mini motor conversion finished

[dinges]

About a week ago, whilst on IRC, I thought to myself "hey, instead of chatting, I'll quickly convert a small motor I have". Expected it would take me 1.5 hours at most. Today, 8 days later, it got finished.

The main purpose of this conversion was proof of concept of the JacquesM-method of offsetting magnets. See these links for more information:

http://www.fieldlines.com/story/2005/1/8/2291/67790

http://www.greenbits.com/images/windmill2/rotor2.png

http://www.fieldlines.com/story/2004/12/25/34425/916

http://www.fieldlines.com/story/2004/12/24/6105/5834

specs of the motor:

1 phase 230VAC

(2nd starting phase, with 2.5uF cap)

0. 23 A
   
1500. RPM (4 pole)
      
      

Shimano-Tokki corp., made in Japan,

X7807-201V

The principle of the Jacques' method of offsetting (skewing would be the wrong word for it) is simple. If you have 16 slots (as in this case), each pole is offset by a fraction of the slot-width (in degrees). I.e., in a 16 slot motor, each slot is 360/16 = 22.5 degrees. The first pole is at 0 deg (i.e, 12 o'clock). The 2nd pole is at 3 o'clock, but a little less: 1/4 * 22.5 = 5.6 deg; i.e. at 84.4 deg. The 3rd pole is at 6 o'clock, but minus 2/4 *22.5 = 11.25 deg; i.e. 168.75 deg. The 4the pole is at 9 o'clok, minus 3/4 * 22.5 = 16.85 deg; i.e. 253.15 deg.

See the drawing below

The downside of this conversion method is that it introduces an imbalance in the rotor, which must be compensated for. I haven't done this yet, but intend to do it by drilling holes in the steel rotor until the assembly balances. BTW, the more stator slots there are, the less the imbalance becomes; in my case of 16 slots, the a-symmetry of the rotor is quite high, thus is imbalance.

The conversion itself consists of 4 round magnets (one per pole), of N42 15x8mm. This gives a total magnetic volume of .345 cubic inch.

A new steel rotor was turned, an aluminium ring crimped over it. The assembly was Loctited (Loctite 638) on the old axle. The bearings were 608-2Z, they were replaced by 608-2RS for better weatherproofness. The pockets for the magnets were milled at 15.2 mm, 10 mm deep. The .2mm play is needed for the epoxy. I used epoxy to glue the magnets, not Loctite 638, because the datasheet is unclear whether Loctite holds to 'inactive' metals. I'm not sure whether nickel is (in)active. Henkel (mfg. of Loctite) didn't bother to reply to the question).

Some more pictures:

The goal was to provide proof of concept of the Jacques M-method of offsetting. It's an utter succes in that respect:

Not even the slightest hint of cogging is present.

As I write this, I'm trying again to turn the axle, but nope: not even the faintest trace of it can be felt. So the offset method works as it should, just as I expected from theory. The advantage is that it's much easier to do on a 3-axial CNC mill, as opposed to Zubbly's 'helix' method. The downside is the imbalance it introduces.

With magnets, it doesn't turn as freely as it did as a motor; there's quite a bit of 'sogginess' or 'stickiness', I suspect this to be the stator losses. When I give it a good spin, it turns freely only for about a second. I will be making torque measurements to get more accurate data of this.

The results, powered by a 300/600 RPM cordless drill (calibrated with a RPM counter before measurement):

300. RPM  -- 20VAC; 70mA
     
600. RPM  -- 45VAC; 140mA
     
     

As you can see, not exactly a powerhouse.

The stator resistance of the original winding is 300 ohm. I don't plan on doing a rewind, but who knows, maybe I will one day.

At 1500 RPM, it should generate 112VAC; whilst it took 230VAC to drive as a motor. I suspect I could have used more magnetic volume to get more V/RPM. However, rotor size prevented me from adding another row of magnets.

Power output is about 26W/cubic inch. Zubbly's rule states 150W/cubic inch. Mine is much less, then again, Zubbly's values were gathered from 'real' motorconversions (several hp's), whilst my mini-motor is just a plaything in comparison.

The output will be at about max. 90VAC. I have several DC-DC converters that accept 38...90VDC in, and give 12V at a few amp output. Maybe I'll let this motor drive such a converter.

Ok, that's it. The motor was never intended to be married to a prop and put into service, but maybe I will. The most important conclusion being, 1) JacquesM's method of offsetting works (just as others have stated it should) and 2) completely coggless motor conversions are not an illusion.

Thanks to Zubbly, RonB, JacquesM and everyone else on this board. Without Fieldlines, I'd probably still be messing with car alternators and DC motors.

Any questions, don't hesitate to shoot.

[dinges]

correction:

the motor is rated at 0.23 A, NOT 1.23 A.

Scoop thinks it knows more about my motorconversion than I did...

[wdyasq]

Maybe Scoop knows which engineer IS the idiot.....

Ron

[oztules]

Nice work dinges,

have you got any figures (calculated or tested) on the electrical attenuation from the 90 degree (or proper, but cogging arrangement) compared to this cogless and slightly less(?) powerful arrangement..... ie what power losses are thought to occur due to this mismatch of magnets..... (feel free to make another rotor to test this )

oztules

[dinges]

Oz,

It's a good idea. I've got about 8 of these motors (8" FDD motors...) lying about.

There's little investment in materials (4 magnets of only .5E/pce or so).

However, I can't mill myself; I need help from a friend with a machine shop, and he's over his head in work. Last thing he needs is me nagging him for my little experiments.

What I do have in mind though, was building another rotor, but using the Zubbly-method of skewing, that I've slightly modified (i.e., skew slightly less than one slot; see me diary for a (long) discussion on that).

Maybe I'll try to find a 15.2 mm dril and grind it down so I can use it as a poor man's mill, to arrive at a flat bottom, after having drilled with a plain 15.2mm drill.

But, the loss in output probably goes by the cosine or the error in angle; error angle is quite small. So will be the reduction in output (cos (angle))? I expect it to be very small.

[gizmo]

Well done dinges.

I posted something about JacquesM-method on my web page a few weeks ago, didn't get a lot of feedback at the time, but I still believed it would work.

http://www.thebackshed.com/Windmill/FORUM1/forum_posts.asp?TID=222&PN=2

I play around with the F&P motors, 56 magnets on 42 poles, and I believed they would gain the most from this methode of decogging. The magnet hub for the F&P is plastic with ceramic magnets. When we replace the magnets with Neo's, the output power is almost doubled, but cogging is real bad and the plastic hubs start to buckle. A forum member, Trev, is working on a aluminium magnet hub to replace the plastic unit, and I think we should look into using JacquesM-method to eliminate cogging.

Its good to see someone gave the JacquesM-method a go. You need to get the maths right, but it solves one of the biggest problems with home made windmills.

Glenn

[vawtman]

Peter 1 round mag per pole is a form of skewing.Wouldnt you think.No offset needed.

 Zubbly has the right idea for these.I think.

[dinges]

Hi Glenn,

Had a look at the link.

I think it's very easy: on that F&P, I count 42 stator slots. This means each slot occupies 360/42 = 8.57 deg.

Now, the goal is to divide those 42 magnets over a circumference of 360 - 8.57 = 351.43 deg.

This means that the angle between each magnet is 351.43/42 = 8.367 deg.

That's the theory, easy enough as you can see.

Now, the practice, of mounting these magnets at the precise position, is up to you.... That would be the hardest part, I think. Because the corrected position of the magnets is 8.57 - 8.367 = 0.203 degrees. (.2 deg would be fine, I guess)

Admittedly, F&Ps have a relatively large diameter, I think (have only seen pictures of them). Assuming it's 30 cm (just my guess!), the magnets would be spaced at:

circumference * 8.367/360 = 300mm * pi * 8.367 / 360 = 21.9mm.

Without offsetting/decogging, they'd be at 300 * pi * 8.57/360 = 22.43mm.

(these are heartlines of the magnets)

Doable, I'd think, but you've got to work accurately.

[dinges]

Don't think no offsetting/skewing will do the trick, not with round magnets, but certainly not with square ones. Maybe Zubbly can chime in.

If you try to visualize what happens when the rotor rotates & magnets pass the stator slots, you'll see that round magnets will have a 'thumpiness' to it: they don't see the same amount of stator iron as they turn, thus the cogging.

Zubbly definitely has worked out a few solutions: skewing round & rectangular magnets and skewing the stator. There are other solutions too: JacquesM used two, one was the offsetting method described above, another one was two identical & symmetric rotors, that can be twisted w.r.t. eachother on the axle, to compensate for any cog. Adjust the angle between the two rotors until cog is zero or minimal. He didn't have to use the last method, since the offsetting method worked fine & eliminated cog.

[ghurd]

Looks good!

Lots of thoughts, other than that 300 ohms is killer.

Are you useing both windings?

It would be interesting to me to see both windings, in series, on a scope.

It looks like the air gap is pretty large if the left hole in the photo has a magnet in it.  The volts drop fast in these little ones with a slightly larger gap.

I would have expected more Vopen.

'sogginess' or 'stickiness'

I don't know about bearings.

The magnets being 'off' will also effect how thay pull toward the laminates, correct?

Maybe part of the drag is caused by pressure on the bearings being not so even?

Has anyone compared this to Zubbly's in this way?

The 12 o'clock magnet is Zubbly's 1st magnet in the row, near one bearing,

the 9 o'clock is Zubbly's last magnet in the same row, near the other bearing,

the others are in between.

Sort of looks like different sides of the same coin to me.

You must put some blades on it.  I think that is a law...

G-

[commanda]

The output will be at about max. 90VAC. I have several DC-DC converters that accept 38...90VDC in, and give 12V at a few amp output. Maybe I'll let this motor drive such a converter.

Peter,

I hope you realise it's just not that easy. The output of this converter will regulate to exactly 12 volts, and drive as much current as it can to achieve that.  Your nominal 12 volt battery will vary from 10.5 volts dead flat to 14 volts at the peak of bulk charge.

What you need to do is electronically adjust the output voltage of the converter to get a battery current such that the converters input (power in = power out) will match the generators output at whatever rpm it is doing at the moment.

I really hope this makes sense.

Amanda

[dinges]

Ghurd,

I'm only using the running winding. There are only 3 wires coming out, the start-winding is internally connected to the running winding. I could perhaps disconnect, if I really wanted to. So, I can't give you a scope picture of outputs in series. The output voltage of the start winding is less than the running winding, BTW.

When I short the two running wires, you can definitely feel that it's a single phase generator. When I short the 3rd wire (starting winding) as well, it's smooth again, like 3phase, but only very heavily loaded, thus hard to turn.

The airgap is about 1 mm, not much more. The magnets really aren't deeply into the rotor, not more than .5mm, whilst airgap is .4mm (1mm difference in diameter; stator is 41.0mm, rotor is 40.2mm).

I expected more Vopen as well... Then again, the stator is 25mm wide, whilst my magnets are only 15mm diameter; so, 10mm of stator is basically left unused.

The stickiness/sogginess is not the bearings, nor the uneven drag of the magnets. You'd have to load the bearings very heavily to get that amount of drag, LOL. I always test my bearings (all are 'pre-used', thus some have a grinding sound or got damaged while removing; these passed my rigorous testing (I gave them a spin with my fingers...)) BTW, with the old 608-2Z bearings the drag was the same. Any sideways force on the rotor, due to magnets being off, is bound to be pretty small. Not enough for this kind of drag of the bearings.

It's a completely different method of reducing cog from Zubbly's method, but has the same effect: no more cog. It's theoretically sound too (if I weren't pretty sure that it would have worked, I wouldn't have bothered to try).

Feel free to donate me a good set of blades, and I'll see what I can do...

Seriously, haven't given the matter much thought. Pmax=40W; RPMs, see original post. What kind of blade (dia, TSR) do you suggest, if I want this thing to be a non-furling genny?

BTW, I just found out from the nameplate of another, identical motor, that it's a 6W motor... (I was assuming that 230V * .23A = 50W would be the max. power)

[dinges]

Ghurd, you're right (ouch, why does it hurt when I say that?...)

airgap is more than the .9-1mm; it's 1mm at the edge of the magnets. In the middle, it's more.

Didn't know that esp. small motor conversions were that sensitive to airgap, as you claim.

[dinges]

You're totally right of course (ouch, there's that stinging pain again...)

I had another issue in mind: output of the converters is 12.25 V, but not adjustable by potmeter, nor by changing resistors (cast in resin). 12.25V is too little to be of use as a charger.

Also, I'd have to make something (windowcomparator and some electronics) to connect it:

0. -38V --> go to linear regulator (LM317 or something) to get 14.4V and/or constant current.
   
38. -90V --> go to DC-DC converter (but: 12.25V output...)
    
    

>90V --> add dumpload AT genny to get below 90V again.

Quite a bit of stuff and not worth it for such a plaything, IMO. But something that definitely needs to be done with the 130W & 3hp conversions that are WIP.

Now, if you'd have designed a decent MPPT with DC-DC converter, I might do it. So, what you waiting for? Get to it! If not, I'll have to rewind that 3HP. And I'll be blaming you for that, as I'm getting cuts in my hands as I remove the old coils, and I'll be blaming you again as I put the new coils in and get the just-healed wounds in my hands open again. And I'll make pictures of my blooded hands and post them in the diary, to give you a guilty conscience...

[commanda]

Hmmmm......

The mppt is basically done, short of real-world testing.

Dc-Dc converters are a new field to me. Steep learning curve.

I've got a commercial 240 volt in, 24 volt out 500 watt smps that I'm currently playing with. It has an adjustment pot for the output voltage, so I should be able to hack the voltage feedback to make it electronically adjustable.

What I've really got to do is build a flying wind turbine for some real testing. Might have to build a vawt similar to what hangar51 recently posted. Maybe with Lenz2 blades.

Amanda

[Flux]

Peter the sogginess is iron loss, at last someone has discovered that it exists with this type of alternator.

With your small magnets and low flux it is not likely to be very great but if you used the maximum amount of magnet and brought the flux level up to a reasonable value it would increase quite a lot.

With such a small motor it would be very inefficient as a motor so you may never be able to throw enough magnet at it to reach the nominal volts at nominal speed, but even so I don't think your flux density is likely to be very high.

I doubt that there is any significant difference in output between your displaced magnets and evenly spaced ones.

Flux

[dinges]

"Peter the sogginess is iron loss, at last someone has discovered that it exists with this type of alternator."

I've read about it before, just didn't expect it to be this large, i.e. very noticeable, also at small RPMs.

In fact, whether I turn fast or slow, the torque required seems to stay the same. Whether I turn at 0.1 RPM or much faster. I had expected that, as you turned faster, the torque would increase.

It seems that the torque is constant, but as RPMs increase, powerloss increases (P=F*v; or P=F*omega*r  with omega = 2*pi*freq), I think. Still, I need to make a measurement of the torque. Then, at various RPMs, stator loss could be calculated.

Anyway, this is a constant drag/torque, NOT cogging. Thus, during startup, no extra hurdle would have to be taken to get the prop to rotate.

[Flux]

Correct, with no cogging there is no problem getting the prop to rotate.

There may still be a problem getting it up to speed through stall at low wind speeds.

Until the stall torque overcomes the drag torque it will sit rotating slowly.

unless cogging torque is bad enough to lock the shaft I suspect this drag is as much an issue as the actual cog. It was in the case of magnet steels with iron poles.

Without the iron poles you may have significantly less drag if the neos are not as conducting as mild steel poles.

Flux

[ghurd]

Still have some HD magnets?

Just for fun, snap 40% off each end (the middle can have strange properties) of 2.

Place those 4 pieces on the center of the existing magnets.

Maybe a wrap of black tape to hold them.

Then check it again.

Blades for this kind of thing is easy.  PVC.

After a week, you will want to change blades or put up the next mini-conversion anyway.

G-

[dinges]

"There may still be a problem getting it up to speed through stall at low wind speeds."

Is this really an issue? Even if it would rotate in low winds, there would hardly be any energy in these low winds to generate anything useful.

It's true, axial fluxes rotate more freely & easily. My mini axial flux genny gen freewheels for about a minute (then again, more moment of inertia due to large rotor dia). But load it and it'll stop fairly quickly too.

If you'd like to extract every bit of energy from very low winds, then maybe one should use an axial flux. But as far as durability, reliability, weatherproofness, etc. goes, that there are not many homebuilt axial fluxes that can compete with a professionally engineering motor (-conversion). Not that it's not doable, but I doubt many can and do build their axial fluxes like that.

It's probably a compromise, like many things in life. We can (and should) reduce cog, but stator losses are there to stay. Unless I find myself some fancy stator iron with low losses....

(BTW, there exist high-efficiency motors; do they use special stator iron? That can perhaps stand more magnetisation, 1.7T instead of the usual 1 T? With low hysteresis losses? What kinds of motors would have the smallest stator losses? If anyone knows, I'd be interested)

"Without the iron poles you may have significantly less drag if the neos are not as conducting as mild steel poles." Me no understand. No iron poles = air core genny. What has conductivity of neos to do with it?

[altosack]

Hello Amanda,

The mppt is basically done, short of real-world testing.

Dc-Dc converters are a new field to me. Steep learning curve.

How can the MPPT be done but not the DC-DC converter ? Aren't they integrated, i.e., you are varying the duty cycle of your P-MOSFET in the converter to get your MPPT ? I've seen your description of your 2-stage converter before; I'll ask more about why you're doing that some other time.

Also, how are you implementing MPPT ?  When I look at a feedback loop for a wind generator with power (or current) as the input, and the duty cycle (for example, of a buck converter) as the output, there is feed-forward, i.e., as soon as you increase the duty cycle to get more current, you instantly get more current before the rotor RPM (or TSR, if you will) has a chance to react. Of course, the RPM (or TSR) is what you're trying to control, but it reacts quite slowly, relatively, and the only way I think you could overcome this is by doing a lot of trial and error with the timing of your duty cycle changes. However, then you may get the problem of a system that doesn't react well to gusts of wind (conjecture, I haven't tested it).

The only way I've been able to overcome this (in the design that's in my head; I haven't build a wind generator yet !) is to use RPM as the input. Well, RPM, battery voltage, and a bunch of constants specific to the wind generator it's hooked to. It's pretty computation intensive, but it's well within the capabilities of an Atmel ATmega48, which is what I'm planning on trying to use (I believe a PicAxe is too slow with its interpreted BASIC, but I haven't timed it out, and I have a bit of experience with both C and assembly, anyway).

Using RPM, I still run into the problem of electronically limiting the RPM or the current (both are things I'd like to do if I'm going to the trouble of having an electronic controller !). The problem is the same feed forward as described above; as soon as you increase the duty cycle to push it towards stall and limit power, you get more power instantly, until the momentum of the rotor backs off. I think I could live with slowing the duty cycle changes way down in this situation, though, but I think it would take quite a bit of trial and error, and I'm not sure it would end up being a robust and safe way of limiting power (there's always furling !).

I've looked at the web quite a bit trying to find info on how the "professional" guys do it, but I haven't been successful. Any ideas you could give me or point me to would be appreciated.

Thanks,

Dave

[dinges]

I think I'll convert another one of those motors, but now by the Zubbly method. See drawing below. Only difference being, the magnets are not exactly under (different) slots, as is commonly said to have to be done. One magnet is exactly under a stator slot, the next one (on the same pole) is exactly midway between two statorpoles.

This way, when the rotor rotates, it should see the same amount of stator iron. It was discussed in another diary entry or mine.

Anyway, here's the preliminary drawing. There's slightly more magnetic volume, but a bit of it is ineffective as it sticks outside of the stator.

[commanda]

Dave,

How can the MPPT be done but not the DC-DC converter ?

Black box design approach. Mppt prototype in my diary in this thread.

http://www.fieldlines.com/story/2006/3/12/14840/1315

Outputs a modulated dc voltage to control the pwm of the converter.

Copied & pasted from the other thread.



A 4046 PLL is used as an oscillator to modulate the control voltage for the pwm. The 3 small op-amps give a signal derived from the generators output power. The first is a current amplifier, the second is an OTA analog multiplier (current times voltage = power), and the third is a differentiator. The 4046 includes a phase comparator. The 2 inputs to the phase comparator are the modulation signal and the power signal. The output of the phase comparator is integrated & becomes the dc control voltage for the pwm. There's also an input from the tacho, to quickly push the control voltage in the right direction as the wind changes.

So, if the pwm modulation and the power feedback signal are in phase, we keep increasing the pwm control voltage. If they're out of phase, we decrease the pwm control voltage.

I think it may help conceptually if you treat the issue of limiting power at the top end as a seperate issue, at least initially.

Considering that I'm running the modulation signal at about 2Hz, I'd say a picaxe was more than fast enough.

I haven't published the circuit of the mppt controller on here because of a lack of real world testing; as you rightly point out there are issues to do with rotor inertia, etc. I am, however, happy to share it with those who are interested and display some grasp of electronics ability. Let me know if you want a copy.

Amanda

[ghurd]

When I said "After a week, you will want to...

put up the next mini-conversion anyway",

I figured it would take you a few days to decide to make another.

G-

[Flux]

I agree with Amanda, don't confuse the power limiting with the operational mppt.

If you are using speed, then above a certain speed, drive the pwm signal forward regardless of what the mppt is asking, to bring the prop into stall.

I doubt that you will hold it above 50 mph in stall without a monster alternator so if you reject variable pitch I think you will need to furl.

If the alternator is big enough and you can accept the thump, then a final stage could be dynamic brake.

Flux

[Flux]

You would do me a favour by measuring or calculating this iron loss. Preferably with a modest amount of magnet and also with the maximum magnet you can get in.

Loosing 30W in low winds from iron loss is not a big deal as long as your prop can pull out of stall with that 30W of drag at a wind speed where it can generate when it is out of stall. It is annoying to see it crawling round all day with winds that would generate power if you bump started it and got it above stall.

Neos are short and of such shape that you can fit them directly so they don't need or benefit from pole shoes.

If you used earlier magnet materials the magnets would be long and thin and you would need steel pole shoes to bring the area up to a suitable point to match the coil span in the air gap. These pole shoes are an extension of the rotor iron circuit and nothing to do with the presence of iron or lack of it in the stator.

From the magnet point of view it wouldn't matter if they were on the pole tips or on the rotor core below the magnets. At first sight they are part of the rotating magnetic field and a rotor made of neos without shoes or of Alnico with shoes would behave exactly the same.

The ripple in gap flux caused by the slots (with their high reluctance)means that there is an eddy current loss in these poles even though they are rotating with the field.

The same is true of neos, but apart from the nickel plating, I suspect their resistivity is higher than mild steel. Also with neos it is easier to get a reasonable flux with a larger air gap and the size of gap has a considerable influence on this induced loss.

There is likely a strong case for using thicker neos than actually required and working with a larger air gap if this surface loss is significant as it is with steel pole shoes.

It is all these issues and their relative effect that I am trying to sort out so that you fellows can make better motor conversions. I am not trying to discourage you from doing it.

I am sure high efficiency motor cores are better, but core material was only part of the change. Core material has a significant effect on iron loss in the stator but if you are causing significant surface loss in the face of the magnets, stator core material will not affect this.

Flux

[dinges]

Ok,

Due to overwhelming demand (Ron & Flux) I've made a torque measurement of this mini-conversion. It's a complete write-up, so anyone else wanting to measure & calculate stator losses can simply copy what's done here. For those only interested in the results, skip to the end.

The measurement setup:

I've temporarily installed a small wheel (diameter 22 mm; radius 11 mm) on the axle. With a black sharpie I put a marker on the white nylon wheel. I took a piece of string, about 1 m long, and wound it around the wheel. Then weight was added to the wheel. The trick here is to add so much weight that it unwinds slowly, evenly. Shouldn't accelerate or decelarate. I used various Neo magnets as weights, they're easy to clamp onto the string.

In this case, 2 20x20x10mm neos were needed, that's 70 gram of weight.

The assumption is made that all power is needed for overcoming statorlosses. Any other losses (in the bearings, air resistance...) are assumed to be negligeable.

The calculations:

1)  P = F * v

P [W]  (power)

F [N]  (force)

v [m/s](velocity)

2)  v = omega * r

omega [rad/s] (angular velocity)

r [m] (radius of the wheel; here 11 mm = 0.011m)

Combining 1) and 2) gives

3.  P = F * omega * r
   
4.  F= m * g
   
   

m [kg] (mass of the weights, here 70 gram = 0.07kg)

g = 9.81 m/s^2

So, F = 0.7N

r = 0.011m

Entering this in equation 3) gives:

P = 0.7 * omega * 0.011

P = 0.077 * omega

So, we now have a function of Power as a function of omega, the angular velocity.

we need another equation:

5)  omega = 2 * pi * f

where f [Hz] is the rotational frequency, in rev/s. To convert from RPM, we need

6)  n = f * 60  (or: f = n/60)

n [RPM]

f [Hz]

For various RPMs, we can now calculate omega and then power.

n = 100  ; omega = 10.5 rad/s; P = 0.08W

n = 200  ; omega = 21 rad/s  ; P = 0.16W

n = 500  ; omega = 53 rad/s  ; P = 0.41W

n = 1000 ; omega = 106 rad/s ; P = 0.81W

n = 1500 ; omega = 159 rad/s ; P = 1.22W

n = 2000 ; omega = 212 rad/s ; P = 1.63W

Above we see that, at 100RPM, the loss in the stator is .08W; at 500RPM it's 0.41W, etc. For all other RPMs we can interpolate, since the relation between RPM and stator losses is linear.

At 600RPM, I got about 5W output from the generator (I should make better & more accurate measurements of this), so stator loss is .45/5 *100% = 9% of output. I assume that this is a linear function of output: I.e., double power output (because of double RPMs) leads to double stator losses, but the stator loss stays the same percentage of total output.

There's quite a bit of uncertainty in these values, esp. the measured output power of the genny. All the other measurements I expect to be pretty accurate.

At the moment, I'm using 10% stator losses as 'the' value for this genny.

In conclusion, it's a pretty simple measurement technique, quick to set up. The calculations are also straightforward. I'll be making these measurements on all motorconversions from now on, as it gives quite a bit of insight in the performance of the genny. It would be nice to know the stator losses in other people's conversions too. 10% loss seems a reasonable price to pay, but it's 10% extra loss that you don't have with axial flux gennies...

Thanks to Ron for pointing this technique out to me.

Any errors or corrections, don't hesitate to correct me (as if I need saying that on this board...)

[dinges]

Your wish is my command...

See last post for the measurements and calculations for this particular genny.

[willib]

Peter , which one of the RPM Vs Power loss calculations , were actual measurements ?

it is nice/essential to have at least two measurements to make a graph

nice work BTW

[ghurd]

Part of this drag is from the coils, not much I expect.  

I couldn't measure any difference between the motor with coils or without.  But that one cogged a bit, and I quickly gave up trying to test it.

It was recommended any unused coils be removed to reduce drag.

After one of these motors gets a little more magnet, I expect you will decide it is a good idea to attempt to re connect the coils for less ohms,

meaning soon you will have a couple of those motors with no coils. Like I do.

BTW, The air gap is 100% great.

Old eyes, tiny laptop, '8' looked like '3', and a strange calibrated metric ruler had me looking a 2mm while thinking 1mm.

G-

[altosack]

Hi Flux,

If you are using speed, then above a certain speed, drive the pwm signal forward regardless of what the mppt is asking, to bring the prop into stall.

I can certainly do this, but in my opinion this is really no better than furling, which will certainly be a part of the equation, and furling is more fail-safe, albeit not as elegant.

I am probably being greedy here (ref. your previous posts on the subject), but once I have electronic controls, I would like to be able to pick a power level that will be my (safe) maximum, which it will reach with good aerodynamic efficiency at a certain wind/prop speed (say 11-12 m/s, 24-26 mph), then hold that power level to a higher wind speed (say 15 m/s, 33 mph, more if possible) by progressively stalling the rotor, then finally furl beyond that. Yes, greedy.

I doubt that you will hold it above 50 mph in stall without a monster alternator so if you reject variable pitch I think you will need to furl.

And, of course, I can solve it with variable pitch, and I am not completely ruling this out, but I agree with your earlier posts on this subject that this effort is only reasonable for larger rotors and a major engineering effort (if it's going to be worry-free for many years). Electronic controls are considerably cheaper in time and effort (to me; I'm sure others disagree !) than this, and I want to see how far I can get before I need to implement pitch control. Also, to me, an anemometer is cheaper/easier than pitch control, and it would give me the extra input that I need so I don't have to feedback on power (power would be calculated instead); this would be my next step before pitch control.

Part of my greediness is to have an elegant (cheap) solution, so a monster alternator is not in my plans.

Actually, I was planning to present what I'm planning to do in a bit of detail in my own post (sorry, Peter) pretty soon, but I saw Amanda's post which directly pertained to what I want to do, and I jumped a bit prematurely. So why don't we table this until then, because it's getting way away from Peter's post, and your comments, while good, will be much better once I've detailed everything with one go instead of piecemeal.

Best Regards,

Dave

[dinges]

Willib,

The only measurement that was made was the weight of the magnets and the radius of the wheel. All the rest is calculated. If I could measure torque whilst the thing was running at various RPMs and under various loads, I'd be a happy camper...

Ghurd,

The drag was measured whilst the genny was unloaded, i.e. coils were open. Thus, there are assumed to be no copper losses. But I think it'll be hard to get more magnet in the rotor; even with the plan as described in post #26, magnetic volume is the same (surface is slightly larger though).

I don't see how removing the coils would reduce drag? Unless there are eddy currents in the coils, which I doubt.

"soon you will have a couple of those motors with no coils. Like I do. "

Start rewinding already...

[dinges]

Willib, see my reply to Ghurd for a reply to you.

Also, it's about impossible to measure stator losses directly. Only way I can think of is by caloric measurements (i.e., how much does the stator heat up). If there are other ways that I'm not aware of, that allow to measure stator losses directly, I'd like to know.