impedance for 99

[Jory A. Olson]

*Yes, never go into a mechanical engineer's bar alone. Other than that, it not only sounds plausible, I imagine it has been done. We need to go the other way though, and anticipate results from design, rather than measure after the fact.
Jory:  I've never been to an Mechanical Engineer bar.  Is that the same thing as a machine shop? :)   I think the ability to measure the impedance of a soundboard before installing it might be very helpful, if for no other reason that to verify that the modeling works properly.
*It's directional because of the cone isn't it? A flat diaphragm wouldn't be nearly as much so, no?
Jory:  Yes, the directionality has to do with the cone, but also because of the cone breaking up at higher frequencies with different mode appearing on the cone.  Until recently a true flat diaphragm speaker wasn't available.  Electrostatic and Magapan's tried to simulated a flat diaphragm with mixed results.
Higher frequencies are more directional.  Lower frequencies are less directional. Two people sitting in the same room, one on axis and the other off axis, can have completely different impressions of the sound quality because they are in fact hearing different things.  
*True, but a soundboard is also being driven in a pistonic fashion at the driven point, like a speaker cone. Also, a speaker cone is not absolutely rigid, and totally without inertia. There should be a bending wave traveling up the speaker cone too, somewhat like the soundboard, with different modal patterns appearing on that cone at different frequencies. Isn't that what the concentric rings molded into a speaker cone are for? I propose that the higher the driving frequency, the deeper into the cone the modal point forms. This would be a function of the mass and stiffness of the speaker cone (mechanical impedance, the stiffness goes up, and the mass goes down as you get deeper into the cone, stiffness being calibrated by those concentric rings), and account for the tighter directionality of the higher frequencies because they originate closer to the 'focal point' of the cone. That would be the narrow part of the diaphragm, and the frequency response, if not the sound directionality, of a soundboard would correspond very closely with what happens in a speaker cone. My 'sort of' referred more to morphology, driving method, and any peripheral effects resulting from same than in the resultant mechanical phenomena. I actually think they relate pretty well. 
Jory:  For somebody that claims to be a "basement tinker" you have an amazing grasp of the fundamentals.  As you go up in frequency the inertia of the cone (from the fact that it has mass) causes the cone to begin to break up.  That is, it is no longer moving as one unit.  The concentric rings on the cone are an attempt to limit the number of modes that appear on the cone, much like a filter or impedance matching device.  About the only point you left out was the limitations of the linear motor.  At some point the backpressure on the cone, the interia, and so forth overcome the motors ability to control the cone motion.  When this happens distortion increases greatly.  Lower impedance voice coils, bigger magnets on the motor, amplifiers with lower output impedance (read:  big wattage) amplifiers, sophisticated feedback schemes are all attempts to exert more control over the moving cone.  Are you sure you don't have a degree in acoustics from Georgia Tech?
*What does happen to a speaker cone at higher frequencies? 
Jory:  Whole PhD's have been dedicated to this topic.  You had the basics right.
*So it is in a soundboard, for any given frequency.
Jory:  Essentially the NXT speakers are soundboards excited at a single point.
*Ain't technology grand? So how does a single excitation point accommodate a full frequency range in this kind of speaker? The soundboard response is complicated a tad by having (nominally) 88 excitation points scattered all over the place. Since it's unlikely that we can place each string termination at the optimum point on the soundboard for the best response at that particular frequency and string load, FEA or not, we have to accommodate a relatively fixed range of termination placements by modifying the response characteristics of the assembly in any given area. Maybe 'sort of' like tinkering with the stiffness, weight, ribbing (beading, banding, or whatever the concentric stiffeners/springs are called), and angle of a speaker cone. Sorry, couldn't resist.
Jory:  The NXT technology is very new and patented out the wazoo.  I know that the point of excitation is critical to the performance and the bigger the diaphragm the lower the device will go in frequency.  The 4X7 or so model I heard only went down to 400 Hz by itself.  There are some theatre screen size speakers being experimented with.  

I like the stuff you, Doug, and Del are doing.  Maybe you guys should form a startup....

Jory