I believe I've found a simple workaround to allow and MC-E wired in 4S, 2S2P, or 4P and mounted on one of AW's turbo towers.
The tower uses a center channel to run the LED+ and LED- wires from the driver board to the emitter. The wires go up this center channel and out a horizontal feed hole near the top of the tower. All you have to do is enlarge this feed hole to accept multiple wires. In fact, I think the hole is already wide enough to take two 24 gauge hookup wires, which is all you need for 2S2P.
So, the connection options can be:
4S: -1 to +2, -2 to +3, -3 to +4; LED+ from the driver to +1; LED- from the driver to -4
2S2P: (+1 and +2) to (-3 and -4). LED+ from the driver to (+3 and +4). LED- from the driver to (-1 and -2). This requires only one "crossover" wire that connects the + and - sides of the MC-E, and you'd run this wire through the horizontal feed hole in the tower.
This is a better wiring configuration, IMO, than the equivalent -1 to +2, -3 to +4, +1 to +3, and -2 to -4; LED+ from the driver to (+1/+3); LED- from the driver to (-2/-4). This has two crossover wires.
4P: Connect +1, +2, +3, and +4; -1, -2, -3, and -4. LED+ from the driver to any of the + terminals. LED- from the driver to any of the - terminals.
For 0.55" drivers, the following limitations seem to apply. Max input current is about 1.5A. These small driver boards should be limited to about 4W maximum power output.
For 4S, you have an effective voltage drop of about 13V (based on a Vf~3.25V @ 416ma). 416ma corresponds, for example, to a Badboy 400. Each die in the MC-E should get 416ma, assuming identical electrical behavior for each die. That gives an estimated output of 425-495 emitter lumens. The power is 5.4W, which looks high, so I don't know if the BB will work well for continuous operation or over the long term.
The board is about 85% efficient. Can we use a single Li-ion to run the flashlight? Let's assume Vbatt=3.7V. Then we get 0.85*3.7*1.5=4.7W. If we assume Vbatt=4.2V, then the max power is 5.36W, which is close. But it is doubtful that a single Li-ion is going to operate at 4.2V for very long before voltage sag under the 1.5A load occurs. So, it doesn't look like a single Li-ion is a good choice for power source.
How about two Li-ions? Ibatt = 5.4/(0.85*7.40)=0.86A. This looks good. You can use any of the 3.7V Li-ion chemistries and be well within the 2C discharge rate limit.
You probably could also use battery configurations like 3xLi-ion, 2x123A, 3x123A, and 4x123A.
For 2S2P, you could use a step-up (e.g., Badboy) or step-down (e.g., SOB) converter. The voltage drop is about 6.4V-6.5V for 2S2P. The drawback with using a step-up converter seems to be that you can't use 2xLi-ion since the source voltage exceeds the voltage drop. Can you use one Li-ion? A Badboy 750 would deliver 375ma per die, assuming each die is identical and assuming standard current divider behavior. That equates to about 405-475 emitter lumens. The power is 4.8W, again perhaps borderline for these 0.55" drivers. A single Li-ion would have to deliver Iin = 4.8/(0.85*3.7) = 1.53A. So it looks feasible, although a bit close. At ~1.5A, you'd have to use an IMR16340 if you want to use that sized Li-ion. Otherwise, a 17670, 18650, or larger cell should work in terms of discharge rate.
How about the voltage at which the driver falls out of regulation?
Vreg=(6.4*0.750)/(0.85*1.5)=3.8V, so it looks like this is also a bit close.
You could also use 2x123A. Let's assume 5V for that battery combo. The 123A cells would have to deliver about 1.1A, which is feasible.
Now, let's look at an SOB step-down converter for 2S2P. I don't think you can use an SOB 1000. That would mean 6.8W of power, and I don't know if the SOB can handle that. An SOB 750 would mean 4.8W, like the BB 750. Same voltage drop and same emitter lumens as well. The difference is that the acceptable battery configurations look better. The SOB max voltage is 16V, so you can run the light using batteries that range from about 6.4V to 16V. That means typical battery configurations could be 2xLi-ion, 3xLi-ion, 3x123A, and 4x123A.
Finally, let's look at 4P. Voltage drop = Vf now. An SOB 1000 means about 3.1W, which looks good. Unfortunately, each MC-E die gets only 250ma, or about 280-320 emitter lumens. May as well just stick with a single-die LED like a Seoul P4. How about an SOB 1416? That's basically 350ma per die, or about 370-430 emitter lumens. To build an SOB 1416, we need R1=0.05 and R2=0.12. Power would be about 4.5W. Board efficiency is 75% to 85%. Input current using one Li-ion would be about 4.5/(0.75*3.7)=1.6A. So trying to use a single Li-ion looks marginal.
Two Li-ions should be fine; basically, input current would be 0.8A. You could run battery configuration from about 4V up to 16V.
The main question now seems to be whether a Badboy 400 works for 4S, a BB 750 or SOB 750 works for 2S2P, or an SOB 1416 works with 4P. Does anyone know the answer to this, empiricially or otherwise?