I've already attempted and discarded a couple of builds. The first and second were the Savonius turbines I came across on Instructables.com. I've been hampered by a few issues: first, I don't have easy access to a band saw, scrolling saw, or rotating saw, and so splitting the PVC into half-cylinders of consistent height is nearly impossible to accomplish on a reasonable timetable. Second, Savonius turbines aren't particularly powerful, especially when you consider the major increase in the tower's swept surface area that they cause. Third is the difficulty in finding cheap/effective bearings to use to get the turbine spinning.

The third build attempt was a Darrieus turbine with the spinning axis in the center and pylon-shaped airfoils jutting upwards and downwards from it. This would have worked if not for the fact that I used a conic section as my airfoil, yielding a neutral angle of attack and a net outward lift instead of rotational force. 

My latest design is a helical Darrieus turbine with a secondary set of blades which serve as the self-start mechanism by transforming updrafts into the initial rotational momentum. Pictures and build logs to be posted later in the week.