I think the OP may acknowledge this, but his concern (if I understand) was not IF there is damping due to the motor's function but that all drivers do not have identical damping or control. So, I think the OP's idea is/was that the closer to having nil "overshoot" beyond the exciting signal the better the speaker - the goal being to eliminate or systematically limit the amount of "extra" energy produced by any given speaker or driver...
I would like to point to the Hill Plasmatronics as a possible exemplar (at least in principle?).
_-_-bear
One little nagging point: acceleration is ANY change in velocity. "Braking" is acceleration. I know that folks generally take "acceleration" to mean an increase in velocity and "braking" to mean a decrease in velocity. Still, if everyone is interested in technical accuracy, this should be kept in mind.
Driver "braking" is done electrically. Think of an electrodynamic driver as transformers that convert electrical energy to mechanical energy (between VC/magnets and cone/suspension), and mechanical energy to acoustic energy (between cone/suspension and air). There are a bunch of energy storage elements in there that give rise to the observed resonances and impedance (electrical, mechanical and acoustic). When the output of an amplifier is zero (while turned on), think of both output terminals as being shorted to ground. Whatever mechanical energy the moving cone & suspension have is converted to electrical energy (transformers work both ways), and that electrical energy is presented with a very low impedance path to ground. So, mechanical energy (kinetic energy of the moving mass + energy stored in the lossy spring-like suspension) gets converted to electrical energy, which is shunted off to ground right quick.
Driver "braking" is done electrically. Think of an electrodynamic driver as transformers that convert electrical energy to mechanical energy (between VC/magnets and cone/suspension), and mechanical energy to acoustic energy (between cone/suspension and air). There are a bunch of energy storage elements in there that give rise to the observed resonances and impedance (electrical, mechanical and acoustic). When the output of an amplifier is zero (while turned on), think of both output terminals as being shorted to ground. Whatever mechanical energy the moving cone & suspension have is converted to electrical energy (transformers work both ways), and that electrical energy is presented with a very low impedance path to ground. So, mechanical energy (kinetic energy of the moving mass + energy stored in the lossy spring-like suspension) gets converted to electrical energy, which is shunted off to ground right quick.
Yeah, and when you stop something suddenly, what happens to "it"?
Besides the electrical braking is far less than ideal or perfect, no matter what gets shunted to "ground"...
_-_-bear
Besides the electrical braking is far less than ideal or perfect, no matter what gets shunted to "ground"...
_-_-bear
Hi Soundminded,
Interesting point. The voice coil is at the center of the cone and imparts all of its force on the cone in the center, but the restorative force is spread out over a larger area. I guess any distortion reduction will be at its highest when the driver is critically damped?
Is the reduced distortion actually measurable? (e.g. same driver in different cabinets... or same cabinet with different damping factors?)
Yes it can be measured but I don't know of any data comparing them. The change in damping factor will affect the system F3 and Q. FR distortion is linear distortion. Cone breakup results in non linear harmonic distortion. The cone flexes and breaks up into modes like a drumhead. Because the 2nd harmonic often predominates this used to be called (and maybe still is) doubling. So in old speaker reviews you'd see something like "output below that frequency was mostly doubling." Why would the cone lose its rigidity, its ability to move as a unit. Basically it's material failure, the differential force distorting it overcomes its relative rigidity. I think today better materials and more uniform control over manufacturing may have reduced that to a degree. AR's 12" paper cone woofer produced about 5% THD when driven at 30 hz. Still a pretty good figure even by todays standards, unique during its era 55 years ago.
A dome is a very strong shape for a material of a given mass and size. Dome tweeters even made of paper don't break up as easily as comparable cone tweeters. The tweeter has requirements that oppose each other, the need to be strong suggesting higher mass and the need to respond quickly by having low inertia hence low mass. The dome shape also seems to produce wider dispersion, an asset to early designers of acoustic suspension speakers like Villchur and Allison but not viewed as desirable today by most designers.
Though the conversation is about the physical, I find it useful to look at the electrical side. You rarely see things plotted this way anymore, but here's a response plot with reactance and resistance. Frequency is marked at various points around the plot. Remember that a reactance can't dissipate any power; all power must be dissipated in the resistive term, the x axis. Most will go to heat in the voice coil, with a small amount radiated and going to heat elsewhere. Negative reactance is capacitive, positive reactance is inductive. Over most of the range this driver is inductive, as expected, but near resonance it appears capacitive. The plot is from an old GR manual, but you'll see something similar with almost any driver. Just another viewpoint that might be useful.
Attachments
"Doubling" has nothing to do with breakup modes or cone characteristics. Significant LF distortion happens well below the frequency of the first modes while the cone is still moving as a piston. High 2nd harmonic will come from the usual nonlinear elements of B field drop off with excursion and nonlinear compliance (with B field being the greater factor).
Regarding cabinet type impacting breakup mode, whether sealed, vented or open baffle the LF air pressure is constant across the cone surface and has no impact on breakup modes. The mechanical impedance of a cone is quite high relative to air load anyhow.
There seems to be a lot of confusion about air pressure within a vented system. It has been shown that the far field pressure can be measured by measuring pressure in the box and then using a 12dB per Octave correction. If the far field response of a vented system is a flat 4th order highpass, then the double differentiated internal pressure will be the same i.e. pressure on the cone of a vented system is comparable to that in a similar sealed system. Of course the wavelengths we are talking about are quite long and there will be no pressure differences across the surface of the cone.
Regards,
David S.
The old "I believe it but can't prove it" argument. 😉
Cone breakup has NOTHING to do with doubling in the example you cite. The LF doubling is caused by asymmetrical motor nonlinearities, not cone breakup.
A figure completely meaningless without drive level.
So what does a phase plug do? 😉
This takes me back to my school days.
Pole?zero plot - Wikipedia, the free encyclopedia
https://ccrma.stanford.edu/~jos/fp/Pole_Zero_Analysis.html
pole zero diagram - Google Search
You really don't see it presented this way very much anymore. I prefer to look at the system as two seperate entities, the mechanically resonant system and the electrically resonant system. The electrical system is of course dependent on how the mechanical system is configured, the loading on loudspeakers, paritcularly the woofer altering its electrical resonance characterisics. Of the two systems, the mechanical system is the far more difficult to deal with. That system should be optimized first. The AS design makes it easy because all three elements are under the designer's direct control. Also both k and b are relatively frequency independent while in boxes with openings whether ports, horns, convoluted transmission lines they are not. The electrical problem yields easily to classical filter theory and whether actiive or passive the filter can be configured to any desired end result. Personally I prefer the active design, it deals with low level signals that are far easier and faster to control if the goal is precision. Only LF blocking capacitors for protecting the midrange and tweeters from transient thumps is necessary at the passive (amplifier output) end.
The scope of the problem as discussed here is very limited, only to the way the speaker/enclosure mechanism works. The wider problem is the nature of the acoustic field the speaker must produce in a given room, the design accomodating different acoustic properties of different rooms to yield substantially similar results. You never see this in commercial designs, equalization and driver levels controls entirely unsatisfactory in themselves in conventional configurations because they do not address the propagated direct and reverberant field independently.
Then there is the even wider problem of overall system transfer function which varies considerably from recording to recording, the transfer function including both recording and playback systems. This explains why high fidelity is never achieved. Anyone with the skill to solve this problem doesn't find it challenging enough or important enough to make it his profession. That's why for me it's only a hobby and why I would never buy what's called "high end" equipment, certainly nowhere near its asking price.
Pole?zero plot - Wikipedia, the free encyclopedia
https://ccrma.stanford.edu/~jos/fp/Pole_Zero_Analysis.html
pole zero diagram - Google Search
You really don't see it presented this way very much anymore. I prefer to look at the system as two seperate entities, the mechanically resonant system and the electrically resonant system. The electrical system is of course dependent on how the mechanical system is configured, the loading on loudspeakers, paritcularly the woofer altering its electrical resonance characterisics. Of the two systems, the mechanical system is the far more difficult to deal with. That system should be optimized first. The AS design makes it easy because all three elements are under the designer's direct control. Also both k and b are relatively frequency independent while in boxes with openings whether ports, horns, convoluted transmission lines they are not. The electrical problem yields easily to classical filter theory and whether actiive or passive the filter can be configured to any desired end result. Personally I prefer the active design, it deals with low level signals that are far easier and faster to control if the goal is precision. Only LF blocking capacitors for protecting the midrange and tweeters from transient thumps is necessary at the passive (amplifier output) end.
The scope of the problem as discussed here is very limited, only to the way the speaker/enclosure mechanism works. The wider problem is the nature of the acoustic field the speaker must produce in a given room, the design accomodating different acoustic properties of different rooms to yield substantially similar results. You never see this in commercial designs, equalization and driver levels controls entirely unsatisfactory in themselves in conventional configurations because they do not address the propagated direct and reverberant field independently.
Then there is the even wider problem of overall system transfer function which varies considerably from recording to recording, the transfer function including both recording and playback systems. This explains why high fidelity is never achieved. Anyone with the skill to solve this problem doesn't find it challenging enough or important enough to make it his profession. That's why for me it's only a hobby and why I would never buy what's called "high end" equipment, certainly nowhere near its asking price.
Perhaps a nitpick, but non-linear distortion is not inherent in cone breakup (at least conceptually) as many people seem to imply when discussing it.
The amount of distortion produced from cone breakup depends on the linearity of the stress/strain curve of the cone material. If the stress/strain curve were perfectly linear within the operating range then the cone breakup would still introduce standing waves (if imperfectly terminated) and therefore frequency response errors, but it would not introduce non-linear distortion.
Of course all practical materials speaker cones are made from have some degree of non-linearity in their stress/strain curves, therefore some amount of distortion will be produced, but it will vary greatly from one cone material to another.
Some materials such as some metals may be more linear for small stress levels (thus produce lower distortion initially) but then become rapidly non-linear past a certain point, hence the somewhat abrupt and harsh breakup of such materials past a certain drive level.
On the other hand paper may have more initial non-linearity for small signals but becomes increasingly non-linear with increased drive at a much slower and more progressive rate, hence a more "progressive breakup" characteristic without an abrupt transition.
Every material will have its own distortion characteristics in breakup, but its important to realise that the distortion is not inherent in breakup, only inherent in the non-linearity of imperfect materials.
Do you have any references to back that up ?
Its a common fallacy that a convex dome provides improved dispersion over a concave cone - it doesn't, it's actually worse.
For a given radiating diaphragm diameter a concave cone gives the best high frequency dispersion, followed by a flat piston, with a convex dome being worst.
The reason for the curvature is strength, as a soft dome which was flat would be impractically weak. For titanium dome tweeters the material itself is strong enough to be usable in a near-flat profile, so many titanium domes are almost flat - with the result that high frequency dispersion is improved over an equal diameter convex curved dome, not worsened.
The real reason that a dome tweeter has improved dispersion over a cone tweeter is simply that it's much smaller. How many cone tweeters have you seen with a 25mm diameter cone ? Not many.
It would be impractical to make a 25mm cone tweeter with a surround, spider, and tiny voice coil at the middle. Too complicated and delicate, especially the voice coil.
A dome design allows you to use a much bigger voice coil for the same radiating area, but the drawback is that the simple suspension can't control rocking modes, and the concave curvature worsens dispersion, but the net dispersion is still better because the diameter is so much smaller than a practical cone tweeter with comparable power handling and sensitivity.
Last edited:
All other things being equal (and they never are) I'd guess it's harder to sell concave tweeters than convex. Just a matter of what people are used to. I think my old Genesis speakers used a concave, and they worked very well.
+1 on looking at mechanical and electrical as two separate systems, but I'd add that it's almost impossible for the local mechanical issues not to show up in the electrical measurements. The sound field is a whole 'nuther problem that's beyond my pea sized brain.
+1 on looking at mechanical and electrical as two separate systems, but I'd add that it's almost impossible for the local mechanical issues not to show up in the electrical measurements. The sound field is a whole 'nuther problem that's beyond my pea sized brain.
In common conversations, acceleration is faster and faster (more m/s every second), and deceleration slower and slower (less m/s every second).
However, physicists tend to use only acceleration, but affected to a plus or minus sign. Deceleration is then nothing else than an acceleration having a negative value.
Frictions have a breaking action on a moving objet but, despite of them, this object can continue to postively accelerate if submitted to a sufficient force. If these frictions diminish, the positive acceleration will be greater.
This is why I think that the description of breaking as having a decelerating effect in the context of loudspeakers can lead to misconceptions.
The Genesis tweeter was supposed to be a good performer.
Concave domes are more difficult in terms of magnet structure in that you need to cut into the core pole for relief from the dome. A convex dome has a natural real volume over a square corepole. All things being equal you will get a lower resonance with a convex dome.
You can have a hollow core pole and lots of volume behind, but the remaining material still needs to angle back to clear the dome, or a longer voice coil will be needed. You will run into dangers of pole saturation if you shrink the dimensions too much.
Small details, but important ones.
David S.
Flat 4th order? Is that one that follows the roll-off precisely, or something else?
"A dome is a very strong shape for a material of a given mass and size. Dome tweeters even made of paper don't break up as easily as comparable cone tweeters."
"Do you have any references to back that up ?"
Dome - Wikipedia, the free encyclopedia
"A dome can be thought of as an arch which has been rotated around its central vertical axis. Thus domes, like arches, have a great deal of structural strength when properly built "
I really have no patience to constantly find references to defend the most basic of concepts.
I was just thinking in terms of a 4th order Butterworth as being most comparable to a 2nd order Butterworth sealed box. The point is no pressure null at V box.
By the way, 4th order Butterworth is not a Q of .707 (or even two of those).
David S.
I'm afraid thats what happens when you attempt to educate us, the great unwashed.
David S.
I'm sorry. I lost interest in this thread almost immediately. But this caught my interest. First of all the pressure forces in an AS speaker are significant only at low frequency around the box resonance where breakup is not an issue. At higher frequencies, where breakup is an issue, the pressure forces are squat. Furthermore, at theses frequencies the wave lengths are typically shorter than the cone dimensions and the pressure in the box is not uniform nor would the pressure distribution over the cone be uniform. In fact, breakup itself ensures that the pressure on the cone in not uniform. And yes, breakup is largely a linear phenomena. Lastly, in any AS speaker there are a couple of distortion mechanisms that are present at low frequency. One is the nonlinearity of the air compliance. The other is the change in suspension compliance due to the pressure forces acting on the surround. At low frequency, as the cone moves in the press in the box increases pushing out on the surround with an effective stiffening. As the cone moves out the pressure decreases "sucking" the surround inward, effectively reducing stiffness. This adds to the nonlinearity of the suspension already present.
- Status
- Not open for further replies.