Here is some more from Mr. Hansen as he answers another poster's questions and comments related to Hansen's original post above:
Poster: So you believe that designers added those "Helmholz resonators" purposely to degrade tweeter performance?
Hansen: Of course not. But they designed by measuring and not listening.
Adding the Helmholtz resonator makes the *steady-state* frequency response flatter in the top octave, but at the same time audibly degrades the transient response. It's a *resonator*, so it does its job and resonates. But the human ear-brain is *much* more sensitive to time-smear from the resonance than a small rolloff in the top octave. And even that is easily corrected in the crossover during the design phase.
The "phase plug" is also an easy way to protect the relatively fragile metal dome.
Poster: The real purpose was to block the out-of-phase wave emanating from the center of the dome.:
Hansen: This is not true.
A typical 1" metal dome will move as a piston up to somewhere past 25 kHz. The flatter the profile the lower the break-up point, and the higher profiles give a higher break-up point. For example, the metal version of the Scan-Speak Revelator has very high profile that moves the break-up point well beyond 30 kHz.
When the dome "breaks up", the output will exhibit a peak , typically around 10 dB or so. This *cannot* be blocked. You can see this peak in the response of any metal dome tweeter, with or without a "phase plug". Just browse the archives on the Stereophile website.
The problem is that higher profiles create a greater path length difference from the periphery of the dome to the listener and the center of the dome to the listener. At high frequencies this path length difference will cause a partial cancellation in the output. The frequency response will fall off (droop) at high frequencies. (In theory there would be a deep notch when the path length difference reached a half wavelength, but in practice this doesn't happen because the dome is no longer pistonic and all bets are off.)
Poster: Soft domes are less prone to this problem due to internal damping, however, they still radiate out-of-phase from the center of the dome.
Hansen: Sort of.
A typical 1" soft dome breaks up at 5 kHz or so. That means that the top two octaves are being reproduced by a chaotic membrane that is flexing and resonating and adding colorations to the input signal.
Poster: The most radical solution to this problem is to exclude dome center completely by adding mass to it. This is how Scan Speak ring radiator works.
Hansen: No, they don't add mass. They fix the center of the radiating surface. The only part that moves is essentially an oversized surround working in a bending mode (similar to the bending mode of the MBL drivers). It is unclear that this will solve the breakup problems of soft dome tweeters. I suspect not, but have never measured one to see.
Poster: Removing front cup from a metal dome tweeter will sure cause ragged frequency response.
Hansen: Again, this is not true. Instead it will cause the output in the top octave (10 kHz to 20 kHz) to "droop" slightly. For a typical dome profile, the "droop" will reach a few dB by 20 kHz. All of this is easily measurable.