[rick57]

I've accumulated a few drivers and some leave, that I hope to take soon to make some progress on a backlog of projects.

I was going to build a horn midbass (about 100 - 500 Hz), but belatedly realize I don't have the time for such complexity, for quite some time . .

Above all, I'm looking for whatever gives midbass the tightest, best transient response.
Also likely to move to a tube amp, so speaker damping becomes more critical. I'll develop the rest of the system to be coherent from that starting base.

With sealed aligmnents you can handily predict or "choose" step response: varying box volume varies the damping 'spring'. If using a driver in a sealed box, I'd aim for a Qtc of about 0.55.

Using the same driver in a dipole (I'm not buying any more drivers!), the operating passband moves up a little, which needs to be covered by the sub's range going higher, depending on baffle size; room modes are minimized, etc.
Known differences aside, how does the transient response of dipoles compare with sealed boxes?

Over a certain range, is system Q in a dipole = driver Qts?? Can either the Edge or MJK's worksheet calculate how step response relates to driver parameters or baffle size?

Thanks, Richard

[johnk...]

From the point of transient response the dipole is no different than a sealed box. They are both 2nd order high pass systems and if the equalized response of a dipole follow a Q = x, Fc = Y 2nd order high pass response then it will have the same response characteristics as a sealed box with the same Q and Fc. In a given room it may sound different due to different total radiated power, radiation pattern, excitation of room modes and differences in room pressurization effects, but that is the coloration of the room, not a characteristic of the system.

[rick57]

John

> They are both 2nd order high pass systems and if the equalized response of a dipole follow a  Q = x, Fc = Y  2nd order high pass response, then it will have the same response characteristics as a sealed box with the same Q and Fc.

As we know, for a given driver, the Q, Fc and response of a sealed box are set by its size. . .

If the dipole’s baffle is big enough, so goes low enough - eg 80 Hz, and the resonance of the dipole is in its passband but far enough away (eg 340 Hz), the Q of the XO is not relevant, so the Q then is . . the Qts of the driver?

For a flat baffle for midbass, is this the relationship between baffle size and resonance:

resonance corresponds to the shortest wavelength around the baffle, eg a 1 m (c 3.3 feet) wide baffle, ie 1 m from the front of the driver to the back of the driver, 1 m wavelength ie the resonance would be at 343 Hz?

Thanks

[rick57]

From postings on other forums it seems I had this wrong: there is no relationship between baffle size and resonance?
But is the resonance Hz simply equal to driver resonance?

thanks

[johnk...]

Rick,

The bottom line is that if you measure the on axis response of a dipole woofer or mid, the transient response of that driver will be dictated by the response. If, for example a dipole mid is equalized to have an on axis response of a 4th order band pass filter, it will have the same response as the corresponding 4th order filter. If the woofer response is a band pass with 2nd order, Q = 0.5, HP at 20 Hz and a 44th order 100 Hz LR4 LP then the transient will be that of the equivalent BP filter. If there is a large resonant peak in the pass band then the response TP will be the same as a filter with that resonances.

[rick57]

Thanks John,

I had thought of transient response and frequency response as related, but largely separate.

The relationship is sinking in

Cheers

[gitarretyp]

Thanks John,

I had thought of transient response and frequency response as related, but largely separate.

The relationship is sinking in

Cheers

You might take a look at this article.

[johnk...]

You might take a look at this article.

Or the second one down in the references, http://www.geocities.com/kreskovs/Stored-energy3a.html

Or my more recent revisions at http://www.musicanddesign.com/Stored_energy_1.html,
http://www.musicanddesign.com/Stored_energy_2.html and http://www.musicanddesign.com/Stored_energy_3.html

The thing is that if the system is linear (i.e. nonlinear distortion is relatively small) and time invariant (meaning it's the same five minutes (or 10 seconds) form now as it is now) then there is a one to one relationship between its behavior in the frequency and time domains given by the Fourier transformation and its inverse. If you know either the impulse or frequency response for a linear system you know all there is to know about a linear system. You can determine to output of the system for any input.

[scorpion]

johnk,

This is very interesing. But when is nonlinear distortion relatively small, 5 %, 1 %, 0.1 % ?
Frequency dependent, absolutely and also SPL dependent, the distortion.
Time invariant, yes possibly even five years from now speaking of dynamic elements.
So when does this one to one mapping exists in reality ?

/Erling

[johnk...]

johnk,

This is very interesing. But when is nonlinear distortion relatively small, 5 %, 1 %, 0.1 % ?
Frequency dependent, absolutely and also SPL dependent, the distortion.
Time invariant, yes possibly even five years from now speaking of dynamic elements.
So when does this one to one mapping exists in reality ?

/Erling

Well let me put it this way. The entire idea of designing crossovers; passive or active, digital or analog, is nothing more than an exercise in response shaping of linear components. That is, if drivers didn't behave primarily as linear elements crossover design wouldn't be what it is. Crossover design is just applying linear corrections (filters) to obtain the response we want from a linear system that deviates from the desired target.

You can do some tests and see for yourself how linear a driver is. Just start taking frequency response measurements at different levels. You have to pound on them pretty had before anything significant is observed. Even 10% 2nd order HD yields a distortion component 20 dB below the "linear" level.

I have always balked at the term linear distortion. It's a red herring. It's just another way of saying the frequency response isn't flat. Same thing with stored energy. Any  system that doesn't have perfectly flat amplitude and zero phase response has some form of linear distortion or stored energy. The question is defining when it is acceptable and when it is not.

[scorpion]

johnk,

I much like your crossover comments and agree about where would we be without some trust in good theory and measurement.
And as so very much around us, still dictated by linear theory.

But if pressed on an answer, is there an acceptable niveau of distortion ? Or should we just take any published frequency response curve
as a sure evidence about what this speaker would give us in return however we use it.

/Erling

[rick57]

gitarretyp and johnk...
 
I just saw your posts (and it's pretty late here now) I'll have a good read over the next day or so, but they look very helpful, thank you both.

Cheers