Will a Linkwitz transform...transform the driver?
 

I want to horn load a driver but don't have one on hand that is good to work with. If I apply a Linkwitz transform, can I then design with the driver just as if it had the different parameters that I will be simulating?

You can, though with suspect results.

Better to measure the subject driver, design a suitable enclosure, then apply the transform.

Thanks, Ed., but may I also ask the question another way... Can I take a high Q driver, measure and transform it to a (effectively) low Q driver, then design a box for such a low Q driver (ie, smaller) and will the box then do as I expect it to as if the driver were truly low Q?


I can see that I would probably choose your point of view for some situations but with a horn I'd like to reduce the box size a little if possible.

ESP - The Linkwitz Transform Circuit

it's just an EQ curve which is the inverse of a closed box speakers roll off so allows the speaker to operate with flat frequancy response at lower frequancys. It does nothing to the actual speaker just changes the input to the speaker.

You should probably do it the other way around. Design the box, based on the T/S parameters of the high Qts driver, which will give you a lower system Q (than the driver on its own). Applying a Linkwitz transform will extend the lower end response and, within the pass band, give you a lower system Q (you would need the in-box transfer function in order to do LT. But the LT on its own will result in a shelved boost of all LF signals; It should be implemented with a High-pass filter to reduce over-excursion or overdriving the amp.

And then: LT was intended for use with sealed enclosures. However, it is possible still to apply equalisation (generically speaking) to obtain the Q and final transfer function of your choice.

Note that the nice flat response doesnt come for free,your maximum SPL will be limited. But they may be what you want.

I have SPL to spare, no worries there.

I think I get the whole issue now. The transform alters the response level and phase, but doesn't touch the acoustics.

For example, I could appreciate someone taking a high Q woofer, closed boxing it to Chebyshev, then transforming it back to Bessel.... this will give you a smaller box than you otherwise need, but I don't think the transform will fix the Chebyshev ringing.

I could produce a horn that's intentionally smaller than it need to be and transform the resultant response, but then I'd know that issues like reactance nulling were fudged, and mouth response was abused, etc.

Just wanted my cake and to eat it too :p

I did take a dipole rolloff once and reverse it up to flat with a Linkwitz transform. Now that was the right way to go ;)

No. It is the final system response that matters. If you end up with Chebyshev, yes, you'll have ringing. But if the end result is Bessel, then you have the characteristics of Bessel, etc.

LT is not for modifying the driver
 

The LT is not a "driver parameter modifier" circuit. It modifies the behavior of a driver mounted in a closed box (a second order system). It does this by creating two electrical functions, one that "cancels" the existing response of the driver+CB, and one that generated a "new" response.

Because a horn loaded speaker has a 3rd or 4th order response, and that is not what the LT circuit is designed for, this will not work.

The LT is also not just a magnitude "boost" like the knob labeled "bass" on your preamp or receiver. It also has particular phase characteristics that a tone control or parametric equalizer does not.

If you want to modify Qts of a driver, you can do that with feedback or series resistance. Qts can be increased or decreased.

-Charlie

Sorry that is not quite correct and you can do exactly what you want with an LT.

The driver Qts and Fs are your "unequalised" Q and F, and your target Q and F are what you want the parameters to change to. Vas will remain the same. There may be a problem though that just changing those parameters a small order of magnitude results in un-realisable component values or rule violation of the LT.

However, in a box design as has been pointed out already, you would not do it this way but instead correct the resulting box response back to what you want. Because the horn is not a 2nd-order system you can't work it that way. I was under the impression that parameters are of much less importance than the horn design anyway, so you are probably going to waste time and money trying to compensate the basic driver.