Please don't clutter the board with amazing projects.
(kidding).
Very nice.
By the magnet being 2" long, I think you could interchange that with 2"
thick, or 2" deep into the center.
Big magnets sound good, until there is not space down toward the axle. Or until the factory rotor falls into a stack of washers on a rod. I have had both occur more than once.
It is
Always easier to conserve than make power.
Consider reducing the load.
LED bulbs in frosted globes are just as impressive to the novice onlooker while using a tiny fraction of the power. (and can get them on ebay for $5)
Look for a fan with similar CFM at a lower current. I expect you can find something that only uses 100ma.
I have NO doubt the other guy achieved 122V, but I doubt he achieved 122V and 0.5A at the same time.
That's the problem.
The volts are easy, but volts and amps at the same time are very difficult to get from something with that many turns and that much resistance. Trust my failures on that one. Simply the amps are what I would consider out of reach.
There are deeper issues, especially concerning the coil inductance in something like a cieling fan motor. There will be a point where the output current simply goes flat. This chart is of something that weighs 11 pounds and with far less issues than you have,
except (combined total) I expect your will flatten out before 0.7A,
If you use a single rectangular neo per pole, the gap will increase and the flux will decrease.
http://www.otherpower.com/images/scimages/2050/Gap.gifDude. If you built that thing, I bet you can rewind the motor to 12 pole 3-ph. I am not saying it would be quick and easy.
I am not sure about the aesthetic requirements, but I still think you would be infinitely better off starting with a different motor.
AND decreasing the load.
AND adding a battery, even if it is a tiny battery.
Which brings up the issue of voltage control. It may be a relatively simple issue with a fixed RPM and known PMA.
There is more than one way to skin a cat.
If you are dead set on making the K55 into a PMA wired as it is, at 600RPM, might even be able to use a pair of 120:18(?) transformers. No clue if that is a good idea, it is simply a thought.
If you get near Pratt U in NYC, you should stop for a look at their stuff.
Impressive and very friendly.
G-