I think I may opt for converting a canoe initially, just to do some power unit development. I've been giving the stability issue some thought, including some ideas of ways to counteract that "adverse roll" effect that Warren has helpfully pointed out.
I've now started to build an experimental stern tube assembly, using a length of 1" x 16g alloy tube with a machined flange brazed to the end to fit the direct-drive motor. I've ordered a length of 12mm diameter stainless steel tube to act as a prop shaft. This will be directly coupled to the motor shaft and run in a bearing at the prop end, with a simple lip seal to keep the water out. Hopefully this will be relatively low drag; I'm hoping that the slight reduction in motor efficiency from running at low-ish rpm will be offset by the large reduction in drag from not using a reduction drive. If this doesn't work, then I will need to looking at bigger diameter, lower Kv, motors, perhaps a modified bike hub motor or perhaps an "alter-motor" using a motorcycle alternator core.
Here's a photo of the motor mounted on the stern tube.

This assembly weighs just under 1kg (around 2.2 lbs) so is pretty light and on track for my target of no more than 10kg for the complete power system, including battery pack.
As an aside, that machined flange was simply push fitted on to the end of the tube and then brazed with Technoweld (see here: http://www.techno-weld.co.uk), a low-temperature alloy "welding" process that is pretty easy to do at home with a standard blowtorch.
Jeremy








