This idea I got about 2-1/2 years ago of going vertical was to accomplish several goals that are out of the scope of multiple placements on the floor.
1) To eliminate the typical peak/null (depending on where you sit in the room) caused by the floor-to-ceiling standing wave, 2) to explore the results of facing 2 identical and identically powered drivers, 3) to be able to place serious fire power in a tiny footprint and, 4) to see if the theory of adding a 4th virtual point source (the ceiling) had any veracity.
Here are the results of my preliminary test (there have been subsequent tests, but I ain't tellin' just yet):
I picked the worst corner in my 3500 cubes room. I placed a single dual opposed 2x15" module (slightly smaller than gpmbc used to prevent bottoming in a worse-case scenario) in that corner and calibrated. The mic was placed at the LP and not moved throughout the exercise.
Sine sweep from the LP, single module:
I then stacked a 2nd module. They're 8 ohm modules, so the 2 were wired to 4 ohms into the same amp, same calibration.
Sine sweep from the LP, stacked dual module:
I then stacked 2 more modules for a stack of 4 total. About 88" high. I set the bottom module on a circular piece of 3/4" MDF that was just large enough a diameter to catch the feet of the bottom module and leveled it. I did not secure the top plate to anything, so that the column was free-standing.
I wanted to, among the other things I listed, see if dual opposed was all I've cracked it up to be in that, had the stack vibrated at all, it would have easily walked off the MDF base and BAM! Disaster!
Sine sweep, 4 modules, 2-identical amplifiers into identical loads, calibrated the same:
The difference between the 1 module and the 2 module should be +6dB. The difference between the 1 module and the 4 module should be +12dB. Instead, taking into account the difference in frequency response, the average difference was +18dB.
The null at 73 Hz, which perfectly matches my floor-to-ceiling dimension times 2, goes from -30dB to perfectly flat. The overalll FR from 1 module to 4 modules goes from (+/-) 21dB to (+/-) 8dB, all in a footprint of 2.75 square feet.
After days of playing bass-heavy movies at the systems limits... the stack did not budge 1 mm from it's original placement. Besides the fact that the test for stack stability was successful, that also means that when the floor rippled with high level ULF soundtrack pulses from the beginning of Star Trek, et al, it was strictly from the pressure wave and not the subs sending extraneous vibrations into the floor.
After placement and sine sweep measurements, I ran Audyssey:
Actually, all that was required was 2 PEQ filters to smooth the response. I personally don't much care for the way my version of Audyssey manipulates the signal, nor the huge boost it imposes on <25 Hz, so that's what I did and I shut Aud off.
The SSD drivers are well built overhungs for relatively cheap $$ (although the price has gone up about 15% since I bought them), the module size is quite small, the stack looks pretty cool and the results were very much as predicted for a pretty small penny.
here's a shot taken during the tests. The front panels were faced into the corner so we could have easy access to the terminals for wiring. [You can see the MDF base with one of the bottom modules feet right at the edge]: