Zeroing in on Bessler's wheel

[Tarsier79]

The wheel is initially accelerated, then left to decelerate. Each time a roller hits, the rotational speed of the wheel is different, as will be the speed of each ball as it hits.

[Fletcher]

The first spike measured 2.65

The second spike measured 2.55

The third spike measured 2.2

They should be all the same because the roller hit with the same GPE

So why are they get smaller each rotation?

I don't think you understand what the sim is showing you.

I added blue velocity vectors to the Rollers so you could see the directions of the velocities at any given sim frame (set to lowish resolution so I could make an animation of it and post it - therefore not as small steps as high resolution accuracy setting) for each Roller.

First of all I don't turn OFF the push force until the first Roller has moved across and made contact - IOW's it is getting an extra acceleration from the push force still in action.

Then you need to watch the Rollers after the transition across to contact point on the tear-drop around 4.0 to 5.30 o'cl (2nd quadrant). They rock back and forward in their casings and when outermost add some torque and acceleration to the wheel (seen as slight bump in blue rpm tracing). However the wheel trend is already a slowing of wheel rpm and the "rocking" oscillation in the 2nd quadrant is being dampened down because the wheel is under deceleration i.e. slowing rpm.

Watch the velocity vectors for each quadrant, and watch the Rollers movement in each casing.

[Leafy]

It doesn’t matter if the other rollers roll back and forth. The time of collision is almost instant.

Obviously the torque before turn off does not effect much of the first and second spike.

The only real value is the GPE when the roller hit.

The first and second spike have about the same GPE.

The third spike have Half of the GPE.

If half of the GPE means half of the spike, the the value of 2.55 and 2.2 does not match.

[Fletcher]

Wrong .. the others factors are individual inertias of Rollers whilst moving in the confines of the casings and being accelerated and decelerated, and the 'whole of system' Moment of Inertia (MOI) changing due to Roller positions constantly changing. Acceleration vectors added.

Beneath I've made the Rollers smaller diameter and added a time input control for the push torque to go OFF at 4 secs exactly.

I've also changed the graph rpm output scale to 20 secs run time instead of 60 secs beforehand to spread the graph plot out.


These changes makes the dynamics more obvious. If in doubt watch the system COM / COG icon clearly move around when rpm increases and decreases.


** The animation is jerky because it is at low frames of 50 animation steps per second to be able to make and post an animation. At higher accuracy the vectors are more docile. You can still get the idea.


sawmill restaurant alabama

[Leafy]

It’s true that if the rollers are not on the circle during collision, then the dynamic change, but the rollers of the first sim is on the circle.

[Leafy]

Actually you’re correct that if the roller swing too far from its lowest GPE, then the spike is effected.

The first sim speed is relatively slow and the thirst spike is almost still.

So the first sim result stays true.

Only GPE is the factor of spike.

[Fletcher]

Conclusion: more torque can be obtained from the same GPE/ momentum.

Case closed.

Nonsense ..

Actually you’re correct that if the roller swing too far from its lowest GPE, then the spike is effected.

The first sim speed is relatively slow and the thirst spike is almost still.

So the first sim result stays true.

Only GPE is the factor of spike.

No .. it depends on whether the wheel is accelerating or decelerating because it affects the Frame of Reference.

If a block is traveling horizontally to the right at 5 m/s and catches up with and collides with another block traveling in the same direction at 2 m/s then their relative velocities come into play. Same if the other block is traveling towards the first at 2 m/s.

If you don't like what the sim is predicting then do a real-world build for comparison purposes and take it up with the 10,000 engineers, physicists, and programmers who built the kinematic sim software and beta tested it before release to the commercial market.

Or, just start practicing on Algodoo.

[Leafy]

You see,

If the second spike value is 2.55, then the third spike value is 1.8.

The fact that it is 2.2 shows OU.


First, you said Sam’s design is smooth and have no jerking. Now you agree that there is a torque.

[Fletcher]

Don't put words in my mouth .. of course Sam's designs have torque - otherwise the system COM / COG wouldn't move upwards and left or right of the vertical beneath the axle and cause a torque reaction in the wheel as it loses GPE. But energies are lost from the system thru frictions etc and there is no net positive torque i.e. surplus torque to cause a cyclical acceleration and it loses momentum and RKE and slows down, as I've said numerous times.

[Sam Peppiatt]

Hi Fletcher,
Unfortunately, I didn't get any work done today. Only thing I can add at the moment is; the rollers seam to hit really hard. If that means any thing, that is, like they are being thrown down hill, like driving a nail with a hammer-----------------------------Sam

[Fletcher]

Finish the build when you can Sam .. then we will have some hard facts to deal with.

Sometimes a sim can be tweaked to have almost identical dimensions and masses etc. Then it becomes really interesting doing the comparison.

[Sam Peppiatt]

Yes, I'll / we'll, learn some thing, one way or the other. Going back to bed---------------------Sam

[Leafy]

Mornin all .. here is a stripped down sim of Sam's Tri-Roller Wheel, recorded in animation.

I start the sim with a Motor at 10 rpm (there is no jerkiness in the plot line up or down from Roller impacts etc …

[Leafy]

I have to gives you credit for saying it is strictly controlled at 10 rpm. But still, the fact that you don’t realized how important that jerkiness is.

Now that I pointed out the important of that torque, you conclude that it is due to rollers moving around.

And the final argument is sim does not shows OU, let’s build it…

Give the old man a break.

[Leafy]

No .. it depends on whether the wheel is accelerating or decelerating because it affects the Frame of Reference.

If a block is traveling horizontally to the right at 5 m/s and catches up with and collides with another block traveling in the same direction at 2 m/s then their relative velocities come into play. Same if the other block is traveling towards the first at 2 m/s.

Let me counter this point real quick.

The rollers moves with the wheel. If the wheel accelerate with gravity so does the rollers. Pretty much they’re on the same frame.