Some observations on switched flux motors

[Ted Ewert]

I’ve been building switched flux motors for a couple of years now (In this case I mean switched flux from a permanent magnet source).
I have run into many problems along the way, but none more baffling than magnets that appear not to “switch on”.
The Flynn Parallel Path demo illustrates how magnetic flux from permanent magnets is directed through alternate pathways through the use of switching coils. This technique is also used in the Flynn motor.
The motors I built all suffered from a lack of power. No matter how many magnets I used, or what configuration they were in, it made little or no difference in the performance of the motor. I was baffled.
Then I recently did an experiment with a Flynn demo. I wanted to see if the force applied by a switched magnet was really what Flynn claimed it to be. So I built the rig below
I tried different voltages, I tried the magnets in the front, the magnets in back, no magnets, and it all acted just like Flynn and everyone else predicted. The magnets also increased their pull on the steel almost exponentially with an increase in voltage. The pull was so strong I couldn't prevent the steel bar from getting pulled into the stator.
What especially confused me was that I tried the exact same configuration that didn't work at all in my motor, and it worked just fine in this configuration.
That confirmed for me that the problem isn’t in the configuration, or the magnets, or the coils. The problem is a timing issue, or more precisely a delay issue.
If you have ever built a Flynn demo, you noticed that when the flux is switched from one side to the other, one of the steel bars will be snapped into the stator and the other will fall off. It’s the one that falls off that we are concerned with here.
Why does it fall off? Unless there is a perfect balance between the coil force and the magnet's force, that bar should stay in place. Even if the magnet is completely switched, the two coils will send current between them selves since they are both opposite polarities. This alone would keep the bar attached.
When that bar falls off it means that there is zero magnetic current flowing in the iron. Why would the reversal of current take so long?
The only thing I can think of is that for some reason the magnet’s flux won’t flow until the phase of the voltage and the current in the coil are synchronized. This means that the more inductance there is in a switching coil, the longer the delay.
This could be why the Flynn motor replications are so disappointing. Even though they have much less inductance in their windings, the switching speed is a lot higher.
I realized that it is this pause which was killing my motors. By the time the permanent magnet’s current is ready to do some work, the window of opportunity is long gone.
Look at this graph of one of my more successful motors below:
As the speed of the rotor decreases, and the time duration of the on pulse increases, the power starts to climb steeply. This shows the magnets starting to work as the coil becomes in phase. The absence of magnets in the stator is also displayed as a more typical response curve, and to show that the magnets do have an effect.
Switched flux motors are either going to have to run a lot slower, or some other remedy will have to be found to fix this problem. Low inductance coils will help, but they're power hogs.
Nevertheless, when those magnets finally decide to kick in they are brutes. They just need some time.

Cheers,

Ted

[gyulasun]

...
When that bar falls off it means that there is zero magnetic current flowing in the iron. Why would the reversal of current take so long?
The only thing I can think of is that for some reason the magnet’s flux won’t flow until the phase of the voltage and the current in the coil are synchronized. This means that the more inductance there is in a switching coil, the longer the delay.
This could be why the Flynn motor replications are so disappointing. Even though they have much less inductance in their windings, the switching speed is a lot higher.
I realized that it is this pause which was killing my motors. By the time the permanent magnet’s current is ready to do some work, the window of opportunity is long gone.
...

Hi Ted,

Sorry that what I am going to tell you I have not tested in practice as yet for such setup...

I believe a possible solution for the problem you wrote above is using a coil pair what Tesla described in his "Coil for Electromagnets" patent, see here:
http://www.tfcbooks.com/patents/512340.htm

The key sentence is this:

I have found that in every coil there exists a certain relation between its self-induction and capacity that permits, a current of given frequency and potential to pass through it with no other opposition than that of ohmic resistance, or, in other words, as though it possessed no self-induction.

Of course the "certain relation" between the inductance and capacitance is governed by the physical dimensions of the parallel wound (bifilar) wires, and unfortunately at low frequencies of some ten or so Hz (motor RPMs of some thousands) the wires ought to be flat so their surfaces could get much closer to each other than in case of twisted cylindrical wires, because the facing areas and the dielectric insulation between them which constitutes capacitance, i.e. the C value of the resonant LC circuit is formed by the two coils connected as shown in the patent, and the L value of the resonant LC circuit comes from the number of turns as usual.

(A random search brought this link to insulated flat wire sizes:
http://www.hmwire.com/New%20PDFs/Insulated_Flat_Wire_AWG_11-30_R6.07.07.10.pdf )

In the past I wound several impedance matching wideband transformers, using twisted pair bifilar or trifilar wires and found that the higher the number of twists for a given wire length, the bigger the (distributed) capacitance between the twisted wires. These were high frequency transmission line transformers in the several MHz range, and probably for the some 10 Hz frequencies for motors that their some thousand RPMs involve, no amount of tight twists could be enough so most probably the use of flat wires, with their much higher facing surfaces, is needed. Experimentation is needed which is surely costly for flat wires.

What do you think?

rgds,  Gyula

[Ted Ewert]

Thanks for the thoughts Gyula.
I'm familiar with that type of coil, having already built a few in the past. Increasing the capacitive energy in the coil through using Tesla's technique may help lessen the delay. I was planning on winding my next motor coils bifiler anyway so I will test this arrangement out.
I don't know what else effects the length of delay other than the apparent phase relationship between the voltage and the current. I mention the phase differential because it also has an effect on induction.
While experimenting on trying to get a MEG to produce some extra energy, I came across a similar problem. The switched flux from a magnet not only doesn't flow right away, it doesn't produce any current in a secondary winding. Only "flux" produced from a primary coil, where voltage and current are out of phase, will induce current in the secondary. This was defined as a "change in flux" by Faraday, but it is not correct. I can change the flux all day with a magnet and it won't induce anything in the secondary (unless the change in flux is caused by movement of the magnet).
Movement of a magnet's field in relationship to another static field apparently alters the phase relationship of the B and the H fields. I assume this is so since induction takes place under these conditions, and induction only takes place when the voltage and current are out of phase. What happens electrically also happens magnetically?
I have no absolute proof for any of this, only my observations from many experiments. I have had a lot of frustrating moments trying to make a device that, according to accepted theory, should work but doesn't (like the MEG). Consequently I have had to rely on what actually works as opposed to what should work.
Anyway, I though I would pass a little info along so some of you wouldn't have to waste as much time as I have beating my head against a wall.

Cheers,

Ted

[Nali2001]

Maybe you find these links useful / interesting:
http://www.youtube.com/watch?v=VfptuTeRdb0
http://www.youtube.com/watch?v=L9HyyGdnmb0
http://www.youtube.com/watch?v=atUTu8sg94w
http://www.youtube.com/watch?v=SHbQXnXK6Xc
http://www.youtube.com/watch?v=BsN2sr3U0PY
http://www.youtube.com/watch?v=x-EuPGl8JjE

[scianto]

I find your experiment and resulting observations very teaching and important to remember for people building anything with magnet and cores.
Keep us updated, please.
Thank you.