Why do people buy a budget model and then complain that it's a budget model? Magnepan makes speakers with multiple drivers, true ribbon drivers, and with dual magnet assemblies in which the drivers are a good deal more rigid than the single-magnet drivers. They cost more because the components cost more.
Sure, they could use esoteric components, but they'd lose their reputation for high bang-for-the-buck. There are planars out there that are no-holds-barred -- neodynium magnets and what have you -- but you'll be spending in the tens of thousands for them with, frankly, little gain in performance.
Magnepan's tradition is value engineering -- the components they use are chosen because blind listening tests show that they are better for the price. For example, Mark Winey, the company's president, told me that he didn't OK the addition of the supertweeter to the 1.7 -- its predecessor, the 1.6, didn't have one -- until it was preferred in blind listening tests by both expert and non-expert listening panels.
Pretty much everything that goes into the speakers is blind tested and often customers don't grasp this. They do uneconomical things at home (where labor is free and it really doesn't matter) and don't understand why Magnepan doesn't use them. That's why you won't see the likes of esoteric wire -- in blind testing, no one was able to hear it. Ditto (since we're talking crossovers) air core chokes -- they used to use them and then compared them blind with iron core chokes and nobody could hear a difference. (You can bet they would have if the test hadn't been blind, lol.)
Before you redesign the crossover, by the way, be sure you know what they've done -- people frequently mess this up. The crossover from the tweeter to the supertweeter is a 1.5-way crossover, which is to say that the whole tweeter plays up to a certain frequency and then a narrower segment takes over at the very top. This optimizes both power handling and dispersion in a panel of limited size. Also note that the frequency response is designed to decline gradually with frequency rather than being flat -- a standard house curve. So many don't understand the purpose of a house curve and think speakers should be flat, ouch. Also, that in an average room you should use the 1 ohm series resistor -- it is equivalent to the center position on a tweeter pot.
At least one commercial mod gets these things wrong. Just imitate the values with better caps and you'll be fine. Or use a higher order XO between the woofer and tweeter, although you'll probably be losing as much as you gain -- it really depends on how loud you like to listen. Retain the series XO so the impedance drop from the tweeter to the supertweeter doesn't throw off the XO point, and to damp the panels out of band -- that's a lot more important in a planar than it is in a dynamic.
If you increase the strength/mass/damping of the baffle, or improve the mounting of the driver in the baffle -- all good things -- make sure you don't cause diffraction issues. Hell, you might be able to reduce diffraction with half round end pieces and some felt -- I don't know, haven't tried.
Not all of the design considerations at all, but an indication of the thought and experience that have gone into these speakers. Forgive me if you're already familiar with some or all of this -- some DIYers are truly knowledgeable -- but from what I've seen, most people know a lot less about commercial planar design than they think they do.
I assume you kept the original crossovers intact somewhere?