Using a Milkovic type Two-Stage Mechanical Oscillator Pendulum with two magnets

[AquariuZ]

I was not sure in which section to post this, but because it involves a gravity/magnet solution here it is.

While I was fiddling with a model for Gurangax I thought of something.

Wouldn't it be nice if you could somehow keep a TSO Milkovic Pendulum in motion? Not sure if anyone considered placing two strong magnets on the end of both sides and pin it to a wall?

So I modeled two (almost) equal length bars - pinned one in the center to a wall and attached the second bar on the right side as rigid pendulum. Then on the left side I added a magnet with South pole facing down and on the right side I attached a magnet with the South pole facing left. (the pendulum).

Next I estimated the repulsive force to be around 5N between the two during proximity - I know this is a polar non linear force but please bear with me.

I solved the Force by estimation using this proximity formula:

and(abs(body [20].p.x - body [21].p.x) <= 2,abs(body [20].p.y - body [21].p.y) <= 2)
and(abs(body [18].p.x - body [22].p.x) <= 2,abs(body [18].p.y - body [22].p.y) <= 2)


Where body 18, 20, 21 and 22 are the N, N, S, S poles of the  magnets. Force is polar and rotates with the bodies.

And then I let it go.

Well - it tries very hard to find equilibrium, but the repulsive force that is inserted into the structure keeps it going. (in a vacuum I might add).

Then I thought what if I planted this on a wheel just off axis?. First I looked at examples of real magnets and their forces. To calculate repulsive forces between magnets is rather complex. I have found however for instance this:

http://www.supermagnete.es/eng/data_sheet_Q-50-25-10-LN.pdf

This 50 x 25 x 10mm N40 Grade magnet is rated at 12KG or 118N

I decided to set the repulsive force to 25% of that (30N) for all poles and pin the pendulum to a 100 KG wheel. What happens is interesting. It slowly starts rocking back and forth and then starts to rotate. The more the magnets meet the more force is translated to the momentum of the wheel.

It is still going - I have seen rotational speeds of 0.300 Rads/s which is fast. I have tried with higher wheel mass and lower repulsion but the results are the same. Rocking back and forth - then acceleration within limits. I cannot really see if it will accelerate further because the simulation always stops at 32000 frames.

Models attached.

Please forgive the proportions, I will make a more reasonable model soon.

Someone living high on Mount Everest care to try this? The big question is: can it overcome the friction/ air resistance losses? Not in the model, but I have not really tried optimizing weight ratios and location on the wheel - etc.

The left side of the pendulum structure has a fraction more weight btw. I do not want the pendulum to flip to the right. The pendulum itself should be easy enough to build/try. But--- vacuum? friction? I do think the idea is interesting in any case.

Your thoughts please.

AZ

[AquariuZ]

Two screen shots.