WM2D working simulation

[Bill_Mothershead]

Simulation works, but not clear if it could be built to work.

Only one weight...blue...hanging with ropes.
Trying to break symmetry... ascending uses short ropes...
descending uses longer ropes near weight center.

Friction set to zero since weight suspended with ropes
does not touch outer wheel/circular frame.

Out of desperation, uses external wheel/pusher to
nudge weight off center.

Busy and probably not going to work on this anymore.
Will check back here to answer questions.

[Fletcher]

Hi Bill .. looks like the sim file is for an avalanche experiment rather than the single hanging weight pic ?

Anyways, looking at the pic it seems that some of the rope attachment positions vary a little. This may be giving you some rotation (temporary torque conditions).

FWIW when I build something like this I start out with everything centered on the background grid at 0,0. And then rotate everything by equal degree increments, and attach the ropes at grid spatial coordinates one at a time etc.

That way the attachment positions are exact. Then I run the sim and let it settle in its natural position which is usually the keeling position.

Then give it a nudge etc or try to get a ramp to influence it etc.

Good luck with it.

[Bill_Mothershead]

uploaded the wrong file....try running this

[AB Hammer]

Bill_Mothershead

I like the design and it does need to be built. The only fear is, it may turn out to be another form of this one.

[Bill_Mothershead]

Before anybody builds, would like to see anybody simulate it
with something besides WM2D. Would help credibility.

Be warned that dealing with rope lengths and attach points
is tine consuming. Magic happens...sometimes. I have a directory
full of saved files where I adjusted something and the magic
went away. Dead simulations that never worked again.

I worked very hard to get it to rotate faster. Nothing I did
seemed to make any difference. Always slow.

And what is with this "nudge off center" bullshit object?
Desperation...and I don't know why it works.

I will try and upload a simulation with more ropes and a different
version of the nudge device....(hangs from axel).

[Fletcher]

The sim gains System Kinetic Energy as you'd expect. That is from the 10 lb Blue Mass initially falling and losing GPE.

This loss in GPE should be converted into equivalent amount of System Kinetic Energy Joules (BTU) with no frictional losses. That's CoE.

I have taken your sim and made some more Outputs and Inputs to change some critical things, and easily see at a glance what's happening.

The acid test is do the outputs look realistic and are changes to hardware realistic etc (like zero frictions) ?

In my adjusted sim the thing accelerates and the System KE climbs to circa 145 J's. That seems about max speed I can get. From a loss of GPE of about 5 J's (bounces a bit). Well OU from this result.

The bottom line is that having either the Ramp or the Blue Mass with zero static and/or Kinetic Friction may not be realistic from a materials standpoint. Unless you discovered a unicorn thigh bone to use.

Try using very low coefficients and see what happens to the comparison between GPE lost and KE gained; and CoE ?

[Bill_Mothershead]

amended latest to replace ramp with roller,
set all frictions to 100....should kill it but doesn;t

runs slow...but seems perpetual

[andyblues]

Hi bill andyb here,i have built designs recently where causing the center to do as yours is proposing creates a lifting imbalance which in one way is linear and the other is angular increase the problem is that the friction /force angle on to the inner rim /weight pushes back and movement is jammed ,i would love to of been able to put more elegantly as fletcher or many other members do, but i cant ,it has thrown be into the grinder a couple of times and all i can see is the friction is not a problem not enough to stop the movement it has to be angle of leverage ,i hope this helps and we all learn some thing together here ,thanks for sharing Andyb.

[Fletcher]

amended latest to replace ramp with roller,
set all frictions to 100....should kill it but doesn;t

runs slow...but seems perpetual

Hi again Bill .. hate throwing spanners in the works but here goes.

I took your sim and added some outputs (plot graphs etc).

I also thought something wasn't right in those outputs and found a circular mass way down the page (circle[13]) which I deleted. That was better.

Then I enabled view system CoM by turning it on and making the yellow 10 lb disk transparent so we could see the CoM move around.

Then I had a look at your frictions set to 100.00. The range is actually 1.00 to 0.00, so I adjusted both to 0.10. Then I adjusted the elasticity to 0.95 (1.00 is a material called 'Unobtainable' - 1000 x harder than diamond). That was to reduce likelihood of overlap error at 200 frames per sec (could increase to 2,000) which can sometimes force objects to 'shoot' apart adding energy to the result and confusing things.

I suggest you zero in on the System CoM and watch it swing back and forward like a pendulum now. Sorry !

The rest of the Outputs you can take a look at for yourself. They are to do with System Energies mainly (re CoE).

You can shoot the messenger now.


ETA: I should have mentioned that if you get a result that is 'suspect' then I normally just increase the frame rate (1-2,000) so there are smaller integration steps to calculate etc. Slows things down a bit but can avoid objects overlapping and spitting things out. If still 'suspect' then I play around with elasticity for materials I might actually use.

Ultimately if it rotates but is not accelerating (as a trend) then probably it is still suspect and needs further investigation.

[agor95]

Good work Fletcher

Every simulator has it's way to give use what we want.

So quality peer-review is the name of the game.

Regards

[Bill_Mothershead]

Your skills are great Fletcher.

These smooth surfaced circular objects are actually
micro-bouncing off each other.

The simulator starts each frame by computing where each
object should be in the next frame. Then it takes a pass
to see if any objects are within each other (impossible).
The software corrects this by separating the objects.
The software error is that it is a perfect bounce.
Energy is preserved or even amplified depending on
how much the objects overlap.

This explains why it didn't work when I tried to use rods/
ropes/springs to keep the weight off center. No chance of
two objects overlapping so so never hit the recovery software.

I will keep reading this forum but I am so burned out that
I probably won't use WM2D for the rest of the year.

I hope we all learned something and perhaps somebody
gets inspired. Perhaps does a non-WM2D simulation.

[Fletcher]

I empathize Bill .. I was fortunate in the early days of using the sim. I had a patient friend who would say .. try building it another way for cross checking purposes. See if you get the same or different result !

Stand back, take a breath - does that result look reasonable ? Be suspicious ! Trust your physics knowledge first. Do they align ? The sim is a tool to quickly explore things. Reliable if you are using it correctly.

Make it simpler, make it even more simpler ! Get rid of unnecessary complexity. Cull it hard ! Try again !

That's interesting - increase the frames per second, watch for overlap error, especially with polygon tool.

We started a thread in the tech forum of BW.com for these types of 'bugs' and how to mitigate them IIRC. It's still there I believe.

Take a well earned break Bill and recharge the batteries. Someone will solve this dam mystery one of these days and probably use a sim program as part of the proof. Now that would have us scrambling for the code and our text books eh !

[Fletcher]

Here's a couple of threads on other users WM2D experiences and 'fixes'.

https://www.besslerwheel.com/forum/view ... =9597#9597

https://www.besslerwheel.com/forum/view ... 7237#57237


FWIW .. ironically we can make sims that are for all intents and purposes exact. Exact placement of parts like rope connection points etc.

Bessler made his out of wood and metal. They were not as exact as a sim can be. He even says a few oz's here and there made not a jot of difference.

So exactness is not a requirement to success.

He did however imo find a mechanical way to maintain an imbalance of weight force which in turn provided asymmetric torque. Which in turn caused an acceleration up to a working speed (natural resonance speed loaded or unloaded). This manifest as a wheel with momentum and RKE which could do external Work and replenish itself to its innate state.

And this principle of PM once rediscovered should be able to be tested and checked in a sim environment, imo.

[ME]

Force vector decomposition...

Say a tethered ball rests on a single point of an incline. The tether itself is hooked at the other end at some point on a wheel (I'll likely short it to 'hook' from here on)
Gravity pushes the ball straight down, yet the ball rests on a single point. Here it basically doesn't matter (for a momentary analysis) if that point is part of a straight line or a curved incline.

The tethered ball pushes with some contact force against the incline. While the sturdy incline keeps its shape it somehow has the ability to provide a normal force.
In normal words this means that the ball does not penetrate the incline and the only way to move with contact is perpendicular to the incline (or 'tangent' when the incline is spherical). As usual, the amount of gravitational force that's left can be determined by Pythagoras.

With this remaining force the ball stretches the tether, which pulls on the 'hook'-point on the wheel.
The wheel can only rotate, so any force directed towards the axle will be countered by a normal force of the axle (or otherwise your axle will bent). So the hook can only move tangent to the axle.

Eventhough only a part of the gravitational force can/will be used to rotate the wheel, the tethered ball will most likely slide down the ramp plus rotate the wheel with it.

Of course this is just for a single tether, hooked at some random point on the wheel. In this topic there are more tethers hooked to the wheel to evenly split the load.

The important question remains: Is there a tether-hook position where the ball can be forced to go up that ramp?

[andyblues]

ME,thanks for clarifying what is going on with this tether approach, i wish some times i had taken more time at school to get my maths and my English up to a better standard ,i agree with your understanding of this approach and will try and get a better grip on Pythagoras and this back torque going into the pulley, i have been puzzled by this energy lose for a while now and finally believe i am ready for a deeper understanding of just how this is happening ,thanks for sharing and thanks to Bill for sharing as well ,Andy blues