From what I have seen, whirligig makers start by making a hub with slots cut in it to hold the blades. Some say that 35 degrees is the best angle so that is what I used. I wanted 8 blades, so the slots are narrowly spaced. To make it, I cut out an octagon, used my bandsaw to define the slots and then chiseled out the waste.
I chose long thin blades, thin for looks and long for more surface area presented to the wind. Here's what the propeller looks like:
It is 22" in diameter and way overbuilt (the blades are 5/16" thick), which makes it heavy and therefore difficult to start because of inertia. I'm going for a traditional windmill look rather than a whirligig so that's why it is much larger than normal. Because of the weight I needed a stout shaft, so I cut off the head of a 1/4" stainless steel bolt and epoxied it into the back of the windmill. The propeller shaft rides in some little bronze bearings I found.
Before someone else says it, this isn't an efficient windmill; it's primarily decorative and designed to look good to my eye. This was confirmed when I tested it with a fan. The blades are quite thin. Because of it's mass, the propeller takes a while to spool up but this inertia also keeps it running at a fairly constant speed in a variable wind. It runs surprisingly true and doesn't seem to be grossly out of balance. The bearings are new and I don't want to lubricate them until after I apply finish so I suspect it will turn more freely in use.
This windmill would be great if I lived on the coast but I'd like to come up with one that would turn in a light wind. It needs to be lighter and the blades can be thinner by half. Wider blades would be better because they have more "lift" but I like the look of the thin blades. I did some research and it appears that the 35 degree pitch isn't right because the blades "stall" in light wind. Surprising to me at least is that scientific research shows an angle more like 5 degrees to be optimal. The latter presents a whopping 22% more frontal area to the wind. This creates a construction issue because the adjacent slots won't clear, so some new way of making the hub is necessary. A 4-bladed windmill wouldn't have this problem but I like the look of 8 blades. I have an idea, although this is starting to feel like my stool saga. In my defense, I did end up with a unique design that works very well. :-(