Alternatorless... or not

I recently (two weeks ago) tested the effects of disconnecting my serpentine belt while driving. Others have done this (or similar things) with much success:
Orange4boy's Toyota Previa
MetroMPG's Firefly (metro)
Daox's Paseo
The belt on this vehicle runs the a/c, p/s, and alt. The timing belt runs the water pump, conveniently.
I had tried monitoring tank-to-tank runs earlier this year without much repeatable success. This day I decided to do some A-B-A runs.
I did a 10-mile round trip run, all highway with a turn halfway, and did no pulse -and- glide, so as to maximize the effect of having the accessories running or not. I did the first couple of trips with the belt on, then did two with it off. Average FE went from 42.5 mpg to 43.2 mpg. Not convincing.
I went home and decided to do some monitoring of my voltage while driving beltless, and measured it to be hovering around 12.2-12.5V. In an effort to boost that, I connected another battery in the car to my normal battery via a 6A battery charger, hoping to get voltage up to 13.5V or higher. I repeated my test run, but alas, the voltage remained below 12.5V even with 6A coming in from the charger.
I'm not ready to call it quits, because I know there is a way to solve this somehow, in a cost-effective fashion. I know there is energy being wasted by my belt and accessories, even without the alternator being included. This should lie somewhere around 8% or higher with no belt at all, and it should show up easily in testing. Time to re-formulate a plan to delete the darned alternator system!