[ ... ]
[ ... ]
^^^
Don't you mean 'Y' connected? If spelled out, it shows as "wye" not "why". (Or do you mean "'why' does it work?" :-)
I would not be surprised if it spun up -- but I would be amazed if it delivered anywhere near the nameplate horsepower. Each of the two split windings contributes half of the magnetic field -- either when they are in series or when they are in parallel. This should be sufficient to spin an idler for a rotary converter. But since you get double the voltage at the output, you will need double the current at the input (ignoring losses in the motor) -- just as with a transformer.
And your maximum *safe* output current (again neglecting losses) will be 1/2 the nameplate current at the desired output voltage. This is because you are limited to the higher voltage nameplate current into the half winding, and there will be a 2:1 ratio of drop in current corresponding to the 1:2 ratio of increase in output voltage.
Yes -- it can handle more than that for a short time, such as starting another motor, but long-term would burn it out.
And all this presumes that the two windings are around the same set of poles. If the windings are scattered around the motor, you will not get the transformer effect, so you get nothing.
I'm not sure, but I suspect that the motor on which Fitch tried it had physically split windings. There are two opposite sets of physical poles in the motor which make up one phase. If there is a separate winding on each pole, instead of both windings on both poles, then it will be useless as a step-up device.
Frankly, if I were designing the motor, I would probably use a separate winding on each physical pole -- to get a little better insulation between them when used for 440V operation. But I don't know which makers use which design, or even if the same maker changes designs from time (or model) to time.
A quick and dirty test of this for a specific motor would be to take one of the pairs of 220V windings and hook it to a 220V light bulb (or similar load), and apply 220V to the other of the pair for a short time to see what the light bulb does. If it lights at near full brightness, then it is acting as a transformer, and will probably work for the needed task -- with *lots* of derating. 1/4 the HP for the fact that it is acting as a transformer, and then another 2/3 for the fact that it is working as a rotary converter -- plus a bit more for enough muscle to start the load. So -- I would suggest at least ten times the needed load horsepower. All in all -- not very efficient.
Enjoy, DoN.