Magnet thickness vs. force

[willib]

i am going to test (i hope) the effect of magnet thickness on a generator , my present testbed allows me to switch magnets fairly easily.

i have fourty eight - 1/2"  thick x 7/8" dia.,, and fourty eight 1" thick x 7/8" dia. N40 neos.

I have several goals in all this .

 1) to build a better generator

 2) determin the effect of magnet thickness  on a generator.

 2a) determine the change in RPM per volt of the two thicknesses

 3) what effect the two different thicknesses will have on several different coils , all of the same outer dia ,some have slightly different inner diameters, and one is thicker than the others..

stay tuned it should be interesting.

any sugestions will be welcome?

http://www.otherpower.com/images/scimages/2965/mline_B2C5D48.jpeg

[willib]

forgot to say , the distance in the graph is in cm.

[Titantornado]

Sounds good.  I look forward to the overall results of this one.  There's got to be a gain, but how much?  I doubt doubling magnet thickness will double voltage, but that would be nice.

Impatiently awaiting your results.  ;-)

[willib]

That was my origional question, would doulbling the thickness of the magnets double the voltage?

i have some data on this which i still  have to put together.

[finnsawyer]

The effects your seeing are in air.  When you use the magnets in an alternator you are using them in a magnetic circuit, which involves iron.  It is possible to drive the iron into saturation, which will keep the flux from increasing beyond a certain point.  Nevertheless, what you are doing is definitely useful, if no other reason than to set possible limits on how strong the magnetic fields can get.  That is, when do we hit the point of diminishing returns?

[willib]

meter reading 1.26A , so far so good..

http://www.otherpower.com/images/scimages/2965/IM001617.JPG

scope image of the gate input , nice clean waveform , no distorted ends or anything.. one over 6.6 -- the fet is on 15% of the time.

http://www.otherpower.com/images/scimages/2965/IM001618.JPG

Meter reading 3.59A. still good ! load is starting to heat up at this point.

http://www.otherpower.com/images/scimages/2965/IM001619.JPG

2.5 divisions  over 6.6 divisions ..scope is saying the FET is "on" for 38% of the time..

http://www.otherpower.com/images/scimages/2965/IM001620.JPG

meter is 5.26A and the fet is still not hot , just warm , with no heat sink!

http://www.otherpower.com/images/scimages/2965/IM001622.JPG

approx 3.7 div. over 6.6 div. FET is "on" for 56% of the time

scope is on .5 mS per division so the frequency of the pwm is 303 times a second (HZ)..not pictured is the sealed lead acid battery supplying the power to the load through the FET..

i'll try to draw up a schematic of the setup later on..

[willib]

can i assume that you are talking about a zubly machine , radial flux?

where there is iron in the path of the magnets?

[finnsawyer]

Yes, I assumed you would follow the designs one finds here, where the magnets and coils are backed by iron.  There wouldn't be much point in doing the test with only non magnetic materials as the results then wouldn't apply to what people are doing.

The B versus H curve of iron looks something like the top half of an S.  For low values of H, B increases at a linear rate, but at high values this rate of increase becomes less.  The magnets provide the H, which in this context is called magnetomotive force and is analogous to voltage.  The flattening of the B versus H curve for the iron means that doubling the magnetic field strength will not necessarily double the flux in the iron and hence the flux seen by the coils.  What you propose to do would be useful to ascertain whether the magnet strengths now available have pushed alternator design close to this limit.  It will be necessary to use the same alternator design for both tests (allowing, of course, for the different magnet thicknesses - you can't escape that).

[willib]

i have no means to perform a test with a radial flux machine .

i only do axial flux.'maan'

and as such the only purpose of having  iron rotors ,is to hold the magnets in place. as i see it.

because as a North pole sucks the flux from the south pole (or visa versa) the rotors have very little to do , in the larger scheme of things ,but hold the magnets in place.

whatever flux flows in the rotor itself has no bearing on the machine as a whole.

[jimovonz]

I think you may have missed one of the critical elements of design here. The iron behind the magnets in one of our axial flux dual rotor alts certainly does more than provide something to hold the magnets in place. Magnetic flux always flows in a circuit N-S. The same amount of flux leaving the surface on one side of the magnet returns to the other side. You can't increase the flux on one side (in the air gap) without increasing the flux on the other side (iron rotor). Just like electricity, the total flux(current) in the circuit is determined by the resistance/impedance. Iron has a very low magnetic impedance (~1000 times less than air) and consequently the greater the proportion of iron, the higher the flux in the circuit and consequently in the air gap where we want it (up to the point of saturation).

[SparWeb]

I've played around with different magnet arrangements using FEMM, a finite-element modelling package of software for magnetic fields.  (Free to download if you want to try yourself).  Doubling the thickness of a magnet affects the field by a factor of the square root of the change in thickness,  or 2^(1/2)

For example: 2 x thickness:   sqrt(2) = 1.41 times the original flux

               1/2 x thickness: sqrt(0.5) = 0.71 times the original flux

Your graph agrees with that, to some degree.  There are other factors at play, like flux saturation that I can't quantify for you.  You have to picture that if you keep the space between magnet face and stator the same, the amount of magnet you're adding to the system is getting farther and farther away from the stator.  You simply can't double the flux by doubling the thickness of the magnet.

From your particular starting point, you probably could find a thickness that bosts your voltage per RPM to the amount you desire.  The number you calculate may not agree with you, tho.

[willib]

"You can't increase the flux on one side (in the air gap) without increasing the flux on the other side (iron rotor)."

how do you know this?

can this be tested?

[jimovonz]

Hi willib, I have never done any experiment to specifically test this. My knowledge of magnetics is relatively recent and growing all the time. I cannot pinpoint exactly where this piece of knowledge came from but is fundamental to my understanding. It is always  possible that I am wrong but obviously I believe that there is a high possibility that I am not. (you of course are free to believe otherwise - and in either case our belief has no bearing what so ever on the truth!) Philosophy aside, you are right to demand proof but unfortunately I don't have any at hand. Off the top of my head, I could ask you if you have ever seen a flux diagram where the flux lines don't make a complete path from the north pole face to the south? i.e a flux line the leaves one face and stops somewhere short of returning to the opposite face?

P.S. A quick Google on 'Flux conservation' returns some of the following:

http://www.magnetweb.com/Sect4A.htm

http://www.cs.princeton.edu/~kazad/resources/math/Gauss/gauss.htm

And there we have it from the man himself: Gauss' Law

Googling 'Gauss's Law' returns as much info as you could desire

 

[willib]

"Flux must always be conserved in any complete magnetic circuit. If this were not so, it would mean that there must exist a point source of magnetic flux, such as a magnetic monopole, which is not possible; all magnetic north poles must be paired with comparable south poles."

i get you now ,this leads back to the non existance of the  magnetic monopole

[willib]

Thank you! that is the kind of relationship between magnet thickness i was looking for..

btw i've tried FEMM but couldnt get it going..maybe you have some examples i could use?

about flux saturation...

in your opinion

could a coil get so saturated with flux that it no longer produces more voltage?

i have one more test to perform , i need to test the severly loaded coil with the 1" magnets..

[kitno455]

TEK 466- i thought i was the only guy still using one of these! darn thing is older than i am.

anyway- i think with the rise in leakage flux from the longer (taller/thicker?) magnets, you have shown that it makes more sense to go with more surface area, if the cost of the volume of magnet is the same.

so the next question is- does it make more sense to have lots of smaller poles, or a few larger poles on the same diameter plate...

allan

[willib]

"so the next question is- does it make more sense to have lots of smaller poles, or a few larger poles on the same diameter plate..."

that is a very interesting question..

ive designed on paper a 12"dia. rotor  machine with 1.5" long x1/2" dia magnets , but i wasnt happy with the strength of the magnets.

 yeah i'm starting a collection i suppose..

they are supposed to be 64lbs pulling force , but i am much happier with the 74lbs of the 1" long x 7/8"dia. ones..they have awesome power...although the strength of the 1/2" thick ones isnt anything to sneeze at , at 37 lbs ..

"anyway- i think with the rise in leakage flux from the longer (taller/thicker?) magnets, you have shown that it makes more sense to go with more surface area, if the cost of the volume of magnet is the same."

i'm not sure this is the case, as far as cost $1.50 vs $2.20 a piece

i have two photos you should see , which shows the leakage flux of long , thick magnets vs thin wide ones ..

anyway i hope i havnt conveyed that conclusion about thicker magnets having more leakage flux ,i have not noticed this at all..

in fact the opposite seems to be the case as seen in the pictures below..

[kitno455]

i think you need to stop looking at flux plots with the mag in free air. it will be on a steel backer, with another mag of equal strength on its own backer just across the way.

you did not label those pics. what am i looking at?

allan

[Kwazai]

in terms of the emf induced in a piece of steel (or copper?) would 1.41 thicker copper/steel double the available electrical energy force (doubled magnet...) (or vice versa?) ...not an area I have studied to any detail (hysteresis affects).

Mike

[finnsawyer]

Yes, it can be tested.  Every time you plug in a transformer this is happening.  By using an oscilloscope in X-Y mode and putting an ac  voltage proportional to the source current in the X input and the output voltage in the Y input you get a plot of B versus H for the transformer material.  The magnetic field H is proportional to the input current and the output voltage is proportional to the B field or magnetic flux.  This simple experiment, which requires that the transformer material form a closed path of simple geometry, is done quite commonly in college labs.  Of course, we just hand the students some wire and a core and let them make their own transformers. Once magnetized the iron doesn't return instantly to a non magnetized state, so the B versus H curve for decreasing current does not follow exactly the B versus H curve for increasing current.  An elongated loop will be observed, which is called hysteresis. The area of this loop is proportional to the energy lost as heat for each cycle.  This effect will also be a function of frequency.  I hope this helps.    

[willib]

" Once magnetized the iron doesn't return instantly to a non magnetized state, so the B versus H curve for decreasing current does not follow exactly the B versus H curve for increasing current."

Interesting...so the laminations could actually hinder the operation of a machine at higher RPMs..?

By holding on to the flux from the previous pole when the next pole comes by , unless the timeframe is very small..

i'm glad i/we dont have to deal with laminations in axial flux machines

shall we call you professor GeoM ;-)

[willib]

the first one is a long skinny magnet , and the second one a short wide one.

[kitno455]

ok- and look at all the leakage flux lines on the first one. no way to get those dark red a blue lines to jump across the gap to another mag.

put those multiple of those mags on a steel plate in your software, and put a like assembly on the other side, and try to compare. then do the same for the other type of mag.

allan

[jimovonz]

I'm not sure that you can derrive any information regarding leakage from a flux diagram  of a magnet in free space... You must first define a closed system for there to be any chance that something might leak from it. I believe that the length(thickness) of a magnet has little to do with leakage, but rather the width of the air gap relative to the distance to the adjacent magnet determines the level of leakage. If the adjacent magnet is closer than the opposite one then thats where the flux heads.

[SparWeb]

I'm not exactly a whiz, myself, but FEMM has made things much easier.

...could a coil get so saturated with flux that it no longer produces more voltage?

Copper is conducting electricity, and the current flowing in the coil generates the hated back-emf that eats away at your voltage.  The more heavily you load the alternator, the more back-emf there is, and in most systems this brings the alternator into some kind of equilibrium - where the power fed in through the rotor blades and all is taken out by either losses or the load.  You say yourself that you are heavily loading your stator.  Then back-emf is what you're causing, and I bet you're seeing voltages well below the open-circuit voltage.  A dead short, then, gives zero voltage, for the same reason: back-emf is equal to open circuit voltage at that point.

If however, you're asking about the generation of eddy currents in the copper, irrespective of the current, or back-emf, then I really don't know what to say there, yet.  I dug up my old university physics textbooks to get myself this far, and that wasn't covered.

So far, the only consequences of "flux saturation" that I know of are the increased torque associated with generating these useless currents, and the heat produced, but this has always been described in reference to stator plate materials, or too much hardware mounting the stator, etc., not in the wire iteslf.

One of the local experts could step in and correct me on that part, if necessary.

[willib]

you just added another piece to the puzzle..for me anyway..

that is if the air gap is wider than the distance between adjacent magnets , then thats where the flux goes

but does it go straight across to an adjacent magnet , or does it loop up a little, i wonder..

for me the distance between adjacent magnets is .4375" (7/16") therefor my air gap should be .4375 maximum....

i would like to test it out with iron fileings

that kinda stinks because i was wanting to use a larger gap than .625" (5/8 "),

i had my eye on .75".. hmmm

[jimovonz]

If my visimag simulations are anything to go by (many in my files) then the flux loops over between magnets. The more direct the path, the 'flatter' the loop.

[jimovonz]

Here is one specifically showing leakage:

http://www.otherpower.com/images/scimages/2210/Leakage.gif

I could have illistrated it better perhaps if I kept the distance between the magnets constant and varied the distance between adjacent ones.

[willib]

"and I bet you're seeing voltages well below the open-circuit voltage"

yes , @ 250 rpm the open circuit voltage is around 1.5 Vpeak =>166 rpm / volt for one coil.

i see in my notes that i got .6 Vpeak at 166RPM loaded with the 0.00583 ohm load resistor. =>  277 rpm / volt peak for one coil.

but i didnt record the exact length of the gap  or either of the above readings

a lot of fun but frustrating sometimes..

[finnsawyer]

The laminations are there to break up the eddy currents.  They do not have any effect on the relaxation time of the induced magnetic field.  That is a property of the crystal lattice of the iron on the molecular level or atomic level.  In most of the PM alternator designs the magnets move with the iron backing.  This minimizes hysteresis.  The induced currents flowing in the coils can cause it, however, as well as eddy currents.  These effects plus resistive losses will put an upper limit on the alternator RPM, where the thing burns out, but until that point is reached the rapidly increasing power in the wind makes the losses due to these effects tolerable.  And yes, the use of laminations in PM alternators would help, but is probably not worth the cost or trouble for home brew units.

[Titantornado]

That's actually a pretty neat visualization.  Is it drawn from real results, or from a hypothosis?

[SparWeb]

I strongly suggest to go through the FEMM tutorials one step at a time.  Then re-do the steps, but for you own case, draw the cross-section of a flat plate (backing plate) with an array of magnets on it, and then another plate above that.  Vary the distance between the plates, vary the thicknesses of the magnets, try whatever you like, the computer will do the math for you.

All of the questions you are asking will be answered by this process.