Oh, so this is what you mean by "linear". I have not ran into the situation where I had to test at different levels, but from some FEA results, it seems quite interesting showing there might be such possibility that lower cone breakup modes exists that I don't see in the measured data. But this is one area which is interesting.gedlee said:
soongsc said:
In the world of physics (and applied mathematics and engineering) "linear" is a technical term that means precisely "this" when applied to systems like loudspeakers. Linearity is important because it is a basic assumption behind so many models and methods of analysis.
The use of "linear" to mean that a graph, such as a frequency response, is straight or even flat (usually on a log-log plot for heaven's sake) is a little irritating because it must eventually lead to this kind of confusion. (But complaining ain't going to fix that so micro-rant over.)
Nearly all materials behave in an elastic (linear) fashion for small stresses and strains (and down to extremely small levels in most cases), and I think that includes everything commonly used to make a loudspeaker. Breakup is usually at high frequency, where as already stated displacements are small compared with the dimensions of the parts. So linearity can be assumed from tiny displacements (far to small to hear) up to a level that will depend on the material, but for any sensible choice should correspond to quite loud sound. Above some level there will be plastic deformation, which is non-linear and irreversible, and will eventually lead to failure of the part. (Plastic behaviour is much more obvious at low frequency, where suspensions show measurable non-linearity for large displacements.)
Ken
http://www.seas.no/index.php?option=com_content&task=view&id=191&Itemid=187
maybe Thympany would create a L(inn)O(lson) variant (without whizzer cone for example) on request as they did with the L16RN S(iegfried)L(inkwitz) ?
----------
kstrain said:
Maybe one should add that the non-linearity of "plastic deformation" does NOT start at a certain level - it starts right away - at least with materials used in speakers and at the forces applying there.
This is important to know as "plastic deformation" does create damping – dumping energy of movement into an increase of entropy (=producing heat
). Whereas "elastic (linear) deformantion" only stores energy that is given back, when movement changes in direction.
The behaviour of "plastic deformation" is a material constant to a wide extent intentionally used in membrane and spider coatings and surrounds or for the membrane material itself.
What makes it questionable and directly links to the "grey out" effect Lynn mentioned is that "plastic deformation" dumps energy NOT necessarily related to the SPEED of movement but rather related to the DISTANCE of movement.
This makes the hell of a difference for speakers.
A SPEED related dampening would be something like provided by a cars shock absorber or the dampening mechanism by shortening the VC by the amp's R-out at Fres
DISTANCE related dampening is basically the same as simple friction (like a misaligned VC rubbing in the gap) causing a change for the decay envelope from exponential to linear.
SPEED related dampening is most efficient at high excursion whereas DISTANCE related dampening is "more efficient" at the end of the tail leaving a random offset to the position of the membrane.
Greetings
Michael
Charles Hansen said:
Charles
Thanks.
I am quite familiar with the NXT approach, but it is not really the same thing as what I am say. I am saying that one wants to randomize the "errors" in a system, not randomize the whole system. Take a car tire for example. If its not round it won't work, but once its round, randomizing the tread pattern is common practice. So I wouldn't want to interfere with a drivers main design, only randomize its unwanted secondary aberations.
The goal and the approach are quite different than NXT. (I read their patent once, there were 168 claims stretching over about 20 pages, not one of them was intelligable.)
Charles Hansen said:
Hi Charles, with above you bring in some confusion again.
Are we in agreement that break up does NOT occur as long as there is pistonic movement.
Are we in agreement that break up is basically what is shown below?
http://youtube.com/watch?v=Zkox6niJ1Wc
Beautiful and well behaved geometric patterns of movement and areas not moveing (where the rice lies) as they are the nulls between areas of different direction of movement.
Are we in agreement that cone break up results in ripples in the FR as a result of
1) comb filtering
2) dampening effects of the membrane material
Why - exactly - do you now state that 1000Hz is a limit for a 15" speakers pistonic movement?
http://www.diyaudio.com/forums/showthread.php?postid=1502948#post1502948
What does all this have to do - in your opinion - whether break up resonances are "suppressed by 40, 60, or even 80 dB" - on not ?
-------------
Lynn, brake up forms the characteristics of many instruments.
Its obvious to everyone to hear that they don't change their characteristic heavily with SPL.
I guess the effect you are describing isn't that directly related to cone brake up but rather loss of control and "linearity" between electrical input and SPL output due to compression
VC leaving the gap gradually (=usual compression) OR power compression effects (=heating the VC making it high restive) I mentioned above would make good candidates for that, no?.
Geetings
Michael
Charles Hansen said:
I missed this part of your post at first. The two paragraphs seem to me to be contradictory? Is it doable in a 15" or isn't it? I don't think that it is and I've never seen a cone that can go above about ka=2 that doesn't show some sign of breakup at much more than "40, 60, or even 80 dB". So I'm confused by your comments here too.
If you refering to my use of a 15" in the larger ESP15? That driver is very well behaved up just over about 1200 Hz - where it looses control - and I cross it over at about 800 Hz. So its not used in a region of cone breakup. This driver has a spider resonance at below 1 kHz, and I complain about that regularly. You can see it as a dip in the response at about 600-700 Hz. Most drivers have this. I could make this go away by randomizing the spider pattern.
OK, let's try and clarify some basic concepts here. I'm surprised that there is so much confusion, as this not rocket science.
Let's start with a table fan. It's a hot day and you want a cool breeze on your face. So you go down to the store and look at different models. They are mostly similar. They will have a motor and a fan blade and a protective grille and a base. Typically the fan blade has 4 blades.
Now some of those different fans may have plastic blades. Some may have metal blades. And of the plastic, some may be polystyrene, some may be polypropylene. And of the metal, some may be steel and some may be aluminum. If you are lucky, you might even find one with wooden blades.
But assuming the same motor and the same fan blade design, do you think that you could tell the difference between a blade made from polystyrene versus one from aluminum? In other words, you close your eyes, turn on the fan, feel the breeze on your face. Wouldn't all the materials produce the same airflow? The answer, or course, is that they will all be the same.
So now we have established the fact that it is possible to move air in such a way that the material used to move the air has no influence on the airflow.
Next we will turn off the motor, remove the guards from the fans, and tap on the blades. Do you think they will create the same noise?
And the answer is equally obvious. No. Each material will sound different because it is resonating and has its own particular sonic signature. And of course this noise is also created by the motion of air.
So now we have established the fact that it is possible to move air is such a way that the material used to move the air has a *large* influence on the air flow. (Yes, I know that "flow" is not the correct word for this particular example, but that is irrelevant to this discussion.)
Yes, that is the definition of pistonic movement.
That's not what I said. I said that there is no 15" speaker in the world that has pistonic motion of the cone all the way up to 1000 Hz.
First of all, I can assure you that the diaphragm of any 15" woofer is *not* pistonic up to 1200 Hz. For a paper cone, a more realistic figure is around 300 Hz.
But let's ignore that for now, because you probably don't believe me. Instead, let's take your assertion at face value.
So if you were using a woofer that had a diaphragm resonance at 1200 Hz and you crossed it over to a tweeter at 800 Hz, and assuming for the sake of argument that you were using an LR4 crossover (about the steepest that is commonly used), then the drive signal to the woofer would be down around 15 dB or so at 1200 Hz.
This means that the signature coloration produced by that 1200 Hz resonance will be reduced about 15 dB below the signal that is produce by the tweeter. If you go back in the literature, I believe it was Barlow of the BBC who studied the audibility of "buried" resonances. And the human ear can quite easily hear the effect of the "buried" resonance in this example.
Of course, the real situation is much worse, as the cone is in break up mode for about the upper 1-1/2 octaves of its operating range.
~~~~~~~~~~
Let's use a different example to make this clearer. Remember the original B&W Nautilus that looked like a giant snail? This was designed by Laurence Dickey. It had a 1" metal dome tweeter and a 2" metal dome upper midrange driver.
Now we are all familiar with the resonant peak created by a 1" metal dome tweeter. For common materials such as aluminum or titanium, the first break up mode will occur at ~25 kHz and create a peak of around 10 dB or so.
Less well known is what happens with a 2" metal dome tweeter. Well, I will tell you. It will have pretty much the exact same behavior, except one octave lower. So there will be a 10 dB peak at around 12 kHz.
Now for the sake of argument, let's say that the B&W Nautilus used LR4 crossovers slopes. Further let's assume that the crossover from the upper midrange to the tweeter was at 3 kHz (a nice convenient number). It is clear to see that the first resonant mode of the 2" upper midrange driver is two octaves above the crossover point. So in this example, the drive level to the 2" dome is reduced -48 dB below the drive level to the tweeter (assuming they are of equal sensitivity).
And now you can see that the signature coloration due to the upper midrange driver is reduced approximately -40 dB below the output of the tweeter. This too, is a "buried" resonance. But in this case the resonance is "buried" much more deeply than in the previous example (that wasn't even a real-world situation). Furthermore, the audibility of a "buried" resonance at 12 kHz, is going to be much less than that of one at 1.2 kHz.
~~~~~~~~~~
The bottom line is that 99.9% of all speakers ever made suffer from diaphragm resonances in their operating range that will grossly color the reproduced sound. These resonances are not even "buried" resonances. They are right out there in the open. Since that is the only type of music reproduction that 99.9% of the people have ever heard, this has become accepted as "normal".
There are a handful of loudspeakers (such as the B&W Nautilus) where the diaphragms operate pistonically throughout their operating range. Of course at some point the diaphragms will break up and create resonant colorations. The degree to which these colorations are suppressed depends on the skill of the designer. In the case of the B&W Nautilus, the "buried" resonances are in the neighborhood of -40 dB below the signal level.
With modern materials (eg, beryllium, diamond, et cetera), it would be possible to build a loudspeaker where the buried resonances are more like -60 dB to -80 dB below the signal level. This would allow for an accuracy of sound reproduction that is literally unprecedented.
~~~~~~~~~~
Of course at that point, there are other sources of resonant colorations. For example, Barlow's original work was looking at the resonances of the cabinet walls. So there is no point to get the diaphragm resonances down to -60 dB if the cabinet walls are resonating at only -20 dB below the signal.
All I am saying is that we are a long way from making an accurate reproducer. But the technology to make a huge improvement in accuracy is within reach.
Let's start with a table fan. It's a hot day and you want a cool breeze on your face. So you go down to the store and look at different models. They are mostly similar. They will have a motor and a fan blade and a protective grille and a base. Typically the fan blade has 4 blades.
Now some of those different fans may have plastic blades. Some may have metal blades. And of the plastic, some may be polystyrene, some may be polypropylene. And of the metal, some may be steel and some may be aluminum. If you are lucky, you might even find one with wooden blades.
But assuming the same motor and the same fan blade design, do you think that you could tell the difference between a blade made from polystyrene versus one from aluminum? In other words, you close your eyes, turn on the fan, feel the breeze on your face. Wouldn't all the materials produce the same airflow? The answer, or course, is that they will all be the same.
So now we have established the fact that it is possible to move air in such a way that the material used to move the air has no influence on the airflow.
Next we will turn off the motor, remove the guards from the fans, and tap on the blades. Do you think they will create the same noise?
And the answer is equally obvious. No. Each material will sound different because it is resonating and has its own particular sonic signature. And of course this noise is also created by the motion of air.
So now we have established the fact that it is possible to move air is such a way that the material used to move the air has a *large* influence on the air flow. (Yes, I know that "flow" is not the correct word for this particular example, but that is irrelevant to this discussion.)
mige0 said:
Yes, that is the definition of pistonic movement.
mige0 said:
That's not what I said. I said that there is no 15" speaker in the world that has pistonic motion of the cone all the way up to 1000 Hz.
gedlee said:
First of all, I can assure you that the diaphragm of any 15" woofer is *not* pistonic up to 1200 Hz. For a paper cone, a more realistic figure is around 300 Hz.
But let's ignore that for now, because you probably don't believe me. Instead, let's take your assertion at face value.
So if you were using a woofer that had a diaphragm resonance at 1200 Hz and you crossed it over to a tweeter at 800 Hz, and assuming for the sake of argument that you were using an LR4 crossover (about the steepest that is commonly used), then the drive signal to the woofer would be down around 15 dB or so at 1200 Hz.
This means that the signature coloration produced by that 1200 Hz resonance will be reduced about 15 dB below the signal that is produce by the tweeter. If you go back in the literature, I believe it was Barlow of the BBC who studied the audibility of "buried" resonances. And the human ear can quite easily hear the effect of the "buried" resonance in this example.
Of course, the real situation is much worse, as the cone is in break up mode for about the upper 1-1/2 octaves of its operating range.
~~~~~~~~~~
Let's use a different example to make this clearer. Remember the original B&W Nautilus that looked like a giant snail? This was designed by Laurence Dickey. It had a 1" metal dome tweeter and a 2" metal dome upper midrange driver.
Now we are all familiar with the resonant peak created by a 1" metal dome tweeter. For common materials such as aluminum or titanium, the first break up mode will occur at ~25 kHz and create a peak of around 10 dB or so.
Less well known is what happens with a 2" metal dome tweeter. Well, I will tell you. It will have pretty much the exact same behavior, except one octave lower. So there will be a 10 dB peak at around 12 kHz.
Now for the sake of argument, let's say that the B&W Nautilus used LR4 crossovers slopes. Further let's assume that the crossover from the upper midrange to the tweeter was at 3 kHz (a nice convenient number). It is clear to see that the first resonant mode of the 2" upper midrange driver is two octaves above the crossover point. So in this example, the drive level to the 2" dome is reduced -48 dB below the drive level to the tweeter (assuming they are of equal sensitivity).
And now you can see that the signature coloration due to the upper midrange driver is reduced approximately -40 dB below the output of the tweeter. This too, is a "buried" resonance. But in this case the resonance is "buried" much more deeply than in the previous example (that wasn't even a real-world situation). Furthermore, the audibility of a "buried" resonance at 12 kHz, is going to be much less than that of one at 1.2 kHz.
~~~~~~~~~~
The bottom line is that 99.9% of all speakers ever made suffer from diaphragm resonances in their operating range that will grossly color the reproduced sound. These resonances are not even "buried" resonances. They are right out there in the open. Since that is the only type of music reproduction that 99.9% of the people have ever heard, this has become accepted as "normal".
There are a handful of loudspeakers (such as the B&W Nautilus) where the diaphragms operate pistonically throughout their operating range. Of course at some point the diaphragms will break up and create resonant colorations. The degree to which these colorations are suppressed depends on the skill of the designer. In the case of the B&W Nautilus, the "buried" resonances are in the neighborhood of -40 dB below the signal level.
With modern materials (eg, beryllium, diamond, et cetera), it would be possible to build a loudspeaker where the buried resonances are more like -60 dB to -80 dB below the signal level. This would allow for an accuracy of sound reproduction that is literally unprecedented.
~~~~~~~~~~
Of course at that point, there are other sources of resonant colorations. For example, Barlow's original work was looking at the resonances of the cabinet walls. So there is no point to get the diaphragm resonances down to -60 dB if the cabinet walls are resonating at only -20 dB below the signal.
All I am saying is that we are a long way from making an accurate reproducer. But the technology to make a huge improvement in accuracy is within reach.
Charles Hansen said:
Charles, remember, I already displayed a 12" paper cone some posings ago telling diametral different!
Hence - have a prove for that?
gedlee said:
Earl, I am surprised - don't you consider "plastic deformation" as nonlinear?
Don't you agree that "plastic deformation" is highly involved at break-up? – most obvious with dampened membrane materials like bextrene, polyprop by itself - or heavily coated paper, kevlar etc.
Greetings
Michael
Charles Hansen said:
Charles, well ONE point we both agree to at least.
Starting from that point we can say that there is no break up below a certain frequency – the geometric rice pictures start at a frequency ABOVE that – otherwise we would see – nothing.
In the CSD of several metal cones it is very obvious ("gross errors" in your terms) that the cone break-up is a SINGLE frequency – NOTHING below that – still agree?
Following the line of my argumentation I don't see any evidence to state cone brake up that low you do.
Greetings
Michael
Charles....
B&W aren't my favorite loudspeakers....but since you invoked them I wanted to point out the 350hz crossover point used in the current 801D which uses a 15" woofer. I believe it was 400 in the original N801. I'm sure the engineers at B&W know what the breakups are in their woofer, just as I'm sure you knew what the breakups were in the Eton's for your old speakers.....as long as this is the right Charles Hansen. I'm sure if I dig around I can find higher crossover points for similarly sized woofers in well engineered product.
Best,
B&W aren't my favorite loudspeakers....but since you invoked them I wanted to point out the 350hz crossover point used in the current 801D which uses a 15" woofer. I believe it was 400 in the original N801. I'm sure the engineers at B&W know what the breakups are in their woofer, just as I'm sure you knew what the breakups were in the Eton's for your old speakers.....as long as this is the right Charles Hansen. I'm sure if I dig around I can find higher crossover points for similarly sized woofers in well engineered product.
Best,
Charles Hansen said:
Charles
I'm not going to argue with your point that 99.9% of all loudspeakers have colorations in them because from the extremist point of view you seem to be taking one could almost say 100%. But I think that you over simplfy the problem of coloration by implying that it is only system resonances that cause this. Room reflections cabinet diffractions and a whole array of other phenomina comtribute and in no small way. So clearly the directivity issue enters into the coloration problem because it is so important to the early reflection problem. In any loudspeaker there are tradeoffs and one has to balance the tradeoffs of cone resonances against directivity control. I think that making this tradeoff assesment is where we might differ.
I have not found the smaller resonances in drivers and the enclosure to be nearly as important as the control of the room early reflection situation. This later aspects requires a strict control of directivity which piston sources cannot provide. So am I willing to allow a 15" drivers resonance to be 20 dB below the compression driver in order to get a high degree of control over the directivty - ABSOLUTLY. Would I use a perfect 15" driver without said resonance if it cost the same amount of money - absolutely again. The issue is giving up what is less significant to control that which is more significant. And doing so within a reasonable cost constraint.
I do disgree with you that the B&C 15 has "significant" cone breakup down at 300 Hz. I can't measure it if it does. It does have a spider resonance at about 600-700 Hz which is not desirable, but the cone seems to be well under control up to 1 kHz. And I never said that this 15 was "pistonic" up to 1200 Hz. I said it was "well controlled". A resonance doesn't matter if it does not radiate to be heard. So "control" is sufficient even if NO resonance is not possible.
Our differences all seem to be about degree not about the physics.
mige0 said:
Michael
I think that you might be confused with the term "plastic deformation" because that means that it does not return to its original shape. As such, no, I certainly don't believe that "plastic deformation" is involved with cone breakup.
"Plastic" as used in this context does not mean plastic as a material. Its a poor complication with the english, but it is what it is. See http://en.wikipedia.org/wiki/Deformation and note how far up the deformation curve the stress has to go before it becomes "plastic". If it is going "plastic" then the cone would break and this takes a whole lot of SPL.
gedlee said:
No Earl, no confusion at my side concerning "plastic deformation".
The pix in the wiki point out what I excluded in my postings explicitly.
What I refer to is NOT metal cones here - its all sorts of plastic (material wise) and its NO confusion that almost all types of plastic ain't behave like metal.
With plastic (material wise) there isn't that strict border between an area of elastic and plastic deformation like with metal (some types like fibre glass and kevlar excluded though they need some bond to keep the shape and quite often kind of high dampening coating material is applied).
In fact, plastic deformation starts right away with almost all sorts of plastic (material wise).
This behaviour is the main reason for the – desired – dampening effect of all sorts of plastic membrane materials. Hence the veeeery "gross defects" seen in CSD's of metal cones.
I guess your 30 years of confidence is "in the bin" – sorry for giving you hard times.
Greetings
Michael
ScottG said:
Hi ScottG, don't think brake-up has anything to do with the coupling of the VC.
Its simply the rigidity provided by a material constant and the shape of the membrane that counts in addition to – as far as I know – the speed of sound in that membrane material.
Basically you could compare a cone speaker to a cymbal.
The fundamental of that cymbal is the first brake-up frequency – nothing below that – except swinging at the stand, that gives a beautiful modulation due to directivity
Greetings
Michael