Feasibility: magnetic stir plate with no moving parts?

Thread Starter

Heaan Lasai

Joined Jun 8, 2017
4
Hi

I need some guidance on determining the precise behavior of the flow of current, and by extension the magnetic field through a small electromagnet.

By asking nicely, perhaps someone here can help me with figuring this out? As many other people, I learn best by combining practice and theory. In this case, I have all the theory behind me, except for inductor calculations. It becomes so much easier to grasp the theory after doing a simple practical experimentation.

During the stirring process, the magnetic flee should rotate at 1500 RPM, or 25 revolutions per second at its fastest. The primary prototype will have 20 coils. As the flee's south pole passes one coil (A), coil A is turned off, and the next one, B, is turned on. With the inductive resistance, there will be a period where both A and B are producing a magnetic field, allowing the flea to rotate smoothly. At any given time one coil will be fed current producing a north pole up, and the coils on the exact opposite side will do the same but with reversed current.

But for simplicity, let's only look at the process of producing a magnetic field with its north pole up right now.

25 revolutions per second, and during each revolution the flea's magnetic south pole end will pass over 20 coils.
  • 1000 ms / (25 revolutions * 20 coils) = 1000 ms / 500 coil activation cycles.
This means powering each coil once for only 2 ms. In continuous operation that would be comparable to 125Hz (Why 125 Hz? 2 ms for rise section of sinusoid wave, 2 ms to return to 0 current flow, 2 ms flow of electricity in reverse direction for the falling lower half of the sinusoid and another 2 ms to return to 0 current flow. 8 ms for one sinusoid cycle, 125 Hz.)

Calculating the dynamic flow of electricity through an inductor challenging, and I would greatly appreciate your help! Currently haven't read that course yet, but I'm excited to get to learn more and try more construction projects!
How do I calculate the current flow through an inductor, so that if that coil is given power for 2 ms it will resist the flow, but still have reached full flow after those 2 ms?

Power source will be 24 volt, but can of course be lowered if needed.
Coil wire will be wound with whichever enameled copper wire the lab amanuensis will let me use, probably c:a 1 mm2.
Core will be either air or ferrite.
Each coil will have a short, cylindrical shape. Outer diameter 20 mm, ID 10 mm and length will be determined by how many windings I should use.



Backstory:
I've been around entry level micro controllers (mostly ATMega and ATtiny), and I've taken an introduction class to analogue circuits.

For fun I want to see if I can build the following:
As you all know, in labs a container (beaker etc) can be stirred using a magnetic stir bar. The usual solution is to place magnets on an electric motor, inside a chemically resistant casing.

However I've been toying with the idea of using electromagnets to spin the magnetic flea instead. In short I would set up a number of electromagnetic coils in a circle. A microcontroller will, using logic level MOSFETs cause each coil (here coil B) to activate just as the previous coil A is disconnected. (Flywheel diode will of course be present). This way the magnetic flea is never floating around, there is always a magnetic field controlling it. On the opposite side, the controller will activate that coil but with opposite current direction, and thus opposite magnetic field direction, just like in a mechanical stir plate.

Have just finished testing the box I'm going to use. It is chemically resistant and using the common method of gluing magnets to a computer fan worked like a charm.

I have a few different ATMega lying around, I'll try using them. Or I'll just use my Uno as proof of concept.
 
Last edited:

wayneh

Joined Sep 9, 2010
16,399
However I've been toying with the idea of using electromagnets to spin the magnetic flea instead. In short I would set up a number of electromagnetic coils in a circle. A microcontroller will, using logic level MOSFETs cause each coil (here coil B) to activate just as the previous coil A is disconnected. (Flywheel diode will of course be present). This way the magnetic flea is never floating around, there is always a magnetic field controlling it. On the opposite side, the controller will activate that coil but with opposite current direction, and thus opposite magnetic field direction, just like in a mechanical stir plate.
There's no reason you can't do this. The stir bar doesn't know or care what is producing the moving field. Before I went much further I'd do a search of commercial products to see if there's already such a device on the market. I suspect there is but can't say for sure.

A non-mechanical solution might lend itself to better feedback and control. The problem with all magnetic stirrers is that, if you attempt to turn up the speed too far, a point is reached where the stir bar cannot keep up with the rotating field and the synch is lost. The problem is worst with viscous liquids. The user has to fiddle with the speed control to slowly accelerate to the maximum speed that can sustain the synch. Having a stir plate that could do this automatically would be awesome.
 

ebeowulf17

Joined Aug 12, 2014
3,274
It would require expertise well beyond my own skill level, but I imagine you could also read values from adjacent coils which are about to be powered, or are no longer powered, and detect where the magnet is. I believe something similar is done in BLDC motors? Maybe it was a different type of motor, but there's definitely some type of motor that uses the same coils for powering the motor and sensing shaft rotation.

I don't have a clear picture in my head of exactly how this would work, but I'm pretty sure there's potential there. This would allow you to keep the moving fields always ahead of the magnet while accelerating, but never far enough ahead to lose sync like @wayneh was describing.
 
Top