Clipping an amp??

Discussion in 'Sound, Music, and Intercom' started by nd925a, May 23, 2009.

1. nd925aMember

Messages:
82
3
Location:
Western New York
What is clipping an amp.
I got the impression that it's when you push an amp too far but what happens when you "clip the amp"

2. achstechdirectorActive Member

Messages:
136
0
Location:
Booneville MS
I really cant define what it means but i was taught that is was a bad thing and never to do it. On occasion, the amps clip at high points in the source such as a crescendo in the music or when someone screams but it should not be clipping constantly

3. waynehoskinsActive Member

Messages:
782
49
Location:
Arlington, TX
I'll go technical a bit here, sort of.

Imagine a sine wave with an amplitude of, say, 10. Now let's suppose that you want to draw this sine wave on a piece of graph paper that only goes to 8. Once the instantaneous amplitude reaches 8, it can't go any higher, so it stays at 8 until the input is less than 8. The wave you just drew has its peaks cut off -- it is "clipped".

All of our pieces of audio equipment have operational amplifiers (op-amps) in them doing various things -- in a mixing console's front end, it's voltage gain. In a power amplifier, it's both voltage gain and current gain. These audio amplifier circuits are run from usually a split DC power rail of +15 and -15 volts. An op-amp can typically get within something like 3 volts of its rail voltages -- so maximum output of this op amp would be 12 volts peak.

Let's imagine a voltage gain stage with a gain (Av) = 10. One volt input, 10 volts output. No problems. If we hit it with 2 volts on the input, it would try to make 20 volts on the output. But its rails are +-15, which means it can go to +-12 maximum, so the resulting output waveform is clipped.

Clipping is most prone to occur in the front-end circuit of a device, especially if there's a level mismatch.

Clear as mud?

4. nd925aMember

Messages:
82
3
Location:
Western New York
For the most part it makes sense, so when it cuts off the top of the wave you don't get sound??

5. Eboy87Well-Known Member

Messages:
1,206
49
Location:
Chicago, IL
You still get sound, but what you get is called a square wave. Imagine the sine wave again, when the line moves towards the top of the page, that's the speaker cone moving out. As the line moves down, the speaker cone moves the opposite direction. In theory, with a square wave, there can be no sound, as the flat line at the top would mean that the cone is held in place at that point. In actuality, the line is not flat, and contains a lot of things called harmonics. The frequencies the harmonics are made up of make it look like the line is flat, but are still able to be reproduced by the speaker. Download a freeware tone generator for your computer, and you can experiment with different square waves.

6. David AshtonActive Member

Messages:
828
88
Occupation:
truck driver
Location:
perth W Australia
The important effect of this is that the power is still heating the coil but there isn't enough movement to move the air which cools it, so you have more heat and less coolig = blown speakers.

7. TimmyP1955Active Member

Messages:
793
54
Location:
Indianapolis
This is a common fallacy, born from an assumption based on what a square wave looks like on an oscilloscope. A square wave is nothing more than the summation of a bunch of sine waves. The speaker does not move less than with a clean signal. The problem is that a square wave just has a lot more power than a sine wave of the same peak voltage - twice as much if it is a "complete" square wave (pushed to the max). This is what kills speakers - just plain too much power. Clipping also sounds bad (unless it is used as an effect, such as a dimed guitar amp).

8. David AshtonActive Member

Messages:
828
88
Occupation:
truck driver
Location:
perth W Australia
Sorry but your idea that a square wave is a summation of sine waves is plain wrong, unless you can back up this suggestion with some evidence I suggest it be ignored.If plain power blew up speakers then an amp of twice the speaker rating would blow speakers and an amp of half the speaker rating could never blow it up.Both of these statements are also wrong.As the output gets more and more clipped it heads towards d.c.

9. Eboy87Well-Known Member

Messages:
1,206
49
Location:
Chicago, IL
Assuming an infinite amount of pure sine waves, it's theoretically possible that you can create a square wave. The odd harmonics add up to the square. I could counter your argument by saying that according to Fourier's theorem, all signals that are not pure sine waves, are made up of pure sine waves. If you get enough sine waves close enough in frequency together, the tops of their peaks form a line across the top of the graph, creating a square wave.

Mr. Ashton, if your argument is that a pure square wave isn't the summation of sine waves, I would concede the point. But a pure square wave can never be created by anything other than mathematics; we can't build anything that instantaneously switches phase like a square wave, and there will always be ringing at the edges of the square caused by the harmonics forming the square wave.

And no, it's heat that kills speakers. When high wattage DC goes through the coil for an extended period of time, it heats up. Eventually, either the glue holding it to the former melts, making a mess in the gap and freezing the speaker in place, or it heats up to the point that the cardboard former catches fire. Overpowering a speaker (especially with transients) results in over-excursion, and the voice coil popping out of the gap.

Apologies for the detour.

10. David AshtonActive Member

Messages:
828
88
Occupation:
truck driver
Location:
perth W Australia
Of course it's the heat that kills the speaker, the illogic of your idea lies in the build of the amplifier, it has 2 dc rails above and below 0 and the transistors control that dc to form the ac signals, so as the signal is increasingly clipped it becomes closer and closer to dc and because it is not moving to move air past it, the coils burn out. That is how you can burn out a 100watt speaker with a 50 watt amp.You don't need high wattage dc low wattage is quite sufficient. a 100 watt speaker is only good as a 15 watt heater.I will admit to being baffled how Fouriers theorem can be applied to an amplifier, which is basically a dc device.You do sound very convincing, but I still think your argument is spurious and unsubstantiated.

11. DerrickMember

Messages:
58
4
Location:
USA
The question here is regarding clipping of an amplifier. Unless you are driving an amp WAY too hard, I mean like intentionally driving the crap outta it, you are NOT going to produce a SQUARE WAVE. In what I will refer to here as normal clipping, i.e. more than just flashing the clip indicators every now and then but less that intentionally severely overdriving the amp. You will only produce a CLIPPED output signal.

Simply stated, a clipped indication means you have over reached the power supply limits of the amp and the signal is limited at that level until the input is reduced so that the output drops below clipping.

This increases the average power output presented to the speaker load, thus resulting in heating and high frequency harmonics presented to the speaker. And yes distortion is also a product but distortion in itself is not always damaging to a speaker. Sometimes it is a desired effect and speakers can reproduce it quite well.

Last edited: May 26, 2009
12. waynehoskinsActive Member

Messages:
782
49
Location:
Arlington, TX
One thing that gets missed a lot is that the waveform represents not loudspeaker-cone position, where the cone stops moving at a clipped peak, but velocity. I think that's right. It's been a long time.

If an amplifier were a DC (and DC-only) device it would be useless at audio, which is inherently AC.

Oh, and Mr. Ashton, a square wave in practice IS BY DEFINITION the summation of odd-order sine waves. We covered that in my first-semester Intro To Engineering class ten years ago. Very very basic principle.

A "50 watt" amplifier blows up a "200 watt" speaker because, basically, nobody can agree how to measure power. Power output is also dependent on signal input.

13. David AshtonActive Member

Messages:
828
88
Occupation:
truck driver
Location:
perth W Australia
Oh, and Mr. Ashton, a square wave in practice IS BY DEFINITION the summation of odd-order sine waves.

It may be a definition, I can't find it, but certainly not THE definition.
And an amplifier is essentially a dc device which switches and controls that dc to produce ac but as it clips the dc content progressively increases.When your output transistor blows it does become a pure dc device.

Last edited: May 26, 2009
14. Eboy87Well-Known Member

Messages:
1,206
49
Location:
Chicago, IL
One of the sources I thought more reputable.

Square Waves From Sine Waves Demo

Wikipedia, of course, also has more information on this, Fourier, and Gibbs Phenomenon (ringing).

15. David AshtonActive Member

Messages:
828
88
Occupation:
truck driver
Location:
perth W Australia
"For a finale, we will go from the fundamental to the 19th harmonic, creating vectors of successively more harmonics, and saving all intermediate steps as the rows of a matrix.

These vectors are plotted on the same figure to show the evolution of the square wave. Note that Gibbs' effect says that it will never really get there. "

Hardly "the" definition, especially as it relies on summing odd harmonics, not exactly a problem the average sound guy is ever going to see.

16. museavCBModCB ModsDeparted Member

Messages:
4,017
562
Occupation:
Acoustical, audio and audiovisual consultant
Location:
Marietta, GA
I agree with Derrick, the discussion of square waves and DC is relevant to extreme clipping but the practical effects of clipping can start long before reaching those extremes.

Clipping of an amp typically occurs when trying to get more than the amp can give. The basic issues as I understand it are that a) clipping is overdriving the input of the amplifier, b) clipping can audibly alter the signal, c) a clipped waveform has an increased average level and thus power delivered and d) clipping increases the relative level of the odd harmonics. C and D can kill a speaker long before you get to actual square wave or DC conditions.

17. waynehoskinsActive Member

Messages:
782
49
Location:
Arlington, TX
When ever were we discussing that in regard to square waves?

It is more correct of me to state that the limit, as the number of harmonics approaches infinity, of the sum of a function of the odd-order harmonics of a sine wave of a given frequency is a square wave of the same given frequency. That function is linear: the inverse of the harmonic series number. Fundamental + (1/3 * 3rd harmonic) + (1/5 * 5th harmonic) + ... + (1/n * nth harmonic) where n is an odd integer. This is covered in the CRC book. The equation is even on freaking Wikipedia, along with an animation of the additive synthesis.

As to practical realities of clipping, Brad's right on the money. Increase in power (power being the integral of, I think, the voltage waveform), especially increase in average power. Increase in high-order harmonic content means more power delivered to the weak high-frequency drivers. It has nothing to do with DC. DC on a waveform would be measured as a bias, not as a flatline.

18. gpforetActive Member

Messages:
156
24
Location:
Atlanta Metro
Some really good, and also some very self-indugent answers here but this may be what you're looking for:

An amplifier clips when the input signal drives the amplifer beyond it's ability to deliver the signal to the speakers (at the specified amplified ratio). When the amplifer is driven to this point, and beyond, the output waveform distorts (specifically, clips at the top and bottom), and the speaker is now attempting to respond to this hard edged (clipped) waveform. Yes, some can call it a square wave, but the bottom line is the speaker is now being asked to move in a fashion which is non-musical, non-sinusoidal, and outside of the design specifications for the speaker.

This can be avoided by proper gain staging throughout the signal path. Granted, the further upstream the clipping occurs, the less damaging it is to the speakers because each amplifier stage downstream will have limits on it's ability to reproduce the wave and have a tendancy to slightly round off these square edges of the waveform (a product of the limitations of amplifiers slew rate). The closer to the power amplifier the clipping occurs, the less likely that an in-line amplifier will round off the corners, so if clipping occurs at the amplifier itself, there is nothing left to soften these corners before they get to the speakers. The result is typically heat build up in the voice coil and premature failure of the speaker.

Hope this helps.

19. TimmyP1955Active Member

Messages:
793
54
Location:
Indianapolis
See my previous post. There is NO DC component to a square wave. If there were, many guitarists would blow their speakers every gig. (That's not to sat that there aren't amps that put out DC when clipped, but it's not because DC is the nature of a square or clipped waveform - if there's DC, it's because the amp is crap.

Messages:
828