mige0 said:
You apparantly did not read down to the section on plastic deformation where it says
"Plastic deformation -
This type of deformation is not reversible."
And
"Hard thermosetting plastics, ... have minimal plastic deformation ranges. "
No loudspeaker cone is going into irreversible plastic deformation under normal circumstances - it can only do that once. The cone motion is elastic, with a hysteresis effect perhaps, but it is not plastic. The hysteresis is what causes the damping.
mige0 said:
I deleted it because I just *knew* that I'd be spending more time writing.. Ah well, best plans and all that.
Anyway -
Break-up is essentially a subset of non-pistonic operation in a typical transducer. For instance when such a driver is operating in a pistonic manner it is therefor NOT in break-up. Conversely the portion of a driver's operation that is in "break-up" is not operating pistonically. The reason for the break-up..thats a structural and material issue and the structure has a *relationship* to pistonic operation.
gedlee said:
Earl, haven't studied underwater sonics but have studied that kind of stuff once...
What else than "plastic deformation" would you say is hysteresis – your "hysteresis" exactly IS "plastic deformation" !
Actually "plastic deformation" of most sorts of plastic (material wise) is always mixed with "elastic deformation" behaviour AND is highly dependant on temperature, moisture and sometimes provides sort of material-memory.
Greetings
Michael
mige0 said:
This is semantic and not consistant with the terminology as used in the industry. To be "plastic" it has to always "provides sort of material-memory", not "sometimes" or its not "plastic deformation" its elastic. Read the definition. I have never heard of anyone defining hysteresis loss as being "a plastic deformation that returns to its elastic state on removal of the load" as you seem to be implying.
I think the weak aspects in the Nautilus is not just the cone breakup modes, but the residual energy on the diaphragm in whatever frequency range, and especially in the mid frequencies and higher. I would love to get a pair to measure them, but it seems that the owners love them so much that nobody else can touch them.Charles Hansen said:
If we look at the Jordan metal drivers, the breakup characteristics are generally much better than what we will normally see in drivers of similar sizes, but the longer decays in the CSD still show it's coloration. If the overall decay time were further reduced, we will hear less coloration, but individual requencies with weak resonances that extend a bit longer will become more detectable.
mige0 said:
I purchased a Crown TEF analyzer over 20 years ago. It was the first analyzer that could perform the "waterfall" plot. It was invented by and developed with the aid of Richard Heyser, who is one of the all time great audio geniuses. At that time it cost over $12,000, plus you needed a good measurement microphone. I bought a B&K 1/4" that had flat frequency response up to 40 kHz. The mic system (capsule, preamp, power supply was another $4,000 or so).
Even though this was all very expensive gear (especially in 1985 dollars), I thought I was getting off cheap. What I really wanted was a laser interferometer that could make animations of the diaphragm's behavior. (Celestion used them when they developed their copper-dome tweeter for the SL-6.) But a laser interferometer was $50,000 and I couldn't afford one.
I figured with the waterfall plots that I would be able to see the cone resonances that resulted from various break-up modes. But as I posted previously, you can't really. You can see really gross problems. But you could see those kind of problems with a pulse generator or tone-burst generator.
I eventually developed a way to look at diaphragm behavior that didn't require a $50,000 laser interferometer. But I consider that proprietary. Independently, Keith Johnson made a similar invention that was described in Speaker Builder (I believe in 1988), but his method is much more cumbersome to build, use, and interpret.
All I can say is that you can either believe me or not. You can buy a laser interferometer or duplicate Keith's apparatus or not. But whatever I say or whatever you believe, it won't change the behavior of the diaphragm. It will either move pistonically or it will break up.
And there is no paper cone 15" diaphragm that behaves pistonically up to 1000 Hz. As I said, you are lucky to get to 300 Hz pistonically. A few drivers *might* get up to 350 or 400 Hz, but that would be pretty much the limit.
From my point of view I am giving away free knowledge that was earned at great expense. I don't feel that I am under any obligation to give everything away. Sorry if that sound arrogant to you.
gedlee said:
I'm with Earl on this one. Hysteresis is not normally referred to as some subset of "plastic deformation". There is plastic deformation, elastic deformation, and hysteresis. They are all different phenomena, which is why they have different names.
I can see why Mige0 is getting confused, as many plastics exhibit enough hysteresis that it might become difficult to distinguish plastic deformation from elastic deformation. You would potentially have to include an explicit time scale. But it will be easier to communicate if we use the standard definitions and terminology.
As I rethink the discussion of "plastic deformation", I'd have to say now that this is concept does not apply to dynamic loads and hence it is easy to get things mixed up. It would be best not to try and apply a concept that is intended for static loading conditions to a dynamic situation because it is ill defined in that context.
gedlee said:
My old designs were piston sources. While the directivity was not "controlled" per se, it was quite uniform with frequency. This places a bigger burden on the owner to make sure that the listening room's decay time is relatively constant with frequency. Otherwise the power response will diverge from the axial response.
If I were to design a speaker today, I would definitely include techniques to control (ie, narrow) the radiation pattern so as to make the room less of an issue. I think we are on the same page there.
gedlee said:
Yes. And I think you would be surprised to hear what is possible when the diaphragm resonances are removed from the equation.
Re: Charles....
The thing you have to keep in mind is that B&W is a large company with many engineers. And many of them have come and gone over the years.
The only product I was referring to was the original Nautilus that looked like a giant snail. That product was designed by Laurence Dickey. The subsequent "Nautilus 80X" series of speakers are a completely different design that has almost nothing in common with the original Nautilus. Laurence Dickey did not design those products.
And actually, I doubt that the current B&W engineers know what the breakups are in their woofers. If they did, they wouldn't be designing the products they do. On the other hand Laurence Dickey does. Go look at his current work -- Vivid is the brand name. He understands pistonic motion and cone resonances. That is clear from examining his designs.
chrismercurio said:
The thing you have to keep in mind is that B&W is a large company with many engineers. And many of them have come and gone over the years.
The only product I was referring to was the original Nautilus that looked like a giant snail. That product was designed by Laurence Dickey. The subsequent "Nautilus 80X" series of speakers are a completely different design that has almost nothing in common with the original Nautilus. Laurence Dickey did not design those products.
And actually, I doubt that the current B&W engineers know what the breakups are in their woofers. If they did, they wouldn't be designing the products they do. On the other hand Laurence Dickey does. Go look at his current work -- Vivid is the brand name. He understands pistonic motion and cone resonances. That is clear from examining his designs.
Charles Hansen said:
We're down to about -50 to -60 db being easily heard in tweeter crossover impedance compensation networks, or in the network for the top of the woofer's response. Of course, this can be related to many complex issues, so it's not a blanket statement. Just a seeming observation that needs be addressed, in our particular case. It's damn annoying, quality capacitors for compensation networks gets expensive, and fast.
As for woofers with severe peaks in their upper ranges: bad news. Like robbing Peter to pay Paul. Find a better driver. Ignore that one, and they'll finally stop making them.
Charles Hansen said:
That hasn't been my experience. I've found that directivity and early reflection control, along with lowering the system and room diffractions (a form of very early reflections) are the key to low coloration and good imaging. The system resonances come in pretty far down the list although I don't ignore them. People who have heard this are in general agreement here. There is a post at
http://www.diyaudio.com/forums/showthread.php?postid=1507485#post1507485
which describes it.
Charles Hansen said:
Charles,
What is your opinion on using steeper slopes like analog LR4 mid-tweet and LR8 bass-mid, or like 96db steep digital slopes to put speaker buried resonances -60 db to -80 db below signal level? Are steep analog or digital crossover slopes worst than cone break-up?
Residual energy is energy that remains in the driver and is dissipated with time, normally viewed in the form of CSD. When the VC drives the diaphragm, part of energy is transferred into acoustic energy, part is heat, part remains electic, part is mechanical. We hope that energy turns to acoustic as much as possible, or dissipated as heat, this way residual mechanical energy and electric energy does not reflect in a way that it continuously transfers into acoustic energy when we really don't want it to do so. If you look at the CSD, it is most desireable to see the whole spectrum decay as fast as possible. Unfortunately there is a limit. To make things worse, the cone breakup mode stores energy and releases it much slower.Charles Hansen said:
soongsc said:
In your previous post you wrote:
soongsc said:
So I am still unclear what you meant in your previous post. It would seem that you are claiming that there is "residual energy" in the "mid frequencies and higher", but that this "residual energy" is not from break up modes.
1) What leads you to make the claim that there is "residual energy" in the "mid frequencies and higher"?
2) If the "residual energy" is not from the break up modes, what is it from?
LineSource said:
It's a good question, but I don't have a good answer. I haven't designed any speakers for nearly 20 years. Here are a few general comments:
a) The steeper the slope, the worse the transient response. (Actually, the sharper the corner, but these two generally go hand-in-hand.) There are some who claim that anything steeper than first-order creates audible ringing. I have never tried anything beyond LR4. Obviously the cost and complexity increases radically.
b) Digital approaches impose their own set of limitations. If a person *only* uses digital sources, then it becomes less objectionable. But there is no real advantage to digital over analog except lower cost.
c) Joseph Audio used to use Richard Modaferri's "Infinite Slope" crossovers as a way to keep the high-frequency energy out of the metal-coned woofers. I know that they dropped this for at least part of their crossovers, but I don't know if it was for the high pass, the low pass, or both. I know that the change was made because the more traditional approach produced a better overall sound quality than the "Infinite Slope" approach.
If you use good drivers (well better than good) that behave well - as in no nasty peaks or valleys and cross them over well within their optimum range (one to two octaves away from the little nasties) with a good quality 4th order electronic crossover (with variable damping) these out of band nasties become pretty much trivial.
8th order does seem to have a bit of a 'ring' to it though. After a while the tend to become un listenable.
8th order does seem to have a bit of a 'ring' to it though. After a while the tend to become un listenable.