# Energy harvesting using a brushless generator

#### cmartinez

Joined Jan 17, 2007
7,552
Say I have a toroidal coil, and I insert a diametric magnet in it and start spinning it.

Is this an efficient way of generating power? Would it be better to partition the coil in several segments, and use several magnets instead of just one? Would it be better to make the coil's outer diameter larger, so as to step further away from the magnetic field?

That last question is because I'm guessing that the flow of current induced in the coil's outer diameter would have an opposite direction than the one induced in the inner diameter, and the result would be heat in the coil, instead of net current flow through the load.

The purpose of my question is to find if a more or less efficient generator could be built by spinning a magnet, instead of a wire-wound rotor that would need brushes. The contraption's outer diameter needs to be no larger than 35mm, and its thickness 15mm or less. The power I'd like to generate would be about 2 mA, at 6 to 9V if possible. In the end, I'd like to calculate the magnet strength, number of wire turns, and rotational speed needed to obtain such power.

And yes... I did some googling around before posting my question here. But the results were a bunch of crap promoting free energy and overunity nonsense. Maybe I used the wrong keywords for my search.

To those new to this forum: The discussion of overunitiy (a.k.a. free energy crap) goes against the T.O.S. of A.A.C. Please refrain from posting comments related to that subject in this, or any other thread.

#### crutschow

Joined Mar 14, 2008
28,545
Your scheme will not generate any output voltage because, as you noted, one-half will be generating voltage and in direction and the other half in the other direction, so the net voltage will be zero.
There would be no heat because there would be no current.

You need to look at how generator windings are wound.
I believe you can put two windings, one on each half of the toroid, and wire then in inverse series to get an AC output.

And there already are permanent magnet generators in use, such as most motorcycles, and I think the inverter type AC generators.
The reason they are not used more is because their is no easy way to control their voltage.

Motorcycles do it by simply dumping the excess to ground. I've burned out three regulators on my cycle because they weren't properly designed to dissipate all the heat and overheated.

I believe some large truck alternators use a permanent magnet rotor and then externally apply an opposite magnetic field from an electromagnet to subtract from the rotor magnetic field to control the voltage.

Last edited:

#### cmartinez

Joined Jan 17, 2007
7,552
Your scheme will not generate any output voltage because, as you noted, one-half will be generating voltage and in direction and the other half in the other direction, so the net voltage will be zero.
Right ... at this moment I'm trying to visualize what the current would look like. And at first I thought that the inner and outer diameters of the coil would have opposing directions in current, which is true. But then after reading your post I also realized that the North pole of the magnet would be generating current in one direction, and the South pole in the other ... resulting in zero current flow, as you've just said.

So ... what if I were to replace the single diametric magnet with many smaller ones? I'd install them with their south poles facing inwards, and the north ones facing outwards. That is, their north poles would be facing the coil. The disadvantage to this arrangement would be that attaching the magnets in such matter would be a bit laborious.

#### crutschow

Joined Mar 14, 2008
28,545
So ... what if I were to replace the single diametric magnet with many smaller ones? I'd install them with their south poles facing inwards, and the north ones facing outwards. That is, their north poles would be facing the coil
Then you'd get little voltage output the coil since there would be no change in polarity.

What's wrong with my suggestion to wind two windings connected in inverse polarity, one on each half of the stator?
That way the two opposite polarities in each half will add together, giving the desired AC output.

#### cmartinez

Joined Jan 17, 2007
7,552
Then you'd get little voltage output the coil since there would be no change in polarity.

What's wrong with my suggestion to wind two windings connected in inverse polarity, one on each half of the stator?
That way the two opposite polarities in each half will add together, giving the desired AC output.
I think I understand ... if things are arranged the way you're suggesting, the same amount of current will be obtained from the device, but at twice the voltage? So twice the power would be obtained at the same RPM's? With a corresponding increase in the torque applied to the shaft, of course.

#### GopherT

Joined Nov 23, 2012
8,012
Toroidal cores are used to PREVENT any magnetic fields from forming outside of the core. That is, there is no access to a North or South magnetic pole because the core itself completes the magnetic circuit.

Toroidal inductors are used when you want limit EMI crossover or emissions.

That being said, I don't see how they will be helpful for a motor unless you wind them in a lace pattern instead of your way.even those were more "novelty items" rather than efficient motors.

#### crutschow

Joined Mar 14, 2008
28,545
the same amount of current will be obtained from the device, but at twice the voltage
I believe that's true.

#### crutschow

Joined Mar 14, 2008
28,545
Toroidal cores are used to PREVENT any magnetic fields from forming outside of the core. That is, there is no access to a North or South magnetic pole because the core itself completes the magnetic circuit.
Really?
Then how can a toroidal transformer work, since the wires must see a magnetic field to induce the voltage in the secondary coil?
The core concentrates the core and minimizes stray fields, but it does not prevent them from forming outside the coil which is necessary for transformer action.

#### BR-549

Joined Sep 22, 2013
4,938
Only the flux thru the center of a loop counts.

#### ebp

Joined Feb 8, 2018
2,332
A toroid core basically bends the magnetic field into a continuous ring, which largely keeps it from coupling to adjacent things but in no way prevents multiple winding from coupling to each other. It is much like a straight solenoid core, except there is a complete, closed path for the magnetic flux. A solenoid core slops mag flux all over the landscape.

On permanent magnet generators: One of my first significant projects as an insultant was the design of an SCR-based regulator for charging lead acid batteries from a 2 kW permanent-magnet explosion proof generator that ran at variable speed. It was a whole lot of analog circuity. If I did it today I'd use a fast 32 bit microcontroller and do almost everything but the power path in firmware.

Though the brushes are a wearout problem, a small permanent magnet motor, as might be used in a toy, makes quite a good generator. If you want to avoid brushes, you build it "inside out" using a permanent magnet multi-pole (2 poles might be enough) armature and a wound stator. A small brushless DC motor might work quite well - something that might be used for an optical or mag disk spindle or perhaps even a small fan. You might do what you want with a some small strong magnets fastened to an armature and some small coils wound on individual iron cores arranged as the stator. You could probably make it spin fast enough to get 20 mW just by glaring menacingly at it. If there is risk of overspeed, use a simple shunt regulator. You might be able to get that much power from a small variable-reluctance generator - the stator consists of coils wound around small cylindrical magnets, the armature a sort of (steel) gear with sparse teeth. Each time a tooth passes a magnet, the coil produces an AC "pulse". Gather up all the pulses with some small diodes. Any of these might pose some torque ripple problems for whatever is providing the mechanical input.

#### cmartinez

Joined Jan 17, 2007
7,552
A toroid core basically bends the magnetic field into a continuous ring, which largely keeps it from coupling to adjacent things but in no way prevents multiple winding from coupling to each other. It is much like a straight solenoid core, except there is a complete, closed path for the magnetic flux. A solenoid core slops mag flux all over the landscape.

On permanent magnet generators: One of my first significant projects as an insultant was the design of an SCR-based regulator for charging lead acid batteries from a 2 kW permanent-magnet explosion proof generator that ran at variable speed. It was a whole lot of analog circuity. If I did it today I'd use a fast 32 bit microcontroller and do almost everything but the power path in firmware.

Though the brushes are a wearout problem, a small permanent magnet motor, as might be used in a toy, makes quite a good generator. If you want to avoid brushes, you build it "inside out" using a permanent magnet multi-pole (2 poles might be enough) armature and a wound stator. A small brushless DC motor might work quite well - something that might be used for an optical or mag disk spindle or perhaps even a small fan. You might do what you want with a some small strong magnets fastened to an armature and some small coils wound on individual iron cores arranged as the stator. You could probably make it spin fast enough to get 20 mW just by glaring menacingly at it. If there is risk of overspeed, use a simple shunt regulator. You might be able to get that much power from a small variable-reluctance generator - the stator consists of coils wound around small cylindrical magnets, the armature a sort of (steel) gear with sparse teeth. Each time a tooth passes a magnet, the coil produces an AC "pulse". Gather up all the pulses with some small diodes. Any of these might pose some torque ripple problems for whatever is providing the mechanical input.
Excellent lecture, ebp. I'll look up what you're suggesting. Many thanks!

#### cmartinez

Joined Jan 17, 2007
7,552
Alright, I'm back.

Here's what I want to do. I want to build a very small (about 2 in diam) generator that uses no brushes. The output can be A.C. but it will have to be rectified. It should produce enough energy at about 60 rpm for it to charge a capacitor bank in 20 minutes. I do not know the amount of charge I'll be needing yet, but it should be enough to actuate a very small solenoid that would generate a force of about 30 grams for one second.

I've been thinking about using a simple diode-capacitor voltage multiplier at its output for it to raise the input voltage to around 6 to 8V, which will then be used to charge the aforementioned capacitors.

Voltage in excess of the 6 volts already mentioned would simply be shunted to ground through a zener and a resistor, as @ebp has already suggested.

One of the things that worries me is the circuit's efficiency, and that if I get too small an output voltage from the generator it won't be enough to go through the first diode due to it's voltage drop... even if I were to use schottky diodes... I'm guessing LTspice should be able to calculate those losses. But the generator itself is something that I think I'm gonna have to build anyway so I can measure how it performs.

Question, would this be the best structure to use in the generator?:

Would it be best to use a magnet (or magnets) at the rotor than at the stator?

And yes, I've already read this site's article about the reluctance motor. Very interesting stuff...

#### cmartinez

Joined Jan 17, 2007
7,552
I've just changed the thread's title for it to more accurately describe what it's really about.

#### cmartinez

Joined Jan 17, 2007
7,552
There seem to be several possible configurations to accomplish what I want. So far the most attractive one is called "permanent magnet reluctance generator":

I have a few miniature telephone transformers laying around, maybe I can tweak them into becoming the stator's coils. The challenge is going to be how to fabricate the stator core without investing to much time/money in it.

#### cmartinez

Joined Jan 17, 2007
7,552
Interestingly, there's a subtle difference between the PMRG and what's called a "flow switching alternator." The FSA seems easier for me to build, but I wonder which one is more efficient... and would work best at the speeds I want...

#### cmartinez

Joined Jan 17, 2007
7,552
I think I've found what I need to charge the capacitor to the levels I want, and it's a Joule thief. I've seen that circuit before, but until now I've only considered it a toy to be played with and do some learning in the process. Not anymore...

I feel like I'm talking to myself in a lecture hall full of empty seats ... but I'll keep documenting my progress here anyway...

#### ebp

Joined Feb 8, 2018
2,332
Oh, I bin payin' attention.

I had no idea that variable reluctance generators actually are a thing. That flow-switching beast is also interesting.

I suspect trying to manage (minimize) the gap between the poles of the rotor and stator is quite important. If you aren't producing much power, run at variable speed and don't care about wasting what you don't need, a shunt regulator may have merit, in which case there may be advantage in not trying for the very best magnetic coupling. If you are planning on spinning the gen with something that can go very fast if unloaded, voltage can get out of hand.

In terms of losses, eddy current and hysteresis losses probably dominate the no-load losses, with bearings contributing some minor amount. Copper losses will come into play when there is a load. I would guess that it might be easier to manage in the flow switcher, but I really am just guessing.

#### ebp

Joined Feb 8, 2018
2,332
OK, I wasn't paying good attention . Forgot I mentioned shunt reg before. I read what you'd posted at 12 way back on Thursday. I'm old. I can't be expected to remember that far into the past.

Also forgot to mention - Linear Tech produces at least one IC for energy harvesting. I looked at it briefly a few years ago, so I don't remember any details, but it is probably worth looking into.

#### cmartinez

Joined Jan 17, 2007
7,552
Also forgot to mention - Linear Tech produces at least one IC for energy harvesting. I looked at it briefly a few years ago, so I don't remember any details, but it is probably worth looking into.
Definitely worth looking into it... mind sharing any clues as to how I should be looking for it? i.e. what would the right search keywords be for me to use?

#### ebp

Joined Feb 8, 2018
2,332