[barryt]

The problems is a physics issue.  Your speaker is called a transducer because trans-duct energy.  Electrical to Mechanical to Electrical.  It is a linear electric motor.  It trans-ducts electrical energy into physical displacement of a diagram and vis-a-vis the back other way.
is
The electrical energy that does not become mechanical energy becomes caloric Energy, photons if you like, in the form what we call heat.  When the voice coil can not move, and energy is applied, it is nothing but a Kelvinic Heater.  Unfortunately we have to used flexible glues to glue the voice coil to the speaker 'cone'.
 
When the coil heats up, the glue of softens, frequently the voile coil will shift enough to let the voice coil (now out of alignment) touch or scrape against the walls of the magnetic gap it is in.  Then it goes to hell until the solder joints from the wire that is the actual voice coil simply melts, temperature then goes up fast until something in the conduction path melts and we have an open voice coil failure.
 
Electrical energy that did not become sound became heat.
 
So much to that root failure mode.
 
DC offset from a power source simply heats the voice coil because it is not doing any work.  It is just cooking away.
 
When the speaker is working the energy from the source becomes sound and not heat.  It the core can't move it becomes hate and not sound.
 
If you try to move the voice coil faster than inertial will let it move produces heat.
 
Typically audio signal energy is the sum of a group usually related energies.  Any given single audio source is a collection of various 'notes' each of which as a timbre of it's own.  It is typically perceived as an identifiable sound, the pluck of a guitar is an example, they are brought together in an ensemble and will call it music.
 

 
Here is guitar note, you have heard it many times.                   The next two suggest that 'clipping' a sine wave is produced through the addition of energies, in this case harmonically related but of warring phases relationships.  Feed this to a loudspeaker and you will produce heat.
 
So  . . .  this is all a bunch of crap!
 
It is just talk about corner-cases in audio done in smoky back room.
 
You asked a practical question so you are owed a practical answer,  
 
Try this general line of thinking .  Make the first thing in your audio system be the one to run out of headroom.  That is the Tape deck, PC, Cable box, Disc player, I-whatever, phono preamp, tuner, what ever you listen to, make it the first to clip.
 
You may have to pad the signal, you may have to amplify it, but in the end make it the first stage in the process that runs out of headroom and you be cool!             

 

[pinnahertz]

  The problems is a physics issue.  Your speaker is called a transducer because trans-duct energy.  Electrical to Mechanical to Electrical.  It is a linear electric motor.  It trans-ducts electrical energy into physical displacement of a diagram and vis-a-vis the back other way.
is
The electrical energy that does not become mechanical energy becomes caloric Energy, photons if you like, in the form what we call heat.  When the voice coil can not move, and energy is applied, it is nothing but a Kelvinic Heater.  Unfortunately we have to used flexible glues to glue the voice coil to the speaker 'cone'.
 
When the coil heats up, the glue of softens, frequently the voile coil will shift enough to let the voice coil (now out of alignment) touch or scrape against the walls of the magnetic gap it is in.  Then it goes to hell until the solder joints from the wire that is the actual voice coil simply melts, temperature then goes up fast until something in the conduction path melts and we have an open voice coil failure.
 
Electrical energy that did not become sound became heat.
 
So much to that root failure mode.

We agree on that then.

DC offset from a power source simply heats the voice coil because it is not doing any work.  It is just cooking away.

A DC offset is a failure mode, it doesn't occur in audio, clipped or not. But a DC offset does do work, it applies steady opposing force resulting from two similarly polarized magnetic fields. Kind of a side point, though, since there is no DC in audio.

When the speaker is working the energy from the source becomes sound and not heat.

Yes, but very, very little energy becomes sound. It's mostly converted to heat. Speakers are inefficient motors/transducers.

It the core can't move it becomes hate and not sound.  
 
If you try to move the voice coil faster than inertial will let it move produces heat.

This is an extremely momentary situation in a clipped waveform. No speaker is so perfectly mechanically damped that it is not moving and held in position. If you look at the amount of time, duty cycle, of that moment you find a massive cooling cycle around it.

Typically audio signal energy is the sum of a group usually related energies.  Any given single audio source is a collection of various 'notes' each of which as a timbre of it's own.  It is typically perceived as an identifiable sound, the pluck of a guitar is an example, they are brought together in an ensemble and will call it music.
 
 
 
Here is guitar note, you have heard it many times.                   The next two suggest that 'clipping' a sine wave is produced through the addition of energies, in this case harmonically related but of warring phases relationships.  Feed this to a loudspeaker and you will produce heat.

Agreed, but that's not an explanation. You have not explained why a clipped wave with less RMS energy in it than a non clipped wave of the same base amplitude would produce more heat.

So  . . .  this is all a bunch of crap!
 
It is just talk about corner-cases in audio done in smoky back room.
 
You asked a practical question so you are owed a practical answer,  

Thanks. It will be nice when we get that.

Try this general line of thinking .  Make the first thing in your audio system be the one to run out of headroom.  That is the Tape deck, PC, Cable box, Disc player, I-whatever, phono preamp, tuner, what ever you listen to, make it the first to clip.
 
You may have to pad the signal, you may have to amplify it, but in the end make it the first stage in the process that runs out of headroom and you be cool!             

Ok, now if you're going down that road, please explain the difference between the heat produced by a 100W amp driven by a clipping source device that results in the signal being hard clipped at the 25W level, and a 25W amp being driven by a clean signal, but being driven into clipping by the same amount.  The output waveform to the driver is identical in both cases. 
 
And, if you don't mind, a bit less condescension would be appreciated.  "Your speaker is called a transducer because trans-duct energy. " " ...they are brought together in an ensemble and will call it music."    Really not required to communicate a technical concept.  

 

[barryt]

Ahh, now I understand.  I am bit slow being new to this forum stuff, very interesting for your are apparently this forum's, and I will admit in my nairvety,  my first experience with what is:
 
 
In Internet slang, a troll (/ˈtroʊl/, /ˈtrɒl/) is a person who sows discord on the Internet by starting arguments or upsetting people, by posting inflammatory,^[1] extraneous, or off-topic messages in an online community (such as a newsgroup, forum, chat room, or blog) with the intent of provoking readers into an emotional response^[2] or of otherwise disrupting normal, on-topic discussion,^[3] often for the troll's amusement.
 
Thank your for the education, it was novel, enjoy your domain, good by.

 

[pinnahertz]

  Ahh, now I understand.  I am bit slow being new to this forum stuff, very interesting for your are apparently this forum's, and I will admit in my nairvety,  my first experience with what is:
 
 
In Internet slang, a troll (/ˈtroʊl/, /ˈtrɒl/) is a person who sows discord on the Internet by starting arguments or upsetting people, by posting inflammatory,^[1] extraneous, or off-topic messages in an online community (such as a newsgroup, forum, chat room, or blog) with the intent of provoking readers into an emotional response^[2] or of otherwise disrupting normal, on-topic discussion,^[3] often for the troll's amusement.
 
Thank your for the education, it was novel, enjoy your domain, good by.

No, I'm not trolling you.  I've spent quite a bit of time and effort to analyze the clipping/speaker damage issue. I've simulated, tested, and acquired data which I then have analyzed and presented it.  
 
You've claimed that I'm wrong, but you haven't substantiated your claims, all you've done is restate theory, some of which is incorrect.  Your basis for these claims is your vast years of experience, and your view that I'm apparently naive.   None of that proves your theories or substantiates your claims.  
 
Now, I'm being very fair here.  I really would like a good answer to this question:
 
Why would an amp driven into clipping blow a speaker faster than an amp not driven into clipping, but operating a the same signal level?  
 
That question was the basis for my investigation.  You've refuted my findings, but have supplied no proof to the contrary, only a spoon-full of condescension.  
 
Until good analysis and proof is posted, I would postulate that the definition you've provided above may apply to yourself.  I find your posts insulting, lacking new information, inflamatory, and insulting. 
 
But I can overlook all of that.  You have the opportunity to inform and educate many more than just me.  I welcome that, if there is indeed something to learn.   If not, then I completely understand your withdrawal.  

 

[jagwap]

  The problems is a physics issue.  Your speaker is called a transducer because trans-duct energy.  Electrical to Mechanical to Electrical.  It is a linear electric motor.  It trans-ducts electrical energy into physical displacement of a diagram and vis-a-vis the back other way.
is
The electrical energy that does not become mechanical energy becomes caloric Energy, photons if you like, in the form what we call heat.  When the voice coil can not move, and energy is applied, it is nothing but a Kelvinic Heater.  Unfortunately we have to used flexible glues to glue the voice coil to the speaker 'cone'.
 
When the coil heats up, the glue of softens, frequently the voile coil will shift enough to let the voice coil (now out of alignment) touch or scrape against the walls of the magnetic gap it is in.  Then it goes to hell until the solder joints from the wire that is the actual voice coil simply melts, temperature then goes up fast until something in the conduction path melts and we have an open voice coil failure.
 
Electrical energy that did not become sound became heat.
 
So much to that root failure mode.
 
DC offset from a power source simply heats the voice coil because it is not doing any work.  It is just cooking away.
 
When the speaker is working the energy from the source becomes sound and not heat.  It the core can't move it becomes hate and not sound.
 
If you try to move the voice coil faster than inertial will let it move produces heat.
 
Typically audio signal energy is the sum of a group usually related energies.  Any given single audio source is a collection of various 'notes' each of which as a timbre of it's own.  It is typically perceived as an identifiable sound, the pluck of a guitar is an example, they are brought together in an ensemble and will call it music.
 

 
Here is guitar note, you have heard it many times.                   The next two suggest that 'clipping' a sine wave is produced through the addition of energies, in this case harmonically related but of warring phases relationships.  Feed this to a loudspeaker and you will produce heat.
 
So  . . .  this is all a bunch of crap!
 
It is just talk about corner-cases in audio done in smoky back room.
 
You asked a practical question so you are owed a practical answer,  
 
Try this general line of thinking .  Make the first thing in your audio system be the one to run out of headroom.  That is the Tape deck, PC, Cable box, Disc player, I-whatever, phono preamp, tuner, what ever you listen to, make it the first to clip.
 
You may have to pad the signal, you may have to amplify it, but in the end make it the first stage in the process that runs out of headroom and you be cool!             

 
This is generally true, particularly for dynamic headphones without crossovers, so agreed, see below.

  No, I'm not trolling you.  I've spent quite a bit of time and effort to analyze the clipping/speaker damage issue. I've simulated, tested, and acquired data which I then have analyzed and presented it.  
 
You've claimed that I'm wrong, but you haven't substantiated your claims, all you've done is restate theory, some of which is incorrect.  Your basis for these claims is your vast years of experience, and your view that I'm apparently naive.   None of that proves your theories or substantiates your claims.  
 
Now, I'm being very fair here.  I really would like a good answer to this question:
 
Why would an amp driven into clipping blow a speaker faster than an amp not driven into clipping, but operating a the same signal level?  
 
That question was the basis for my investigation.  You've refuted my findings, but have supplied no proof to the contrary, only a spoon-full of condescension.  
 
Until good analysis and proof is posted, I would postulate that the definition you've provided above may apply to yourself.  I find your posts insulting, lacking new information, inflamatory, and insulting. 
 
But I can overlook all of that.  You have the opportunity to inform and educate many more than just me.  I welcome that, if there is indeed something to learn.   If not, then I completely understand your withdrawal.  

 
 
You are both right, but not equally.
 
Clipping a large lower frequency signal in a speaker does send more energy to drive units through the cross-over in the form of odd order harmonics, which can put them over the edge, as they almost always have a fraction of the power handling.  However with a music signal, most of the excess power comes from the spectrum of that drive simply being too loud for it to handle.  Usually much larger power than the clipping harmonics.  Each system varies a lot, so it cannot be disputed entirely.  However as you generally do not take peoples word for it, Doug Self did a bunch of work on this, and he concluded, it is not the clipping which causes most of the damage.  Maybe you believe him?  He generally never states anything without proof.  
 
However as stated above, when we are talking about single dynamic drivers of headphones, it is not a major contribution to the problem.  I do not have power handling figure for IEMs with crossovers, so here it may be a minor issue again.

 

[pinnahertz]

   
This is generally true, particularly for dynamic headphones without crossovers, so agreed, see below.
 
 
You are both right, but not equally.
 
Clipping a large lower frequency signal in a speaker does send more energy to drive units through the cross-over in the form of odd order harmonics, which can put them over the edge, as they almost always have a fraction of the power handling.

"Drive units" being high frequency drivers, correct?

 However with a music signal, most of the excess power comes from the spectrum of that drive simply being too loud for it to handle.  Usually much larger power than the clipping harmonics.

Yes, that's what my tests showed. Sure, clipping generates harmonics, but they are pretty much always below musical content, except for some special cases. My second spectrum graph shows exactly that (the one with the black background). Even well into clipping the HF content didn't change.

Each system varies a lot, so it cannot be disputed entirely.

If you drive a sine wave past clipping by 10dB, relative to the fundamental the resulting third harmonic is a -8dB, the 5th is at -16dB, and so on. If the amp clipped at 100W, -8dB is just under 16W, and well within power handling lf typical HF drivers. 10dB past clipping sounds absolutely horrible and wouldn't be tolerated very long. I therefore conclude that harmonics resulting from clipping do not contribute significantly to the HF spectrum. And again, that's what my graphs show.

However as you generally do not take peoples word for it, Doug Self did a bunch of work on this, and he concluded, it is not the clipping which causes most of the damage.  Maybe you believe him?  He generally never states anything without proof.    

Got the link handy? That's one I'd like to read.

 

[jagwap]

"Drive units" being high frequency drivers, correct?
Yes, that's what my tests showed. Sure, clipping generates harmonics, but they are pretty much always below musical content, except for some special cases. My second spectrum graph shows exactly that (the one with the black background). Even well into clipping the HF content didn't change.
If you drive a sine wave past clipping by 10dB, relative to the fundamental the resulting third harmonic is a -8dB, the 5th is at -16dB, and so on. If the amp clipped at 100W, -8dB is just under 16W, and well within power handling lf typical HF drivers. 10dB past clipping sounds absolutely horrible and wouldn't be tolerated very long. I therefore conclude that harmonics resulting from clipping do not contribute significantly to the HF spectrum. And again, that's what my graphs show.
Got the link handy? That's one I'd like to read.

I don't. It was verbal as I worked with him a couple of times. I'm not sure if he wrote it up now I recollect. But it is a common misconception in the industry.

We would discuss all, and it took over a year for him to take my listening tests as useful seriously, even if he didn't like the concept.

But he threw a great fireworks night party...