Disconnect the woofer

The 20.1 crossover is jumpered, and as long as you are removing the LF signal from the input, via an active x-over(with jumpers removed); you shouldn't have any issues. I'm, using a modded TacT 2.2X to actively bi-amp my Maggies, with tubes above 250Hz/SS below w/a 10th order slope and transmission line woofers. Open baffled woofs would(of course) be accurate/fast enough to seamlessly blend with planars also, if driven with amp that can control them, and avoid excessive overhang(amp and woofer have to trace the signal faithfully). An interesting balancing act.

Many crossovers have a simple 1st order Butterworth section(an inductor) in series with the woofer Some let the woofer's natural inductance roll off the mids/highs. If there's more than that(a cap to ground after an inductor, balancing or EQ circuit resistors/caps/coils, etc), you will experience some loss/signal degradation. What's the low freq circuit of the x-over look like? A X-over that's designed for bi-amping/bi-wiring separates the entire hi/low-end section, via removing it's jumpers. If you can trace the low freq circuit of your x-over, and clip the positive circuit at the input terminal, you'll not have any further concerns.

Perhaps you've never seen an EQing X-over, such as that in the Rogers, Harbeth, Chartwell, Spendor or KEF LS3/5A, LS5/9 and other BBC approved monitors, the Harbeth HL-P3(with it's 21 element crossover) and so on. There are a number of trim resistors, shunt coils(some that act as autotransformers), and caps that are in the notching/shelving circuits and go to ground from the positive input(the main reason they are so stinking inefficient). If these are not removed from the LF signal path, they WILL dissipate energy as designed, but even more as the woofer will not be in the circuit as a load. That's why I asked what his crossover looked like. There are too many variables to make a blanket(uneducated) statement. ie: What would occur in this system, if the woofer were simply removed: (http://www.syer.net/images/DAHLQUIST_DQ10Schems.jpg) YES- The best/most accurate way to bi-amp is via an active x-over, before the power amps. The second best, is to separate the HF/LF sections of the existing crossover, install two new terminals and use identical amps on both sections(either horizontally or vertically).

What amount of power would be dissipated would not concern me. You brought that up. The signal degradation, resulting from by those devices dissipating energy, and the differences in a circuit's impedances brought about by removing a reactive device(the woofer) would be my concern. How accurately do you suppose the midwoofer in the system(schematic) I posted would perform, were the woofer removed from the circuit? What I'm pointing out is that it's foolish to make blanket statements, without knowing the design of the OP's crossover.

You're correct on both those counts. I was stating what would be MY concerns in such a situation, and (again)addressing the blanket statements. It would be nice to hear from the OP, regarding the complexity or simplicity of his X-over, so his question could actually/accurately be answered.

All that's why I replaced my modded Dahlquist DQ-LP1, with a modded TacT RCS 2.2X, after 23 years. It was Linkwitz's writings, in the early eighties, that got me started bi-amping at home. I'd been doing it for years before that, in live venues, but his articles inspired listening room experimentation, and the first LS3/5A's that I built.

The 20.1 X-over's Hi-Pass section has 120.3uF of capacitance after the jumpers and before the signal is further split into MF/HF. If he's not removing the LF via an active x-over b4 the main amp, he's basically "passively bi-amping". Not the cleanest or best way, but it'll work.

Are you the first owner of those Maggies? Perhaps someone modded the X-overs? If they are original, and you get no output, correctly hooked up/jumpers removed; something's amiss.