AC vs. DC

[SteveThackery]

Alternating current (AC) is an electric current that periodically reverses direction and changes its magnitude continuously with time, in contrast to direct current (DC), which flows only in one direction.

Yeah - polarity reversal is part of my definition of "alternating", too. However, not everyone agrees.....

[Aldo22]

If we simplify it by having a 2V p-p sine wave added to 6V DC, is the signal now AC or DC?  I still argue that you cannot apply either label alone because it clearly contains aspects of both.  I did mention earlier that instead of AC and DC we could legitimately refer to "the AC component" and "the DC component". This is much better because it embraces the the observed phenomenon more completely: the signal comprises two components summed: 6V DC and 2V p-p AC.

To summarise: I suggest that a voltage varying sinusoidally between 5V and 7V should not be described as DC or AC. Rather, it should be described as having both a DC component (6V DC) and an alternating component (2Vp-p AC), and the actual signal is the sum of the two components.

No more need to argue about what counts as DC and what counts as AC.

Yes, isn't that exactly what I was trying to say/ask/illustrate here?

[Terry Bites]

Homework??

[SteveThackery]

Yes, isn't that exactly what I was trying to say/ask/illustrate here?

Yep. I'm adding my voice to those of us who think it's unhelpful, and possibly silly, to declare it to be AC or DC. It's a simplification too far.  I particularly loved the contribution arguing that it is all AC, because DC is just AC with a frequency of 0Hz. 

[Zero999]

As is the case with many terms, it's dependant on the context.

If it's a rectified waveform, from a transformer, then it makes sense to consider it to be a DC signal, with some ripple.

If it's an audio signal, superimposed on a DC offset, say in an audio amplifier, then it's AC superimposed on a DC offset.

[Analog Kid]

Homework??

Was that somehow aimed at me, the OP here?
Well, yes, this is homework for me, in the sense that I'm always trying to properly educate myself in the ways of electronics.
But no, it's not homework for any college course.
Happy now?

[golden_labels]

Was that somehow aimed at me, the OP here?
Well, yes, this is homework for me, in the sense that I'm always trying to properly educate myself in the ways of electronics.
But no, it's not homework for any college course.
Happy now?

Don’t get offended. I believe Terry Bites just needs their detector to be recalibrated. Please remember there are many people, who come for nothing more than to have their homework solved for them. I believe you’re not one of them, but it’s easy to misjudge.

Homework??

Appreciated, but don’t you think it’s worth first checking OP’s behavior before jumping into conclusions? The line between spotting a troublemaker and being a jerk is thin. Analog Kid clearly does participate in this thread way beyond the original question.

[radiolistener]

But you're still wrong when you claim that my waveform 3 is my waveform 1 shifted by a DC bias. It's not, which is obvious from just looking at the pictures.

Where did you see that I claimed that "3 is 1 shifted by DC"?
I never made such a claim.
What I said is that 3 is an AC waveform with a DC bias shift.

[radiolistener]

Actually, even DC can be considered as a special case of an AC sine wave with zero frequency, where its value corresponds to the sine wave's amplitude at a fixed phase. Essentially, it represents a sine wave frozen in time.

Sophistry.

This is not sophistry, it’s simply mathematics. For example, consider the Fourier transform, which decomposes signals from the time domain into the frequency domain. In this context, a constant DC level is represented as a sine wave with zero frequency.

When you take the Fourier transform of a signal, the DC component appears as a spike at zero frequency in the spectrum. This demonstrates that a constant value in the time domain corresponds mathematically to a sine wave with zero frequency in the frequency domain. It's a fundamental principle, not a rhetorical argument.

[Analog Kid]

But you're still wrong when you claim that my waveform 3 is my waveform 1 shifted by a DC bias. It's not, which is obvious from just looking at the pictures.

Where did you see that I claimed that "3 is 1 shifted by DC"?
I never made such a claim.
What I said is that 3 is an AC waveform with a DC bias shift.

I don't think even that is true. Can you show us how that's the case? (Draw a picture on the back of a napkin if you like.)

An AC waveform with a DC bias shift would look like what I drew in reply #17 above, not like the half-wave rectified waveform.

[Simmed]

how about using math depiction ?
V*sinωt ?

[radiolistener]

I don't think even that is true. Can you show us how that's the case? (Draw a picture on the back of a napkin if you like.)

No problem:

1) AC sine - dotted blue line
2) Rectified AC sine - red line
3) DC bias + Rectified AC sine - solid blue line
4) Rectified AC sine without DC bias - solid green line

[SteveThackery]

When you take the Fourier transform of a signal, the DC component appears as a spike at zero frequency in the spectrum. This demonstrates that a constant value in the time domain corresponds mathematically to a sine wave with zero frequency in the frequency domain. It's a fundamental principle, not a rhetorical argument.

No, I entirely agree with that. My point is that, in the context of finding suitable words to describe AC, DC and their combination, I don't think pointing out that DC is AC with zero frequency helps much.  Mathematically it is totally correct, but I think it is useful to keep the concepts separate at this stage.  Fourier transforms are for later.

At least, that's my opinion.  I wasn't expecting anyone to agree with me. 😄

[EEVblog]

The term DC on its own, without any other context, implies just what it says, a "straight" DC voltage without any variation.
If it has variation and you don't specify that then the fault is on the laziness of the person using the term.
So if you call a half or full wave rectifed AC signal "DC" just because the current doesn't reverse polarity, then expect to be taken to task for that poor description.

Of course, engineering is a practical field, so in the practical real world "DC" almost means "DC with acceptably low ripple/noise", and no one will evert take you to task for that.
e.g. There is no need to call a DC plugpack "DC with ripple and noise".
Any competent engineer who uses it knows that a plugpack will have ripple and noise and other specs, and is going to have the weigh up that later. But they won't be upset that you said it's a "DC plugpack".

[EEVblog]

DC is a straight line at a fixed level, while everything else is AC. 

That's your preferred definition, but not mine. I don't believe you have any authority to declare your definition as the correct one.
The ONLY authority in the English language is common usage, and - as we can see from this debate - that is far from settled.

Electronics engineering is ultimately a practical field, so the only thing that matters is whether or not you adequately convey the required practical information to another engineer.

To quote you:

"A test point on an electrical circuit has a voltage which varies between 5V and 7V, and this variation is random.  It contains no data or patterns.  Over the long term the random variations are centred around 5V, so they contain no DC offset.  Think of it as noise, if you will.At this test point, is the signal AC or DC?

If you say just AC or DC then you are guilty of poor engineering communication, and hence be prepared to be branded "wrong" and expect to get called out for it.
Your boss and fellow engineers are going to be pretty pissed off if you said your module outputs "5V DC" and it jumps around like crazy randomly by several volts.

[Analog Kid]

Let me try something else here, which may (or may not) shed some more light on the subject.

Seems to me there are basically 3 main "flavors" of waveforms:



(obviously the bottom-left quadrant, alternating but not varying, is nonexistent)

There don't seem to be any accepted terms for these categories, except that alternating/varying is, in its pure form, AC, and not-alternating/not-varying is, again in its pure form, DC.

But of course there are variants of each of these "flavors" which can be disputed.

Maybe we need some more terms (or categories) for this?
Perhaps for some well-known examples (half- or full-wave rectified sine wave, positive- or negative-going square-wave pulses, etc.)?
Or do we only need, as some have suggested, to be precise in our description of any given waveform?

[shabaz]

Let me try something else here, which may (or may not) shed some more light on the subject.

Seems to me there are basically 3 main "flavors" of waveforms:

(Attachment Link)

(obviously the bottom-left quadrant, alternating but not varying, is nonexistent)

There don't seem to be any accepted terms for these categories, except that alternating/varying is, in its pure form, AC, and not-alternating/not-varying is, again in its pure form, DC.

But of course there are variants of each of these "flavors" which can be disputed.

Maybe we need some more terms (or categories) for this?
Perhaps for some well-known examples (half- or full-wave rectified sine wave, positive- or negative-going square-wave pulses, etc.)?
Or do we only need, as some have suggested, to be precise in our description of any given waveform?

DC is not just a straight line at a fixed level for many people; DC isn't an acronym owned by anyone, and for millions of people who are in technical engineering vocations and are indeed engineers, DC means some sort of connection with a pair of wires where one is always "plus" with respect to the other one, and if the signal doesn't visibly change level much on the multimeter, then that's maybe a bonus but not always.

Even for design engineers who may well use equipment to observe signals in extreme detail, DC can still include a signal that varies over a period that realistically might not even be "long".

Part of successful comprehension for engineers relies on recognizing context and recognizing when they need to give or seek additional information and knowing how to do that.

Part of successfully conveying information comes from recognizing your audience.

And, if you recognize that your audience needs particular levels of accuracy, then there's still really no point in trying to be accurate with written words alone, because if the audience is deeply technical, they will simply expect you to have provided things like diagrams, or maybe even photos or maths formulae, or additional material such as specifications or schematics, to understand the full detail of the signal of interest, to the level that they want or need to know at.

My Proxxon hobby tools come with a power supply with the DC symbol on it, and that's absolutely the correct description for their customers, some of whom will be very technical if they are using such tools. The output, of course, looks nothing like a flat line by any stretch (another example would be old-school battery chargers, I expect, from memory).

The manufacturer was still right to label it simply with a "DC" symbol with no further description.

I don't think every situation needs to have precise well-defined words (which might not even translate well to hundreds of languages), and will still not be as good as annotated diagrams and numbers/maths, specifications, etc.

[Xena E]

Electronics engineering is ultimately a practical field, so the only thing that matters is whether or not you adequately convey the required practical information to another engineer.

This.

When using language to describe a generalisation such as this, the context has to be stated as well as just trying to qualify electricity as AC, DC, Zero frequency AC, slowly varying AC, or fairydust cotton candy farts, as one would when defining a voltage source in a simulation for instance.

OTOH, youre not going to put "Zero frequency AC output adaptor" on the label of your wall wart case, because you will look like a complete twat, and Joe public won't understand what that means.

This is another waste of time. This thread is just an example of the TS trying to start controversy for it's own sake by picking abstract subjects, claiming confusion and wanting clarification which is a trait that is now abundantly clear, (and hats off to him, he's very good at it, and playing you all like a fiddle).

X

[Analog Kid]

This is another waste of time. This thread is just an example of the TS trying to start controversy for it's own sake by picking abstract subjects, claiming confusion and wanting clarification which is a trait that is now abundantly clear, (and hats off to him, he's very good at it, and playing you all like a fiddle).

Oh yeah, you nailed it: I only start these threads because I'm a stupid troll who doesn't give a shit about what I'm writing about, only the reaction it will elicit. Not.

Methinks that perhaps you're the twat here ...

[Xena E]

Truth hurts, dont it?

[EEVblog]

Let me try something else here, which may (or may not) shed some more light on the subject.
Seems to me there are basically 3 main "flavors" of waveforms:
(Attachment Link)
(obviously the bottom-left quadrant, alternating but not varying, is nonexistent)
There don't seem to be any accepted terms for these categories, except that alternating/varying is, in its pure form, AC, and not-alternating/not-varying is, again in its pure form, DC.
But of course there are variants of each of these "flavors" which can be disputed.
Maybe we need some more terms (or categories) for this?
Perhaps for some well-known examples (half- or full-wave rectified sine wave, positive- or negative-going square-wave pulses, etc.)?
Or do we only need, as some have suggested, to be precise in our description of any given waveform?

Again, it's about what you want to and need to convey as an engineer, often taking into account who the recipient is, and the context.
Again, as discussed, that might include putting the onus on the recipient to determine if the description is adequate.

Same with all aspects of electronics engineering. Take a multimeter spec for example, it's usually sufficient to say it's a 0.05% accuracy multimeter. But some people need to know it's a "0.05% +2 digits @20degC +/- 5degC with 1ppm/year aging" meter.

[Analog Kid]

OK, so I'll stick with the last part of what I wrote there, edited:

We need to be precise in our description of any given waveform (but only as precise as required).

Works for me.