Rookie learning through DIY, prototyping alternator

[SimonMester]

Hello!

I have looked around these forums, and if something like this has been answered before, apologies.
I learned a bit from just searching around here, but I think people are doing a bit more advanced designs/ directly useful turbines.
I am simply learning about electronics by building from scratch. Or at least as close to it as possible. I have been trying to build a generator, and after a small prototype one (with 2 coils on sides) I decided to build a more proper prototype. I have 100mm long NdFeB magnets on a 80mm diameter rotor, and 12 stator coils around.
I have been, dissappointingly unable to get any decent current or voltage going from the setup. I know this must be due to not really knowing how to wire the 12 coils together.
I tried getting some information on that, but its either wired into rectifiers, or talking about 3 phase generators.
Logically speaking, my original idea was to just wire the ends into the starts of the next coil, essentially creating one big coil.
I tried many setups but nothing seems to get any real power going.
Even more strangely, the faster I spun the rotor, the less power I got out of it. If I just jerked the rotor back and forth, rocking it back and forth a fraction of a full rotation, I got multiple times as much power.

I just learned about parasitic capacitance, and wonder if that might be the entire reason? Or is it the wiring? I use notably thicker wire (0.85 mm) than in my first small prototype (0.2 mm). I knew this would influence the voltage/current, but not much more than that.

I don't use iron cores, most elements are tough 3D printed resin I designed in CAD.
The coils are hand wound, and thus not exactly even, despite using the 3D printed uniform backings to wind the wire around.

Bonus question: is lightly scratching the enamel coating on the coils of any significant effect? I had to remove them from their original scaffolds, so some of them got tiny scrapes here and there.

Any pointers are welcome!
Thank you!

[Bruce S]

Welcome to the forum, pull up a chair a sit a spell.

There are many more people here that will help you along this journey (aka addiction ) that are knowledgeable than I.
Having said that, maybe a good start would be to disconnect all the coils from each other (if possible) and start with getting readings from each of the coils to see how each one stands.
There is also a FAQ section that very knowledgeable people have contributed to if you are so inclined to do a bit more reading.

Cheers
Bruce S

[electrondady1]

what are the 3 dimensions of your magnets and how many do you have?

[SimonMester]

They are not large magnets, 100mm long, 10mm wide and deep. 8 of them on the rotor.
They are N52 neodymium.
My problem isn't the 'absolute' output of it, but I played around with the bar magnets before I put the coils together, and the comparatively the overall structure is not much better than when I was just doing that with a single coil or 2. Same goes for the first prototype which only had 2 coils and 4 small ferrite magnets, it actually functioned better.
So I'm fairly certain I'm doing something wrong in terms of having the coils too close or wiring them wrong.

[SimonMester]

Attached some pictures. The coil on the original backing, I have been removing them so I can play around with them more easily.
You can also see the embedded magnets.

[Mary B]

Small nicks could cause shorts, check each coil with an ohm meter to make sure they are identical.

Wired in series like that you get high voltage/low current. Wired in star(most common) you get lower voltage closer to battery voltage and higher current. Either can work if you use MPPT(Maximum Power Point Tracking) especially with the first.

Wiring diagram for the coils




This gives you 3 phase AC, you need a 3 phase rectifier to change it to DC so you can use it.

[SimonMester]

Small nicks could cause shorts, check each coil with an ohm meter to make sure they are identical.

Wired in series like that you get high voltage/low current. Wired in star(most common) you get lower voltage closer to battery voltage and higher current. Either can work if you use MPPT(Maximum Power Point Tracking) especially with the first.

Wiring diagram for the coils



This gives you 3 phase AC, you need a 3 phase rectifier to change it to DC so you can use it.

I have seen these diagrams, but I wasn't looking to do 3 phase, I was just looking to do single phase first, and understand what I'm doing wrong. I like to move from simple to more complex designs as I learn more.

I'm sorry what do you mean? I can measure resistance with my multimeter, but what would I see if the coils are okay, and what would I see if they weren't? Not sure how you mean I should check the coils.

[Bruce S]

That depends on your meter.
IF it's the old reliable swing pointer type and "good" coil will show a dead short, unless you have one that can read down into the single ohms range. And broken coil will show nothing, it'll be as if the leads aren't touching anything.

Newer digital type for a "good" coil will show something real low even when you switch it to lowest ohm setting, broken one will show all zeros like the leads aren't touching anything. Give us the size of the wire mm or awg and length, we can let you know what the Ohm (resistance) should be.

Bruce S

[electrondady1]

those are nice magnets, but the spacing between them is too wide .
 first, they should be sitting on an iron core .
the magnets should be alternating polarity.
the width between the magnets should be the same as the width of the magnet.
the hole in the center of the coil should be about the same width as the magnet or slightly wider.
the width of the legs of your coils should be the same as the width of the magnet.
you want to have one leg of a coil crossing a negative pole and the other leg crossing a positive pole.
as the mechanism rotates the polarity changes and electricity is generated in your coil.
it is the change of polarity that makes the alternating current

[SimonMester]

Thank you for that! I have gone through them on the lowest Ohm setting (which is 200 on my multimeter)
It just showed fairly random jumpy numbers from single to double digits on all of them.
As per above the current wire I am using is 0.85mm enameled copper wire.

Some news in the meantime!
I have liberated the coils from their resin backing. Was hard since I used a tough, flexible resin (which I instantly regret now )
I have been testing the coils with my rotor, and the numbers are really low, compared to the small ceramic magnet I have. I removed the rotor rod, which is mild steel, in case that messes with the magnets somehow, but no change.

I have considered maybe I ruined the NdFeB magnets with heat, as they are sensitive, but I tested the same NdFeB magnets (ones I have never used) and they induce the same low current. The magnets are strong, doesn't feel like there should be anything wrong with them, and yet the ceramics are seemingly better?

For context, each NdFeB is 100mm x 10mm x 10mm, and the ceramic magnet is 48 x 22 x 10 mm. Not on a rotor, just passing it above the coil, at the same speed and distance it produces about the same current.

Does this mean the limiting factor is not the magnet but the wire? Or is it do with something else?
Also, in any of these cases, it still seems to apply that the faster the magnets move the weaker current I get. Is this just me misunderstanding the readings, or some actual physical effect in the coils, like parasitic capacitance?

Thank you. ^^

[SimonMester]

those are nice magnets, but the spacing between them is too wide .
 first, they should be sitting on an iron core .
the magnets should be alternating polarity.
the width between the magnets should be the same as the width of the magnet.
the hole in the center of the coil should be about the same width as the magnet or slightly wider.
the width of the legs of your coils should be the same as the width of the magnet.
you want to have one leg of a coil crossing a negative pole and the other leg crossing a positive pole.

Thank you for the reply.
You mean the magnets should be sitting on iron? I have only heard about the coils having iron core so far, but I can't really do that, and I have seen that you can make air cored coils, so I thought I'd start there for learning.
The magnets are alternating polarity, I made sure of that. Since I have been having issues I double-triple checked with my spare magnet.
Ah I see! That is interesting information. Maybe my first prototype worked better despite being ceramic because it had very little space between the magnets on the rotor? Due to those magnets being bulkier. (for the same amount of money spent.)
I think I made the center holes in the coil 10mm which is the same as the width of the magnets.

"the width of the legs of your coils should be the same as the width of the magnet."
Could you rephrase that one please, I don't think I understand. :p

"you want to have one leg of a coil crossing a negative pole and the other leg crossing a positive pole."

Hold on. You mean the starting bit of the wire, of the coil, and the ending bit of the same coil, has to be aligned in a specific way? I haven't heard of this anywhere. They would need to be far apart to be "under" the field of two different magnets?
I'm sorry if I'm misunderstanding.

[Bruce S]

Looking at the pics of both the magnets and wire along with the information about magnets.
The ceramic ones are 2x+ wider than the Neos, this is why you're seeing a "higher" number.
Also the reason why it was stated the spacing between the Neos was too wide.
Just as one leg of the coil is passing over the N side of the mag the other should be passing the S side of the next mag.

This "will work" without iron, just not nearly as good since the iron will give the flux a return path.

Since you printed via 3D printer you could print a new one using closer slots for the magnets (also using more mags) or you could open up the slots and add another mag to each of the ones you have in there.

Bruce S

[SimonMester]

Looking at the pics of both the magnets and wire along with the information about magnets.
The ceramic ones are 2x+ wider than the Neos, this is why you're seeing a "higher" number.
Also the reason why it was stated the spacing between the Neos was too wide.
Just as one leg of the coil is passing over the N side of the mag the other should be passing the S side of the next mag.

This "will work" without iron, just not nearly as good since the iron will give the flux a return path.

Since you printed via 3D printer you could print a new one using closer slots for the magnets (also using more mags) or you could open up the slots and add another mag to each of the ones you have in there.

Bruce S

Aaah, that makes sense, I understand now.
Unfortunately I just can't find anything to act as a core. It can't just be any lump of iron, that I know, and getting electrical steel or laminated iron, especially in the site that I'd want, would be close to impossible.
Tbh even if it could be any lump of iron, even pure iron is hard to find, everything is steel.
But first things first I just want to get it to work and then I can think about iterating more.
Exciting! I get to get back to planning.

Thank you for the pointers I'll redesign the rotor and find more magnets.

EDIT: Having re-read all the replies 3 times over I'm starting to understand my own ignorance. Each side "leg" of the coil, should be as wide as the magnet, and be under the influence of a magnet at the same time (different poles).
See my poorly drawn paint explanation of how I understand it now.
My brain is expanding.
My coils are too skinny and my magnets are too lonely.

[electrondady1]

you can fasten a steel sleeve to go over  your existing axial.
 you could anchor it with a set screw or a  tack weld.
the sleeve must give you the proper diameter so that your magnets can sit on the sleeve and have the proper spacing.
you need to do a cross section / end view drawing to discover that dimension.
you may get lucky and find  a piece of steel pipe that is just right or close.
the steel pipe/sleeve needs enough wall thickness to absorb the magnetic flux of the mags.
 i like the way the mags are sitting in an indentation of the blue/gray material.
but the mags should be sitting directly on mild steel or iron .

in addition ,  speaking about the coil leg / magnet comparison diagram  you created,  the coil legs could be longer. 
so that the magnet could actually slide through the hole in the coil.

if you can create something  that rotates  the mag rotor ether by a crank or motor  or what ever   and create something to hold your coils. it gives you a platform for experimentation with  wire diameter and the number of turns in the coil.

[MattM]

The field on a neodymium is stronger and more intense, but from what I've seen from my own magnets is the intensity drops off fast with distance.  Ceramics are much less intense but seem to reach further out before they drop off in magnetic strength.  When you stick them on the correct type of magnetic iron the fields seem to reach further distances with pretty decent intensity.  But if you choose a poor type of iron its not very much improved at all.  Some iron alloys are better than others and a high nickel content is generally desirable.  Not everyone seems to agree with me here, but this is what I have observed with my untrained eye.

[SimonMester]

you can fasten a steel sleeve to go over  your existing axial.
 you could anchor it with a set screw or a  tack weld.
the sleeve must give you the proper diameter so that your magnets can sit on the sleeve and have the proper spacing.
you need to do a cross section / end view drawing to discover that dimension.
you may get lucky and find  a piece of steel pipe that is just right or close.
the steel pipe/sleeve needs enough wall thickness to absorb the magnetic flux of the mags.
 i like the way the mags are sitting in an indentation of the blue/gray material.
but the mags should be sitting directly on mild steel or iron .

in addition ,  speaking about the coil leg / magnet comparison diagram  you created,  the coil legs could be longer. 
so that the magnet could actually slide through the hole in the coil.

if you can create something  that rotates  the mag rotor ether by a crank or motor  or what ever   and create something to hold your coils. it gives you a platform for experimentation with  wire diameter and the number of turns in the coil.

The actual coil hole and magnet align decently well I think? The legs are a bit skinny, but for first try I'll just have to go with this, as long as the spacing works so the magnets align with them it should still work, just less so I take it.
I'm going to really struggle finding anything with pure iron, its really hard to find. I have this steel foil, 0.2mm, which is perfect size but its stainless steel.
Any rule of thumb what thickness the iron should be, depending on the magnet size?

I might have found a company that sells flat bars and pipes in 'mild steel'. How optimal is that compared to actual iron (or soft ferrite) or just not having any metal? Just wondering if there is any general idea or consensus how good they are compared to each other.

Thank you for your answers!

[SimonMester]

That depends on your meter.
IF it's the old reliable swing pointer type and "good" coil will show a dead short, unless you have one that can read down into the single ohms range. And broken coil will show nothing, it'll be as if the leads aren't touching anything.

Newer digital type for a "good" coil will show something real low even when you switch it to lowest ohm setting, broken one will show all zeros like the leads aren't touching anything. Give us the size of the wire mm or awg and length, we can let you know what the Ohm (resistance) should be.

Bruce S

I'm not exactly sure what length the wire is, but I used about 100grams of copper for a coil. It's 0.85mm wire, so napkin math would suggest about 5 meters of wire roughly?
How do you calculate what's a healthy resistance for that? Also, I have a digital multimeter.

[electrondady1]

now is the time to make friends with your local machine shop and scrap yard.
you need a piece of round bar stock with a hole in the middle for your axle to go through.
as well as being turned down to the proper diameter to sit you mags on .

that steel foil might be a good material to mount your coils to .

[Bruce S]

That depends on your meter.
IF it's the old reliable swing pointer type and "good" coil will show a dead short, unless you have one that can read down into the single ohms range. And broken coil will show nothing, it'll be as if the leads aren't touching anything.

Newer digital type for a "good" coil will show something real low even when you switch it to lowest ohm setting, broken one will show all zeros like the leads aren't touching anything. Give us the size of the wire mm or awg and length, we can let you know what the Ohm (resistance) should be.

Bruce S

I'm not exactly sure what length the wire is, but I used about 100grams of copper for a coil. It's 0.85mm wire, so napkin math would suggest about 5 meters of wire roughly?
How do you calculate what's a healthy resistance for that? Also, I have a digital multimeter.

That is roughly 0.1ohm. This is why it'll read as a near short or go wild with numbers. This is a good thing!
Means that your wire isn't broken inside of the coil.

[SimonMester]

That depends on your meter.
IF it's the old reliable swing pointer type and "good" coil will show a dead short, unless you have one that can read down into the single ohms range. And broken coil will show nothing, it'll be as if the leads aren't touching anything.

Newer digital type for a "good" coil will show something real low even when you switch it to lowest ohm setting, broken one will show all zeros like the leads aren't touching anything. Give us the size of the wire mm or awg and length, we can let you know what the Ohm (resistance) should be.

Bruce S

I'm not exactly sure what length the wire is, but I used about 100grams of copper for a coil. It's 0.85mm wire, so napkin math would suggest about 5 meters of wire roughly?
How do you calculate what's a healthy resistance for that? Also, I have a digital multimeter.

That is roughly 0.1ohm. This is why it'll read as a near short or go wild with numbers. This is a good thing!
Means that your wire isn't broken inside of the coil.

I seeeee!
Some of them were fairly stable tiny ohm readings, some were jumping wildly but went to near zero as I moved the probes around a bit. Good news and thank you for the information I'm learning a lot from you guys!
I was worried but the scratches are tiny so it makes sense.

I thought about using the steel foil as backing for my coils now that I stripped them from their original resin molds. It is the perfect size.
I don't know of a machine shop nearby but I found a company selling some mild steel bits cheap so I'll get some to experiment with. They do cut them to size if you ask nicely.

I have also managed to talk to an electrical engineer friend and he has some ferrite winding cores. Not the right size exactly, but I'll take the freebies. I'll do some readings for:

- air core coil, bare magnet
- air core coil, mild steel backing magnet
- mild steel core coil, bare magnet
- mild steel core coil, mild steel backing magnet
- soft ferrite core coil, bare magnet
- soft ferrite core coil, mild steel backing magnet

I'm just really curious to see what effect these factors have.

[Bruce S]

Another place to look for fairly high-grade steel.
Old broken microwaves 
DO BE CAREFUL!!! The Capacitors can be dangerous and don't mess around with the magnetron!!  <=whole lot of nasties inside that shouldn't be opened!!!
The coils however should still made of silicon steel and will be a good grade steel will be of an "E" configuration and can easily be cut. Best of all free99 

Cheers
Bruce S

[SimonMester]

What is the reason for having to put empty space between the magnets? Are the two different poles next to each other weakening the field in some way if they are closer?

[Bruce S]

You need the space for the field of one part moving to collapse before the other builds.
Easiest way I can explain it.

 

[joestue]

call up your local HVAC companies and ask them for broken ECM fan motors.

almost all of them are fixable for use as generators.

the cores, magnets, copper windings are all optimized to reduce cost. i suppose if you were to replace the ferrite magnets with neodymium, you could double the torque density, at the cost of increased losses due to core saturation.

[MattM]

Be aware your magnetic field is directional and permeates space in every direction.  It also induces eddy currents that have their own magnetic fields, therefore the goal is not to simply add iron.  The goal is to add iron that can generate useful magnetic fields.  Round stock, cut into small nubs spaced into the empty spaces is easier to find than wedges that perfectly fit.  But you get even better results the more granular size of the iron emoloyed, because it more directionally reacts to passing magnetic fields.
BBs work, but there are even smaller options out there.  People have imbedded iron filings into epoxy for excellent results.

[SimonMester]

Be aware your magnetic field is directional and permeates space in every direction.  It also induces eddy currents that have their own magnetic fields, therefore the goal is not to simply add iron.  The goal is to add iron that can generate useful magnetic fields.  Round stock, cut into small nubs spaced into the empty spaces is easier to find than wedges that perfectly fit.  But you get even better results the more granular size of the iron employed, because it more directionally reacts to passing magnetic fields.
BBs work, but there are even smaller options out there.  People have imbedded iron filings into epoxy for excellent results.

I know about Eddy currents and the resulting iron losses (well I know they exist, not much in detail ).
The two reasons I wanted to do entirely air core at first, is because getting silicon steel or laminated core is not really possible for me, and the second reason being I imagined just making the basics first, so I have a working prototype, and then adding some form of iron to see how things change and can be iterated on.
I'm going to try mild steel as I got some of that for cheap, and see if it helps at all.
Iron filings in epoxy sounds like a great idea! I dont know what density of iron it should have, but if its anything like the laminated core, it should be very little epoxy I imagine?
In any case, that should be fairly easy to cast into some kinda mold. Just need some epoxy that doesnt stick too hard to it.

Thank you for your comment.

[SimonMester]

Okay so I have made a lot of progress in random directions, however I am once more in the need of wisdom.
I am trying to measure a lot of different setups and materials, to see what is best for getting the most out of coils and cores.
To this end I use a multimeter and try to make an educated guess what is better.
Since AC is produced, in the range of miliamps and milivolts, with the numbers jumping quite a lot, it's sometimes difficult to say if one things was better or another.
Is there an instrument that could record these readings in time so I can better analyse than my shoddy 8$ multimeter?
Would there be any sense in maybe using capacitors instead to concentrate the generated energy, so I get a better solid reading, and just factor in the time it took to charge the capacitor?

I am quite a rookie, and my googling in electronics equipment hasn't really turned anything up, as it is a specific field, and if you don't know what you are searching for it doesn't really pop up.
Thank you in advance!

[Mary B]

Okay so I have made a lot of progress in random directions, however I am once more in the need of wisdom.
I am trying to measure a lot of different setups and materials, to see what is best for getting the most out of coils and cores.
To this end I use a multimeter and try to make an educated guess what is better.
Since AC is produced, in the range of miliamps and milivolts, with the numbers jumping quite a lot, it's sometimes difficult to say if one things was better or another.
Is there an instrument that could record these readings in time so I can better analyse than my shoddy 8$ multimeter?
Would there be any sense in maybe using capacitors instead to concentrate the generated energy, so I get a better solid reading, and just factor in the time it took to charge the capacitor?

I am quite a rookie, and my googling in electronics equipment hasn't really turned anything up, as it is a specific field, and if you don't know what you are searching for it doesn't really pop up.
Thank you in advance!

A modern oscilloscope would do averaging BUT a steep learning curve and around $300 for a good one, data logging multimeter would work but expensive... O-scope I have https://www.amazon.com/gp/product/B07XSH2NR2/ref=ppx_yo_dt_b_asin_title_o06_s01?ie=UTF8&psc=1 you could probably get by with 1 or 2 model levels lower... 100mhz bandwidth would be plenty... but you only save about $60...

an analog multimeter with a needle that moves is easier to read on rapidly changing inputs, they are relatively cheap, think the Harbor Freight freebie they used to give away. Not very accurate though!

[SimonMester]

Okay so I have made a lot of progress in random directions, however I am once more in the need of wisdom.
I am trying to measure a lot of different setups and materials, to see what is best for getting the most out of coils and cores.
To this end I use a multimeter and try to make an educated guess what is better.
Since AC is produced, in the range of miliamps and milivolts, with the numbers jumping quite a lot, it's sometimes difficult to say if one things was better or another.
Is there an instrument that could record these readings in time so I can better analyse than my shoddy 8$ multimeter?
Would there be any sense in maybe using capacitors instead to concentrate the generated energy, so I get a better solid reading, and just factor in the time it took to charge the capacitor?

I am quite a rookie, and my googling in electronics equipment hasn't really turned anything up, as it is a specific field, and if you don't know what you are searching for it doesn't really pop up.
Thank you in advance!

A modern oscilloscope would do averaging BUT a steep learning curve and around $300 for a good one, data logging multimeter would work but expensive... O-scope I have https://www.amazon.com/gp/product/B07XSH2NR2/ref=ppx_yo_dt_b_asin_title_o06_s01?ie=UTF8&psc=1 you could probably get by with 1 or 2 model levels lower... 100mhz bandwidth would be plenty... but you only save about $60...

an analog multimeter with a needle that moves is easier to read on rapidly changing inputs, they are relatively cheap, think the Harbor Freight freebie they used to give away. Not very accurate though!

Well, I only really need to be able to compare setups against each other, in that sense accuracy is not the gravest of issues, as long as they are all inaccurate to the same degree.
I can get an analog one for 5$ so I'll def give that a go.
An oscilloscope is definitely out of the budget for now.

[SimonMester]

Okay so I have made a lot of progress in random directions, however I am once more in the need of wisdom.
I am trying to measure a lot of different setups and materials, to see what is best for getting the most out of coils and cores.
To this end I use a multimeter and try to make an educated guess what is better.
Since AC is produced, in the range of miliamps and milivolts, with the numbers jumping quite a lot, it's sometimes difficult to say if one things was better or another.
Is there an instrument that could record these readings in time so I can better analyse than my shoddy 8$ multimeter?
Would there be any sense in maybe using capacitors instead to concentrate the generated energy, so I get a better solid reading, and just factor in the time it took to charge the capacitor?

I am quite a rookie, and my googling in electronics equipment hasn't really turned anything up, as it is a specific field, and if you don't know what you are searching for it doesn't really pop up.
Thank you in advance!

A modern oscilloscope would do averaging BUT a steep learning curve and around $300 for a good one, data logging multimeter would work but expensive... O-scope I have https://www.amazon.com/gp/product/B07XSH2NR2/ref=ppx_yo_dt_b_asin_title_o06_s01?ie=UTF8&psc=1 you could probably get by with 1 or 2 model levels lower... 100mhz bandwidth would be plenty... but you only save about $60...

an analog multimeter with a needle that moves is easier to read on rapidly changing inputs, they are relatively cheap, think the Harbor Freight freebie they used to give away. Not very accurate though!

Wouldn't something like this work? : https://www.ebay.co.uk/itm/285182708844?var=587059028930

[JW]

Here follow this  link    https://www.fieldlines.com/index.php?topic=128519.0

There certain harmonics to winding frame the stator.

I want to point out that the 9 coils by 12 magnets we have worked on this forever its a good arrangement, and it works in theory.