High Efficiency Speaker Asylum

Posted by Wayne Parham (M) on July 11, 2002 at 23:31:48

In Reply to: Re: Horn stuff and a post back for Tom posted by tomservo on July 10, 2002 at 19:32:50:

Hello again Tom!

You wrote:

[about the phase change of a driver or system at resonance]

>> non linear normally means a departure from linear operation and is
>> evidenced by generation of harmonics or non proportional change in
>> loudness with a change in power, neither of which are related to
>> being "At resonance".

You are referring to frequency linearity. I'm describing phase, and the fact that it changes rapidly as a system enters resonance.

Phase is not a straight line in a system at resonance. It's non-linear. That's what I'm saying. And I would not consider a system with moving phase - delay - to be considered to be "time corrected."

>> If one looked at a unity impedance curve, one would see the
>> impedance of the crossover network, it has several peaks.

And with each one, there is a change of phase, or delay. It moves, with the peak. It isn't a fixed delay, but one that moves as the peak does.

I asked:

That brings us back to the question, so I ask again - How far are
the midrange orifices from the compression tweeter diaphragm, at
the apex?

>> Do you mean physically, acoustically (with respect to the
>> respective drive signals) or in time?

With this question, I am asking about a physical distance. But in other questions in our discussion, I've asked about the crossover slope. Being a passive crossover, phase is easy to determine. And I've also asked if the resonant frequency of the diaphragms were within the overlap region, because that is significant.

>> Ideally the acoustic phase would be at zero degrees over the
>> entire spectrum, this in real life is only possible below about
>> 2500 Hz to what ever the low cutoff is.

I see. Above 2.5Khz, the tweeter subsystem has unloaded; That is to say it has become inefficient, or has passed mass rolloff.

>> As the mids are not 50% efficient and the compression driver is
>> even less (around 20-25%), there is mass reactance showing in
>> the acoustic phase.

Yes, I understand. And the compensation circuit provides EQ above 2.5Khz.

>> One see's that the acoustic phase is essentially that of the
>> drivers. While not a straight line at zero, it is a lot closer
>> to it than a direct radiators -90 degrees.

I expect it is very good.

>> In the range below about 800 Hz it becomes easy to make the "right"
>> drivers for the ranges below and there one can get 50% and acoustic
>> phase around zero.

Yes. The octave from 300Hz to 600Hz is "friendly" for many devices.

>> Remember we are dealing with phase, one can intercept the woofers
>> response where the phase is a straight line (starting above the
>> low cutoff), at Rmin where the acoustic phase has risen to zero
>> degrees (and at zero has NO change in time or phase with frequency).

Agreed. Until we reach the resonance of a driver, it "behaves." But at or near midrange and tweeter crossover points, we run into resonant behaviour again.

>> By unload, that would imply the acoustic load has changed which it
>> has not, the acoustic load IS essentially constant above the high
>> pass freq.

Perhaps you do not like my choice of words, like my choice of the term "non-linear" for describing phase at resonance. What I am saying when I use the word "unload," is that the system becomes "inefficient." Like an internal combustion engine has a torque curve, a motor/diaphragm/horn has a response curve. In both of these systems, efficiency drops on both ends of the rotational velocity scale. Since little useful work is done outside the boundaries of the system, I describe this as an "unloaded" condition.

>> You keep saying there is an upper limit based on throat size, why
>> do you say that and give me some examples.

OK. I'll use your own example:

>> At 20 kHz, that "impedance matching" is done well before the sound
>> has reached the exit of the driver. 1WL in circumference is .21 inches!!.
>> Even at 5 kHz, the passage way in the compression driver is large
>> enough to fully load the driver. From that view point, one could
>> say that ALL compression drivers do not load any horn above 5K
>> (1 exit.inch driver)

Another example is that of a 4" diaphragm driving a 2" hole without a phase plug. In this case, efficient horn loading of the system is limited to the vocal frequencies. Above these frequencies, the system becomes inefficient. It is like the horn were not even there, because the scale of the orifice is too large compared to the wavelength. There is a mismatch of scale that prevents acoustic loading.

You wrote: [about offsetting a diaphragm from its horn apex]

>> I would say though that if the dimensions were acoustically larger
>> than they are, a "non ideal" situation exists.

Yes. This is what I was calling a "mal-formed" horn. I believe you called it a "lossy" horn, yes?

>> When the spacing is 1/2 wl, a forward lobe is produced as used in
>> several EAW speakers and patented by Rocha (sp?) While useful for
>> producing directivity, it is too large a spacing to have horn
>> acoustic transformation, somewhere between 1/4 and 1/2 wl spacing
>> wall to wall is the max size limit for horn loading because if it
>> is already "that large" the acoustic loading has already reached
>> the final impedance of air.

Yes. It is as if the horn flare weren't even there.

>> I do have to say you are tenacious and many of the things your
>> concerned about are valid issues.

And I would say that you are persistent and conscientious too, and I'm impressed.

After a few of our dialogs on this section of the thread, I became more and more impressed with you. Honestly, even though the device cannot have "perfect phase" throughout its intended range, I am certain that it has good phase performance, and that has merit.

Thank you very much for your time and I hope one day we will meet.

Sincerely,

Wayne Parham
π Speakers