Author Topic: Question about turns calculator (Read 3175 times)

solarengineer · « **on:** November 07, 2006, 10:38:38 PM »

I have a question about the number_of_turns2.xls calculator

when i run various magnet sizes for existing machines ive seen on this board, the calculator doesn't quite spit out the right answer unless the number of poles includes magnets on both rotors..

i.e for 12 magnets per rotor you enter 24 in the poles field. for the right number of turns.

does this sound right?

for existing machines like DanB's machines, the number of turns the calculator spits out is only close if you include the magnets on the second rotor as well. so again you put 24 in the poles field.

so, is it correct that if i'm making a machine thats dual rotors with 12 magnets per rotor i would put 24 in the poles field in order to calulate the number of turns correctly?

I know the rest of it. it's kust the # of poles to enter thats confusing me.

thanks!!

Jamie

windstuffnow · « **Reply #1 on:** November 07, 2006, 04:22:49 PM »

If all the input values are correct it works quite well in predicting the outcome of an alternator. You only use the pole count of 1 disc. Different magnets at different distances give you a different Tesla (gauss) reading. For instance my wedge magnets with a 3/8" gap have an average reading of 7000 gauss ( .7 T ) and with a gap of 5/8" the reading falls to 5000 gauss ( .5 T ). The 1/2" thick block magnets ( 1 x 2 x .5 ) will give you about .6T with a 5/8" gap and .75 T with a 3/8" gap.

.

solarengineer · « **Reply #2 on:** November 07, 2006, 05:08:42 PM »

Whats odd though is for 2" x 1" x .5" neos and 12 poles it's saying to use

65 turns per coil for star and I've use 37 turns per coil on my tims turbine clone

so if i went based on the calculator i would end up with a 24v machine??

but if you enter 24 poles it comes out right.

please help me to understand my error so i can use it properly.

Turns of wire

volts    13    Open voltage

poles    12    total magnets

rps    2.27    revolutions per second (rpm/60)

tesla    0.55    gauss / 10000

area    0.00129    sq in x .000645 of 1 magnet

coils    3    default is 3

# turns for delta    112    turns per coil

# turns for star    65    turns per coil

thanks!!!!

windstuffnow · « **Reply #3 on:** November 07, 2006, 06:55:36 PM »

I would wonder what grade the magnets are, that might make a considerable difference in Tesla input on the spreadsheet. If all else is correct then I would suspect the average gauss in the air gap is higher than .55 T.

.

SparWeb · « **Reply #4 on:** November 07, 2006, 11:35:54 PM »

I've got a question: Where did you get the spreadsheet? I'd be interested in reading it, now that I've gone to all the trouble of developing my own program!

:^)

Flux · « **Reply #5 on:** November 08, 2006, 11:14:21 AM »

Not sure where this spreadsheet comes from but something is odd.

Doing it from basics

2 sq inch is 12.9sq cm or 12.9/10,000 sqM

Total flux per pole is .55 x 12.9/10 = .71mWb

Speed is 2.27 x 60 = 136 rpm or 13.6 Hz for 12 poles.

Using the 65 turns, turns per phase is 65 x 3 =195 turns.

Rms volts/ phase is 4.4 x flux per pole x f x no. of turns

4.4 x (.71/1000) x 13.6 x 195 = 8.28v

Line volts = 8.28 x 1.73 and dc volts = 8.28 x 1.73 x 1.4 = 20vdc.

Well above the intended 13v

"area 0.00129 sq in x .000645 of 1 magnet"

No idea where this comes from but one of these numbers must be suspect.

I don't trust things with constants in them unless I can see where they come from.

Flux

windstuffnow · « **Reply #6 on:** November 08, 2006, 01:41:56 PM »

Flux, the .000645 is simply a conversion from sq inches to sq meters. The result is the same as your 12.9/10000.

2 sq in x .000645 = .00129 sq meter area of 1 magnet.

Can you elaborate on the above formula?

Flux · « **Reply #7 on:** November 08, 2006, 02:35:10 PM »

Thanks Ed. I assume from this it must be your spreadsheet.Thanks for clarifying the conversion to sq M. I am piggy in the middle here, I haven't seen the spreadsheet, just have the data in the question to go on.

I assume it is the emf formula that you want me to elaborate on.

Total flux per pole (phi) is B x A. B is flux density in Tesla and area is square Meters if we use SI units.

In each rotation each conductor cuts (phi) lines of flux and p is the number of poles.

Each turn has 2 active sides so the average emf is 2(phi)Pn where n is in revs/sec.

I prefer to work in frequency and f=pn/2

That gives the average emf as 4(phi)f volts per turn

The average volts per phase is 4(phi)f x N where N is the total number of turns per phase.

These type machines give a close approximation to a sine wave and we are justified to use the sine wave form factor of 1.1

That gives the rms volts per phase as 4.4(phi)fN.

For star connection the line volts ( between a pair of terminals) will be root3 times this. When used with a 3 phase rectifier charging a battery the dc will be almost the peak value of the rms.

Near enough, the dc voltage will be 4.4(phi)fN x 1.7 x 1.4

I find that this works out fairly closely with these air gap alternators. It works for slotted iron cores but you need to take various distribution factors into account depending on the winding.

Although slightly different, the same emf equation still works with the transformer. That is why I prefer to work in frequency then one formula does the lot.

Flux

windstuffnow · « **Reply #8 on:** November 08, 2006, 04:26:07 PM »

Thanks Flux ! I'll play around with it tonight when I get some free moments.

The spreadsheet is based on the formula; NARBP*2 and doesn't take into account the conversion from AC to DC. I see your formula gives this a 40% increase which would drop the turns considerably. I never added it to the formula but used 30% to adjust my windings after the fact. When I plug in the 1.4 it comes out pretty close to your calculated output with 195 turns per phase. Interesting stuff !

Anyway, thanks again ! I like playing with numbers and formula's should be fun.

.

solarengineer · « **Reply #9 on:** November 08, 2006, 04:32:39 PM »

HI Ed, i was wondering if you do modify the sheet if it would be ok to request a copy.

Thanks for everyones help!

Very interesting stuff!!

Jamie

solarengineer · « **Reply #10 on:** November 08, 2006, 04:41:51 PM »

Hi Flux, thanks for the input. It still strikes me curious on the tims turbine clone using 37 turns per coil and 9 coils. = 111 turns per phase not 195.

I've build a clone of this turbine with 2x1x.5" n40 neos using 9 coils with 37 turns 2 in hand #14 guage.

The turbine works fine for 12 volts. but your calculations come up with almost twice as many turns. wouldnt that give a 24 volt machine??

the spreadsheet agrees with your results for the most part but still shows me twice as many turns as DanB used on the tims turbine.. and twice as many as i used on mine.

I'm just trying to figure this out so i can reliably use the spreadsheet to design different stators.

Thanks

Jamie

windstuffnow · « **Reply #11 on:** November 08, 2006, 06:56:32 PM »

Jamie,

The N40's are probably giving you a Tesla of around .7 in the air gap which would bring the turns down to what you are using. This is what I was getting at in an earlier post.

My spreadsheet doesn't consider the rectified DC output as does Flux's formula. You an add that in, in the single phase/delta calculation you would divide by 1.4 to get the number of turns. For example in the # turns row the calculation would look like this...

=ROUND(B2/(B3*B4*B5*B6*2)/B7/1.4,0)

The star will automatically adjust the turns as its based on the single/delta number

.

Flux · « **Reply #12 on:** November 09, 2006, 12:32:23 AM »

I ran the figures that you gave for the spreadsheet just to see what happened.

I came up with 20v cut in at the 136 rpm. with the 65 turns. For cut in at 12v that would have given

12/20 x 65 = 39 turns.

I would expect above .55T with a reasonable grade of magnet so your 37 turns seems quite in line with normal expectations.

I think Ed has found the problem in not allowing for the 1.4 factor with dc being near peak with the 3 phase rectifier. When this is sorted it should be very useful.

As Ed mentioned earlier the biggest error will usually be in estimating the flux density.

Flux

solarengineer · « **Reply #13 on:** November 09, 2006, 09:51:29 AM »

just wanted to say how much i appreciate the time and effort given to my post..

Thank you all VERY MUCH!!

Jamie

wayne · « **Reply #14 on:** November 09, 2006, 04:07:03 PM »

Hi Ed

I ran your program with the change using your 6" radial air core. And more confussed with the amount of coils 18 vs your 11 turns. Did I run the numbers right of did I miss something. Trying to build a small one but no real luck yet. Thought I would try this and work it from here. Or this would not work for radial generators!

Thks

Wayne

Turns of wire        For D.C output Ed's Demo

volts    13    Open voltage

poles    18    total magnets poles

rps    3    revolutions per second (rpm/60) 136rpm's 3=180 rpm's

tesla    0.65    gauss / 10000 at .25 gap

area    0.000129    sq in x .000645 of 1 magnet

coils    18    54 holes or 18 per phase

# turns delta    57

windstuffnow · « **Reply #15 on:** November 09, 2006, 06:03:13 PM »

Hi Wayne,

Since it's an air core the Tesla number would be much lower. Somewhere in the range of .25 to .3 . I believe the surface reading was about .55 but at 1/4" away it drops quickly with no return path. I'm using N35 blocks in that one, N40's or higher would definately increase its output although it does quite nicely for what it is.

That was a fun project overall, quick and simple... as long as you have the patience for drilling lots of holes!

.

wayne · « **Reply #16 on:** November 10, 2006, 08:27:35 AM »

Hi Ed

I entered a higher tesla number to bring down the #turns star to (33). And if I lower the tesla number to .3 example then # turns star jumps up to 70. Maybe I am missing the boat but with your 11 turns you had 12 volts at about 120 rpms. I am more stumped than ever now and don't understand. I have my rotor built and in my files now.

Thanks

Wayne

windstuffnow · « **Reply #17 on:** November 10, 2006, 10:13:23 AM »

This is what I get when I plug in the numbers... Not sure how your comming out different.

Turns of wire

volts    12    Open voltage

poles    18    total magnets - one disc

rps    2.25    revolutions per second (rpm/60)

tesla    0.25    gauss / 10000

area    0.00129    square inches x .000645 of 1 magnet

coils    18    per phase

# turns    18    single phase or delta

star    10

2 phase    13

rpm 135

.

wayne · « **Reply #18 on:** November 10, 2006, 01:53:24 PM »

Ed

I downloaded another copy and did work like yours. I feel really silly and maybe my old hands did something wrong with change. But all is good and thank you for your help again.

Have a Good Weekend

Wayne

spinner · « **Reply #19 on:** December 03, 2006, 10:01:54 AM »

flux

is there a reasonably acurate way to predict # of turns/size of wire IF gauss is unknown? like many others on this board, i know of no way to measure gauss if i am trying to work with used magnets.......

thanx

spinner

News:

Author Topic: Question about turns calculator (Read 3175 times)