MYTH: Clipping causes your amplifier to send DC voltage to your subs

[ChrisB]

I can't tell you how many times I heard that clipping will send DC voltage to your subs which will in turn ruin them rather quickly. I recently had 4 Orion HCCA 225 Digital reference amplifiers on my test bench and I measured DC voltage in the millivolt range at clipping with my Fluke 117. Furthermore, there is DC Voltage in the millivolt range at power levels below clipping.

What gives? Where is this DC voltage that will damage subwoofers quickly?

[OSN]

Some people think that squared off waves are DC voltage for short periods of time before ALTERNATING.

[onebadmonte]

Originally Posted by OSN

Some people think that squared off waves are DC voltage for short periods of time before ALTERNATING.

It's not?

[BeatsDownLow]

Same as ripples in DC current being called AC possibly?

[onebadmonte]

Originally Posted by BeatsDownLow

Same as ripples in DC current being called AC possibly?

Whaaaa? Holy smokes, getting this is going to be as bad as trying to understand what the hell was going on in that movie "Inception"

[OSN]

Originally Posted by BeatsDownLow

Same as ripples in DC current being called AC possibly?

The answer hinges on: do these ripples cross the x axis?

[MacLeod]

AC, DC....doesnt really matter. Its the big blast of power coming down the line at double or more the amps rated power thats doing the damage!

[Andy Wehmeyer]

too much power burns coils. doesn't matter whether it's dc or ac. dc that provides less power than the speaker is rated to handle continuously won't burn the coil. Hook up a 6V battery to even a cheap woofer and leave it. You'll drain the battery, the coil will get warm and that'll be it.

[sonikaccord]

But with DC the coil is also stationary and not cooling itself like it would with AC.

[ChrisB]

Originally Posted by sonikaccord

But with DC the coil is also stationary and not cooling itself like it would with AC.

So where is this evidence of DC produced during clipping?

One of my friends had an Elemental Designs amplifier blow and send direct DC to his subs therefore taking them out. Aside from serious malfunction, I have yet to measure any significant DC voltage delivered by an amplifier to the speaker outputs at clipping.

[SQ_Bronco]

Originally Posted by sonikaccord

But with DC the coil is also stationary and not cooling itself like it would with AC.

The distance a cone moves is based on the p-p voltage it sees. For a given peak-to-peak voltage, a clipped signal has twice as much electrical energy as a pure sine wave, but the woofer cone will move the same distance (function of p-p voltage) whether it is getting a sine wave or a square wave. Also note, the amount of heat produced by electrical resistance increases with the square of current (joule's 1st law, Q = I^2 * R), so doubling current creates (at least *) 4x the heat dissipation requirement.

(change temp of the conductor also increases resistance, so it's probably more than 4x).

The primary means to cool woofer coils is air forced across the voice coil by the cone motion. More energy with the same amount of cooling can easily lead to failure. This is clearly not a myth.

Anyway, a square wave is identical to a switched DC signal. ITRW, a clipped signal will probably not really be a square wave, since that would sound like aural sandpaper, but it should be easily visible on an oscilloscope.

ChrisB, can you please describe your test setup?

[ChrisB]

Iota DLS 55 power supply, Velleman HPS40 oscilloscope to verify clipping, 4 ohm 250 watt resistive loads to save my hearing, and a Fluke 117 set to measure DC voltage at clipping. The Fluke was purchased BNIB December 2010, so I am sure it shouldn't be out of calibration so soon.

EDIT: I also used a 68 Hz test tone.

Regardless, I have been told that clipping produces DC voltage, which in turn damages speakers, by many audio sales associates and I have measured no such thing.

[SQ_Bronco]

ChrisB- one good way to visualize this is to look at a graph of amplifier power output vs THD. Zapco did a great job of this with their c2 series- look in one of the manuals for a picture. Output is basically linear/slightly increasing up to the point the amp clips, at which the THD goes WAY up while "power" output (calculated based on p-p voltage, not current) increases very little.

[SQ_Bronco]

I'm confused- how do you "verify clipping" if you don't see the waveform change? If the waveform changes, there is a DC component of the signal.

The issue may be that the velleman is looking for a constant DC signal, whereas it's going to be a switched DC signal. but mathematically, the amount of heat transfer the woofer will need to dissipate is going to be exactly the same whether you are sending 50% duty cycle 2x VDC or +/- x VDC to the woofer. The point is, at clipping, "rms" power dissipation calculations no longer apply, since "rms" is higher for clipped signals vs unclipped signals.

[ChrisB]

Originally Posted by SQ_Bronco

I'm confused- how do you "verify clipping" if you don't see the waveform change? If the waveform changes, there is a DC component of the signal.

The issue may be that the velleman is looking for a constant DC signal, whereas it's going to be a switched DC signal. but mathematically, the amount of heat transfer the woofer will need to dissipate is going to be exactly the same whether you are sending 50% duty cycle 2x VDC or +/- x VDC to the woofer. The point is, at clipping, "rms" power dissipation calculations no longer apply, since "rms" is higher for clipped signals vs unclipped signals.

So why doesn't my Fluke 117 measure DC voltage on the amplifier's outputs as the wave squares off?

[sonikaccord]

Originally Posted by SQ_Bronco

The distance a cone moves is based on the p-p voltage it sees. For a given peak-to-peak voltage, a clipped signal has twice as much electrical energy as a pure sine wave, but the woofer cone will move the same distance (function of p-p voltage) whether it is getting a sine wave or a square wave. Also note, the amount of heat produced by electrical resistance increases with the square of current (joule's 1st law, Q = I^2 * R), so doubling current creates (at least *) 4x the heat dissipation requirement.

(change temp of the conductor also increases resistance, so it's probably more than 4x).

The primary means to cool woofer coils is air forced across the voice coil by the cone motion. More energy with the same amount of cooling can easily lead to failure. This is clearly not a myth.

Anyway, a square wave is identical to a switched DC signal. ITRW, a clipped signal will probably not really be a square wave, since that would sound like aural sandpaper, but it should be easily visible on an oscilloscope.

ChrisB, can you please describe your test setup?

I was referring to a pure DC signal at first. But I agree. A square wave has more "area under the curve" than a sine wave which means it has more power than a sine wave. And as you said, power=heat which leads to blown VC's.

Add the fact that inductors are shorted with DC sources and you basically turn your speaker into a resistor connected across the amp terminals at DC.

OSN put it the best. You can still have an AC signal above the x-axis, but would it be AC or DC?

[alm001]

Originally Posted by ChrisB

So why doesn't my Fluke 117 measure DC voltage on the amplifier's outputs as the wave squares off?

Because half the time the voltage is negative. 68hz is pretty fast. If you had an analog meter, it would be hard to see as well (if the needle could even keep up).
Are you really surprised you didn't find DC (as it is universally understood)?
If you want to measure the voltage peaks of the wave use a scope.

[alm001]

I am not trying to defend the DC argument. An oscillating square wave is not DC.
This is just a matter of ignorance.


But it is (almost) constant current

[ChrisB]

Originally Posted by alm001

Are you really surprised you didn't find DC (as it is universally understood)?

Nope. I only performed the test to put an electrical engineering student in his place on another forum. He was quoting book theory but had no idea of how things measure in the real world. Watching him try to explain my results was comical at best.

[DBfan187]

good read on this subject

Clipping: damage from power or cone movement - CARSOUND.COM Forum

[sqshoestring]

Someone record me a sound that looks like a square wave, where it is flat at some voltage that is not zero.

IMO DC is a flat line not at zero, if you turn the knob on the scope you will see it. Its just a matter of time (division). Call it AC it does not really matter it is power either way, and it will blow the speaker faster. It is also not an electrical signal any natural sound would make, and the speaker does not make sound with static voltage. Its really a major screw up due to amp design that is not limited to its capacity. I still don't know why with all our technology hardly anyone can make an amp that does not clip or even soft clips like a tube, I guess they are too busy making toasters or something. Pro sound can make something clip any way you want, has for decades. Another one of those things that makes me wonder what year I'm living in.

[finkster]

Originally Posted by sonikaccord

But with DC the coil is also stationary and not cooling itself like it would with AC.

i highly doubt the motor has the strength to stay stationary for any period of time when it is still moving more than 30 times in most cases.

Think about a power to movement ratio.

[ChrisB]

Originally Posted by sqshoestring

Someone record me a sound that looks like a square wave, where it is flat at some voltage that is not zero.

IMO DC is a flat line not at zero, if you turn the knob on the scope you will see it. Its just a matter of time (division). Call it AC it does not really matter it is power either way, and it will blow the speaker faster. It is also not an electrical signal any natural sound would make, and the speaker does not make sound with static voltage. Its really a major screw up due to amp design that is not limited to its capacity. I still don't know why with all our technology hardly anyone can make an amp that does not clip or even soft clips like a tube, I guess they are too busy making toasters or something. Pro sound can make something clip any way you want, has for decades. Another one of those things that makes me wonder what year I'm living in.

I remember the DDT lights on my Peavey CS amplifiers when I used to DJ. If you want to talk about an amp that sounds like ass without equalization yet is rock solid rugged, those were the ones.

A little flicker of yellow from the DDT circuit wasn't a bad thing, but if you let it stay lit up in yellow too long, the next color was red because you were in thermal shutdown.

Oh, and NEVER, under any circumstances, let your business partner see you bring one of the amps out of thermal with a CO2 canister blowing air through the amp. I thought I would never hear the end of his complaining the first time he saw me do that to "our" equipment.

[sonikaccord]

Originally Posted by finkster

i highly doubt the motor has the strength to stay stationary for any period of time when it is still moving more than 30 times in most cases.

Think about a power to movement ratio.

Huh? It will be stationary for an infinite amount of time if DC is applied to the coil.
EDIT:as long as the dc is being applied an the coil doesn't burn up

In the case of clipping, the speaker attempts to play the n amount of harmonics that are created during clipping. No it wont't stay in the 'exact' same spot but the movement will be so small that you might as well say it is.

Power to movement ratio? You mean like voltage applied= x amount of excursion?

[sqshoestring]

Originally Posted by ChrisB

Oh, and NEVER, under any circumstances, let your business partner see you bring one of the amps out of thermal with a CO2 canister blowing air through the amp. I thought I would never hear the end of his complaining the first time he saw me do that to "our" equipment.

LOL! Good one

Originally Posted by sonikaccord

In the case of clipping, the speaker attempts to play the n amount of harmonics that are created during clipping. No it wont't stay in the 'exact' same spot but the movement will be so small that you might as well say it is.

I don't know exactly what way you mean this, but far as playing music I thought I read someone tested this and found the cone movement (xmax) was nearly the same with massive amounts of distortion, to the point of the music hardly being recognizable. Same with a square wave. However that does not discount the fact that more energy is contained in a square or clipped wave, so while the cooling may be the same/similar power is still increased. In my non-scientific experience clipping might cause a little more xmax, if there is more left.