage is even through the day and it gets charged overnight, filling up well before morning \/‾\/‾. Any top-up charging will further break the periodicity – Chris H Jun 22 '21 at 08:34
  • 7
    I've always felt that calling anything "AC" or "DC" without qualifications is a bit misleading unless it's purely AC or purely DC; everything else has both an AC component and a DC component. It's like asking whether 1+j is a real number or an imaginary number; it's neither, it's a complex number. – Hearth Jun 22 '21 at 13:05
  • 1
    Theoretically, a pure DC signal has an infinite-time support (domain) and so does a pure sine wave (which I assume is what is meant by an AC signal). So either way, neither really exist, as all causal signals exist starting at some time-point (and our observations of them will end at some point in time). Also, as pointed out above, some inconsistencies will exist in real-world signals, e.g. for electrical signals there are stray capacitances and unwanted resistance present, affecting the signals as well as the equipment. – JosephDoggie Jun 22 '21 at 16:07
  • Here's a recent one on this: https://electronics.stackexchange.com/q/553412/110857 – ilkkachu Jun 22 '21 at 16:50
  • ...and another old one: [Why graphical representation of a DC current/voltage is a straight line?](https://electronics.stackexchange.com/a/90817/7036) – Nick Alexeev Jun 22 '21 at 20:45
  • Was it designed to be AC but has a large DC offset? Call it AC. Was it designed to be DC but has a large AC ripple? Call is DC, even if it has zero crossings. – Thomas Weller Jun 23 '21 at 08:07
  • "The output voltage is always positive or zero" --> It a question about _current_ (AC: alternating current), not voltage. If the current sign alternates, sounds like AC to me. – chux - Reinstate Monica Jun 23 '21 at 23:02
  • Is a zebra black with white stripes or white with black stripes? That is to say, calling something AC or DC is only to give people an idea of how it will behave with certain components like capacitors. Beyond that, the platonic meaning of the terms are pointless. So it is 100%, as you say, semantics. – Joel Keene Jun 24 '21 at 01:05

9 Answers9

57

This has been discussed (argued) several times on this site and the result is always both yes and no. My own take is:

  1. If the polarity never actually reverses then it's not AC.
  2. At the same time we can represent it as AC with a large DC offset.
  3. Alternately we can represent it as DC with an AC ripple waveform superimposed.

I wouldn't bother getting into an argument about it - commenters please note. Take whichever view best suits the analysis you are doing or that seems most intuitive for a particular situation.

Transistor
  • 168,990
  • 12
  • 186
  • 385
  • 2
    Thanks for the perspective. My main goal was to point out that context is everything. Given the context of "solar power from panels" I felt it was prudent to point out that it's DC. I did search for previous coverage of this question; if you know of any particularly good posts, I'd appreciate links. – JYelton Jun 21 '21 at 19:51
  • 2
    Technically item 1 is correct when applying a statistical model for the net motion of charge carriers in a conductor or semiconductor medium. However item 2 is not incorrect because by convention it is common to refer to the transient or time-varying components as alternating current even when the net motion of charge carriers would remain in the same direction with a time-varying component (AC with large DC offset). – SystemTheory Jun 21 '21 at 21:06
  • 4
    I'm not going to argue. He-he-he! – Transistor Jun 21 '21 at 21:08
  • 1
    1. So you could change a non AC signal to an AC signal by just DC-shifting the potential? – lalala Jun 22 '21 at 19:40
  • 6
    @lalala, sure. A DC decoupling (blocking) capacitor in an audio amplifier does that. – Transistor Jun 22 '21 at 19:59
  • Every periodic signal can be decomposed in two orthogonal components, an AC and a DC component. The L2-norm of these two components is the RMS value. Based on this fact, I would agree with your points 2 or 3, but not with 1: Any time varying signal, even without zero crossings, has an AC component. – Bart Jun 29 '21 at 14:00
  • In the case of, say, DC solar panels, we're not talking about a signal but about energy transfer. In that case, it is better to regard it as time-varying DC, as the net transfer is what you care about and typically you integrate over the time-varying component anyway. Like @Transistor says, it depends what's useful in a given circumstance. – Dan Sheppard Apr 11 '23 at 18:31
9

My own take on this is to look at the actual current at the node in the circuit that you are interested in.

If the voltage wiggles up and down and the current intermittently reverses direction then it is AC.

If the voltage wiggles up and down, but the current stays in the same direction with its amplitude varying then it is pulsating DC.

AC & DC

Peter Mortensen
  • 1,676
  • 3
  • 17
  • 23
  • Hm. Consider a measurement line, 5 VDC with maybe 0.1 mV noise superimposed, into a purely capacitive load. Surely this is a bona fide DC situation‽ But if you consider the current, that'll _only_ respond to the noise part, so by your definition we're dealing with AC. – leftaroundabout Jun 22 '21 at 13:14
  • 1
    @leftaroundabout Sorry for the delay in responding to your comment, I've been out for the afternoon, just got back in. Your described situation into a resistive load would be varying dc. The dc level would just be varying slightly with the current increasing and decreasing but always in the same direction so this would be pulsating dc. Your described situation into a purely capacitive load would be a small amount of ac biased up to +5V. The noise current is repeatedly switching directions as the capacitor charges and discharges and so we have ac. –  Jun 22 '21 at 15:54
  • Well, no need to apologize for that! — Yes, that's what I mean: we're performing a _DC voltage measurement_ – but you want it classified as AC, on the basis that the DC resistance is (by design!) so high that all remaining current is alternating? – leftaroundabout Jun 22 '21 at 21:23
9

Ok, I'll bite.

tl; dr: no. The PV output is intermittent DC, but it never reverses (that is, alternates) so it's not a true AC.

This all hinges on the meaning of ‘alternate’ in ‘alternating current’. Strictly speaking, AC means that the current changes direction over the duration of a cycle. A PV panel does not do that: current flows only in one direction.

We often speak of composited DC signals that have an ‘AC’ (time -varying) element that can be separated and examined by itself, but that doesn’t necessarily make the composited signal an AC signal.

There's a range of cases / categories to consider:

  1. 'pure' DC (no time-varying component at all - practically impossible)
  2. DC with a time-varying ‘AC’ component (no current reversals - describes most real-world DC signals)
  3. AC with a DC bias (some current reversals, but with a net DC current)
  4. 'pure' AC with no DC bias (complete current reversals with zero DC net current)

The described signal (PV panel daily output variation) doesn't fit into either 'pure' category 1 or 4. Further, as we know its current never reverses, 'biased AC' category 3 doesn’t apply either.

For the PV output that leaves us with category 2, a DC current with a time-varying ‘AC’ component.

I agree with your conclusion that it's not correct to label the PV output as 'AC', but for a different reason: PV panel current flow is always the same direction. It varies but never reverses outright. That is, the PV panel current never alternates.

Let’s talk about a couple of things in a PV system that could ‘alternate’.

A rechargeable storage battery reverses its current direction over its charge/discharge cycle, so it is kind of an AC device. It nonetheless will have a net DC bias in favor of in-flow, owing to the battery losses. Thus, it would be category 3, AC with a DC bias.

Now consider a theoretical ideal storage device, like a perfect no-loss capacitor or a superconducting magnet. Being ideal, their charge/discharge behavior will alternate back and forth with no loss-induced bias. These could be regarded as category 4 - pure AC. No such device exists however.

hacktastical
  • 49,832
  • 2
  • 47
  • 138
5

in any case, if I measure a signal where no polarity reversal occurs (at any time scale), should it be considered AC?

It depends on the context. If the 'signal' represents the absolute output voltage or current of a solar panel then it is DC. But if the 'signal' you are looking for is (eg.) the change in voltage relative to the solar panel's average output, then it could be considered as AC riding on a DC baseline. This classification might never be needed for solar power installations, but in other circuits it is often necessary.

Analog video is one example of a signal that is 'DC' in that the level is relative to a fixed voltage rather than the average voltage of the waveform. This is required because the average brightness of an image is dependent on the amount of light and dark parts in it, which may vary depending on the composition.

On computer video displays the usual range for RGB video is from 0 V (black) to 0.7 V at maximum brightness. This works fine if the video card can send the varying DC voltages directly to the display, but is a problem if the signal is sent over a medium that don't preserve the DC level. Although the RGB signals are considered to be DC, when passing through amplifiers etc. they may actually be AC. Then when the signals are received by the monitor they must be converted back to DC using a 'DC restorer' which uses the (originally 0 V) signal voltage in the blanking area at left and right hand sides of the screen as a reference.

The key difference between whether a signal is considered to be DC or AC is whether it has a fixed DC reference voltage or is aligned to the average value of the waveform. Signal voltages transferred via capacitors or transformers are called AC even though the actual voltage may always be above (or below) ground. And of course the current going through these coupling components is always AC because it goes backwards and forwards alternately.

For power it is different. AC literally means 'Alternating Current' which originally referred to electrical power generators. AC power is expected to alternate from positive to negative voltage, and the equipment that uses it must be designed for it. DC power is expected to have the same polarity all the time, though not necessarily the same voltage all the time. A DC or AC power supply may also accept current as well as provide it. In a DC power supply this 'reverse' current is not classed as AC because it doesn't alternate from one direction to the other.

Since a solar panel produces DC voltage and current its power output is classed as DC, not AC. But the varying component of the output could be classed as an AC signal if the absolute voltage is not important, eg. for a small solar panel used as a modulated light receiver.

Bruce Abbott
  • 55,540
  • 1
  • 47
  • 89
  • 1
    The best answer so far. And the most important is the first sentence. – fraxinus Jun 22 '21 at 14:21
  • I'd like to stress that it is alternating *current*, not alternating *voltage*. AC power is expected to alternate from positive to negative current. Voltage sign change depends on the reference voltage you choose, current sign change does not. – Astrinus Jun 23 '21 at 08:33
  • 1
    @Astrinus not true. Mains power is supplied as AC **voltage**. Current is determined by the load's response to that voltage. If you connect a device with a half wave rectifier the current drawn will be of one polarity only. So the term 'Alternating Current' is a bit of a misnomer wrt mains power, which may be why the acronym 'AC' is invariably used. – Bruce Abbott Jun 23 '21 at 08:59
2

Generally speaking, I think you have to "Filter out" frequencies below a certain cut-off, otherwise signal analysis would necessarily begin at the big bang.

Putting aside batteries for one second: let us consider typical components used in signal analysis: resistors, capacitors, and inductors. If, like me, you've forgotten the math(s) you can find impedance calculators online. I put in your frequency of 11 microhertz and got for a 1 farad capacitor a value of 10K ish. For a 1 pF capacitor, multiply that by 1012, and you are now looking at an absurdly high value which of course would be negligible compared to the leakage resistance - for small capacitors like this, the answer is that it is effectively DC.

The same for inductors. The signal is moving too slowly for them to create any significant magnetic field, and they just act like very low value resistors. Again, it looks like DC to an inductor.

In the time frame of charging a battery, the periodic nature of the "signal" is of course significant, and in many ways the day is not long enough to charge a lead acid battery properly (assuming it started flat). However, you would not apply traditional signal analysis to this. You would apply a battery-specific sort of analysis, so maybe then the terms "AC" and "DC" become moot.

Peter Mortensen
  • 1,676
  • 3
  • 17
  • 23
Rodney
  • 332
  • 1
  • 6
1

It's very easy to miss the main point here: it's about communication. How you call something is, in this case, purely about being able to communicate the idea to others.

How likely, do you think, is anyone told about solar panel output to use the term "AC" to describe it?

AC and DC aren't universally defined in some formalized axiomatic system of reasoning. If you discuss non-negative integers in mathematical terms, there's a couple common sets of axioms that let you build up the concept of such numbers, and effectively reason about them. Everyone will then agree about what is meant, as soon as you mention what axiom system is in use.

Alas, with AC and DC, there's no such common ground, and using those term in a "wiseass" manner is unwise. Remember: it's not about how clever you are, but about how clearly you communicate. And, frankly, someone using AC to describe daily variation in solar panel output just seems not to get what clarity in communications means.

That's where I'd end such discussions: they are a waste of time. There's no "technically" here. There's only about you being clear or hard to understand. If you're willing to be hard to understand just to stay on some higher "moral" ground... well, it's your hill to die on.

TL;DR: This has nothing to do with technology, everything to do with the fact that we are humans, and our notions of AC and DC die with us. We are the only reason those terms exist, in their current cultural framework. Some other intelligence, if such exists, may have similar terms, but we'd have not a clue initially about what their scope may be. They may perceive "now" at quite different time scales than we do.

Kuba hasn't forgotten Monica
  • 32,734
  • 1
  • 38
  • 103
0

AC and DC are ideal theoretical concepts. If you do DC analysis on a circuit, then all voltages are constant, period. Add a bit of noise or fluctuation, no matter how small, and it's not DC analysis anymore. AC is expected to be periodic and have a constant amplitude, so only signals which are infinite in time are AC. If you have a sine wave which dies out, you can't describe it with just period and amplitude.

Now, when you call a real-life signal DC, it means that for your purposes the non-DC part of a signal is insignificant. If it is significant, then you have a signal which is not DC. This is the same with pretty much any ideal concept: you call a conductor with a significant resistance "resistor" when you're interested in this particular property, even though every conductor has inductance as well. If your circuit relies on the inductance to be there instead, you might want to call it "inductor" even though it still has resistance.

Furthermore AC and DC are not exhaustive categories. You cannot say "this is not DC, therefore it is AC". Of course it's a question of definition, but if you define AC as "anything that is not DC", then the whole concept of AC becomes useless because no useful properties of such signals could be asserted. If you have a complex signal, labelling it as AC or DC will not help you understand it or explain to others what's going on. It's just like describing the Moon crescent with just "square" and "circle".

In a typical solar panel application this signal would be described as "DC", because the load you would connect to the solar panel would behave the same if you connect it to an ideal DC source. On the other hand, if you connect it to the windings of a motor doing one revolution per day, you would probably want to call it AC. Finally, if both the constant level and the variation are significant, you just call it "periodic signal" and provide a plot.

Dmitry Grigoryev
  • 25,576
  • 5
  • 45
  • 106
  • Re *"AC is expected to be periodic and have a constant amplitude"*: Perhaps make it clearer this does ***not*** restrict it to a sine wave (that is how it can be read at first glance)? – Peter Mortensen Jun 24 '21 at 13:54
0

Depends on if the instructor/teacher is playing tricks with words. AC stands for Alternating "Current", so you could have a sine wave that is all voltage and no current and it wouldn't fall under the term AC, it would simply be a sine wave of a voltage source. In electrical/physics based science, we tend to treat electricity similar to the flow of water through a system. Voltage is the pressure of the water, and current is the amount of water flowing. So you could have a tank of water that goes up and down in pressure but the water never leaves the tank, This would be analogous to a sine wave on an oscilloscope showing voltage but no current is flowing. To have AC, you would have to have current flowing and not just the pressure of voltage.

Luke Dyson
  • 11
  • How does the pressure change in a tank of water if no water ever leaves the tank? – Elliot Alderson Jun 29 '21 at 14:10
  • Pressure could change due to acceleration (including gravity). And if it is a sealed tank, changes in temperature would result in expansion and contraction and thus pressure change. If the tank is NOT sealed, then water vapor would leak out. – PStechPaul Apr 11 '23 at 21:29
0

To alternate means to go in a different direction. In a circuit, the only other way to a forward current is reverse current, therefore AC would refer to current that flows in both directions; forward from 0 and reverse from 0 volts.

DC flows in one direction only, like current from PV cell, because photons release electrons from doped semiconductors, which is a one way process.

Different light frequencies and intensities would vary the DC and voltage from the PV cell, therefore the fluctuating DC from a PV cell would still be DC.

An inverter circuit converts the DC to AC, with a voltage regulator to stabilize the AC around a nominal value.

To just isolate the DC offset from the fluctuating DC, connect a capacitor in series to get fluctuating DC from 0 volts.

Note that this would still be DC, although varying levels of current, since current would only flow in one direction.

To make AC, the current has to reverse, ie. go in reverse direction, below 0 volts, then go forwards again. A cirçuit similar to the one below stops forward current flow on rising voltage, then pulls current in reverse on falling voltage, resulting in a negative voltage at output.

negative voltage circuit

Image source: Maker Pro - Producing Negative Voltage, by Darshil Patel

This negative fluctuating voltage before the capacitor bridge can be coupled and phase matched to the fluctuating DC voltage to give a fluctuating forward and reverse current, effectively an AC.

SamGibson
  • 17,231
  • 5
  • 37
  • 58
Zimba
  • 117
  • 4