Extremely Disappointed

[SubSam 21]

If all the amp will do is 200 volts, and 20 amps..that is all it will do. Even if you do normalize power at say 100 volts of output without "clipping" and then drive it with a square wave the sub still gets PISSED

Exactly. With even that small of an amp if you run a square wave esp at low low frequency the coil will not like it. I can't say it will destroy your coil in every case, but why try. Overhead power is what would help this guy, I'm pretty sure he isn't testing for numbers.

4 ohm power ftw

[SubSam 21]

That clipped/squared wave at it's peak is direct current.

No, it's not.

Nick: Have you managed to determine a higher (quicker?) rate of failure for a clipped signal operating at equivalent average power levels as a non-clipped signal? Obviously a heavily clipped signal of the same RMS voltage as a non-clipped signal is going to cause a driver to fail sooner (if you breach the thermal or mechanical thresholds) as the clipped signal will deliver significantly more power over time. But I've not yet seen a test that determined a quicker failure rate for a clipped signal vs a non-clipped signal at the same average power level.....and I've never personally had enough drivers on hand to intentionally blow stuff up just for shits and giggles

Yep...which is the only reason why I'm debating it.

Same 4kw crown amp in the shop...sine wave, very cool sub is not pissed. Flip the signal generator over to square wave form...not so cool.

The time value of the 'flat' part of the wave at the two extremes literally makes the coil sit there..and do nothing, it is not in motion. It's just hanging out burning, like bouncing the Kia off of the rev limiter and dropping the tranny in reverse.

I can't shift anywhere close to 20 Hz tho lol. It's too frequency specific to really put it that way, although its entirely an accurate comparison, in the aspect of the non-rev limit period.

Obviously that analogy can apply well to visualizing the scenario, but in a waveform at very low frequency. I guess if we are sticking to SUBWOOFER frequencies, which I wasn't but ok, it would be appropriate to think of it that way. When comparing 20-80Hz to15kHz or 5Hz square waves the normal subwoofer frequencies have more in common with the 5 Hz wave, although even cracking the car back and forth at 5 Hz would be close to impossible.

But if you look at the frequency as it approaches 0Hz (dc) the cone obviously cannot generate instantaneous acceleration so there is a "rpm lag cycle" but as frequency approaches 0 even from 15 kHz the cone spends more and more time in a "dc current cycle", which could even be calculated by the wave period minus the rpm lag cycle. By this argument you could prove that clipping is worse for subwoofers than any other speaker since the higher the frequency, the less time the vc spends under straight current. This reinforces Nicks claims towards his speakers, as he builds SUBWOOFERs.

[SubSam 21]

Double post football sucks!

Must be an American thing

Edited February 6, 2011 by SubSam

[kirill007 52]

I'm sure it makes a difference, but the test just goes to show clipping isn't nearly as bad as everyone makes it out to be. Additionally, the excursion with those small drivers is proportional to a few inches in a large driver...

I haven't seen many other people go to this length to test this either...

There's a million "what ifs", but when you break it down, it's hard to dispute hard data...

Everyone says a clipped signal produces more heat, I proved it doesn't...

You're test is invalid, equipment(not actually subwoofers) and tones(freqeuncy) were wrong.

I like that you atleast try it, but you shouldn't us this as a good scientific test, because it wasn't.

And subsam(and nick with the pictures) have already explained why, sqaure wave vs sine wave are just physics, the wave is more "powerfull" with a square wave.(in terms that everyone understands)

EDIT: what do you mean with the next comment? (That's awesome.. Lol.. ^^^ It obviously wasn't scientific....... )

Yes you tried to be scientific, but it wasn't scientific enough, you need to have the correct enviromnemt.(and a controlled one.)

Edited February 6, 2011 by kirill007

[95Honda 486]

That's awesome.. Lol.. ^^^ It obviously wasn't scientific.......

[///M5 2,833]

Still going to disagree..

You've doubled if not more the time per division of the wave form...hence a ton more heat.

It can not double the output power if it is already running at maximum output to begin with.

If all the amp will do is 200 volts, and 20 amps..that is all it will do. Even if you do normalize power at say 100 volts of output without "clipping" and then drive it with a square wave the sub still gets PISSED...which realistically would be less power on the coil that is driven with the square wave form on the per watt basis because impedance is much higher (voltage constant, current lower because of impedance rise due to more heat being induced)..if you normalize to 100 volts of output on square vs. sine wave. The "wattage" by normalizing power definition is no different, all the amp is making at that point is 100 volts, lower current with the square wave if anything, and the square wave form induces a shit ton more heat?

It is because it is going to the high point of the wave, stopping..burning. Immediately reversing and doing the exact same thing. It is hell on soft parts..and it burns the coils up...because it induces a TON more heat. Even using a fluke meter to test heat after driving the sub for 10 minutes..there is a very significant difference in heat in the coil with the square wave form.

Don't normalize in voltage but in power. Confusing to read when we are discussing the effects of power and clipping as the waveform has a huge effect on the amount of power.

I'm sure it makes a difference, but the test just goes to show clipping isn't nearly as bad as everyone makes it out to be. Additionally, the excursion with those small drivers is proportional to a few inches in a large driver...

I haven't seen many other people go to this length to test this either...

There's a million "what ifs", but when you break it down, it's hard to dispute hard data...

Everyone says a clipped signal produces more heat, I proved it doesn't...

You're test is invalid, equipment(not actually subwoofers) and tones(freqeuncy) were wrong.

I like that you atleast try it, but you shouldn't us this as a good scientific test, because it wasn't.

And subsam(and nick with the pictures) have already explained why, sqaure wave vs sine wave are just physics, the wave is more "powerfull" with a square wave.(in terms that everyone understands)

EDIT: what do you mean with the next comment? (That's awesome.. Lol.. ^^^ It obviously wasn't scientific....... )

Yes you tried to be scientific, but it wasn't scientific enough, you need to have the correct enviromnemt.(and a controlled one.)

Huh?? Invalid? Perhaps science eludes you. Test was spot on and way better from a scientific perspective than I have ever seen ANY manufacturer do.

[Impious 1,476]

I'm sure it makes a difference, but the test just goes to show clipping isn't nearly as bad as everyone makes it out to be. Additionally, the excursion with those small drivers is proportional to a few inches in a large driver...

I haven't seen many other people go to this length to test this either...

There's a million "what ifs", but when you break it down, it's hard to dispute hard data...

Everyone says a clipped signal produces more heat, I proved it doesn't...

You're test is invalid, equipment(not actually subwoofers) and tones(freqeuncy) were wrong.

I like that you atleast try it, but you shouldn't us this as a good scientific test, because it wasn't.

And subsam(and nick with the pictures) have already explained why, sqaure wave vs sine wave are just physics, the wave is more "powerfull" with a square wave.(in terms that everyone understands)

EDIT: what do you mean with the next comment? (That's awesome.. Lol.. ^^^ It obviously wasn't scientific....... )

Yes you tried to be scientific, but it wasn't scientific enough, you need to have the correct enviromnemt.(and a controlled one.)

The pictures shows normalized voltage, not normalized power.

No one would disagree that a squarewave will deliver significantly more power over time than a sinewave of a normalized voltage. That's not what is at issue.

What is at issue is a difference with normalized power. Nick explained his experiment in short detail, and 95Honda is spot on that flipping a switch, without normalizing power, doesn't answer the question. If he's operating the amplifier at 200V peak voltage, then the RMS voltage with a sinewave is ~141V. If you flip a switch to generate a squarewave, that same 200V peak voltage equates to 200V RMS voltage, significantly increasing average power (doubling it, in fact). We all agree, that's bad. But that's not what is the disagreement is over. The disagreement is over normalized power, which I don't believe has actually been settled yet. Nick's stated his opinion, I'm not sure if he's actually tested it as what he described wasn't comparing normalized power (I could be wrong, maybe he has tested it but just didn't describe in detail how).

[nem 2,019]

I think this is a start to a Great Sticky in the Advanced Discussion section.

The info you guys are presenting is Awesome!

[kirill007 52]

Still going to disagree..

You've doubled if not more the time per division of the wave form...hence a ton more heat.

It can not double the output power if it is already running at maximum output to begin with.

If all the amp will do is 200 volts, and 20 amps..that is all it will do. Even if you do normalize power at say 100 volts of output without "clipping" and then drive it with a square wave the sub still gets PISSED...which realistically would be less power on the coil that is driven with the square wave form on the per watt basis because impedance is much higher (voltage constant, current lower because of impedance rise due to more heat being induced)..if you normalize to 100 volts of output on square vs. sine wave. The "wattage" by normalizing power definition is no different, all the amp is making at that point is 100 volts, lower current with the square wave if anything, and the square wave form induces a shit ton more heat?

It is because it is going to the high point of the wave, stopping..burning. Immediately reversing and doing the exact same thing. It is hell on soft parts..and it burns the coils up...because it induces a TON more heat. Even using a fluke meter to test heat after driving the sub for 10 minutes..there is a very significant difference in heat in the coil with the square wave form.

Don't normalize in voltage but in power. Confusing to read when we are discussing the effects of power and clipping as the waveform has a huge effect on the amount of power.

I'm sure it makes a difference, but the test just goes to show clipping isn't nearly as bad as everyone makes it out to be. Additionally, the excursion with those small drivers is proportional to a few inches in a large driver...

I haven't seen many other people go to this length to test this either...

There's a million "what ifs", but when you break it down, it's hard to dispute hard data...

Everyone says a clipped signal produces more heat, I proved it doesn't...

You're test is invalid, equipment(not actually subwoofers) and tones(freqeuncy) were wrong.

I like that you atleast try it, but you shouldn't us this as a good scientific test, because it wasn't.

And subsam(and nick with the pictures) have already explained why, sqaure wave vs sine wave are just physics, the wave is more "powerfull" with a square wave.(in terms that everyone understands)

EDIT: what do you mean with the next comment? (That's awesome.. Lol.. ^^^ It obviously wasn't scientific....... )

Yes you tried to be scientific, but it wasn't scientific enough, you need to have the correct enviromnemt.(and a controlled one.)

Huh?? Invalid? Perhaps science eludes you. Test was spot on and way better from a scientific perspective than I have ever seen ANY manufacturer do.

Well excuse me, but if he uses midrange/tweeters to test instead of subwoofers where this is all about.

And he uses too high freqeuncies.(170hz and 1000hz)

The subwoofers don't use any of the cooling methods a "decent" LF driver has.

Good job for him trying to do it, but you can't use the results on SUBwoofers with SUBfreqeuncies.

The higher the freqeuncies the less it matters if the wave is square or a sine.(in heat that is.)

DIRECT QUOTE from the link honda95 gave : "A speaker that fails at 100 watts fails at 100 watts. 100 watts of a sine, 100 watts of a square, 100 watts of music averaged..... This is a fact..."

He tries to tell square wave=sine wave , that this is a fact, which it isn't.(he only concluded it from his tests.)

And somebody else on this forum had a test with square vs sine waves, although it was a different subject.(SPL instead of heat)

His conclusions were that on lower power the square wave is louder, if lots of power is applied the sine wave is louder.(out of that we can conclude that power compression occured, or because it doesn't moves smooth pressure doesn't get's build up enough. And not enough pressure will build up in the forced cooling from the subwoofer.)

Edited February 7, 2011 by kirill007

[NDMstang65 403]

If...amplitude is given at say 5 volts, and current remains constant we have these two graphs below.

The area below the graph line is the time constant that it is considered 'on'.

The Square wave form does not see more power. It sees 5 volts for a longer period of time. The sine wave only saw the 5 volts for a very short amount of time. For those examples 15.078/10 = 1.5078..

It sees over 50% more energy for that given period of time, that one cycle. That builds up over an extended period of time more..and more..and more. It is all about time, not the power itself. The time frame is 0 to pi, amplitude is the same..which is 5. The only thing that varied is the wave form. 5 volts, 50 volts, 500 volts..5 million volts...

Time per division is the issue here...

[SNIPER-92 0]

Heat is a voicecoil's foe.

understatement... Heat is the devil to all electronics... there is nothing worse in the world than heat build up on a component no matter the task it is intended for.. Good luck with your situation holley... hopefully it is quick..

true ****

[BenHemp 1]

Bump that voltage up even higher

[BenHemp 1]

http://www.wolframalpha.com/input/?i=integrate+30sin+x+dx+from+x%3D0+to+pi

http://www.wolframalpha.com/input/?i=integrate+30+dx+from+x%3D0+to+pi&asynchronous=false&equal=Submit

94.2/60

= 1.57

[BenHemp 1]

100 volts, 314.159/200 = 1.57 = 57% more area under the curve of the square wave.

[ZMX 0]

The Square wave form does not see more power. It sees 5 volts for a longer period of time.

[...]

It sees over 50% more energy for that given period of time, that one cycle. That builds up over an extended period of time more..and more..and more. It is all about time, not the power itself.

Energy over time is the definition of power.

Edited February 7, 2011 by ZMX

[95Honda 486]

DIRECT QUOTE from the link honda95 gave : "A speaker that fails at 100 watts fails at 100 watts. 100 watts of a sine, 100 watts of a square, 100 watts of music averaged..... This is a fact..."

He tries to tell square wave=sine wave , that this is a fact, which it isn't.(he only concluded it from his tests.)

You don't even understand the test. And now you are just making shit up, lol... If you had a clue, you would understand that I never said all waveforms were the same, quite the opposite really.

This is why you, and most other people who arm chair quater back this are full of BS... You just make crap up out of gut feelings.

But anyway, this is truly a dead horse... I am done beating it...

[Impious 1,476]

If...amplitude is given at say 5 volts, and current remains constant we have these two graphs below.

The area below the graph line is the time constant that it is considered 'on'.

http://a8.sphotos.ak...9_7319515_n.jpg

http://a3.sphotos.ak...0_6586465_n.jpg

The Square wave form does not see more power. It sees 5 volts for a longer period of time. The sine wave only saw the 5 volts for a very short amount of time. For those examples 15.078/10 = 1.5078..

It sees over 50% more energy for that given period of time, that one cycle. That builds up over an extended period of time more..and more..and more. It is all about time, not the power itself. The time frame is 0 to pi, amplitude is the same..which is 5. The only thing that varied is the wave form. 5 volts, 50 volts, 500 volts..5 million volts...

Time per division is the issue here...

I'm no electrical engineer or mathematician, so I very well could be wrong, but I do not believe you are factoring it correctly. Peak and RMS voltages are not calculated the same for a sinewave and squarewave. What you have represented there would be a sinewave and a squarewave of equivalent peak voltage, not RMS voltage (or average power).

If you have a 5V peak sinewave (the area under the curve there is peak, not RMS), then you would need to compare it to a 3.535V RMS squarewave to arrive at equivalent average power. (For a sinewave, RMS voltage is Peak voltage * .707, for a squarewave RMS voltage = Peak Voltage. Since average power is Vrms * Irms, we would need to compare RMS voltages not peak voltages for an accurate comparison of normalized average power)

http://www.wolframal...rom+x%3D0+to+pi

http://www.wolframal...rom+x%3D0+to+pi

The difference there is 10%, not 50%, for a sinewave and squarewave at normalized average power. With a 50hz signal, the subwoofer is completing one cycle every 20ms. How much more quickly will a 10% increase in heat ("area under the curve") build in a coil that's being actively cooled via coil movement every 20ms, or 50 times per second? The difference in area under the curve would be the same as comparing a 5V peak sinewave and a 5.5V peak sinewave.

If I'm wrong here, please show me where, as it's possible I'm misunderstanding. But I do not believe I am.

[NDMstang65 403]

No..

The area under the curve is the energy applied to the coil that entire cycle time..

The thing that is where the coil is changing position, to where it then starts traveling south. Rolling off and back south if it is a sine wave, or slamming immediately if it is a square wave. There is no 'down time' in the square wave form...hence the heat being induced and things burning up because the coil is stuck at 100% north or 100% south for that given point of amplitude. There is very little time that it is not at an arbitrary 5 volts. The sine wave, has a great deal of time that it is not at the 5 volts...which is where your RMS figures comes into play.

As you stated, in square wave form..your peak is indeed your RMS...which completely makes sense as to why these things are burning up. Because what you are forgetting about is the coil is not in motion at this point. It does not move. Sine wave..it moves in a rolling form as it should where a great deal less heat is built up for the same input power.

It is north..burning, immediately south, burning, then back again because the coil is getting to the X point north and the Z point south in the amount of time it takes for the fets to change direction of the coil. (Remember the car bouncing off of the rev limiter analogy)

It is factored in over time per division..where the coil literally stops, builds heat, shoots south, builds heat, shoots north..builds heat..100 watts, is 100 watts, is 100 watts, But when the coil is burning at the extreme north and south points...because that is what it is being told to do, is where we run into issues. Again, time is the problem here. Not more power, more time.

The coil sees no 'down' time or 'rest' time with this square wave form. It only gets worse when you are dumping 2kw into the sub...trying to kick off 1800 watts of heat. Then smack it with a square wave form where it is not in motion for the period of time that the gates are switching on the output transistors..and you have a burned up sub.

[NDMstang65 403]

The Square wave form does not see more power. It sees 5 volts for a longer period of time.

[...]

It sees over 50% more energy for that given period of time, that one cycle. That builds up over an extended period of time more..and more..and more. It is all about time, not the power itself.

Energy over time is the definition of power.

Yes

But 100 watts, is still 100 watts. Sine, Square, Triangle, Clipped.

Still 100 watts.

What changes, is the time the 100 watts is applied.

[ssh 924]

No..

The area under the curve is the energy applied to the coil that entire cycle time..

The thing that is where the coil is changing position, to where it then starts traveling south. Rolling off and back south if it is a sine wave, or slamming immediately if it is a square wave. There is no 'down time' in the square wave form...hence the heat being induced and things burning up because the coil is stuck at 100% north or 100% south for that given point of amplitude. There is very little time that it is not at an arbitrary 5 volts. The sine wave, has a great deal of time that it is not at the 5 volts...which is where your RMS figures comes into play.

As you stated, in square wave form..your peak is indeed your RMS...which completely makes sense as to why these things are burning up. Because what you are forgetting about is the coil is not in motion at this point. It does not move. Sine wave..it moves in a rolling form as it should where a great deal less heat is built up for the same input power.

It is north..burning, immediately south, burning, then back again because the coil is getting to the X point north and the Z point south in the amount of time it takes for the fets to change direction of the coil. (Remember the car bouncing off of the rev limiter analogy)

It is factored in over time per division..where the coil literally stops, builds heat, shoots south, builds heat, shoots north..builds heat..100 watts, is 100 watts, is 100 watts, But when the coil is burning at the extreme north and south points...because that is what it is being told to do, is where we run into issues. Again, time is the problem here. Not more power, more time.

The coil sees no 'down' time or 'rest' time with this square wave form. It only gets worse when you are dumping 2kw into the sub...trying to kick off 1800 watts of heat. Then smack it with a square wave form where it is not in motion for the period of time that the gates are switching on the output transistors..and you have a burned up sub.

and it's the manufacturers fault right?

Good read.

But seriously this is a perfect example as to why many high end speakers come with no warranty, they don't want to deal with people blowing everything up and saying the drivers were faulty.

[Impious 1,476]

No..

The area under the curve is the energy applied to the coil that entire cycle time..

Right. The value of the graph would be in volts. The coil would track the motion of the graph. If you were wanting to compare average power levels, as we are, you would have to factor the RMS voltage of the sinewave and make that value the voltage of the graph for the squarewave, as I did in my graph. Which would equate to a 10% difference.

[NDMstang65 403]

You took voltage away from the square wave...to make it "equal" using the assumption of .707 is the RMS figure (the 3.535dx from x=0 to pi), the sine wave you assigned 5 volts to, the square wave where rms=peak you took power away from it to make it "equal" by RMS figures.

I definitely see what you are saying...which begs the question, does voltage actually drop with sine vs. square with same wattage? But at that point how could the same 100 watts for example be the same? If voltage is down on the square wave, and current is definitely down due to a higher resistance of the coil due to induced heat...you are not getting anywhere near the 100 watts that you were with the sine wave, you are getting far less..and there is way more heat present?

[///M5 2,833]

The Square wave form does not see more power. It sees 5 volts for a longer period of time.

[...]

It sees over 50% more energy for that given period of time, that one cycle. That builds up over an extended period of time more..and more..and more. It is all about time, not the power itself.

Energy over time is the definition of power.

Indeed and power should always be described by instantaneous or average. When using the word power on its own it is intrinsically assumed that you just left off the average portion. This of course means there is a rather stark difference in power in those two graphs. .707....

[///M5 2,833]

Ok, I posted that before I realized there was another page.

You took voltage away from the square wave...to make it "equal" using the assumption of .707 is the RMS figure (the 3.535dx from x=0 to pi), the sine wave you assigned 5 volts to, the square wave where rms=peak you took power away from it to make it "equal" by RMS figures.

Exactly as that is how you describe power. Again, average is implied. Has to be unless you are referring to instantaneous but then it wouldn't be a stress test showing heat as you can't measure instantaneous heat changes accurately.

I definitely see what you are saying...which begs the question, does voltage actually drop with sine vs. square with same wattage? But at that point how could the same 100 watts for example be the same? If voltage is down on the square wave, and current is definitely down due to a higher resistance of the coil due to induced heat...you are not getting anywhere near the 100 watts that you were with the sine wave, you are getting far less..and there is way more heat present?

You should think about your calculation of watts. By definition it is average power. You are taking a shortcut of course and using a simple measurement device that only measures instantaneous which is confusing this as it is the wrong unit of measurement for the task at hand. You need to account for the delta T otherwise the whole discussion of heat and cooling makes no sense.

[BenHemp 1]

Kids, stop frying your voicecoils. It makes for long argumentative discussions.