By a notch filter do you mean a BSC circuit or are you talking about something else?

Martin

Basically, whatever you do to modify the amplitude response also modifies the phase response. So if you do something that equalizes the characteristic response curve, you have also modified its phase response too. It really doesn't matter if the equalizer is an electrical, mechanical or acoustic filter, the laws of physics apply to them all the same. If you change the amplitude response by any means, then you must have also changed the phase response.

In the frequency range where the baffle step occurs there is a smaller phase shift due to the summation of the "sources" along the edge of the box combining with the driver's reponse. For my Lowther DX3 ML TL enclosure this amounted to about 30 degrees maximum. If you apply a Zobel across the driver, the combination behaves as a pure resistor. Then placing a BSC circuit in series the baffle step is corrected including the phase. The BSC circuit has the opposite phase shift whne compared to the baffle step response. Above the baffle step region the BSC circuit acts as a pure resistor so you have a simple voltage division with the driver/Zobel to pad down the SPL magnitude. There is no impact on the phase.

This is how the circuit worked for my Lowther DX3 ML TL design. I measured the response and plotted the SPL and phase (after subtracting out the time of flight phase shift). The phase response was improved with the BSC circuit in place! I verified this with some calculations in MathCad. This is just one data point but I think it is accurate for most situations, I am sure somebody could dream up a situation where this was not the case. They probably reside over at AA. The purists at AA have completely closed minds and very little technical understanding beyond the rhetoric. Our frined TC is a prime example.

Hope that helps,

Martin

I understand the idea of minimalist solutions and low parts count. But remembering that the speaker motor/diaphragm/cabinet is a system that has many filters, sometimes having a part of two in the circuit counteracts reactances and makes the system as a whole more pure.

I really like keeping it simple, and sometimes that's best. Phil's little speakers sound great just bolting the speaker in the box. But there are other systems that are definitely improved with passive compensation. And it's not necessarily a matter of quality, or of artificially "fixing" a driver's response. It's just the way linear systems act.