Dan, I am doing exactly that right now. It turns out that while Adire's passive XO looks simple, it really is optimized to work better than more complicated topologoes. So I came up with active filters in LspCAD and I'll be building one up with JFET buffers, though opamps could also be used.. hopefully in a month or two. A preliminary schematic:
And this is what the frequency response should look like (I modeled the driver response characteristics with a few active filters). There is about 4.5db of baffle step compensation.
Xover looks good. I admire your originality, using FETS.
You may want to look at designing a notch at around 2k2 to help tame the AV8s peak.
Also, from what I have read, MTM designs usually exhibit a peak in frequency response in relation to the centres of the M the T and the M which calculate to around 2k2 as well.
I am aware of that, but an active notch filter would increase the complexity of the circuit significantly, and also would exacerbate the off axis drop in response around that frequency. So I'll probably try it as a tweak later on.