Since I have this turbine basically completed I decided this is the time to do some testing with different stator and magnet rotor combinations to see what results I get. What I have found so far is pretty interesting when you try to spin a permanent magnet generator at high speed and still match a slower turning rotor's power output. Flux and some of the other guys that build generators might find this interesting too. I had to re-think my generator design for this thing again, but I got enough different stators and mag rotors just laying around the shop that I can come up with a half dozen different combination just by re-wiring them delta and see what they do.
What I've decided is that I can cut the cost of the generator to about 2/3's of a direct drive unit. You still need two generator rotors - but only one has to have magnets on it. So I started calling this rotor combination a "mag rotor" and a "core rotor", since the rotor without magnets acts as a core to channel the flux lines.
When I tested this on the lathe I didn't get the results I expected. But I used two rotors that were the same size. The flex density was .3510 tesla in the air gap (.800") vs .6075 with two mag rotors, or about 58%.
Tonight I got a wild hair and tested this using a 10" mag rotor and an 11" core rotor. The flux density rose to .4135 tesla, or about 68%. Evidently you need to use a bigger core rotor than mag rotor to capture flux lines that extend outside the perimeter of the mag rotor to properly concentrate them in the air gap.
Even the combination using a 10" mag rotor and 11" core rotor is way too powerful. So I opened the air gap to a full inch figuring this would throw my cut-in way off. It didn't. 5 rpm on the input shaft is 10.7 rpm on the generator and the rpm band of the generator increased considerably with the wider air gap, and it got less "stiff" and less hard to turn at higher output over 50 amps.
A little more resistance in the generator winding actually helps the power curve. The extra resistance doesn't seem to hardly make any difference below 50 amps, and it makes the generator easier to turn at outputs over 50 amps. The easiest way to get that needed resistance, and still get the ampacity needed, is to reduce the flux strength and use more turns of large wire in delta configuration. This is a 12 volt machine, so it has to be able to handle 100+ amps.
This is only a .467:1 overdrive, meaning for every turn of the input shaft, the generator spins 2.14 times. I'm pretty sure one of these could be built using ceramic magnets instead of neos. Better results would probably be achieved with ceramic mags (and less cost).
After several hours of running time testing in the shop, the transmission is now "broken in" and it spins so free that if I spin it up to 300 rpm shaft input speed I have to grab the input shaft to stop it because it will spin for almost a minute before it comes to a stop by itself. The "losses" in this type transmission drive are negligible.
Also, thanks to Brian for the comments on lubricant. I'm going to stick with the half and half #2 diesel and ATF. We get winters here at 30 below F and with #2 diesel in the gear case it will spin freely. I added the ATF because it has a lot of zinc in it and gives the diesel fuel better lubricity. I'm not worried about the hygroscopic properties of diesel fuel because I decided to build the transmission hermetically sealed so no moisture can get in thru a breather. The temperature rise of the transmission at high outputs around 75 amps for an hour is zero, or at least not measurable with my infrared meter that I use to test diesel engine exhaust port temp. And that meter is accurate +/- 1 degree F. I imagine there's some heat generated in the gear train, but it's absorbed by steel mass and dissipated before it can be detected. An oil bath chain transmission is an incredibly efficient drive, especially at the input shaft speeds that a 12 foot turbine rotor can run at.
This has been one of the more fun turbine projects I've done and I'm extremely pleased with the testing results. My new blades should be here next week, and then I can complete the rest of it.
--
Chris