Author Topic: Magnets vs Coil design (Read 4844 times)

vacuum1313 · « **on:** February 17, 2008, 03:34:56 PM »

Is there a relationship between magnet size (physical dimensions), magnet strength (N rating) and either optimal wire size (guage) or coil dimensions (length, width, thickness). I have 20 x N48 discs, 3/4" diameter x 1/4" thick I want to use to build a scale version and make my mistakes there before making a full sized model. Assume for the model it would be charging 1.2V NiMH batteries, either 2 or 4 at a time, as a 1/10 scale of 12V batteries. I am looking to get another 4 magnets to maintain the 3 coils to 4 magnets ratio I have seen elsewhere and am looking to build a dual rotor model, 9 coils x 12 magnets per rotor. I could determine the coils empirically, lots of trial and error, but am looking for a reasonable starting point. I have tried looking up magnet size, magnet strength, wire size and coil design on the site but have had little luck. I have often seen, and been guilty in an original post, of not providing enough information and hope this is sufficient.

finnsawyer · « **Reply #1 on:** February 17, 2008, 09:12:45 AM »

I don't see any point in using those magnets any way other than the final way you intend to use them. That is, two rotors having 12 magnets each. O.K., you don't have 24 magnets currently, a bit of a bummer, but you could set up a single rotor of twelve magnets for now. What spacing to use? Probably 3/8 inch, half the magnet diameter. That seems to be standard, but others might advise you differently. From the magnet placement you can determine the maximum coil diameter. If you already have some magnet wire, then wind a test coil with it and see what voltage you get. You also need to keep track of the coil resistance. Coil thickness is also a variable. Don't commit to a final coil design or rotor spacing until you have built the second rotor and repeated your tests. With the second rotor you would expect to at least double your output voltage, but that may not happen due to magnetic saturation of the rotors. Of course, someone may have built just this alternator and may have the answers you seek.

vacuum1313 · « **Reply #2 on:** February 17, 2008, 09:41:02 AM »

Thanks, I did leave an important piece of info out I see. I have to buy the wire, copper is expensive, and would like to limit the number of sizes I buy for the experiments so am looking for info that will at least put me in the ballpark. I do understand the relationship between turns and volts / wire guage and resistance. I do not know the relationship between magnet strength and coil dimensions or wire size. When you mention spacing above I am understanding that to be the distance between the edges of the magnets on a rotor.

Instead of using resin to hold the magnets in place could I use a wood disk with appropriately spaced holes bolted to a mild steel backing disk?

Flux · « **Reply #3 on:** February 17, 2008, 09:59:50 AM »

You will get in a mess. Wind turbines don't scale very well. I think you will cause so much confusion that you will end up wasting time copper and magnets. If you get the scale thing right it will give you no information for the full size one.

Flux

vacuum1313 · « **Reply #4 on:** February 17, 2008, 10:29:31 AM »

Are you saying there is no math for these things? I do understand relationships are often not linear but logarithmic or based on a power relationship. A typical example would be light intensity decreasing by a factor of 4 as the distance doubles. The modeling is not for power output but for practice in making correct connections of the wiring, making the coils, keeping the process straight. Though I am a highschool science teacher my area is the biological sciences not the physical. My primary purpose is to build a full sized unit for a lakeside home I am building but I also want to understand the theory as well so I can take aspects into the classroom. Model building is also applicable to the classroom.

Flux · « **Reply #5 on:** February 17, 2008, 10:56:11 AM »

Yes you can reduce it all to maths if you understand all the variables.

Volts are proportional to total flux, number of turns and rotational speed. That is perfectly easy. Sorting out resistance and power delivered to a battery is also possible but nothing like as easy.

Now what do you do with the aerodynamics, you would hardly use the same size prop for a model and a full size machine. Power in the wind varies as wind speed cubed. Power of a prop varies as diameter squared. Even that is perfectly manageable but you also need to match the power curve of the prop to the input power of the alternator. The whole thing is a bit more tricky.

Now how do you want to scale things, on rotor size, power out, tsr, wind speed or what.

I see your dilemma , schools always want to reduce things to figures and maths, as long as you produce numbers everyone is happy. In real life a description of all the issues and showing an understanding of principles would be worth all the maths but it wouldn't get you good marks.

Try to work out what you want to do, come up with a few ideas and I am happy to see if you are going in the right direction, but at present you have far too many routes to start from.

Flux

TheCasualTraveler · « **Reply #6 on:** February 17, 2008, 01:56:04 PM »

Vacume,

I'm new at this also, but with my trials and tribulations still quite fresh let me tell you what I would do. First, are you using the magnets on a test generator first and then removing them to use on a larger unit OR are these magnets only for the test unit and you plan to get larger magnets for the second one?

If you are set on using those magnets then use them the best way possible at first (as Mr. Sawyer said), and dual rotor, 12 pole (and thats a practical setup) then the next thing you need to do is decide on the size of your magnet rotor. Using 3/4 inch as the width of the magnet and 3/8 as the spacing between them you get,

12 (magnets) x 3/4 inch = 9 inches

12 (spaces) x 3/8 inch = 4.5 inches

so 9 inches + 4.5 inches = 13.5 inches circumfrence

13.5 inches divided by pi (3.14) = 4.25 inches diameter at the center of the magnets.

That will give you roughly a 5 inch diameter rotor. Thats pretty small and your mill will have to spin very fast. I think (and I may very well be wrong) but if you increase the rotor to say 6 or 7 inches what you lose in spacing you may make up in speed as the outside edge of the larger magnet rotor will be moving the magnets faster with the same rpm's as the smaller magnet rotor. I'm sure someone can help determine the best rotor size for you.

Then, build both rotors, Glue the magnets on and mount them together and about 1/2 inch apart on something that can spin them at a given RPM. Then go to Radio Shack, get a cheap $5 pack of magnet wire and wind some test coils. Don't worry about wire gauge yet. You can place the coils between the rotors and measure output and from there determine the # of turns. Once you know how many turns to use you can get the largest size magnet wire that will allow you to get the # of turns in the space allotted for the coil.

Please know I am not an expert but I think I might speak Newbie-ease fairly well.

TheCasualTraveler · « **Reply #7 on:** February 17, 2008, 02:02:48 PM »

Whoops, that should say:

so 9 inches + 4.5 inches = 13.5 inches circumfrence at the center of the magnets.

13.5 inches divided by pi (3.14) = 4.25 inches diameter from the center of 1 magnet to the center of the magnet on the opposite side of the circle.

wpowokal · « **Reply #8 on:** February 17, 2008, 04:59:49 PM »

Vacuum I for one would very much encourage you to build a scaleed down version, the experience gained will be invaluable when building a large one.

The trial work has already been done and published, Hugh's book is a prime example.

Allan down under

blueyonder · « **Reply #9 on:** February 17, 2008, 05:10:13 PM »

hi i think you must first understand that the three phase gen

will make ac electric out. then you need to make it dc.

so you have to rectify it. and here's whare you will get a bit of loss.

i dont know what volts you hope to get. but using a bridge rectifier.

it might leave you nothing to charge the 1.2 battery's you have.

if you used a dc motor for a gen it might work out a bit better.

but good luck in your project

vacuum1313 · « **Reply #10 on:** February 17, 2008, 06:52:22 PM »

I have Hughs plans for an axial generator. I could easily follow it by rote however I am not only trying to make something that works but to understand the underlying principles hence the build a model and play with things there. Thanks.

Flux · « **Reply #11 on:** February 18, 2008, 01:11:53 AM »

I think I may have misunderstood your intentions. By all means start small and learn from the experience, it's better to learn to carve blades on small ones rather than make a mistake with big ones. Certainly you can build small alternators and learn from that.

If you get to grips with the basics of alternator design, you can build one for any size machine.

Somehow I thought you were going to build a 1/10th scale model and scale it up. That would probably not work.

The first starting point is always to decide what you want to get out and at what wind speed. That settles your prop size and cut in and running speed. From that you can choose your winding for correct cut in speed. From wire size you then predict its performance above cut in. If it is not powerful to match the prop you use more or bigger magnets and try again until you reach both requirements ( you can always get cut in speed right with any combination of coils and magnets).

If you change size then you really need to go back to the beginning and settle prop size. With a bit of experience you can predict that alternator output increases roughly with flux squared so you can hit the new target fairly easily, but few of these things are predictable enough to actually scale a design without checking all the steps.

Flux

vacuum1313 · « **Reply #12 on:** February 18, 2008, 07:56:31 AM »

Definitely not trying to reinvent the wheel. Have done enough work on another project, an updraft gasifier to combined heat and power, to have realised that knowing the theory, reading the books and actually hands on making it are very different things. I went full sized first on that one and it still isn't working and thats just the gasifier part. I restarted with a 1/10 model and then a 1/4 scale model and have learned a LOT! I now think I have a handle on what to do to make the full scale work. Hoping the 1/10 scale here will help me learn some of the practical concerns in making the full sized one.

Here's what I have come up with.

Coil formula = NARBP*2 = voltage

N     number of turns ( total turns )    40

A    area of 1 magnet in meters ( not mm )    0.000285

R    revolutions per second ( not rpm )    4

B    magnet strength in tesla (gauss/10000)    1.37

P    number of poles ( one disc )    12

Voltage 1.5

Using 24 gauge wire

08973 = coil width in inches (of wire not including central hole)
18750 = coil thickness
75000 = magnet diameter
92946 = total coil diameter (round magnets, round coils)
5 = circumference at magnet centers (inches)
37 = circumference needed for 9 coils (inches)

therefore lots of space for coils (could have up to 300 turns in this space)

Note I am designing this coil for powering by pedalling. Cadence 60, gear ratio 4:1 hence the 4 rev/sec. I may also get out some of my retired large scale model aircraft props and try those on it. This exercise though is just to build the axial alternator so I plan to do a blade matching exercise next. Probably use the same rotors but rework the coils to match a blade.

Comments certainly welcome

finnsawyer · « **Reply #13 on:** February 18, 2008, 08:32:47 AM »

Correct on the spacing. I believe someone stated that Radio Shack sells small packets of Magnet wire. Perhaps you could get enough wire, say #18 gauge to wind one test coil. From that you would get a voltage from which you could calculate the total output of the alternator (single rotor at this point) and its resistance. If that output was too high you could then go to heavier wire or the #18 wire two in hand for the coils. Two in hand would give you half the voltage and 1/4 the resistance. In any case you would be well advised to refine your specs and make use of Flux's offer.

If you don't want to glue down the magnets, a stainless steel ring bolted down that encompasses all 12 magnets should hold them in place for test purposes. More work, but safety should be paramount. Don't give the magnets a way to be flung off.

ghurd · « **Reply #14 on:** February 18, 2008, 08:35:12 AM »

Wow!

All that for 1.5V and pedal powered?

I expect the battery will blow up from over current if it's done even semi decent.

Might consider 12V, or maybe 6 coil /8 magnet, or both?

Might consider more RPMs (750 would change the numbers for the better), and a fender to catch flying magnets. Just in case.

There is plenty of magnet wire (#20~22) in the degaussing coil of junk TVs, or CRT monitors. I'm told they are in LCDs too.

I just have a feeling you are making it more complicated than it needs to be for a classroom model.

G-

Flux · « **Reply #15 on:** February 18, 2008, 09:11:34 AM »

Not sure I follow some of your figures. What is the 40 turns, is it turns per coil ( you say total turns).

Your flux density figure is not realistic, that is about Br for neo in closed circuit, impossible in any air gap.

I don't follow your tiny coil thickness but again you give no magnet length.

What is the 1.5v is that per coil? How do you intend to connect.

Taking a 3 phase star case assuming 40 turns per coil and your impossible flux density I get about 12v dc at 240 rpm. For a realistic flux density we should be looking at 80 turns per coil. If the magnets were 1/2" thick and coils 1/2" thick then I would expect this to work reasonably well.

Wind rating would be perhaps 150W. If you can pedal indefinitely then about 80W would probably be the thermal limit with about #16 wire.

Flux

vacuum1313 · « **Reply #16 on:** February 18, 2008, 09:32:11 AM »

Sorry some of the numbers didn't come out right when I copied them from the spreadsheet. When I previwed looked OK but there were a lot of &nbsp's that confused things. lets try again.

08973 = coil width in inches of wire not including central hole
18750 = coil thickness
75000 = magnet diameter
92946 = total coil width
5 = circumference at magnet centers (inches)
37 = circumference needed for 9 coils (inches)

therefore lots of space for coils

vacuum1313 · « **Reply #17 on:** February 18, 2008, 09:45:36 AM »

OK the numbers aren't coming out right, anything with a 0 in front of the decimal seems to think it's counting up, any suggestions to correct this? As for the impossible B value I thought that would be a problem. Was working on this about 1am last night and couldn't find a way I was confident of converting what I know about these magnets, N48 and size to tesla's but figured someone out there would be willing to point out what a dimbulb I am and set me to rights ;-))))! It was a choice between that value and a rule of thumb value I've seen a couple of times of 0.5T. Hadn't gotten far enough to figure out the effect of multiple coils, as I said I'm a biologist and am partly doing this for the fun of learning and not having almost all the answers my students have questions about. I really enjoy pushing my boundaries so please be patient I will learn eventually. Am planning on connecting as a normal 3 phase star format. For these magnets does the AWG 24 wire size at least sound right even if my power calculations bear no relation to reality?

vacuum1313 · « **Reply #18 on:** February 18, 2008, 09:50:18 AM »

Is there a way to easily pull in the comment one is replying to other than copy/paste to wordpad then back to the reply? Makes it easier to address comments.

TomW · « **Reply #19 on:** February 18, 2008, 10:08:29 AM »

13;

Is there a way to easily pull in the comment one is replying to other than copy/paste to wordpad then back to the reply? Makes it easier to address comments.

No. Manual only. I just copy it into my buffer and paste into the comment. No in between program. I add italics tags to get it to look different than my text.

TomW

Flux · « **Reply #20 on:** February 18, 2008, 10:14:27 AM »

I find that I can use back and forward on the browser to get back to the original from the new post.

Starting a line with 0 seems to screw things up. One way is to call the lines a, b c etc leave a space and then add the number beginning with 0.

Simplest way is to work the magnet at about 1/2 Br. for N48 that would be near 700mT. That way you have a decent gap for the coils without loosing half of it on mechanical clearance. It also works your magnet near BHmax. Usually an air gap somewhat less than two magnet thickness works out ok with coils about magnet thickness or a bit over.

I think your formula is ok for voltage per coil if N is number of turns per coil. I think strictly it gives average volts, the factor would be 2.2 for rms, but either way it is near enough as you will not know the exact flux density or waveform.

Assuming it gives you rms ac per coil, then you have 3 coils in a phase. The line volts in star will be 1.7 times phase volts. The dc mean will be about 1.4 times that.

Assuming 1.5v per coil, you have 4.5v phase. 7.78v line and 10.9v dc.

If you take your chosen diameter of disc, wind the coils with a hole about .7" diameter and wind the outside so that they just touch you should be about right. Unless your magnets are really thin you should manage a lot thicker than #24.

Flux

vacuum1313 · « **Reply #21 on:** February 18, 2008, 11:41:44 AM »

Okay lets see if I've learned anything.

Not sure I follow some of your figures. What is the 40 turns, is it turns per coil ( you say total turns).

I got the formulae from something I saw Hugh post. I had been thinking it was turns per coil, if that is in error please let me know. (the italics works, yay, thanks for the tip)

Your flux density figure is not realistic, that is about Br for neo in closed circuit, impossible in any air gap.

I was thinking this but couldn't find a good way to convert I was seeing 13700 gauss for these magnets and the conversion from gauss to Tesla was divide by 10000 hence the 1.37 originally.

I don't follow your tiny coil thickness but again you give no magnet length.

Not sure where I got it but somehow had the idea that the space between magnet faces on the rotors should be about the thickness of the magnet which are 1/4". That is 0.25". Take 1/4 of that, 0.0625", for the clearance between magnet face and coil (1/32" on either side of the coil) which leaves 0.1875" for the coil. If my reasoning is off again please kick and tell me so. The 1/32" clearance does seem to be tight to me but this was a 1st kick at the cat and one of those practical details I'm wanting to learn from the model.

Simplest way is to work the magnet at about 1/2 Br. for N48 that would be near 700mT.

For the formulae I make that 0.7T as the value for B?

If you can pedal indefinitely then about 80W would probably be the thermal limit with about #16 wire.

Another consideration I hadn't reached yet but was aware of, how does one figure the thermal limit of wire of a given size? Am I to take it you are suggesting using #16 wire

By the way, in case it isn't obvious I really do appreciate the time it takes for you guys to respond. I'm having a ball learning something completely outside my field.

Flux · « **Reply #22 on:** February 18, 2008, 11:55:32 AM »

I think it is turns per coil, seems to work out right. Area is is in M^2 not M but I think you realised that.

If those magnets are 1/4" thick then your gap and coil thickness is ok. You will get just under 700mT in that gap. That will severely restrict the wire size so you may be forced to use #24 and the power out will be drastically down on my prediction for 1/2" thick magnets.

Thermal limit is nearly impossible to predict. Wind ratings are typically 30% duty cycle and it is cooled with a good wind. With pedal power and no cooling you will be very restricted.

With those thin magnets I think you should aim for a higher speed and change the gear ratio if you want to use it seriously, if it is for play then it doesn't matter.

Flux

vacuum1313 · « **Reply #23 on:** February 18, 2008, 02:28:26 PM »

Magnets, yep 3/4" diameter (yes area calculated as m^2) by 1/4" thick. When I plug that into the formulae I get 0.77v per coil. I think I'm near the go and build it stage but still need to learn about bridge rectifiers. I see lots at 35 amps and upwards of 700v and am thinking this is overkill. Would it be long term a good idea to get one that will work with a 100 - 150 watt bike generator (not these magnets or coils), I plan on that being the 1st practical one. I'll charge a battery, run a tv off an inverter and get loads of exercise plus be ready for cycling season. Will oversizing a bridge rectifier result in loss of so much power the output won't be measurable? Yes this current one is principally for learning and play, if I can recharge batteries great, if not I'd at least like to light an LED. Thanks for all the help flux.

wooferhound · « **Reply #24 on:** February 18, 2008, 07:04:06 PM »

Quote

"Is there a way to easily pull in the comment one is replying to other than copy/paste to wordpad"

When you reply to a comment you are placed onto a page with a text box to compose your response. The entire comment that you are replying to is above the the Text box. just scroll up the page above the text box and copy your text from there.

Just like I did for this reply . . .

If you use the "Post a Comment" button, then the Original Post will appear above the text box for entering the comment. You can just scroll up to refer to the original post, or to Copy/Paste from it.

vacuum1313 · « **Reply #25 on:** February 19, 2008, 10:49:03 AM »

OK, have found that as a rule of thumb a bridge rectifier should be rated approximately double expected outputs to account for spikes. Haven't yet found if there is a large negative to massively oversizing the rectifier. For volts I'm thinking in the 24v range (yes that's probably really over optimistic) and am in the process of figuring the amps.

ghurd · « **Reply #26 on:** February 19, 2008, 11:02:52 AM »

No such thing as too big for a bridge.

A 400V is only a few cents more than a 100V.

A 25A is only a $1 more than a 10A.

If you know what I mean.

G-

PS- Notice how the numbers have "A" before them, above.

No goofy number problems because of the "A".

Flux · « **Reply #27 on:** February 19, 2008, 11:04:39 AM »

It is a good idea to rate the bridge for a minimum of twice dc volts. Normally this is no issue as the lowest ratings are usually 200v and the higher voltage ones only cost marginally more.

Current rating doesn't matter as long as it is adequate for the job. Again the claimed rating will be with an enormous heat sink and you will do better with something more highly rated. Within sensible reason there is no disadvantage with using something vastly over rated, it may start to be some issue if you chose things with ratings in the thousands of amps as the leakage might start to become significant.

What is more important for low voltage is the forward volt drop. The only thing that is significantly different are the schottky varieties. These are considerably better regarding volt drop. They unfortunately have limited voltage ratings. They are worth considering for 12v but I wouldn't bother above 24v. You will not get bridges, you will need to use 6 diodes.

Flux

vacuum1313 · « **Reply #28 on:** February 19, 2008, 05:36:50 PM »

You will not get bridges, you will need to use 6 diodes.

OK, a little more research and I believe I understand your reference here. I am unsure if any of the diode ratings are additive with respect to the overall rating of the system. For the little project I am looking at here would the diodes found in ebay item# 180215691212 be useful, I know its only for 5 but the poster indicates he would be able to make other custom postings. From the ratings I am thinking these would also be useful in the 100-150 watt TV generator I mentioned earlier as the 1st upscale useful sized alternator. If I am understanding the markings it would also only be necessary to wire up 2 of the 3 connectors on each diode. Comments?

madkane · « **Reply #29 on:** February 03, 2009, 02:33:48 PM »

the power of the mags some times not what you think as of the space needed for coils but i did well with a 10 turn coil and just run disc to a set speed and then can guess the teslas

you might have as an example 1.2volt at 400 rpm wiht 10 turns

looks like 1200/400/10= teslas then with that number play with speed or turns to reach ther desired point

News:

Author Topic: Magnets vs Coil design (Read 4844 times)