Dennis used his years of experience and his crossover design software to work up a new crossover. While listening to it via his crossover emulation software, he sent me this rather provoking email:
“It sounds freaking great to me. The tweeter isn’t state of the art, but it gets the job done…”
For those who don’t know Dennis Murphy, he avoids colloquial exaggerations ands is usually rather understated. When he gets excited, I sit up and take notice.
The predicted frequency response with the redesigned crossover is shown below, first with proper driver polarity and then below that with the midrange’s polarity reversed. The latter curve shows the new crossover points, and demonstrates that the drivers are in phase around the crossover frequencies when the polarities of the connections are correct.
The woofer-mid crossover, at ~950 Hz, involves Linkwitz-Riley 4th order crossover slopes. The mid-tweeter crossover, at 5 kHz is also LR 4th order. The glaring 6-7 kHz peak is essentially eliminated, and the high frequency comb filter cancellations are also gone! According to Dennis,
“It took more than adding a low pass filter for the midrange driver above 5 kHz because the big peak was not caused by driver break-up. It’s actually an additive diffraction artifact caused by the wide baffle and the goofy layout of the drivers. Getting rid of it wasn’t easy, and certainly wouldn’t have been possible using the design technology of the '70s.”
The profile from 10 to 20 kHz remains uneven, and is probably the best the tweeter can do – looking just like the unfiltered tweeter response curve.