I believe that Birgir is trying to express that basically the motion of the membrane is really small even in comparison to the height of the spacer. Oppositely, an electro-dynamic type transducer is supposed to have a much larger displacement range, in hopefully "piston" fashion. The truth is that any membrane has so-called breakup modes, where the surface no longer moves uniformly. These "modes" are basically the result of interaction of waves that are being reflected off the boundary of the cone.
Think of a wave like the ripples on a calm lake when you through a stone, the wave moves outward in circular shape. When it reaches an edge, it will reflect back. In a finite size structure, you get these waves reflecting off the edges and interacting which leads to the establishment of "modes". Each mode has an associated "resonance frequency", that is if you excited the structure at that frequency and can observe the vibration on the surface, you will see exactly the shape of that mode. I have pasted a link here to illustrate the modes of a circular disk: http://paws.kettering.edu/~drussell/Demos/MembraneCircle/Circle.html
At low frequency, any cone will behave like a piston. At some higher frequency, it will have the first mode (1,0 in the link above), higher order modes will occur as you move up in frequency. Ideally, you want the cone to have no resonances in the audio range and behave effectively like a piston. It is seldom the case so people pay attention to the damping of the material / surround / spider (or the air layer in front / back of the membrane in the case of electrostatic driver) in order to limit the effect of the resonance on the acoustic response (peaks and valleys, change in the phase). Fundamentally, the resonance frequency of a mode is proportional to the bending stiffness of the cone and inversely proportional to its surface mass. So, the stiffer and lighter the better. Accessorily, the lighter the cone, the higher the frequency its response will start to roll-off, which in the time domain will translate into very quick transient response.
For a stax electrostatic transducer, the membrane is very thin and low mass. Its stiffness comes from the tensioning as it's basically a thin sheet of plastic as is... It isn't allowed to move like a piston (what birgir is trying to explain) so basically it's movement is regulated by the frequency at which the first mode (1,0 above) occurs and the tensioning of the membrane. This first mode is the one causing the largest displacement (at the center of the membrane) and I thought it could be an issue (contact with the stator under high SPL or if the spacer is too thin) as referenced in cmoy's explanation. But from what birgir is saying, actually it isn't. I don't have any idea how much displacement you get for say reaching 100dB at the ear for the size and thickness / stiffness of diaphragm.
One thing I mentioned in a previous message is that the advantage of the electrostatic transducer is that the force is applied uniformly on the whole surface. Because of this, ( I believe ) the effects of the resonances (the modes) are much less visible than when they occur on a traditional electrodynamic transducer with the forces applied on the rim of the voice coil.
Nor do I know how much the stat diaphragms move and certainly it is less than a dynamic speaker. However I recall that a number of stat systems use plastic spacers on the stators to minimize arcing, the first I encounter were the old B &W and I thought that Stax might have also done so. This would seem to be done to physically stop the diaphragm from getting too close to the stators, thus implying a fair bit of movement, at least at the center of the diaphragm.
Edited by edstrelow - 4/12/11 at 11:22pm