I had actually considered this approach myself- the four-box monoblocks.
I think physical implementation is critical for it to work well, and not end up a problem. Bottom box (with transformer(s)) top, and top box (everything else) bottom, should be steel, and thick enough for the expected emissions to be squelched, or, why bother? Counterpoint felt that copper plated steel was the hot ticket. The air gap between them is good as EMF is diminished very rapidly with distance, so that takes care of itself.
Female sockets on the top of the bottom, and bottom of the top, exist, and are in alignment such that distance for DC to travel is reduced as short as possible, and massive interconnection reduces loss. Copper "jumper rods" interconnect the chassis to one another. Perhaps integrate these into the feet, so they are "invisible" from the front.
Additional interconnection is necessary to avoid some Darwin award nominee sticking their snot-moistened finger into one of the DC sockets of the bottom cabinet and enjoying the light show of their ignited hair immediately prior to death...
J!m,
I consider myself guiltier than most when it comes to the dreaded 'analysis leads to paralysis' syndrome.
But sometimes Necessity is the Mother of Real-World Concentrated Learning. Back in the day I was trying to help a buddy who wanted some affordable "Big-Fi" for the house he had just built. (With an enormous 2-story main living room) Using a deft touch while employing a triple-dog-dare that I couldn't refuse, he set a ridiculously low budget & bet that I couldn't deliver the goods.
Long story short? First I found a used
Sansui 9090 receiver for ~ $1/Lb in working condition.
I then
lucked into a pair of used VMPS towers for short money. (Too big to sell? Very poor WAF factor eliminated the buying competition? Or some combo of the 2? I simply don't know.)
At my house I put them together...and although there was plenty of 'there' there, it just wasn't right. Even with classical music with a wide dynamic range, it sounded loud...but at the same time, all the music sounded like it had been run through a compressor? Time was running out, and since I wasn't exactly working with 'legacy equipment', after reading around a bit I popped the cover off the 9090 & scoped the DC rails while playing organ music.
Sure enough, if I connected the 'A' channel of the scope to the + of one of the speaker outputs, and the 'B' channel to the DC rail (+ or -) I watched the
rail voltage *sag severely* whenever a righteous transient occurred. Naturally, it seemed that I needed to stiffen the DC rails as much as possible. I installed new/larger DC power supply caps (got a little greedy, nothing dangerous) and bypassed them as the fashion dictates of the day would have you do. I then did some further electrolyic cap bypassing right at a handful of locations. Not too much math/science...just whatever seemed to make for a stiffer rail/flatter B channel scope display under load. (NOTE: I had the VMPS speakers being driven while testing my hackery -- you know, the no-BS testing where the rubber meets the road...)
****
When I was done, I had come in well under budget. (He already had a CD player to use as his source.) And to tell you the truth, this quick-n-dirty fix ended up sounding *way better* than it had any right to. (Part of this must have been the inherent acoustics of the large 2-story main room.)
Oh yeah, when that 1st deep organ note hit, (and it sounded like it came up from some subterranean place) the new owner could not have been any happier.
Sometimes, the cheap stuff ends up being a lot more fun than the big $ toys.
****
The point I was trying to illustrate is that with a basic plan + some ad-hoc experimentation, I'd be willing to bet that a '4-box monoblock' installation could be dialed in so that it delivers an amazing soundstage + lower noise floor + life-size effortless reproduction.
Disclaimer -- The above is a sample size of 1, your mileage may vary, etc., etc. But man oh man, the above was the essence of *fun*.
For what it's worth --