Putting Toshi to the Test

[SparWeb]

Hi,

Months ago, I obtained an old Toshiba 7.5 HP three-phase motor.  I began converting it for use as a wind-turbine generator by installing magnets on the rotor back in the fall, but since the assembly I haven't been able to get back to the project and finish it off.  The complete motor weighs 125 pounds, so even moving it around is difficult.

Here is a LINK to some earlier forum discussion about the assembly process.  24 magnets = BIG FORCES!   :-)

The cogging is noticeable, but not excessive.  It seems bad when turning it by hand, but with a suitably sized propeller, there will not be enough resistance to prevent early start-up.  I get a laugh out of my GE motor-conversion, whenever I see it tick-tick-tick to a stop, so I expect to see the same from this one, if I ever get it up a tower.

Early bench tests showed me that cut-in will be very slow in this one.  Wired in series-star, the 24V cut in is just 40 RPM!  It just begs to be wired in parallel, or operated with a 48V battery bank.  Later tests show that very high voltages are present at high RPM's, making direct electrical water heat a serious consideration.  I had to take unusual precautions while running it up to avoid electric shock.  More on this later.

To give Toshi a complete run-up test, I used a lathe at work.  This lathe has a 5 HP motor, which sure seemed like enough before I started, but I was wrong!  I chucked the main shaft into the jaws and the tail (fan) shaft into the tailstock.  Then I bolted a board over the mounting lugs to serve as a torque beam.  Sorry if the messy pile of stuff beside the lathe makes it hard to tell how the torque was measured.  I just dropped a board on the floor, put an eyebolt through it, and weighed it down with bags of lead shot.  It was just a convenient way of anchoring something to the floor under the torque beam.  A spring scale connects the torque beam to the floor.  Load on the scale indicated torque (force X arm).  In all tests the arm was 36.0" (91 cm).  If you can see it, sticking out the back was another arm to which I taped a counterweight to get the scale to read somewhere close to zero when the machine was stopped.  I still had to deduct a "tare" from the scale reading.

I brought a pile of batteries with me for this test; enough for 24V and 48V trials.  Quite a mish-mash here.  The smallest battery, from my tractor, was far too small.  It suffered a lot of gassing and spilling during the test!  This actually affected the test results, so I had to make allowance for the over-voltage when calculating the output that it would actually show if a charge controller was regulating the voltage.

There were a lot of combinations available to test.  Unfortunately, I did not have time, nor the extra rectifier, to test "Jerry" connections.  I really regret this because it was the perfect opportunity to make comparisons for myself.  Not since Flux wrote his "Matching the Load" thread on Otherpower have I seen such a clear comparison of Star, Delta, and Jerry connections, and he wasn't using a motor conversion to do then.

The choices I did have were 24 & 48 volts, series & parallel, star(Y) & delta(D).  In the end, I had time for parallel-Y in 24+48V, series-D in 24+48V, and series-Y in 48V.  I did not test in parallel-D because the vibration in series-Delta was frightening!  The lathe was putting out more noise than I've ever heard it make.  After realizing that I should have checked the data plate on the lathe's motor, I discovered that I was on the verge of locking it up!  In the end, fear, not power, was the limit to my run-up tests.  Hehehe.

It makes for quite a mess to plot all of the curves on one graph.  Here they are anyway.  Splitting them into separate graphs for Inputs and Outputs will make more sense.

Some results aren't surprising, like the output curves at 24V in parallel-Y and 48V in series-Y, which match up very well, because the increase in voltage is proportional to the decrease in current.  What WAS a surprise was more output power in series-delta at 24V than parallel-Y.  I did not expect that.  Nor would I have guessed that the Parallel-Y curve at 48V would line up with the Parallel-Y curve at 24V like it does at 300 RPM!

Another surprise is how closely the input power required for 24V series-Y and 48V parallel-Y are.  You would think that the higher current necessary to make the same power at lower voltage would incur a resistance loss, which would show up in greater input power demand.  And yet, the two curves are nearly the same.  If anything, it took more power to turn at higher voltage.  I'm still scratching my head about that.

None of the output power curves are straight lines.  They bend down at higher RPM.  This is probably a characteristic of motor conversions, where the copper is wound around iron teeth.  The windings have a certain amount of reactance to current, due to the iron laminations, which increases when the AC frequency increases.  Delta was least affected by this.  It is possible that the output would plateau at some speed.  I couldn't explore such high speeds due to the immense power required to do so.

Delta-connections seem to be the winner looking at the curves, but something I haven't mentioned yet is the vibration!  Sure there was noise during all of the tests, but nothing compared to the heavy vibrations running in Delta!  The torque measurements are, in fact, averages of the scale readings, which were often 20 pounds +/- 4 or 5.  Windmills have enough trouble with vibrations coming from blade imbalances, tip tracking differentials, cogging, and wind turbulence, I don't think I really need to add any more...  Too bad.  I can only imagine what parallel-Delta would have yielded, but alas, I didn't dare!  Perhaps if I decide to try some Jerry connections, and haul the genny and batteries back for more tests, I will risk a few parallel-delta tests to see what happens.

Here's the last graph, where the rubber meets the road, so to speak.  After wondering if Parallel-Y would be worth anything at all, here it rises to the top, though only under certain conditions.  The P-Y connection offers the highest efficiency in both 48V and 24 Volts.  There may be other combinations that give the same or more output power, but more power at the prop is needed to get there.

Which leaves me with the next question.  What size and TSR would make a good prop for each of these combinations?  Higher TSR is needed for the faster connections, bigger diameter for the less efficient ones.

Well, it seems like I've gone on and on, yet there is so much more detail to tell.  Skew, open-circuit volts, there's so much I could ramble on about.  If anyone wants more, just ask!

[FishbonzWV]

Glad to see this finished Steven.

There is one thing with the magnets that is confusing me, or is it just an optical illusion. It looks as though there is no space between the poles to allow for the width of the copper windings as per "size your magnets so they are the the same size as the hole in the coils". If this motor is wound 6-8-10 shouldn't the space be equal to the three legs of the coils? When I do conversions I measure the distance between the inner legs of the coils and size the mag pole to this distance. But I must admit I was going to experiment with adding more mag width until the power out dropped, just to see where the limit is.

Vibration.

I have my 1hp on a test stand about six feet tall and it is just weighted down with no guys attached, it is Jerry rectified and I can tell when it hits cut in because the mill starts to vibrate and then it smooths out as the rpm increases. When the weather breaks I want to switch to star to see if this happens in that configuration also.

Flux will probably chime in, but here's what he told me "forget delta...Jerry rig".

Bonz

P.S.

I'm working on a 3hp now, rotor is turned, 4 29's & 4 skewed 29's await installation, but it will accept more than that so I will have order some 'filler' mags.

[Flux]

I can't help much more here. I used wound field machines for years and in some ways they will behave in the same way as pmg conversions but in other ways they differ.

I found that with wound fields and the necessary small air gaps to keep field current reasonable the things always reactance limited and at very low speeds the things became very large and heavy so I resorted to gearboxes and all their evils ( they really are evil).

I was surprised by what Zubbly was managing to squeeze out of these conversions, but he never gave any input power figures and I assumed that the end efficiency was low. Steven's tests confirm my suspicions. The reactance limiting factor is still present but much reduced because of the very large effective air gaps of the saturated magnet. The effect may be more in the nature of armature reaction than reactance but the end effect is the same.

Although it is possible to get a fair % of the original motor rating as a pmg it seems to come at a price of very low efficiency. The fact that they survive must largely be a factor of the low wind power duty. No way is something running at 30% efficiency going to stand a full load heat run on a bench test for many minutes.

The input characteristics are similar to the wound field machines but the output holds up far better. The curves turns over but doesn't become horizontal within the thermal ratings of the windings.

Probably the best magnet to space combination for cost/output is with spaces about 1/2 magnet width and this will roughly be the same as magnet width fitted to the smallest coil in a distributed pole set. Increasing magnet width right up till they touch will probably increase output but at a higher cost. When dealing with available magnets it is difficult to get the best combination and extra magnet will be far better than going the other way.

Regarding vibration on load this depends on many factors. It may be that near cut in Jerry connection will have some of the vibration issues of single phase but it should never be as bad. Star may be the smoothest but with rectifier loads I doubt that it will ever be as smooth as resistive loading onto the ac. Delta could easily be the worst for vibration and it will become much more marked as the rectifiers start to conduct. With jerry there will be no vibration below cut in and that should be true for star.

Although it seems possible to get very large outputs at low speed I think the same issues remain that I found with low pole number wound field alternators, the size and weight gets excessive and if you try to get efficiency up then it becomes even more so.

There is little doubt that iron cores get you more output for less magnet but for low speed the size and shape of a standard high speed motor core leaves much to be desired. The teeth are too long and the space for magnet is too limited, using many more poles with smaller teeth seems the way to go and the things should be short and large diameter. Taking standard motor cores is cheap and convenient but the things are really better suited to higher speeds.

Flux

[Bobbyb]

How did u gather the power input figures, did u measure lathe input power or calculate it with torque an rpm?

[SparWeb]

The case of the generator was free to rotate when the shaft was held in the lathe chuck.  When the board is bolted to the feet of the motor casing, a force is required at the end of it to prevent the board from swinging around.  I measured this force, and the length of the arm from the center of the shaft to the scale.  Force times distance equals Torque.  Torque multiplied by RPM and an appropriate conversion factor equals power.  That is how I measured the power input into the machine.  It's a bit rough because vibration makes the spring scale jiggle around a bit, and spring scales aren't very accurate anyway.  That would be why there are bumps here and there on the input power curves.  

[SparWeb]

Hi Bonz,

...I was going to experiment with adding more mag width until the power out dropped...

That sounds ambitious: measuring the power in successive runs, changing the size of magnet pole every time.  I doubt there is a serious penalty for having magnet poles that are "too wide".

The rows of magnets are all evenly spaced on my conversion.  I agree that the size of the pole would match the stator if the magnets were clustered closer together, but the size of magnets basically did not permit this.  There is only about 1/8" space between each on the rotor.  It's possible, but my mediocre abilities would be hard pressed to do a good job if I was varying the increments of angles for each facet.

Your 1HP vibrates in Jerry-connection?  That's unfortunate.  I was considering experimenting with Jerry once the windmill was up the pole, as a consession to not being able to try it on the lathe.  Thank you for the warning, I will be cautious doing that.

[FishbonzWV]

Quote: The rows of magnets are all evenly spaced on my conversion

Are they grouped as a four pole?

I'm kinda confused about this, I would think that you would have to have the N pole coming to one leg of the coils at the same time the S pole is intersecting the other leg.

I just measured the 3hp stator and the three legs in slots 1,2,3 are an inch and a quarter wide (30 mm), so to me there should be at least that much space between the poles so the flux is not canceled out by the next pole coming around. Is this not correct on a conversion?

Another question comes to mind about the flux density.

I think I recall back when you and Peter were doing the FEMM modeling that one of you mentioned over saturation in one of the models. Do you think that could increase the reactance, limiting power output?

I am only using 1/4 inch thick mags in my conversions even though the coils are 3/4 inch thick. Is there a balancing act to hit the sweet spot?

Hmmm, I think this is going to keep me in the garage (or should I say dog house, with the missus) for months.

The vibration is only right at cut in. As soon as the rpm picks up it smooths right out.

[SparWeb]

..Are they grouped as a four pole?..

Yes.  I didn't get into those details because I've written about them before in other threads.  You could look at those in the link at the top of my posting, if you didn't already see it.

For each pole, there are 6 magnets.  Three around, and two long.  All evenly spaced because it's just easier to machine the flat faces.  That makes each pole cover about 90 degrees of the stator.  Since there's a space between each magnet, the gap to the next magnet in the next pole covers only 81 degrees.

Now there's no simple answer to your question, because your 1HP motor may have 24 slots, your 3HP may have 36, and here is my 7.5 HP conversion, whose stator has 48.  When I looked at the wire, each coil covers 10 teeth, enclosing 75 degrees.

Should I have tightened up the magnets on each pole?  Maybe.  Also note that the wiring coils are all the same size.  Each coil wraps around 10 teeth, and there are two per pole, the second one being one tooth clockwise of the other one.  So actually there are 11 effective teeth per pole, for an angle of about 82.5 degrees.  Now the spacing issue doesn't look so bad any more.

If you have a 36-tooth stator, then the windings may be different.  It could be 10-8-6, like you say, but it also could be 7-7-7, where three coils that each enclose 7 teeth are layed in together, covering 9 teeth total, and again you can have 4 poles like that (4*9=36).  Maybe you would want to squeeze the magnets into 80 degrees, in that case, but the difference is pretty minor, again.

Then there's the skew spreading the magnets out...  I just had to give up on optimizing and just get it built!

I may not have explained this too clearly, so maybe I should make a drawing.  It would probably be a good idea because this can be hard for anyone to visualize.

About the magnetic saturation.  My limited understanding leads me to believe that a strong magnetic field applied in one direction (from the magnet) is counter-acted by a magnetic field in the other direction (current in the coil of wire).  When the current is low, this de-magnetizing field is also low, and there isn't enough of it to bother the magnetic flux sweeping through the coil.  The Neo magnets are just so strong that the small counteracting field is overwhelmed.  When larger and larger currents flow, however, the de-magnetizing field subtracts some flux from the Neo magnet's field, reducing the net flux that can be linked through the coil.  The power no longer increases linearly, and the output curve leans over.

Saturation delays the point where that happens.  It could be this effect happening that you can see in the output curves I posted above.  Then again, the system is much more complicated and already I'm on thin ice with this explanation.

Looking for the sweet spot:  I'd suggest you try your hand at FEMM, or something like it.  You can try new iterations in minutes or hours, not days, and learn a lot about those fields and fluxes as you go.

For the Toshi conversion I ran many FEMM simulations to select the magnets I used from a variety of other options.  I would be happy to help out if you don't get the hang of using FEMM right away.

[vawtman]

Hi Steven

 I hope Toshi survives longer than mine.I built a 5hp4pl conversion a couple years ago using arc seg mags.Kinda like the ones you have(used 4 per pl)i stayed within the coil span for each pole and they were skewed(backspaced like yours)I could turn the thing with 2 strong fingers but when it was hooked up to the tubine geared 1;3 it was amazing how that slight cog reverberated through the blades and tower.It lasted about 2hrs and was gone.I wish you could get a tax rightoff for these things

 Are you following Peters and Jacs method of decogging?I never could understand how that could work without weird circulating currents if one used the factory stator.

 Just me

 Mark

[SparWeb]

...Are you following Peters and Jacs method of decogging?

Well I think it's more like Zubbly's method of skewing magnets, though his later conversions had skewed stator laminations, not skewed magnets.  All I did was machine the rotor to have facets, one every 30 degrees, all around.  With two rows, one of the rows is offset by 7.5 degrees from the other.  The 7.5 comes from dividing the 48 teeth into 360 degrees (360/48=15 deg) and half of that is 7.5.

A friend of mine has suggested that I gear this one DOWN to get it to run at 24V.  That's a surprising idea that I can't shake out of my head!

[Flux]

Yes I agree with all this.

If you are dealing with coils wound in one slot then the magnet should fit within the coil hole for maximum voltage for a given volume of magnet. The waveform will be basically rectangular with modifications due to practical effects.

Large low pole number motors are not wound that way, partly because huge slots would be useless and more so because the rectangular flux distribution would make a noisy motor. Two methods are used in motors, concentric windings are just that,the coils are wound in slots such that the coils of a phase are concentric. There may be perhaps 4 coils in a pole/phase group. You could choose magnet size to match the smallest coil and that may be the least prone to cancellation but it will also give the least flux. If you choose magnets to suit the largest coil then it is less effective for the inner coils but correct for the largest one and it gives much more total flux. To me it would seem logical to make the magnet suit the largest coil of the pole group. Whatever you do you will almost certainly do better with more magnet and more flux. The waveform can no longer be rectangular and you are now dealing with changed waveforms and the actual peak voltage is not the important issue. When you rectify the waveform goes to the dogs anyway so all the nice distribution needed for the motor winding doesn't apply in the same way ( and it may vibrate a bit more but that is no big deal).

The other method of winding uses identical coils and the distribution is obtained by overlapping the coils as a 2 layer winding. The end effect is the same and once again the magnets need to be bigger than the hole in an individual coil.

For conventional alternators the pole arc tends to be about 2/3 of the pole pitch ( gaps equal to magnet width). This makes sense when you look at 3 phase and find that at any instant 2 coils are conducting and one virtually dead. I don't think erring on the too wide a magnet arc does more than make the cost/ output factor of the magnets a bit worse. I think you still gain output.

The basic fact with motor conversions is that there is too little space to get enough magnet in and if you use the perfectionist approach then you are way short of flux especially on a machine that you are trying to screw 5 times its normal rating from ( running it way below intended design speed)

I agree with Steven on the saturation issue, for light load you can reduce the volume of magnet drastically with little or no loss of voltage. If you do this you find the armature reaction acting against your magnet flux and the nett flux falls and also it is distorted and doesn't link all the turns of all the coils and leakage reactance is increased. The output will drop rapidly with load and it will eventually tend to current limit with losses going up with speed but no increase in output with speed.

If you use lots of magnet and completely saturate the teeth then it behaves as an air gap alternator like the axial air gap machines. If you only partially saturate the teeth then you have the characteristics of both with the linear load line of the air gap machine and the curved iron core machine load line and the thing may never reach the constant current point within the winding rating.

I came to the conclusion some time ago that slotted cores are not as wonderful as many believe for wind power now that neo has let us get decent flux in large air gaps. I prefer to spin the magnets on the outside of a laminated iron core where there is more room for decent sized ones and the teeth are a doubtful virtue, contributing low load losses and saturating and wasting winding space. It may make a cheaper machine with teeth but I find that winding the coils in the gap between the magnets and a smooth core ( Hugh Piggot original design for example) gives a better machine.

Axial construction introduces mechanical issues that probably justifies keeping the simple single layer partially wound construction that is generally used. but that is another issue.

Flux

[FishbonzWV]

Steven

I looked at the FEMM program back when you and Peter were discussing it, but I was under the impression that you had to Cad a drawing up and use it for the modeling.

I don't have any Cad programs installed so I didn't pursue any further.

Another question.

I was going to experiment with putting 8 mag poles on one of my 1hp 36 slot 4 pole motors.

Would I be wasting my time?

[Flux]

Yes

If it is wound 4 pole it will only work with a 4 pole rotor. Rewind or don't try.

Flix

[SparWeb]

Woah totally botched math there!  Gotta start again.

360/48 = 7.5 degrees.

Half of 7.5 degrees is 3.75 degrees.

Due to the limitations of my machining skill and to keep track of what I was doing, I actually machined the facets with 4 degrees skew.  

Sorry to have mixed that up.

[ghurd]

Good numbers.  I wish I could do that.

I would be interested in a test of individually rectified.  Regular old "Jerry Rigged".

http://www.otherpower.com/images/scimages/2050/JerryRig.gif

I expect it will not vibrate worse than Star.

G-