A few words about the latest conversion project.
The victim:
The specifications:
type DM 100-80
manufacturer Groschopp & Co, 4060 Viersen 1, Germany.
- 130W motor
- 3 phase, 380V star, 220V delta
- n=1400 RPM (4 pole)
- current: 0.57A (star), 1A (delta)
- cos phi = 0.55
Each phase has a resistance of ~ 40 ohm.
The bearings are 6002-2RSH on the front (prop side), a 6000-2Z at the back. This last bearing takes the full thrust load.
The surgery:
Below is a partial AutoCAD drawing of the new rotor.
The conversion consists of 48 magnets, N45 12.0 mm x 6.0 mm (close, but not quite, to 1/2" x 1/4"). There's 2.0 cubic inch of magnetic material in there; according to Zubbly's rule, this should yield an output of about 2 * 150 = 300W. Since my magnet grades are slightly higher than normal (N45 vs. N38/40) I expect slightly more output.
The magnets are arranged in 4 poles. Each pole has 2 rows of 6 magnets each.
The original shaft was used after the original rotor was removed. Then, a steel core was installed and glued in place with Loctite 638. The steel core was turned to the correct diameter, after which an aluminium sleeve was fitted and Loctited in place. The outside of the aluminium sleeve was dimensioned to correct dimensions after the Loctite had set.
The next step consisted of machining the pockets for the magnets. This was done by printing the layout of the magnet (see CAD drawing) and glueing this temporary on the rotor. Then, the correct position was marked with a centerpunch.
The holes were pre-drilled with 3 mm, then drilled to 8mm and finally drilled to 12.5 mm. The final step consisted of milling (on the drill press) with a 1/2" end mill, to achieve the flat bottom for the magnets to sit on.
The magnets were glued in place using epoxy. I didn't use Loctite for this, since the manufacturer has been unclear whether Loctite 638 has good adherence to nickel coated magnets.
After 3 days of curing the rotor was manually inserted into the stator. This was a relatively straightforward procedure, and despite the presence of 48 strong magnets,
the insertion was fully controlled.
Unlike the previous 3 hp conversion, no other machines were used than a lathe and a drillpress.
No balancing has been done, neither static nor dynamic.
No rewinding of the stator has been done.
The results:
The first test results, by no means complete, indicate that it requires about 7.5 RPM/volt (per phase, i.e. neither star nor delta). These are open volts. Further tests have to be performed by letting it actually charge a battery at various RPMs.
At 300 RPM, Vopen = 40Vac
At 600 RPM, Vopen = 85Vac.
These are volts per phase (neither star nor delta), unloaded.
There is no cogging to speak of. One can feel a little cogging, but it doesn't impair performance. I ascribe this cogging due to the less-than-perfect machining of the pockets and thus the position of the magnets being sometimes slightly off. The actual procedure for the decogging (see my diary of the 3hp conversion) has proven that the theoretical concept is valid. Had I machined it on a CNC mill, I have no doubt it would be 100% cogless.
We all know that the magnetic pole width should be about the size of the inside of the coils. I was aware of this rule, and knew that the price of wider magnetic poles would be cancellation of voltage and thus less than optimal use of the magnets. This seemed like a small price to pay, if it meant it would generate more power and at lower RPMs. So, I crammed all the magnets in there I could (48 pcs), which meant that the magnetic poles are wider than the inside of the coils.
The cancellation can be felt very clearly when the generator is loaded and one rotates the shaft; at the points of cancellation, the generator doesn't generate as much and this can be felt by a sudden drop of load, i.e. the shaft becomes much easier to turn. This happens at the 12 o'clock, 3 o'clock, 6 o'clock and 9 o'clock positions. I fear this will make the genny noisy and make it vibrate. It appears that cancellation has much greater impact than 'just' a less-optimal use of magnets. It's something that should be avoided as much as possible.
The cancellation feels a bit like cogging, but only at 4 positions. Plus, unlike cogging, the 'cogging cancellation' can only be felt in 4 positions and -only- when the generator is being loaded.
Here's a picture of the unloaded shape of the waveform of one phase. As can be seen, there's quite a bit of distortion from a perfect sinewave.
The lessons:
Do as Zubbly says; make your magnetic pole about the size of the inside of the coils. Cancellation doesn't just cause less-optimal use of magnets but also vibration and noise.
The next phases:
The hub is more or less finished. Next step is the carving of a prop. I am considering something of about 1.2m diameter. Probably with a low TSR, since this genny generates at relatively low RPMs. However, no real design work on the prop has yet been done. I'm open to suggestions w.r.t. TSR. As far as system voltage goes, this genny would probably make a nice one for a 48V system, but a lousy one for 12V, without a rewind. I don't intend to rewind.
Conclusion: a few important lessons were learned building this relatively small genny. The previous 3hp conversion was built to theoretical perfection, and machined as near to perfection as I could get. With this 130W genny, I tried to use as much magnet as possible and to cut corners in the machining. The results are cancellation and (slight) cogging.
Important lessons and I'm glad I've learned them on this small (but still very
useable) generator, as opposed to the 10+hp conversion I feel like tackling now...
(more pictures can be found in my IRC gallery, page 9 and further, and this sub-gallery: http://www.anotherpower.com/gallery/album49 )
Regards,
Peter.
