This experiment is my attempt to upgrade from the fairly-free-spinning cereal-box "yard art" toy VAWT, to a larger box on a real generator. Actual level of power output is a secondary consideration to getting something physically stable and safe, that will turn at all and/or in light winds, and that can put out any power at all.
(This all part of my urban wind-power pilot project.)
For this implementation I visited various local DIY/tool shops, and now have nearly a grown-up haul of G-clamps, mallets, roofing bolts, etc, since I needed some extra tools (eg hand-drill and modelling knife) and some newer widgets to make a supporting structure for the generator, which is large (7kg, 17cm/7in diameter) and has its three power wires emerging from the centre of the spindle.
The generator is from FuturEnergy. Handily I encountered them with their PM gen and HAWT turbine at their booth at a trade show a few days after the generator arrived (and had puzzled me). I could see that they were letting the generator's internal bearings take all the load in their turbine design, thus I can do the same.
The first thing I did was to make a support for my PM gen to allow me to build the rotor on top for a VAWT configuration. Cunning use of a plank of wood, 2 G-clamps, and two angle brackets to grasp the flat on the PM gen shaft, take the weight on the end of the spindle, and prevent rotation of the the spindle and its emerging wres, allowing the body to spin above it on the generator's internal bearings, with the rotor to be bolted on to and abive the generator body. It may not last but it is reasonably sturdy for experimental purposes...
This also for the first time allowed me to test when I might get cut-in to charge a 12V battery, using my analogue multimeter on 'AC Volts' across one pair of the delta-wired outputs while I turned the spindle by hand. My estimate is that bridge-rectified output will pass ~13V--14V at ~120rpm, which is probably just about reasonable for a drag-based VAWT design.
Shorting any two of the output wires makes the spindle very hard to turn, indicating that the generator is extracting the mechanical energy efficiently.
Then I attached my biggest-available box (from my scanner) to the PM gen, and cut the vertical slots as for my smaller cereal-box rotor.
Unfortunately the cogging and torque required to start up rotation was too high for conditions on the day, and a larger/flatter box is probably necessary for this to turn in low wind speeds. Maybe 1msq minimum swept area and possibly at least a metre for enough torque.