B&W 803D crossover caps

I have also used the Mundorph M-Resist (20w) with good results.
Have you considered redoing the crossovers on new boards with point to point wiring and better component layout to reduce interaction ?

Dover:

In a word, no.

I had only planned to replace a few caps, not rebuild the entire cross over. I don't think that a rebuild is sonically warranted at this point.

I am hesitant to replace 100W rated components with 20W versions, no matter what the sonic results MIGHT be. Perhaps with a 0.5 ohm resistor a 20W might work, but I have not done the power dissipation analysis to confirm.

Dhl I did move the huge crossover in my 1985 JBL 250Ti's outboard to great effect. Perhaps you can consider ways of isolating your XO within the boxes, although I feel B&W are respectful of this benefit as previously they advertized the benefit of having the XO in the bases of top speakers. By the way what are your feelings re: bypass caps?

The best sounding resistors by far are Path Audio folks. Tried them all and these are special. Not cheap, but very special! Look into them and be impressed.

Ptss:

I can see the potential benefit to moving the XOs outside the box. My early JBL L212 had the crossovers outside of the enclosed driver housings.

Vibration from the drivers will definitely influence the inductors and capacitors in the network, simply "gluing" them to the pc board with lumps of RTV or caulk may not be that effective.

But practically speaking (no pun intended) moving them outboard of the 803D also has it's pitfalls, as you end up with longer speaker wiring between the drivers and the XO, with the accompanying increase in inductance and resistance. Not to mention the ugly appearance. The bass XO is huge, with monster 100 mFd Mundorf polypropylene bypass caps. Placing that outside would look butt ugly IMHO.

Bypassing caps...not sure what they buy you in a crossover. I know folks use them to "tweak" the sound but I have never played with them for that. I would rather buy a single cap with the right dielectric and low ESR in the first place if at all possible. Plus, from a technical viewpoint, things like dielectric adsorption cannot be remedied with parallel caps; only ESR or maybe ESL. The frequency range in a speaker crossover (maxing at maybe 30 KHz) is just not high enough to see a resonant frequency shift by using a 0.1 mFd cap in parallel with a 5 or 10 mFd coupling cap in a tweeter XO, for example.

So for example, the B&W diamond tweeter is crossed at 3500 -4000 Hz. The reactive impedance of the 4.7 mFd coupling cap at 4KHz is about 8 ohms. It would only be 0.8 ohms at 40 KHz! The 4.7 Mundorf has an ESR that is about 10 milli ohms or .01 ohms, still 10 times lower than the capacitive reactance at 40 KHz. Now what good is placing a 0.1 mFd cap in parallel with the 4.7? It would have a capacitive reactance of 50X the 4.7 (at 40 ohms) and who cares what its ESR is, because that would be swamped with the capacitive reactance.

If you are parallel connecting a series of lower value caps that may make sense if you cannot get the total value in a single cap (like a teflon or polystyrene, for example).