'Core Rearrangement' - 'Fin Motor' - Open Tech - OU?

[gyulasun]

Hi Tim,

Just curious on the permeability of your chosen core for your decent sized coil? Because with the sizes you wrote, inductance comes between 4 to 5 mH (from a wire OD of say 3mm) for such air core coil.

Now on your question: "What am I missing?"

Well, when you make a parallel resonant tank circuit (10 H, 10 uF at 16 Hz) it will certainly have an impedance maximum at 16 Hz and you can calculate it by assuming a certain loaded Q for the tank. Say the loaded Q would be only 10, so the resonant AC impedance would be Z=Q*XL i.e. 10*2*Pi*16Hz*10H=10048 Ohm cca 10 kOhm.  If the loaded Q were only 1 (very unlikely, surely higher in practice), then Z=1 kOhm.

Now if you pulse this tank circuit at its resonant 16 Hz frequency and say the rotor would be already spinning at 1000 RPM, the input current would depend on this estimated 1 to 10 kOm impedance AND on your SS relay ON time of course, so if you adjust your supply voltage to say 14V DC, then in the moments of ON time the peak current draw would be 14V/1 kOm=14 mA and average current would be less.
When you feed this tank circuit from a pure 16 Hz sinusoid AC source, the rms AC current draw would be about 0.7*14 mA or less when Q=10.

Of course the current inside the tank circuit would be Q times that of the input current, as usual for parallel LC tanks. Say your loaded Q is 100, and input current would be 10 mA, then the current via the coil or via the capacitor would be 1 Amper.
If you attempt to make an impedance matching between your LC tank and the AC input source, by using either a tap on the coil or using a coupling coil, then your loaded Q could be increased perhaps indeed towards the 100.

You can estimate your unloaded Q for the LC tank by considering the DC resistance of the coil and its inductive reactance at 16 Hz: Q=XL/r i.e. roughly 1000/0.3 or so but this is surely lowered by the actual core eddy current and hysteresis losses to maybe a Q of some hundred in practice. Say Q=300 remains with the core inside the coil, this gives a huge AC impedance at resonance of course and you wish to transform the AC input source impedance to be as high as this huge impedance to get a match and highest power transfer. This match would decrease the Q=300 down to 150 or so, this would mean that to have your 14 Amper current via the coil, your input current should be roughly 14/150=94 mA... 
This sounds very favorable but it all depends on the real Q with the core and correct impedance matching and then you have to feed the down-transformed loaded tank impedance with the initial 48 W input. This may involve further losses due to the lower input impedance.

One more thing you have not indicated yet to consider: this 14 Amper current may influence and toss the core towards magnetic saturation and this means that its calculated 10 Henry inductance may get reduced significantly, causing the resonant frequency of the tank to increase. Should it decrease to say 3 Henry, the frequency would go up to 29 Hz from the 16 Hz.

There still be some other issues I am not aware of yet... 

You may learn on matching LC tank to a H bridge switch and some other useful hints here: http://www.richieburnett.co.uk/indheat.html  and a possible frequency control circuit (PLL= phased lock loop) which is able to follow the change in the tank resonant frequency here:
http://uzzors2k.4hv.org/index.php?page=ihpll1   Unfortunately this must be adopted to your 16-30 Hz or whatever range because it is shown working in the 70-120 kHz frequency range. Maybe such frequency tracking precision circuit is not yet needed for your prototype though.
This link may also be useful in general: http://www.allaboutcircuits.com/vol_2/chpt_6/5.html

Greetings
Gyula

[tim123]

Hi Gyula,
  thank you for that.
So that's what Q means? Wow, cool. At the moment, I just don't get most of what you said, but I'm going to add all this to my coil-calculator program, and learn it. It might take a few weeks... The inductance of the example coil at 10 Henries - was with a core half-filled with iron - as that's more or less the proportion it'll be with the rotor & stator parts.

In the meantime, this is an update for the fin-motor idea. It's designed to provide continuous rotary motion from the changing field inside a coil, in a tank-circuit, but with minimal damping of the resonance.

The magnetic field strength in the core is proportional to the slope of the sine-wave, generated by the tank-circuit. It's plotted as a cosine. As our iron rotor is attracted equally to either polarity, the field strength looks like a rectified cosine wave.

This design works on the following principles:
- The rotor is attracted to the stator at full-field - because:
   a) the common surface area between rotor and stator fins is greater than between the rotor and PMs, and
   b) the magnetisation of the iron stator is more than the permanent magnets (1.5 Tesla vs maybe 1 Tesla).

Full-field is the main power-stroke. It's power is derived from the rotor & stator being fully saturated by the coil - and doing what magnets do...

- When the field strength of the coil drops below that of the permanent magnets, the rotor is attracted to the PMs.
   a) The PMs attract the rotor at it's ends - the arrangement shown provides an almost complete magnetic circuit through the rotor, from one side to the other. (could close the circuit easily enough)
   b) Note that the direction of the field of the internal PMs is at right-angles to the coil's field... This means there should be no generator effect from the presence of the PMs, or from the rotor when it becomes magnetised by them.

In this arrangement, the magnetisation of the rotor plates moves through 90 degrees in each quarter - so it's actually rotating around the x-axis - i.e. into the page. I think that's kinda cool, but have no idea if it's significant. Maybe the motor will levitate too

The rotor - at full-field will be fully saturated by the coil - in the axial direction. So it won't be attracted to the PMs at all. With careful selection of the PMs, the resulting force on the rotor could be balanced so 50% of the time it's attracted to the stator, and 50% to the PMs - thus rotating.

It's a bit more difficult to build, but not much. The PMs could be lots of little ones...

[tim123]

Hi folks, it's a rainy day here in the UK, so I've nothing better to do...

Quarter-Turn-Pulse Fin-Motor.

Attached is a pic of the simple version of this that I intend to build as soon as... It's still designed to run as part of a tuned circuit. It provides pulse power - with a 1/4 turn action.

The idea being to hopefully harvest a large amount of mechanical force from the changing magnetic field, while causing minimum disturbance to it... Hundreds of horsepower out, for a few watts in, would be nice.

- The rotor & stator fins are all identical - cut from half-round mild steel.
- The fins are all attached to the shaft before it's inserted into the housing.
- The rotor fins are firmly attached to the shaft - i.e. welded or bolted.
- The stator fins are attached to the shaft via pillow-block bearings bolted into the rectangular face.
- The stator fins are attached to the casing by bolts.
- The casing could be made of acrylic tube for the prototype.
- The stator fins are aligned vertically.
- The rotor fins are free to move, and return to the bottom under their own weight, though it would probably be beneficial to add return springs too.
- PTO would be via a ratchet gear, and a flywheel.

I'm aiming to test a small version of this ASAP, but I'm still working on the workshop. Almost done, it's taken months...

Things to test: get the tank circuit running, and see if a) the core actually moves, b) whether it destroys the resonance, or not.

PS: PTFE washers between all the fins & a drop of oil - to act as spacers. The fins have to be centered...

[Khwartz]

nice to see you keep going on your idea Good luck and ideas!

[Khwartz]

How is it going, tim?