Besslers prime mover and its enabler.

[daxwc]

JB:

No he does not have gravity off , it affects the results because they bounce instead of just roll , let him turn off elasticity and you will see the same GPE results.

Shouldn’t read things until I had my morning coffees. Duh of course he has gravity on. Fake news propaganda on your sheet there Fletcher ;)))

But in real life why does the hoop start catching the disc going uphill. A point mass at COM going downhill is going actually more distance, but the distance is that much shorter in the hoop going uphill?

[johannesbender]

JB:

No he does not have gravity off , it affects the results because they bounce instead of just roll , let him turn off elasticity and you will see the same GPE results.

Shouldn’t read things until I had my morning coffees. Duh of course he has gravity on. Fake news propaganda on your sheet there Fletcher ;)))

But in real life why does the hoop start catching the disc going uphill. A point mass at COM going downhill is going actually more distance, but the distance is that much shorter in the hoop going uphill?

Difference in velocity and acceleration , the one in front will have accelerated to its velocity quicker so the distance builds up, and the one in front will decelerate and lower its velocity as it nears the climb first before the other so it catches up as the distance decreases.

[daxwc]

I was looking at it straight as a point mass view. Going uphill the distance the point mass takes does shorten. I guess it does shorten more than the disk.

It is like drawing a cycloid at point of mass of center of mass of the two objects.

[Fletcher]

Here are the WM sim comparison tests to test whether PE lost = Total KE gained, where Total KE = KE translational + KE rotational ..

The first animation is of a disk vs a ring (polygon) - as expected the the yellow disk recovers almost all its height, the difference of 1 J being losses to static and rolling frictions .. HOWEVER .. the ring does not recover nearly as much PE (height) .. And its max KE is short of the PE given .. they have the same standard coefficients for static and rolling frictions ..

The Classical Laws say the ring should have less translational KE at the bottom of the curve than the disk, but more rotational KE than the disk .. added together they should be the same Total KE and recover the same height (PE-mgh) - ok, we might expect some slippage with the ring due to it resisting being rotated as per this discussion but I would expect that to balance out on the way up also, approximately ..

I came across this anomaly years ago for disk, and ring comparisons when building sim tests - I don't think it just my program, and is something lacking in the code design ..

..........

In the second animation I swapped out the ring polygon for 4 small disks at the rim of the ring still totaling 1 kg - note the even larger chunk of missing energy showing in the output meters and graphs - almost no KEr at all, when there should be equivalent to the previous polygon KEr .. hence the occasional comments about polygons and bugs, and point masses ..

Relating to MrV's thread, this swapped out ring should be recording much more KEr than it is imo - thus it's Total KE is way off base, imo ..

Sims included .. I can't spend anymore time on it ..

...............

JB: @Fletcher you have to remove elasticity ...

Fletcher has gravity off so how does that change it?

May as well go a step further, what about a viscous fluid in a finned hoop half-filled going down an incline plane.
Or try two substances that repel each other; mercury in a glass tube.

No he does not have gravity off , it affects the results because they bounce instead of just roll , let him turn off elasticity and you will see the same GPE results.

Nice catch jb .. Yes - Gravity is ON, and Air Resistance is OFF - static and rolling frictions are the same for both and standard ..

So I made Elasticity INPUTS for the polygon curved track and the rest of the rolling objects - First I set the track elasticity to zero and both the hoop/ring and the disk climbed back up to just short of their starting GPE's - not exactly the same but close enough to not send up a red flag - that means they did lose similar amounts of energy to static and rolling friction losses - Next I also set the rolling objects elasticity factors to zero and got the exact same result (meaning only the track elasticity need be changed to zero) ..

And this confirms that Total Energy = KEt + KEr (less dissipative losses) as the Laws say .. (nb for this example of comparing rolling rings to disks losing and recovering GPE i.e. energy conservation ..

N.B. ** this is an idealized sim with no elasticity and no air resistances i.e. as pure a test as the sim can do .. in the real-world there is always air resistance and elasticity of objects ..

So jb's description of the workaround was on the money - with elasticity ON the hoop/ring was more prone to "bouncing" down and up the track because even tho I made the track polygon as smooth as I could it is still a series of very short straight lines forming the curve - and things in contact will tend to bounce a bit from one short straight to the next ..

... BUT ...

I also changed out the elasticity of the track to zero in the second sim test for the disk vs the approximated ring with 4 rim weights and the big missing rotational energy was still missing (see the graph) and it DID NOT recover its initial GPE or thereabouts .. why I was thinking of MrV's sim experiments ..

..........

When I come back I intend to compare the hoop/ring roller with the approximated one mentioned above - maybe there is an obvious answer and I'm just not seeing it ..

[johannesbender]

yip , the second sim seems wrong to me , if you observe the RKE it appears wrong compared to a hoop of the same mass , I think if you used the ring of sim1 and the ring of sim2 in the same sim and removed the disk , the difference would be very obvious between the two , it does not appear to sum correct for the mass distributed version.

[Fletcher]

Correct jb .. well observed ..

I built 2 versions today and have been tidying them up ..

To compare apples with apples I built the evenly mass distributed polygon ring/hoop and put it on a standard disk at 1 gm mass and that disk contacted and ran the slope i.e. the polygon ring did not contact anything - the mass distributed (4 wts) version also ran the slope on its 1 gm disk backing .. apples with apples ..

Ran with elasticity OFF etc ..

Funnily enough I'm pretty sure the code is doing what it is supposed to be doing - I'll load the main comparison test tomorrow morning .. I believe I found the problem - but not sure what to do about it atm .. maybe you will have some ideas ..

[johannesbender]

Is the moment of inertia calculated given the radius for the placement of the 4 placed masses relative to the center of rotation, I understand the velocity is going to change given the radius positioning (observable) , but does the moi change correct for the positioning of the evenly distributed masses around the cor .

Anyway was just wondering.

[eccentrically1]

It would seem the version with the 4 weights would have a different MOI because the center of mass of the 4 weights is nearer to the axis than it is in the ring with no weights.
Unless you specified that the 4 weight version was equal. The weights would have to be inside the rim for it to roll; so that version would have to have a bigger radius. That's what I see. Maybe that's why the sim shows it slower than the normal ring.

[Fletcher]

Ok .. here is the cleaned up sim I made yesterday - it compares a blue polygon ring/hoop mounted onto a 1 gm disk in contact with the track, against, a red pseudo ring of 4 weights mounted near the rim - they don't have exactly the same MOI's but close - the track elasticity is set to ZERO as per yesterday ..

Comments .. A couple of things to watch in the sim and animation ..

1. we expect a little static and rolling frictions to erode some of the original GPE at start for final/Total GPE recovered .. but we see the pseudo red wt ring is WAY DOWN of GPE Recovered - and that is primarily because it didn't get any KER going until about a third down the slope (but note both were together until then) .. its KET was about normal ..

2. So, not developing full GPE Recovery nor KER ! .. so it must be SLIPPING the first third, and this doesn't allow the KER the build up steadily to give it the extra angular momentum push to climb back up to similar heights .. and slipping isn't aiding the KET component significantly ..

If the slope was a gentle one, either straight of curved they would regain the same GPE height or similar imo ..

** I have no idea how to fix this slippage anomaly unless they can be rack and pinion type geared (no frictions) ..

Somebody may have a better explanation for the energy loss or workaround ..

** lever and attached weights swinging around are perfectly fine and reliable - 'Buyer Beware' .. this steep rolling situation can be suspect it seems ..

.................



.................

[johannesbender]

True , the other sim showed slipping near the end of the track too , perhaps the energy test sim would have been a better comparison if both were made to translate and rotate the same , I don't know.

[Fletcher]

Here's the other one I made yesterday jb .. a 1 kg blue ring with a disk background of 1 gram running the track .. vs .. a 1.001 kg red ring running straight to track - elasticity of the track ZERO ..

They have consistent comparable rpms right thru the run .. and recover almost identical GPE's ..

So a polygon vs a smooth polygon (make a circle option/disk) running the track makes no difference - in this context ..

fwiw I'm not at all worried about it - the Laws are safe - I hardly use rolling disks down steep tracks and if I do again I'll keep it in mind i.e. slippage causes energy losses bigger than ordinary static or rolling frictional losses it seems - and this reduces KER component and KETotal ..

.............



.............

[johannesbender]

I think it is safe to say , the sim handles those situations good .

[Roxaway59]

I've been experimenting in Algodoo for a few days with weight shifting mechanisms and I will share those soon but first I wanted to show something that is purely theoretical and I don't really know if it could work. Its based on the wheel I have already shown but the mechanisms on it work differently. Instead of the blue lever weights there are springs. The action to over balance is a momentary one and some kind of latching mechanism is needed. If you take the fixate cross off the stage 1 wheel and move the wheel slightly with the motor before disengaging it you can see how well it turns. If you remove the fixate cross from the left hand mechanism you can see that operate. Its just theory so let me know what you think.
Graham

[daxwc]

Fletcher: ** I have no idea how to fix this slippage anomaly unless they can be rack and pinion type geared (no frictions) ..

Somebody may have a better explanation for the energy loss or workaround ..

How do you know it is slippage exactly?
It is not a function of the inertia formula or some formula in the setup?

it compares a blue polygon ring/hoop mounted onto a 1 gm disk in contact with the track, against, a red pseudo ring of 4 weights mounted near the rim - they don't have exactly the same MOI's but close.

I take it you think the MOI’s are close enough. What exactly is a pseudo ring? The mass isn’t evenly distributed so it is slipping because of this?

How many points is it calculating on the hoop?
It is short 19 percent of the velocity at the bottom.
How can the angle be too steep? A hollow sphere has even less contact area.

Seems like a coding issue from averaging something.


Yesterday the hoop had a problem; today Pseudo ring

[Roxaway59]

Just adding a bit more to the thinking about the idea.

Obviously the springs may not behave in an ideal way so its difficult to say what lift they would give to the weight at 9 o’clock but any lift what so ever will over balance the wheel to some degree. The key thing is will it go back to the correct position before 12 o’clock so the wheel can coast? Basically the equivalent to this is someone giving the wheel a slight push twice every cycle. Unfortunately I don’t have many springs available to me so I wont be able to try this for real just yet. Besslers wheel was said to work off a simple principal and I’m all for simplicity but is this just too simple?
Graham