Hello Martin.

I have done a 1st order high pass line level filter for the tweeters. One 560pf polystyrene cap soldered in series with the signal which in conjunction with the 47K resistor to ground that was already on the amplifier board creates the filter. Second order is doable but it starts to get tricky with impedance and insertion losses and I only needed a first order anyway. 

Using the 47K to ground that was already there meant that I wasn't adding anymore resistors so any impedance issues were negated. No buffer required and if there had been I wouldn't have done it. If I had been adding both a series cap and a resistor to ground then things would have been different. 

And the amp board already has a low pass filter to cut out radio frequencies etc but if I wanted I could replace the capacitor to ground to a different value and in effect create a 1st order passive line level low pass filter for a mid bass. But as I use doped mid bass run full range then I don't need to do that. 

Regarding your question about the sub bass amp, both of my mid bass drivers (up and down firing) cover exactly the same frequency range (full range) so they both require the same signal. If I didn't do that then I wouldn't get the full effect of the baffle step compensation and force cancellation that bipole opposing drivers provide. It's a two way with two mid bass and four tweeters.

Slightly off topic but I could have had my speakers as a 3 way or 2.5 way but that introduces complications that go against my ultimate goal. I would have to use low pass filters on the mid bass because it doesn't matter how much dope you throw at a driver you can never get them to roll off at a low enough frequency for 3 or 2.5 way. And I like RD I believe in eliminating electrical filters where possible and doing things mechanically (dope, firing drivers in directions other than the listening position).

So it had to be a 2 way but with multiple drivers covering the same frequency range to reduce inter modulation distortion etc. That was part of the reasoning behind using multiple mid bass and tweeters. Share the work out which reduces the distance the cone has to travel which in turn reduces distortion especially on these cheap Chinese drivers (£10 - £15 a pop). Then I discovered the advantages of having mid bass on opposing panels and wired in phase. It provides BSC for free and does it mechanically without any electrical components and if the magnets of the opposing drivers are rigidly linked mechanically then there is force cancellation which means the enclosures are subject to vastly reduced vibrations which always a good thing. I have also noticed much less energy transferred to the floor.

As I don't want any mid bass pointing at me then the logical thing to do was to have the opposing mid bass firing up and down. Theoretically there is no cabinet "rock" as the cabinet is not trying to move anywhere unlike a conventional speaker with drivers all facing at you. Some of the down firing mid bass' upper frequencies will be absorbed by the carpet but this can be used to advantage by creating a shallow dip at the frequencies which  annoy the hell out of me.

The tweeters are housed in a little 12 cm sq "pod" which sits on top of the mid bass enclosure and behind the up firing mid bass (to try and get some time alignment). I used to have the tweeters mounted in the walls of the mid bass enclosure but i found the little pod behind the up firing mid bass was vastly superior. Also I used a hollowed out foam dice with nice rounded off corners to house the tweeters (forward, up, in & out firing) which will have had a positive effect on diffraction and imaging. 

Trying to keep tweeters to a first order high pass filter is tricky. Usually they must have a low resonance frequency (which must be avoided) and/or a high crossover frequency much above a typical 2 way. Typical 2 ways can't xover too high because the mid bass will be beaming at xover and the tweeter won't be (not good). But luckily i don't have a typical 2 way with the mid bass firing up and down. Also my xover frequency is high and out of the ear's most sensitive region. Even so I have witnessed the big reduction in distortion when using more tweeters. Can't have more than one firing in the same direction because of comb filtering but with them at 90 degrees to each other it works well and enhances the omni effect I was after so my whole family can enjoy it spread out over two sofa's rather than just me in a chair with my head in a vice.

It's not perfect and a compromise like any other speaker but it's my compromise which is bespoke to me and exactly what I want. there is nothing on the market which is comparable. It kind of explains my whole approach to audio and diy. I do what suits me best and typically it is not something a mass market manufacturer could do as it would limit it's use with other components. the line level passive filters are all part of this system wide design - everything is considered at once not just that particular component. Not every one agrees with the low order filter approach saying that LR4 etc are much better. I disagree, especially when other things are addressed in a non invasive way to minimise the bad points. The low phase shift approach (fewer filters the better) seems to make the best music for me. I must be mad or deaf or just special 

Well, this has turned in to a bit of an essay which has probably bored you all to tears. Sorry 

Yes, fascinating.

I only wish I understood more than one sentence in four.

I would certainly be very interested to hear your system!

I'm in Deep Gloucestershire, south westwards.
But happy to travel a fair way but not for a good few months, I suspect. Probably next year.
So you've got 12 months to get it sorted.