Jhula principle - increase speed by applying brakes.

[xnonix]

In the graphics: the latch torque of the lever an pendulum shows you when the latch is on. When is 0 the system is free.

About the formulas they are easy to explain...

If i want a force +F when the body(x) is rotating clockwise:
F * (Body(x).v.r > 0)

the part (Body(x).v.r > 0) is a conditional that is 1 when is true and 0 When is false.

Something like this is what i did on the two motors to simulate the latches. The system seems to work.

Now let's make it in reality.

[prajna]

Yes, nothing is more convincing than something that works in real life.  It looks like it should be very easy to build.  I am still playing around with designs to find out the optimum relationship between the counterweight and pendulum mass, the length of the pendulum and the position of the two bearings.  Small adjustments in any of these seem to make quite a difference in how it performs.

[bluesgtr44]

Hmmmm....interesting. I have used WM2D for quite awhile now. I do not trust the polygon functions very much and try to avoid them. I also believe that if one was to build this that they would find the 3rd deminsion kicking in and it would take away from the system in the form of attempting to also oscillate in and out. I downloaded your set ups and increased the air resistance to what I would think are more realistic levels. The first one made two revolutions and died....you had air resistance set to "none", I changed it to the default low setting of .3....the second one you had a low setting of .07, I raised that to .2 and this made the set up more erratic, but it did keep going and maintain an increase. I feel this was in part due to the use of the polygon function in WM2D...I just don't have a lot of faith in it.

Draw a circle or disk about 4ft. in diameter.....pin it in the center and then use the curved polygon function and draw another circle of 1ft. in diameter.  Set the two to "collide" and then set the polygon circle inside the regular circle. Hit RUN! This thing will accelerate itself to oblivion! Why? My understanding is that the polygon circle thinks it is on the outside of the circle and  rolling off....it just keeps trying to roll off....a software glitch...one reason why I do not trust them....


Steve

[prajna]

Yes, I notice glitches too, Steve; sudden accelerations. I have noticed that there are a number of parameters that need to be balance. It seems to be a balanced unbalance that makes this system work.

[bluesgtr44]

My apologies, Prajna....I do not think this system will work at all. My main point was....there is no 3rd deminsion in WM2D. Very much like Milkovich, the pendulum is going to want to go backwards and forwards to stop the motion....gravity always wants to win. Think about it, a constant downward force will utilize any means to achieve that point of rest. Even if you try to brace it, this produces friction in the hard mechanics of the system and decreases the output and the whole thing comes to a stop.


Steve