Radial Flux Generator Project

[DragonFly III]

I was wondering if it would be worth using powdered inductor cores for moulding stator cores.  When I scrap and strip I get a crap load of electronics that I strip down and salvage what I can.

[ChrisOlson]

The mags  will settle over the slots and unfortunately cog hard there as they bond on the wide core segments .I think you can buy some increased power by cutting a slot in those to lessin the bond .That's why you see small long segments in even the largest motors .   

I do expect to get some cogging.  Whether or not it's an issue with the startup torque of 13 foot blades vs a 7.5" generator rotor remains to be seen.  If I figure it's too much I may make a lot of modifications to the core on the mill   
--
Chris

[tecker]

Sorry Chris I come off as a smart a)(*&  but I messed around with metal filings and magnetite . As soon as the opposing mag touches the wide core you get current .  Similar action when used in axial coils . Really the run out looks good with 12 mags though the leading  mag will be in the slot by the time the trailing mag hits the core segment .

[Fused]

You thought about possible wave winding Chris? Im collecting pieces parts, thinking of a smaller version.

Fused

[ChrisOlson]

You thought about possible wave winding Chris? Im collecting pieces parts, thinking of a smaller version.

I did consider it.  And I may try that with build #2 - assuming I get the magnetic performance from this core that I need and don't have to start over with designing a new core   

The UPS tracking outfit says my magnets are "on schedule" to be delivered today.  So hopefully tonight, later on, I can run this thing up for the first time and see what she does.
--
Chris

[dlenox]

Chris,

I noticed that you did not use any material to insulate the wire from the core.  Is this because you are simply testing coil sizes for appropriate voltage.  I know that I'd be afraid of nicking the insulation from the wire when winding on the abrasive core material.

I'm assuming that you will insulate the finished coils, and wondering what you are intending to use for this purpose?

Dan Lenox

[ChrisOlson]

Chris,

I noticed that you did not use any material to insulate the wire from the core.  Is this because you are simply testing coil sizes for appropriate voltage.  I know that I'd be afraid of nicking the insulation from the wire when winding on the abrasive core material.

I'm assuming that you will insulate the finished coils, and wondering what you are intending to use for this purpose?

Its current configuration is for test only.  The final unit will use those little insulator tubes that they use in electric motors (I can get them at the local motor shop) and the stator windings will be varnished with electric motor winding varnish (also being done for me by the local motor shop).

In lieu of varnishing it, I think one of these could also use fiberglass resin to seal the voids in the windings and still leave them exposed to the air for cooling.  But I'm going the varnish route because if it ever needs rewind it's a simple matter of cutting off one of the winding heads, heating it to soften the varnish, and pulling the old windings out.  When I built it, before I painted it, I heated the stator core ring with a torch to about 500 degrees - it burned off and turned the resin binder black on the outside but otherwise didn't hurt it.  Typically, open-flame burners are used to soften electric motor varnish for rewind and I wanted to make sure the same thing could be done with this thing.  It passed that test with flying colors (although it looked like a burnt pineapple upside down cake after I got done trying to set it on fire).
--
Chris

[ChrisOlson]

Got 'er assembled for a test run:



Got three problems:

• The voltage is way too high.  39 volts @ 100 rpm.  It should be more around 19 volts.  I only need 15 turns of wire instead of 30.
• The cogging is severe.  She locks up every 10 degrees of rotation due to the stator teeth being the same width as the magnets.
• Now for the biggie - the core is saturated.  The flux is going all the way thru the core to the steel backing ring around the core.  In fact, I can hang a 3/4" wrench on that steel ring around the core and it sticks there.  Not good.

I'm back to square one on building a core for this thing.
--
Chris

[RP]

If the voltage is too high and the flux is leaking out of the ring perhaps you could use this platform to test some ceramic magnets from Lowes (hardware store)...

[ChrisOlson]

If the voltage is too high and the flux is leaking out of the ring perhaps you could use this platform to test some ceramic magnets from Lowes (hardware store)...

Then it would end up being an 800 watt generator instead of 2 kW.
--
Chris

[frepdx]

Got 'er assembled for a test run:
Got three problems:

• The voltage is way too high.  39 volts @ 100 rpm.  It should be more around 19 volts.  I only need 15 turns of wire instead of 30.
• The cogging is severe.  She locks up every 10 degrees of rotation due to the stator teeth being the same width as the magnets.
• Now for the biggie - the core is saturated.  The flux is going all the way thru the core to the steel backing ring around the core.  In fact, I can hang a 3/4" wrench on that steel ring around the core and it sticks there.  Not good.

Can you cast some more core, increasing the outside diameter? maybe switch to thinner magnets and skew them? Is the cogging really an issue given the size of the blades?

[ChrisOlson]

Can you cast some more core, increasing the outside diameter? maybe switch to thinner magnets and skew them? Is the cogging really an issue given the size of the blades?

The cogging is not really an issue with a 13 footer.  Using an inch-pound torque wrench it takes 47 inch-pounds of torque to "break away" from one cog to the next.  And as soon as it gets spinning it's not really noticeable.  For a 13 foot rotor it would just stay parked until the wind picked up to cut-in speed, which is not a bad deal anyway.  For a smaller rotor 18 more slots could be cut into the core and it would reduce the cogging considerably.

The bigger issue is the core saturation, and I don't really know about how to fix that.  It's 1.5" from the face of the magnets to the steel backing ring on the core now.  There's 8.2 lbs of iron in that core.  But the flux shouldn't be leaking outside the core.  It's possible I could use a non-magnet backing ring of some sort and maybe it's not really an issue being I have such good magnetic performance at the coil legs.  I don't really know at this point - I have to think about it for a bit.

Edit: Doing the calculations on this, at the 100 rpm that 13 foot PowerMax blades like to cut in at, the ~4 lb-ft of torque required to overcome the cogging and get it turning is 56 watts.  For a 13 footer that's nothing - basically cut-in power.
--
Chris

[frepdx]

The bigger issue is the core saturation, and I don't really know about how to fix that.  It's 1.5" from the face of the magnets to the steel backing ring on the core now.  There's 8.2 lbs of iron in that core.  But the flux shouldn't be leaking outside the core.

doesn't matter how much iron the core has. Look at the crossection through which the flux has to flow - there is a choke point between the drilled holes and the OD. It's less than the crossection of the magnets, meaning that it has to carry a higher flux density than the magnets, which I doubt will happen with un-compressed iron in resin. Increasing the OD with more casting should help, but you might try measuring your losses with the current version, maybe it's not so big a problem.

[jimovonz]

The voltage you can fix - this was only a test winding anyway.
If cogging is a problem you could skew the magnets (if you have clearance) or offset them.
Significant flux outside the core could be saturation or it could just be that it doesn't have the permeability you expected. Either way it means a rethink if you want to get the kind of output your after. You are obviously not making the best use of your magnets and you are also in for a hit due to eddie losses with all that stationary iron in the rotating magnetic field.
When you run additional iron (aka wrench) around the rim (in its current configuration) are there spots that it sticks to better than others? How well do the magnets stick to the core (with no support ring) compared to normal steel?

[DanG]

Enercon out of Germany makes gearless low RPM direct drive megawatt wind turbines with the biggest having 5,281 m2 swept area:- I borrowed (stole) these images of their annular (your radial) generators from their "WindBlatt" magazine for your review...

ArtV suggested staggering the hit order of the flux cog points - skew either the stator or the magnets so there is 3-5-7 degrees cascade to keep all the attractions from happening at once, others here have done it, IIRC Dinges (and others) showed good work reducing conversions' cogging.

An idle question : With only 15 turns per coil to hit your target voltage why not jump to 5-phase?

http://www.enercon.de/en/_home.htm
http://www.enercon.de/www/en/windblatt.nsf/vwwebAnzeige





ALL photo credits owned by ENERCON GmbH · Dreekamp 5 · 26605 Aurich · Germany

[joestue]

I would mix up more core material and add another 3/4 inch all the way around the core at this point.
You could try turning the magnets 10 degrees or so to fix the cogging partially, but you might have to mill a flat in the rotor to do that.

[hiker]

looks cool........hmm..what about  smaller mags? looks like those are a bit oversize.........overkill?
did some tests on my 3hp conversion using 1/4"mags--no skew...12mags per pole..slight cogging..
i hit 1000watts at 28v dc..on my pedgen..at a fairly slow rpm..400 watts was no problem at all..
with your set up --skewing the mags would most likely help out--as far as cogging is concerend..

[scoraigwind]

Got 'er assembled for a test run:

(Attachment Link)

Got three problems:

• The voltage is way too high.  39 volts @ 100 rpm.  It should be more around 19 volts.  I only need 15 turns of wire instead of 30.
• The cogging is severe.  She locks up every 10 degrees of rotation due to the stator teeth being the same width as the magnets.
• Now for the biggie - the core is saturated.  The flux is going all the way thru the core to the steel backing ring around the core.  In fact, I can hang a 3/4" wrench on that steel ring around the core and it sticks there.  Not good.

I'm back to square one on building a core for this thing.
--
Chris

If you are building a new core then why not try 13 poles on it.  That would smooth it out a bit.  Wind one coil on each pole, and run 13 wires to the rectifier in a big star configuration.  Or maybe you could even connect them all in series in a big 'delta' style polygon winding...

I see no advantage to having coils that span 2 stator poles in the 18 pole stator, by the way. 

You could drill the holes a bit closer to the magnets so as to have more core material outside the slots?  Or cast oval slots into the core using a mould?

fun project.

[jeraklidis]

I would mix up more core material and add another 3/4 inch all the way around the core at this point.
You could try turning the magnets 10 degrees or so to fix the cogging partially, but you might have to mill a flat in the rotor to do that.

The easiest way to reduce cogging is to to compute the Least Common Multiple of the slots and poles

thus 12 18 LCM = 36 which is not that great which means that there are 36 cogging periods in one revolution. The more cogging points the less the amplitude of cogging force.

You are right about the tooth shoes the same size as the magnet... that's a no no. Another good tip is to place your magnets no more than half their width apart from eachother.


oh wow you said every 10 degrees it locks up HAHA LCM does work!!!

[prasadbodas2000]

Hi Chris, It is amazing project - no doubts.
Do you have an approximate estimation/number about the stator core effective density. Means if it was full steel (very closely laminated or say solid steel) it would be 7850 kg per cubic meter. Now with iron powder emebedded in fiber what would be the density you think? Will this be 50% of normal steel or more or less? Will this explain the cause of core saturation?
Just curious to know.

With regards,
Prasad

[tecker]

Lay the mags in at an angle so the corners of the opposing mags just barely touch the same core segment and you'll be good.Now to see if ti heats up maybe 20 degrees.

[DragonFly III]

Man so much work to have problems.  It reminds me of my first set of laminates.  Took me 3 days (I could only put in a few hours a day.  I have a 3 yr old, Mommy works.  After all that work I found some serious problems, maybe a good thing for experience.  Second core sucked too.  I then read a post, i think it was here where some manufacturer skewed the laminate layers so I did the same.  No cogging at all.  Since I very rarely hit cut in speed here, my mill is more for show right now and I love to sit and watch it turn.  Anywho, I think that the Idea of skewing your mags is a great Idea.   The idea of thickening the core seems Like a good idea too.  Other than the fact that you are adding more weight.  I am using stacked saw blades for my rotors dual rotor mags only on one side (i don't have enough mags for the other rotor and can't exactly get more since they were salvaged from ancient hard drives.  The platters are 5 1/4 about.  Small mill that will be geared.  I stack the blades till a washer won't stick.  I guess thats why a thicker stator core seems as though it would work.  Or the other option.  Scrap it and start over.  Maybe drilling the coil hole at an angle and making it a little larger.

[Fused]

Motor conversion 101. 

If you make another core, why not just skew the core slots.
Jerry rig that bad boy and fly it.

Fused

[ChrisOlson]

Do you have an approximate estimation/number about the stator core effective density. Means if it was full steel (very closely laminated or say solid steel) it would be 7850 kg per cubic meter

Thanks for all the replies on this latest setback, guys!

I think the above quote outlines a lot of the problems with this design.  The cogging is not a big issue.  It would be nice to have none, but what is there is only going to keep the turbine from "practicing" (spinning without making power) prior to cut-in.  After doing some figuring on torque I decided that's not as huge as I thought.

But back to the above quote - I found a cast iron ring that's a front wheel weight for a tractor.  That ring is roughly the same dimension as my stator core and it weighs 20 lbs.  My core, prior to cutting holes in it, only weighed 10 lbs - half the density of real iron.  Having the core saturated, no matter how elaborate of a scheme I come up with to skew magnets and whatnot, is not going to work.  Making the core thicker is not going to work either.  The magnetic circuit needs to be completed just in back of the coil legs - not 2" away.

Again, I need to lay this one aside for a bit and think about how to build a new core for it.  My best ideas usually come to me after a couple beers (usually in the form of, 'Hey - watch this!!!"   
--
Chris

[ghurd]

My best ideas usually come to me after a couple beers (usually in the form of, 'Hey - watch this!!!"    

When I am in that situation, the next thing out of my mouth is
"Wow.  I didn't know it would catch on fire so fast...
You think this will heal up OK, or does it need a few stitches?  Judy is working the ER today and she always does such a nice job."

G-

[joestue]

Making the core thicker is not going to work either.  The magnetic circuit needs to be completed just in back of the coil legs - not 2" away.
Chris

this is not the case at all. as long as the flux travels through the coil it will work just fine.

[windvision]

Hey Ghurd-

  That's funny! I don't care what the rest of em say!

[ChrisOlson]

this is not the case at all. as long as the flux travels through the coil it will work just fine.

Well, it is - because eventually you have to have something to support that stator ring.  And it has to be pretty stout to handle almost 3 hp worth of torque at full load.  If you let the flux get thru to whatever you're using to support the core you got eddy currents.  Ideally, that magnetic circuit should be completed just behind the coil legs so all the flux goes thru the coil, but none leaks to the outside into whatever supports it.  And that was my goal here - but when I found I can hang a 3/4" wrench on the outside of the steel backing ring I knew I had problems.  Sure, all the flux is going thru the coil legs - but it's also going to set up eddy currents in the support structure big time.  At low speed it's no big deal.  But wind this thing up and it will be.

The core material needs to be more dense.
--
Chris

[SparWeb]

Tried my own calculations, and I got 19.5 V / 100 RPM.  Gee I don't know how you got 39V either!  But when 2 x 19.5 = 39.... it's usually an obvious thing in the math that makes me slap my forehead when I find it.

Anyway it's a moot point because of the cogging and the narrow flux path.  I'd hate to see you quit now, so I thought I'd just shout some encouragement for the next one. 

A couple of other thoughts:  You could put at least 50% more magnet on that rotor.  When you fix the saturation issue settled by designing a flux path with enough cross-section to carry from pole to pole, make sure it's wide enough for a full complement of magnets.  Then increase it again with that density ratio you found between your stator's density and solid iron.

Maybe the thing is getting heavier and more bulky with every idea we add!

Simplicate! And add more lightness!

[jimovonz]

that makes me slap my forehead when I find it.

I'm sure that I've got a permanent red spot on my forehead 

[Ungrounded Lightning Rod]

Regarding the field off the back of the core:  That just means the back of the core part is too thin.  So cast an additional ring of powdered iron around it to thicken it up (until you have negligible "grab" against iron tools) and put it into a bigger support ring.

With powdered iron in a plastic matrix you have a small gap between each grain.  So you have some "air gap" distributed through the core.  Thus you'll have some field "willing" to take a shorter air path through a solid piece of metal on the outside even if your grains aren't saturated.  (But grabbing a wrench through a steel ring is definitely excessive.  B-) )  The distributed air gap will also be reducing your field and the grains in the pole pieces may be saturating.  But it looks to me like you've got enough generation even so and the game is to keep from losing HP heating the support ring.

One thing troubles me:  What is your iron powder and in particular how magnetically "soft" is it?  Eddy currents aren't the only loss that burns HP into heat in a core:  Hysteresis in the core material does it, too.  Powdering the material and insulating the grains does nothing about that loss (or the resultant heating).

[Ungrounded Lightning Rod]

Regarding the field off the back of the core:  That just means the back of the core part is too thin.  So cast an additional ring of powdered iron around it to thicken it up (until you have negligible "grab" against iron tools) and put it into a bigger support ring.

With powdered iron in a plastic matrix you have a small gap between each grain.  So you have some "air gap" distributed through the core.  Thus you'll have some field "willing" to take a shorter air path through a solid piece of metal on the outside even if your grains aren't saturated.  (But grabbing a wrench through a steel ring is definitely excessive.  B-) )  The distributed air gap will also be reducing your field and the grains in the pole pieces may be saturating.  But it looks to me like you've got enough generation even so and the game is to keep from losing HP heating the support ring.

One thing troubles me:  What is your iron powder and in particular how magnetically "soft" is it?  Eddy currents aren't the only loss that burns HP into heat in a core:  Hysteresis in the core material does it, too.  Powdering the material and insulating the grains does nothing about that loss (or the resultant heating).  But insulating the grains with a coat of potting material DOES break the easy path for the heat to leave, as it would along laminations.

[ChrisOlson]

A couple of other thoughts:  You could put at least 50% more magnet on that rotor.

It was too dang hot to do anything else today so I loaded the thing in the back of the Dodge Cummins and went to the motor shop with it.  I showed my buddy that owns the shop what the deal is.  He says, well, why not make it air core?  All the new high-efficiency electric motors are air core, not laminated steel anymore.

So I got a new plan I'm going to try.  I'm going to cast a non-magnetic core out of SmoothCast 300 plastic and using N50's instead of N42's on the rotor.  I checked, and I can get them N50's with 1/8" holes pre-drilled for like 7 bucks each.  It'll be a little time consuming making a mold to cast the plastic but I think it will be a fun experiment.

Here's the deal - I pulled the rotor out of the core, put some iron powder on a piece of plexiglass and held the plexiglass 6" from the magnets, between two poles.  Even at 6" the flux "arranges" that iron powder on the plexi.  Rotate the rotor while holding the plexi above it at a distance of 6" and it really does some cool stuff to that iron powder.

So if I throw enough flux at it, why do I even need a magnetic core?
--
Chris