Clipped Amplifier Signals: What's correct and incorrect about this Video? - AVS Forum
Forum Jump: 
 67Likes
Closed Thread
 
Thread Tools
post #1 of 183 Old 08-06-2014, 03:18 PM - Thread Starter
Senior Member
 
Join Date: Jul 2014
Location: Canada
Posts: 464
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 205 Post(s)
Liked: 69
Clipped Amplifier Signals: What's correct and incorrect about this Video?

Garidy is offline  
Sponsored Links
Advertisement
 
post #2 of 183 Old 08-06-2014, 05:27 PM
AVS Addicted Member
 
amirm's Avatar
 
Join Date: Jan 2002
Location: Washington State
Posts: 18,375
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 981 Post(s)
Liked: 547
If he knew how to dress better than the local gas station attendant, he would have a bright future in video tech reporting .

Amir
Retired Technology Insider
Founder, Madrona Digital
"Insist on Quality Engineering"
amirm is offline  
post #3 of 183 Old 08-06-2014, 08:48 PM
Senior Member
 
89grand's Avatar
 
Join Date: Mar 2013
Location: Phoenix, AZ
Posts: 388
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 171 Post(s)
Liked: 87
The ONLY thing I heard correct is avoiding excess bass boost. I can guarantee I can play a completely clipped signal from a 10 watt amp and I could never blow any subwoofer even remotely worth a damn.

Subwoofers get blown from exceeding it's thermal limits, and burning up the voice coil or exceeding it's excursion and therefore damaging the suspension and or cone. Both of those take excessive power input.

Also, setting a gain control on an amp, and in a sense, it IS a volume control, so that the signal never cliupped, you'd lose quite a bit of output from the amp. A little bit of clipping on the loudest peaks never hurt a speaker, and generally even goes unnoticed by the listener.
89grand is offline  
post #4 of 183 Old 08-07-2014, 07:19 AM
AVS Addicted Member
 
arnyk's Avatar
 
Join Date: Oct 2002
Location: Grosse Pointe Woods, MI
Posts: 14,387
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 762 Post(s)
Liked: 1175
Quote:
Originally Posted by Garidy View Post
Yet another rendition of the audiophile myth that clipping turns AC into DC. Sigh!

Reality is that clipping is a means by which the average signal output of a power amp can significantly exceed its ratings.

The basic rule is that speakers are mostly damaged by receiving too much power, period.

You may think that your 100 WPC AVR is safe with that 100 watt speaker, but if you push that AVR into clipping, it can deliver an average of 200 WPC to the speaker.
Just cruising likes this.
arnyk is offline  
post #5 of 183 Old 08-07-2014, 07:27 AM
AVS Addicted Member
 
Bill Fitzmaurice's Avatar
 
Join Date: Nov 2009
Location: New Hampshire
Posts: 10,168
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 1675
What's correct? Nothing. There is no DC in a clipped waveform, and clipping is no more harmful to a sub than water is to a duck. The cone does not move and stop before moving again. A sub treats a square wave just like any other complex waveform. Besides, the vast majority of the harmonics that make the wave square are filtered out by the inductance of the sub voice coil. The person in the video has no clue what he's talking about, and he certainly hasn't read this:
http://forum.qscaudio.com/forum/view...hp?f=29&t=2736

If he's a 'resident expert' for that company I shudder to think about the quality of their products.

Bill Fitzmaurice Loudspeaker Design

The Laws of Physics aren't swayed by opinion.

Last edited by Bill Fitzmaurice; 08-07-2014 at 07:30 AM.
Bill Fitzmaurice is offline  
post #6 of 183 Old 08-07-2014, 09:00 AM
AVS Special Member
 
JHAz's Avatar
 
Join Date: Mar 2009
Posts: 4,036
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 117 Post(s)
Liked: 165
Agree w all above. Fwiw a square wave contains the fundamental plus only odd harmonics at very specific relative levels and in phase. Afaik it would be a magic amp thelat produces no even harmonics. The waveform with even and odd harmonics in the same kind of ratios is a sawtooth, not anything loke a square wave. Moreover a sine wave is a normal audio test signal. No real instrument emits a sine wave, except maybe in rare cases a synth. But most of the sonic character of any sound is defined by harmonics, which prevent the real wave from being anything like a sine wave to start with. when you distort a real, complex wave you just cannot get to anything like a square wave. And looking at a square wave as a fundamebtal plus distortion, the THD of a square wave is greater than 40 percent. An amp will be said to be cliping way below ten percent THD, except in some musical instrument specs that define power at 10 percent THD.

I can imagine music creation applying well over 10 percent thd to a sine wave from a synth or to the almost sine wave a Hammond organ will produce if only one drawbar is pulled out (although that's really an unusual Hammond registration) but that distortion will include even harmonics and will never result in a square wave. A clean square wave sounds kinda fat-but-hollow which is seldom what real players are going for.

Can't find the addess but if you google stereophile precision bass you will find a lot of pictures of the lowest note from an electric bass. You will see the first 20th of a second of the sound is more than 10 dB louder than the sustain of the note. So the big power requirement lasts a twentieth of a second after which power requirements drop to one tenth or less. The note contains second harmonic, about 80 Hz far higher than the fundamental and significant levels of higher harmonics. The wave looks nothing like a sine wave.

Last edited by JHAz; 08-07-2014 at 09:15 AM.
JHAz is offline  
post #7 of 183 Old 08-07-2014, 09:14 AM
AVS Addicted Member
 
amirm's Avatar
 
Join Date: Jan 2002
Location: Washington State
Posts: 18,375
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 981 Post(s)
Liked: 547
Quote:
Originally Posted by JHAz View Post
Agre w all above. Fwiw a square wave contains the fundamental plus only odd harmonics at very specific relative levels and in phase. Afaik it would be a magic amp thelat produces no even harmonics. The waveform with even and odd harmonics in the same kind of ratios is a sawtooth, not anything loke a square wave.
Can't say that I have seen a sawtooth coming out of a clipped amplifier. But anything is possible because real amplifiers have far more factors causing them to "clip" than the simple paper assumptions we make. Here are clipping waveforms in varying amounts from a number of AVRs that I have measured:



Pioneer with class D amp:






The reason for the complex waveforms instead of what we imagine it to be is due to complex interactions between the feedback circuit, protection circuit, and unregulated power supplies that feed the output stage of the amplifier.

As some of you know, Arny and crew have "proven" that mild clipping can be audible:






Amir
Retired Technology Insider
Founder, Madrona Digital
"Insist on Quality Engineering"
amirm is offline  
post #8 of 183 Old 08-07-2014, 09:23 AM
AVS Addicted Member
 
arnyk's Avatar
 
Join Date: Oct 2002
Location: Grosse Pointe Woods, MI
Posts: 14,387
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 762 Post(s)
Liked: 1175
Quote:
Originally Posted by amirm View Post
Can't say that I have seen a sawtooth coming out of a clipped amplifier. But anything is possible because real amplifiers have far more factors causing them to "clip" than the simple paper assumptions we make. Here are clipping waveforms in varying amounts from a number of AVRs that I have measured:



Pioneer with class D amp:






The reason for the complex waveforms instead of what we imagine it to be is due to complex interactions between the feedback circuit, protection circuit, and unregulated power supplies that feed the output stage of the amplifier.

As some of you know, Arny and crew have "proven" that mild clipping can be audible:





Unfortunately none of the pictures above are relevant to the article that was cited.

I searched around on the net and found this article that was far more relevant the the audible problem that the Audio Research D120 SS amp demonstrated:

http://www.ti.com/lit/an/snaa008b/snaa008b.pdf

The oscilloscope trace from that article that most resembled what I observed back 30 or so years ago was much more like this:

Attached Thumbnails
Click image for larger version

Name:	TI SOA article.png
Views:	312
Size:	457.5 KB
ID:	201097  
arnyk is offline  
post #9 of 183 Old 08-07-2014, 11:52 AM
AVS Addicted Member
 
Ratman's Avatar
 
Join Date: May 2002
Location: Collingswood, N.J.
Posts: 14,646
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 230 Post(s)
Liked: 315
How many times do we have to be subjected to seeing the same attachments over and over and over....
Tack, koturban and andyc56 like this.
Ratman is offline  
post #10 of 183 Old 08-07-2014, 12:01 PM
AVS Special Member
 
Glimmie's Avatar
 
Join Date: Sep 2000
Location: Los Angeles, CA
Posts: 8,017
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 182 Post(s)
Liked: 251
Quote:
Originally Posted by Bill Fitzmaurice View Post
What's correct? Nothing. There is no DC in a clipped waveform,
Except when the "clipper" (teenage party host) keeps it up for a few hours and an output transistor finally develops and emitter to collector short. Then we'll get some DC!

Glimmie's HT Page
Being redone - comming soon!

Glimmie is offline  
post #11 of 183 Old 08-07-2014, 12:29 PM
AVS Special Member
 
JHAz's Avatar
 
Join Date: Mar 2009
Posts: 4,036
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 117 Post(s)
Liked: 165
Quote:
Originally Posted by amirm View Post
Can't say that I have seen a sawtooth coming out of a clipped amplifier. But anything is possible because real amplifiers have far more factors causing them to "clip" than the simple paper assumptions we make. Here are clipping waveforms in varying amounts from a number of AVRs that I have measured:
I never suggested you'd get a sawtooth wave from a clipping amplifier. Sorry I didn't state every negative. Similarly, looks like your broader experience is like mine - never saw a square wave out of a clipping amp, either. None of those test signals looks to be moving cleanly toward a square wave, and obviously none is a square wave.

Even if an amp clipped with only odd harmonics, it would have to (a) maintain phase and (b) just happen to put out voltage of H3 at 1/3 the voltage of the fundamanetal, H5 at 1/5, etc etc, or you just cannot get a square wave out of the thing. Just as unlikely as having an amp that precisely issues H2 at 1/2 the voltage of the fundamental, H3 at 1/3, H4 at 1/4 etc. This strikes me as no great surprise, And demonstrates that no amp is ever going to put out a square wave based on a sine wave input. Nor will it put out a sawtooth. Nor did I state (or imply) that it would.

And to get into square wave range isn't just a little clipping. I am sure I've pushed guitar amp chains past 40 % THD, although I cannot prove it, but I suspect music reproduction at 40% thd would sound downright awful to even a mighty drunk listener. IOW, to be perfectly as clear as I can be, I don't believe (but cannot prove) that in real world sound reproduction anybody ever pushes their amps anywhere close to the 40 percent THD that a square wave would require.

So the whole square wave thing is really a red herring. For speaker damage the question is how much power is getting to the speaker, for how long. The speaker doesn't know or care whether the harmonics come from cleanly reproduced violins or a guitar distortion box, or a distorting power amp. In the end it's power. IMO, it's unlikely that the harmonics from distortion significantly contribute to speaker damage in most real world situations. Even at 40 percent THD, the distortion products are almost 8 dB below the undistorted signal, so they contain something like fifteen percent as much power as the undistorted portion of the signal.
.
JHAz is offline  
post #12 of 183 Old 08-07-2014, 12:54 PM - Thread Starter
Senior Member
 
Join Date: Jul 2014
Location: Canada
Posts: 464
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 205 Post(s)
Liked: 69
Food for Thought:

The terms DC and AC are used to visually describe in the English language, the general nature of a 'currents' state in delivery, as opposed to using the numeric language of mathematics. These are in fact, general terms, not absolutes. DC has ripples, which can be visually perceived on many scopes,especially digital scopes, as having an AC component, at a particular level of magnification. Inversely, the same statement can be made about AC; however,viewing them at a correct level of magnification (one that’s in keeping withthe standard notation) reveals that they each have very different properties,relating to how the current is delivered. One is direct (0Hz) and another is on a carrier frequency.

The math that explains the applied wattage within a DC circuit is different from that of an AC circuit, due to the presences of the carrier frequency(s). DC calculations are not complex; however, AC calculations are extremely complex.

Moving away from the math, just for now, I’d like to return to the visual representation of 'current' as it appears on a scope. Using Amir’s measurement, we discover that both channels are clipped, one more than the other and at different Voltages (let’s not read into the quality of this product at thistime). Note: that the flat top of the Yellow trace resembles the delivery of 'current' in a way that is more direct (closer to DC) than alternating (AC).



During periods of more DC then AC deliveries, the wattage is no longer calculated using an RMS determination (for sine waves); it is determined using peakVoltage. Instead of developing a mean of 70.7% of peak, 100% of the Voltage is applied, respectively.

Now, for some math; but just before I get into that, please also note that for the duration that clip is occurring, for a give portion of a waveform, the impedance, becomes mitigated to a range nearing or equaling the total DC resistance of the voice coil (6.5-Ohms in most 8-Ohm Nominal products).

AC Wattage

RMS Wattage of an AC signal @ 90 & 270 Degrees = (Peak-to-Peak x .707) x Impedance (nominal 8-Ohms)

Peak-to-Peak clipped yellow 81.5 Volts AC x .707 = 57.62 VAC RMS

(Divide Peak-to-Peak score by 2) x Impedance = non-clipped RMSpower = 28.81 VAC x 8-Ohms (Nom.)

RMS Wattage of a non-clipped AC signal equals – 230.48 Watts AC

Note that my Voltage calculations are in good agreement with Amir’s scope, which calculated 29.293 AC RMS – the deviation deriving from the truncating of decimal places and not the formula.

DC Wattage

The Total Wattage Applied for DC is the Voltage x DC resistance voice coil resistance (6.5-Ohms).

81.5 / 2 = 40.75 VDC x 6.5-Ohms DC = 264.875 Watts (delivered more as direct current than alternating), making them mathematically and inreality DC Watts.

Some summary facts (more to be addedalong the way):

1. The presence of DC causes a driver to ‘hang’ or alternate less, for the full duration of its presence (across all degrees of delivery), within a waveform.

2. When hanging occurs, heat increases, driver venting is lessened and therefore cooling decreases, and all manners of distortion increase, while further reducing MTBF scores (Mean Time before Failure).Increased heat weakens adhesives and burns voice coils, but kills sonic goodness's well before driver failure.

3. In Higher Fidelity applications, 0% of clip is acceptable – period.

4. One may crudely determine the % of DC wattage vs. % of AC wattage present within a measured signal, by means of adding up the number of degrees within a waveform that are derived from the delivery of DC, and subtracting them form 360-degrees. The balance will total the wattage derived from AC delivery.

5. Degrees and time are interchangeable – view thegraph.

6. In the presence of a clipped signal, most raw drivers cannot survive long term exposure of more than 25% of their rated RMS AC wattage.

a. I.e. a 200 Watt RMS AC driver is likely to fail with exposure to wattages derived in part or in whole, from 50-watts of DC and onwards, which is to say, that as a rule of thumb, drivers can only withstand approximately 25% DC of a rated AC RMS wattage.

7. If a signal is heavily clipped, say for 270-degree of a waveform, but the clipped voltage is under 25% of the rated RMS power handling of a raw driver, the driver will not fail due to thermal overrun. It will sound poor, and if ran in free air, may suffer mechanical failures.


So these are some truths to consider, before you develop yourfinal stance.


Once we debate through what I have posted above; I will post mathematical evidence, that DC, in the traditional/common sense use of the acronym, is in fact produced at all amplifier outputs and can become problematic to frequency response and become exacerbated during durations of clipping.

A few steps at a time as they say, opposed to one-step at a time, which they also say.
Attached Thumbnails
Click image for larger version

Name:	Clipping 1.jpg
Views:	59
Size:	68.1 KB
ID:	201481  

Last edited by Garidy; 10-12-2014 at 11:00 PM.
Garidy is offline  
post #13 of 183 Old 08-07-2014, 01:11 PM
Senior Member
 
89grand's Avatar
 
Join Date: Mar 2013
Location: Phoenix, AZ
Posts: 388
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 171 Post(s)
Liked: 87
One quick question, did you write all of that from your own thoughts or is it simply a cut and paste from somewhere else?

Just curious.
89grand is offline  
post #14 of 183 Old 08-07-2014, 01:33 PM - Thread Starter
Senior Member
 
Join Date: Jul 2014
Location: Canada
Posts: 464
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 205 Post(s)
Liked: 69
Quote:
Originally Posted by 89grand View Post
One quick question, did you write all of that from your own thoughts or is it simply a cut and paste from somewhere else?

Just curious.
It is safe to assume that it's always original thought, unless I cite otherwise.

This is merely the polarizing tip of a large and poorly explored iceberg.

I'm looking forward to some gentlemanly discussions.
Garidy is offline  
post #15 of 183 Old 08-07-2014, 01:45 PM
AVS Addicted Member
 
amirm's Avatar
 
Join Date: Jan 2002
Location: Washington State
Posts: 18,375
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 981 Post(s)
Liked: 547
Quote:
Originally Posted by arnyk View Post
Unfortunately none of the pictures above are relevant to the article that was cited.

I searched around on the net and found this article that was far more relevant the the audible problem that the Audio Research D120 SS amp demonstrated:

http://www.ti.com/lit/an/snaa008b/snaa008b.pdf

The oscilloscope trace from that article that most resembled what I observed back 30 or so years ago was much more like this:

Hi Arny. You should tell us if we need to wear 3-D or X-ray glasses in order to make sense out of that scope display! That is one crusty old scope snapshot, likely printed from the CRT onto polaroid paper, then scanned and put in that doc.

I am puzzled how after 30 years you remember the measurements. And that 1% clipping looked like that.

Also puzzling is that I thought up to now you held the position that output transistors are so robust as to not hit their SOA per this and many other posts like it:

Quote:
Originally Posted by arnyk View Post
The above comment misses the point which is that good engineers don't spend money on transistors with SOA and then waste it with over-aggressive protection circuitry.
What you are showing from TI, is aggressive multi-factor protection circuit kicking in. It is a simulation of worst case of course but you say that you saw something similar with just 1% clipping in your amplifier double blind test???

Anyway, here is another graph from TI paper where they are showing what happens with just voltage clamping:



Ignore the sine wave looking waveform as that is the current graph. The other that looks like a square wave is pretty much the same as what I showed:



I don't know how if you had read the TI document you would not have seen the above graph and resemblance to clipping waveform I had shown in AVRs.

Amir
Retired Technology Insider
Founder, Madrona Digital
"Insist on Quality Engineering"
amirm is offline  
post #16 of 183 Old 08-07-2014, 02:36 PM
Member
 
Nodscene's Avatar
 
Join Date: Jul 2014
Location: Canada
Posts: 102
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 59 Post(s)
Liked: 14
I don't want to get too far off topic as this thread is going in a different direction than what I am asking (although it was mentioned). I'll try to keep this short but from my limited experience ages ago and from a friend with a fair amount of knowledge (at least as far as I was concerned), I/we always found that speakers (almost exclusively the tweeters mind you) were mostly blown from underpowered amps rather than too much power. There more than a few occasions where an inexpensive amp (Yamaha and the like) would blow some speakers when cranked for a long period of time but would have zero issue with PA amps (Hafler exc). So we are talking about a Yamaha with 150 watts give or take vs a 500 watt Hafler or 800w PV and the only time a speaker would blow was with the former and this is with relatively the same speakers power handling ability of 100-200 watts or so. Of course I never blew anything myself as I actually listened for any distortion but was always able to get better volume out of the same set of speakers using more headroom from a higher powered amp. I'm not saying you can throw unlimited clean power into any speaker but within reason that was the result. As it's been mentioned that clipping an amp won't necessarily blow a speaker then I'm curious about my experience.

Thanks.
Nodscene is offline  
post #17 of 183 Old 08-07-2014, 03:28 PM
AVS Special Member
 
Glimmie's Avatar
 
Join Date: Sep 2000
Location: Los Angeles, CA
Posts: 8,017
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 182 Post(s)
Liked: 251
Quote:
Originally Posted by Nodscene View Post
I/we always found that speakers (almost exclusively the tweeters mind you) were mostly blown from underpowered amps rather than too much power.
The reason for this is that cheap amps will often break into ultrasonic oscillation when driven to clipping. You can't hear it but there is often a lot of energy in that oscillation. It simply burns out the tweeter due to too much power being expended. The crossover blocks this energy to the woofer.

As a teenage hacker I lost both my tweeters at once just listening at low volume to an old reel to reel deck I was given. I was listening while monitoring in record and the bias signal which was around 50khz was leaking into the audio output. Different cause but same issue to the speakers as an amp with ultrasonic oscillation.
Henrick Rennow likes this.

Glimmie's HT Page
Being redone - comming soon!


Last edited by Glimmie; 08-07-2014 at 03:32 PM.
Glimmie is offline  
post #18 of 183 Old 08-07-2014, 03:52 PM - Thread Starter
Senior Member
 
Join Date: Jul 2014
Location: Canada
Posts: 464
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 205 Post(s)
Liked: 69
Quote:
Originally Posted by Nodscene View Post
I don't want to get too far off topic as this thread is going in a different direction than what I am asking (although it was mentioned). I'll try to keep this short but from my limited experience ages ago and from a friend with a fair amount of knowledge (at least as far as I was concerned), I/we always found that speakers (almost exclusively the tweeters mind you) were mostly blown from underpowered amps rather than too much power. There more than a few occasions where an inexpensive amp (Yamaha and the like) would blow some speakers when cranked for a long period of time but would have zero issue with PA amps (Hafler exc). So we are talking about a Yamaha with 150 watts give or take vs a 500 watt Hafler or 800w PV and the only time a speaker would blow was with the former and this is with relatively the same speakers power handling ability of 100-200 watts or so. Of course I never blew anything myself as I actually listened for any distortion but was always able to get better volume out of the same set of speakers using more headroom from a higher powered amp. I'm not saying you can throw unlimited clean power into any speaker but within reason that was the result. As it's been mentioned that clipping an amp won't necessarily blow a speaker then I'm curious about my experience.

Thanks.
I have actually answered this question above.

It has to do with different types of Wattage deliveries: AC vs. DC; when an amplifier is in a clipped state, both delivery methods are present; but the speaker rating only depicts it's tolerance to one of them.


I will add a little more:


Let's, for the purpose of answering your question, assume that all Watts are the same, and that the only difference is the nature of the Watts development, from a 'current' delivery standpoint. Musical Watts, are intended to be derived from pure, alternating currents, not direct current; for what I believe should be obvious reasons, to most AVS readers.

Now musical waveforms are intrinsically dynamic, which is to say, varying in amplitude, which is commonly measured in Volts. These variances occur due to changes in the opposition to 'current' flow, which we have learned mathematically, to be based on frequency. This form of opposition is known as impedance (because it varies). If the resistance varies then so must the 'current' and therefore the Wattage, due to Wattage = 'current' x 'current' x a sampled resistance, within an AC signal (IxIxR).

Extrapolation #1 : Musical AC Wattage is in a continual state of change, throughout each waveform, and related harmonics. To gain an idea of the typical range of AC Wattage output, we use a method known as root, mean and squared (RMS), to determine such (much more on the later on). This method requires that multiple samples be taken, at different degrees/Amplitudes, then that each sampled be squared, then summed, then divided by the number of different sample points, which provides the mean, then calculate the root of the mean, which concludes the final step in RMS Wattage estimation.

Extrapolation #2 : Non-musical DC Wattage is not in a continual state of change, it is essentially constant. The reason being, direct current is deliver via 0Hz, and therefore zero variance in resistances, regardless of time. There is no RMS value for DC, just M or mean. One can very easily add a series of samples, take the sum and divide it by the number of samples, and the result would be the average or mean of the Voltage.

Therefore, they're not electrically similar in the way in which they're produced, calculated or applied. A speakers RMS Wattage rating is based on AC, not DC input Wattage. DC power is constant, not ranging, it's applied at 100% - for 100% of the time. The presence of greater durations of direct current delivery (minimal, to no variance), more quickly brings a drivers voice coil to it's thermal limits, than the delivery of ranging alternating currents. In this regard, a Watt is Watt, and the speaker simply has a different tolerance predicated on how it's delivered: higher for pure AC, lower for mixed current, lowest for pure DC deliveries.

In your examples, the differences are most probably rooted in the presences / occurrences of 'current' that has been delivered in a more direct fashion, which is to say, some amplifiers were probably clipping more than others.

With regards to your tweeter comment: Tweeters simply have smaller gauge voice coils and less cooling potential than that of larger drivers, as such, it's common to see them come undone or 'cook' first, in systems that are being driven heavily into clip, or in systems with a severe, over powered, mismatch, between the amp and speaker system; typically by a factor of no less than 2:1. As a rule of thumb, one should minimally match the RMS Values of their speakers to the max power value of an amplifiers, when high output is the intended application, or much of the amplifiers rated power is likely to be demanded (more on this later as well).
Henrick Rennow likes this.

Last edited by Garidy; 08-07-2014 at 04:21 PM.
Garidy is offline  
post #19 of 183 Old 08-07-2014, 04:27 PM
AVS Special Member
 
Speedskater's Avatar
 
Join Date: Mar 2004
Location: Cleveland, Ohio
Posts: 2,037
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 13 Post(s)
Liked: 20
Quote:
Originally Posted by amirm View Post
Hi Arny. You should tell us if we need to wear 3-D or X-ray glasses in order to make sense out of that scope display! That is one crusty old scope snapshot, likely printed from the CRT onto polaroid paper, then scanned and put in that doc.
............................................
That's what electronic engineering was like 20 or 30 years ago. Maybe you should ask a friend who is knowledgeable in electronic engineering to try to explain it to you.
67jason and Garidy like this.

Kevin
Speedskater is offline  
post #20 of 183 Old 08-07-2014, 04:57 PM
AVS Addicted Member
 
Bill Fitzmaurice's Avatar
 
Join Date: Nov 2009
Location: New Hampshire
Posts: 10,168
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 1675
Quote:
Originally Posted by Garidy View Post
With regards to your tweeter comment: Tweeters simply have smaller gauge voice coils and less cooling potential than that of larger drivers, as such, it's common to see them come undone or 'cook' first, in systems that are being driven heavily into clip, or in systems with a severe, over powered, mismatch, between the amp and speaker system; typically by a factor of no less than 2:1.
Tweeters do get toasted from clipped waveforms, but not for the reasons you stated. They get cooked from overpowering. That's because a clean signal has a 3dB decrease in power density with each octave increase in frequency. A tweeter with a 5kHz crossover will normally see only 5 watts out of a 100 watt broadband signal. When a signal is clipped the harmonic content increases, raising the high frequency power density. That tweeter that normally sees no more than 5 watts of a 100 watt signal may see four times that or more, though not for long, as it will burn out. A clipped signal does not increase the power density in the fundamental frequencies, so subs and woofers see no more power from a clean signal or a clipped signal of the same voltage swing, thus they are not overpowered. And as I already noted they're also protected by the inductance of the voice coil. Any competent engineer, and for that matter any knowledgeable amateur, knows that when you low pass filter a square wave you end up with a sine wave, and the inductance of a voice coil is a low pass filter. That's also why no sub or woofer cone ever moves, stops, then moves again. The voice coil filters out the high frequency components of a square wave that makes it square.

You can blame the myth of under-powering on this document:
http://www.jblpro.com/pub/technote/lowpower.pdf

It correctly explains why clipped waveforms can kill high frequency components, ie., tweeters and midranges. The problem is that with each retelling of the tale the key phrase, 'high frequency components', was lost, probably with the first retelling at that.

Bill Fitzmaurice Loudspeaker Design

The Laws of Physics aren't swayed by opinion.
Bill Fitzmaurice is offline  
post #21 of 183 Old 08-07-2014, 04:57 PM - Thread Starter
Senior Member
 
Join Date: Jul 2014
Location: Canada
Posts: 464
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 205 Post(s)
Liked: 69
Quote:
Originally Posted by Glimmie View Post
Except when the "clipper" (teenage party host) keeps it up for a few hours and an output transistor finally develops and emitter to collector short. Then we'll get some DC!
Yes, straight-up DC!
Garidy is offline  
post #22 of 183 Old 08-07-2014, 05:37 PM
Member
 
Nodscene's Avatar
 
Join Date: Jul 2014
Location: Canada
Posts: 102
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 59 Post(s)
Liked: 14
Thanks guys
Garidy likes this.
Nodscene is offline  
post #23 of 183 Old 08-07-2014, 05:38 PM
AVS Addicted Member
 
amirm's Avatar
 
Join Date: Jan 2002
Location: Washington State
Posts: 18,375
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 981 Post(s)
Liked: 547
Quote:
Originally Posted by Speedskater View Post
That's what electronic engineering was like 20 or 30 years ago. Maybe you should ask a friend who is knowledgeable in electronic engineering to try to explain it to you.
Boy you guys are grumpy today .

I have many friends who "try" to explain such things to me to no avail. I continue to be dumb and happy. So let's not talk about the sun rising out of the east (or is it west? I forget).

That said, I still have my Tektronix CRT scope from circa 1990. For some things, an analog scope is still nice to have even though I use my digital scope for most work as you saw from earlier AVR measurements. The quality of images on the Tek scope is far better than the image in the TI of course since it hasn't gone through a film capture and then scan into a document. Hence me pulling Arny's leg for posting it.

Anyway, if you like to teach me something, knowing full well that it has high chance of falling on my deaf ears, please feel free to do so. Like to know for example if you think that TI graph is representative of the amplifier Arny tested 30 years ago where he found it in double blind ABX tests to sound different, thereby "proving" that amplifiers sound different. If you just wanted to make fun of me, mission accomplished and I hope future contributions are technical in nature.
Garidy likes this.

Amir
Retired Technology Insider
Founder, Madrona Digital
"Insist on Quality Engineering"
amirm is offline  
post #24 of 183 Old 08-07-2014, 05:50 PM
AVS Special Member
 
Glimmie's Avatar
 
Join Date: Sep 2000
Location: Los Angeles, CA
Posts: 8,017
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 182 Post(s)
Liked: 251
Quote:
Originally Posted by amirm View Post

That said, I still have my Tektronix CRT scope from circa 1990. For some things, an analog scope is still nice to have even though I use my digital scope for most work as you saw from earlier AVR measurements. The quality of images on the Tek scope is far better than the image in the TI of course since it hasn't gone through a film capture and then scan into a document. Hence me pulling Arny's leg for posting it.
Ha, I still have my 50mhz Leader from 1984! I also have a 1999 Tek TDS. But sometime that old analog scope is just better!

Glimmie's HT Page
Being redone - comming soon!

Glimmie is offline  
post #25 of 183 Old 08-07-2014, 06:00 PM
AVS Addicted Member
 
amirm's Avatar
 
Join Date: Jan 2002
Location: Washington State
Posts: 18,375
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 981 Post(s)
Liked: 547
Quote:
Originally Posted by Bill Fitzmaurice View Post
Any competent engineer, and for that matter any knowledgeable amateur, knows that when you low pass filter a square wave you end up with a sine wave, and the inductance of a voice coil is a low pass filter.
As was just correctly stated, I am no competent engineer, nor knowledgeable amateur, so forgive me for saying that statement is not correct. A square wave becomes a sine wave only if your filter point is lower than third harmonic and it has sharp enough response to completely filter out the later harmonics.

A 100 Hz square wave will have odd harmonics of 300, 500, 700 Hz, etc, all of which are still in the domain of the woofer. Here is a 100 Hz square wave generated in Audition CC:



Here is the exact signal but now filtered at 4 Khz, using an 18th order ChebyChev low-pass filter:



That is not remotely a sine wave as the math predicts. There is ringing because we took away the harmonics beyond 4 Khz but there are plenty of harmonics below 4K to still let us have a waveform that looks like a square wave and nothing like a pure sine wave.
Garidy likes this.

Amir
Retired Technology Insider
Founder, Madrona Digital
"Insist on Quality Engineering"
amirm is offline  
post #26 of 183 Old 08-07-2014, 06:01 PM
AVS Addicted Member
 
arnyk's Avatar
 
Join Date: Oct 2002
Location: Grosse Pointe Woods, MI
Posts: 14,387
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 762 Post(s)
Liked: 1175
Quote:
Originally Posted by Garidy View Post
Yes, straight-up DC!
Only if there is no protection facility in the amplifier.

Use of protection relays operated by a protection circuit that is specifically looking for DC at the output of the power amps is pretty common. Speaker fuses can also help.
arnyk is offline  
post #27 of 183 Old 08-07-2014, 06:11 PM
AVS Special Member
 
Glimmie's Avatar
 
Join Date: Sep 2000
Location: Los Angeles, CA
Posts: 8,017
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 182 Post(s)
Liked: 251
Quote:
Originally Posted by arnyk View Post
Only if there is no protection facility in the amplifier.

Use of protection relays operated by a protection circuit that is specifically looking for DC at the output of the power amps is pretty common. Speaker fuses can also help.
I'm sure you have heard the old saying, "the output transistors are there to protect the fuses"

Glimmie's HT Page
Being redone - comming soon!

Glimmie is offline  
post #28 of 183 Old 08-07-2014, 06:12 PM
AVS Addicted Member
 
arnyk's Avatar
 
Join Date: Oct 2002
Location: Grosse Pointe Woods, MI
Posts: 14,387
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 762 Post(s)
Liked: 1175
Quote:
Originally Posted by amirm View Post
As was just correctly stated, I am no competent engineer, nor knowledgeable amateur, so forgive me for saying that statement is not correct. A square wave becomes a sine wave only if your filter point is lower than third harmonic and it has sharp enough response to completely filter out the later harmonics.

A 100 Hz square wave will have odd harmonics of 300, 500, 700 Hz, etc, all of which are still in the domain of the woofer. Here is a 100 Hz square wave generated in Audition CC:



Here is the exact signal but now filtered at 4 Khz, using an 18th order ChebyChev low-pass filter:



That is not remotely a sine wave as the math predicts. There is ringing because we took away the harmonics beyond 4 Khz but there are plenty of harmonics below 4K to still let us have a waveform that looks like a square wave and nothing like a pure sine wave.
This is known as the Gibbs Phenomenon, and is often covered in university courses related to Fourier Analysis. A more complete understanding of the relevant math exactly predicts the observed result.

http://en.wikipedia.org/wiki/Gibbs_phenomenon

In fact the apparent ringing is due to the removal of the higher harmonics that would normally add up to produce a regular square wave.

This shows a comparison of the spectral content of the filtered (red) and unfiltered (green) wave:



The red line overlays the green line below 3 KHz. Both spectra are the same below 3 KHz.


As you can see from the thumbnail the 3 Khz harmonic's amplitude is unchanged by the sharp-cutoff filter, but the remaining harmonics are removed. Low pass filtering caused a peaky appearance that may raise concerns among people who are unaware of the effects of filtering square waves, but in fact the filtered wave is less likely to fry tweeters than the unfiltered wave, despite the peaky appearance.
Attached Thumbnails
Click image for larger version

Name:	amirs bogus filtered clipped wave.png
Views:	264
Size:	41.1 KB
ID:	202017  
Garidy likes this.

Last edited by arnyk; 08-08-2014 at 03:10 AM.
arnyk is offline  
post #29 of 183 Old 08-07-2014, 06:17 PM
AVS Special Member
 
CharlesJ's Avatar
 
Join Date: Jan 2006
Posts: 3,409
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 189 Post(s)
Liked: 117
Quote:
Originally Posted by Garidy View Post
[B][ ...

RMS Wattage of an AC signal @ 90 & 270 Degrees = (Peak-to-Peak x .707) x Impedance (nominal 8-Ohms)

Peak-to-Peak clipped yellow 81.5 Volts AC x .707 = 57.62 VAC RMS

(Divide Peak-to-Peak score by 2) x Impedance = non-clipped RMSpower = 28.81 VAC x 8-Ohms (Nom.)

RMS Wattage of a non-clipped AC signal equals – 230.48 Watts AC

...
I have no standing but I do have a Q on this example.
What is the peak AC power, not RMS?
CharlesJ is offline  
post #30 of 183 Old 08-07-2014, 07:15 PM
AVS Special Member
 
PrimeTime's Avatar
 
Join Date: Mar 2003
Location: Lower California
Posts: 1,791
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 53 Post(s)
Liked: 47
Midrange-frequency program material driven past clipping will certainly increase the power density in the tweeter frequencies. Tweeter failure can be precluded by passive overcurrent protection devices in the crossover. "Light-bulb" and PTC devices are the most common ones used in Pro/Musician speaker systems. In those applications, live microphones producing feedback are common tweeter-killers.

However, in domestic use, feedback is rarely a concern. Further, most clipping occurs with the subwoofers trying to demolish the house during a Transformers scene. Hence, manufacturers often decline to add a buck to their materials cost.
PrimeTime is offline  
Closed Thread Audio Theory, Setup, and Chat

Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off