MTM with BW3 and LR4 - Polar response

There have been numerous discussions on MTM. Most people expressed their concerns for the vertical response nulls due to the cancellation from two separated woofers and most say they would stay away from MTM. This includes Lynn Olson, who successfully designed the popular Ariel with a MTM that has a high XO point.

Looking at Dennis Myrphy's measurement of the "Stock vs Modified Gemini", I would be very worried about the vertical off-axis 20dB null between 1k to 4k. The measurement can be seen down at the bottom of the page:

According to the Loudspeaker Design Cookbook, D'Appolito used LR4 instead of BW3 in his later designs.

I have just modelled the polar response using the Xdir and I have the following:


XO point is at 2kHz. Driver (woofer to tweeter) distance is at 180mm. If BW3 is used, there would be a theoretical 90 degree phase lag (or let us say we obtain this by aligning the phase of the drivers, considering the physical offset of 32mm between the tweeter and the woofer). Each driver is -3dB at XO. The polar response seems to be reasonable, at most 4 -3dB points are found. Increasing the woofer to tweeter distance does not deepen the nulls. Not ideal, but quite acceptable.


If LR4 is used, there would be a theoretical 360 (or 0) degree phase lag. Each driver is -6dB at XO. The polar response now exhibits two large (at wider angle) -10dB nulls.

From Xdir, it shows that BW3 is substantially better than LR4 in polar response therefore power response. I would be extremely glad if the modelled response is the actual response.

The question is, does it mean we can solve all the problems of MTM with a BW3? Are there any other issues associated with MTM other than the above?



2006-08-27 11:02 am
Couple things off the top of my head:

BW3 will cause lobing to the sides, that's the price you pay for having the on axis response 90degrees out of phase.

The main argument against MTMs I always thought was not on axis lobing, but comb filtering of the woofers, and lobing/combing off axis.

Those polar plots are not good.

Here is good article on "fixing" the D'Appolito MTM:

Thanks for the link. The article is excellent. It shed some new light on MTM. Previously, the interactions between the M-T and T-M had been considered, but not the M-M. The paper included that into the analysis. I think the result would be more accurate.

From the measurements in the article, they still show that BW3 is substantially better than LR4 in the vertical plan. John K's modelling also showed the power response of a BW3 to be better than LR4.

BW3 will cause lobing to the sides

Do you have any further information to support this? I thought that since drivers are aligned in a vertical plain, there would be little concern for the horizontal responses, which would be determined by the individual driver off-axis responses, with no interactions between the drivers need to be considered. Please correct me if I am wrong.