Thinking out loud here: would this speaker kit perhaps be a good platform to develop a DIY high-value active monitor?
My take of this review is that the kit, with the Sehlin/ScottS/Mudjock perfectionist crossover mod is pretty good, and becomes really good with a little EQ.
So what if the speaker-level crossover is discarded completely, and the crossover is implemented at line/pre-amp level, then fed into three amplifiers (one for each driver)? A variation on this idea is for the speaker input to be digital, and the crossover is implemented via DSP, followed by three amplifiers.
I don't have enough knowledge/experience to know how feasible it is to "convert" a speaker-level XO to line-level. I am assuming the line-level XO would be smaller and cheaper, since the components don't need to be rated for high power.
If my assumption is correct (line-level XO is small), then I think it wouldn't be too hard to have a PCB like this: input buffer -> XO -> two class D amps (one configured for stereo output, one configured for PBTL output). Since the speaker cable lengths would be very short (i.e. just inside the box), the class D amps won't need the typical (large) EMI-supressing LC filter (can get by with a cheap/small ferrite bead solution). The TI tpa3220 and Infineon ma12070 class D chips have simple PCB implementations, and are pad-down ICs that use the PCB as a heatsink. The TI tpa3250 is also a candidate, but has a more complex implementation.
Thoughts?