I know I should answer this one by building and testing a model, but it's been bugging me for a while now, and I still haven't found time to build it, so here goes...
The potential voltage induced is calculated as (this is cut from picoturbine notes)
Vmax = NARPB/2
· N is the number of loops of wire.
· A is the area enclosed by a loop of wire, in square meters.
· R is the rotational velocity of the magnets, in cycles per second.
· P is the number of magnet poles per cycle.
· B is the strength of the magnetic field of each pole, in Tesla.
Given that A is significant, would it make sense to stretch the coils out some? Say for example if the above diagram had only one coil occupying the span of both the coils in the diagram it seems that you'd get almost 3x the voltage for less wire than it would cost to make 3 smaller coils in series. The thing is I am not sure how much current you'd lose to the increased inductive Z, but I strongly suspect that a net power gain is possible.
Has anyone played around with this notion, or can anyone tell me if I am missing something?