I have a working Bessler wheel in my simulation !

[hansvonlieven]

I think he refers to the fact that two outer polygons were incorrectly placed on the grid, something you cannot do real world, thus creating a torque in the program which would not exist real world.

It would be debatable if the program should allow this or not, as it is just that, a simulation. Maybe an idea to add a check model function if the end goal is a symmetrical model?

Very hard to call, but from his reply I understand why it fails. Too bad this "flaw" cannot be built for real.

I don't think this is a valid argument. The guy is just defending his programme. The grid has nothing to do with it as it is just a marker to show where you are, The proof of the pudding is that if you only change marginally the value of the connecting piece the effect stops. There are  a number of other things you can do that will stop the programme misbehaving, I have tried a few. Very small alterations will effect its behaviour profoundly.

I am not buying it. I am also not fussed about his rather aggressive demeanor. I mean the guy was virtually calling us a bunch of idiots for querying his programme, when all we tried to do was help.

Hans von Lieven

[scott_z]

Quote 1:  "It should however not go into an uncontrolled accelerating spin"

Response:    If your model spins uncontrollably (without any inputs to it),  you probably have other issues with your setup...... Just to be clear, all of my previous comments were directed at the (corrected) model where the device was pinned to ground and there were no external inputs applied.  You then hit the Run button and the device should not move at all.  This would be expected considering there is no torque causing it to want to move (per my previous test for this) and the Center of Mass of the entire system resides at the same 0,0 of the main pivot.  When I first grabbed this model from the forum, the body positions (hence, Center of Mass) were not symmetric about the main pivot located at 0,0.  This makes sense as to why it would start to "fall" or "spin" when the Run button was hit, but not uncontrollably.  It did accelerate (due to the imbalance).  And yes, it if friction were presen, it would eventually come to rest.  However, are you confusing an "uncontrolled accelrating spin" with too large of an "Animation Step"    If the original model (with the imbalance) is the one you are still working with, and it is spinning uncontrollably, try decreasing your "Animation Step" to something like .005 s and your "Integrator Error" to something like .001 m.  Many times it may look like a model is going crazy but it is simly because the step size is too large (like drawing a cartoon on 50 pages vs. 500 pages). There are some other things to consider but I would rather post such detail on our forum rather than here. 

I have attached a model (BS7-corrected.wm2d).  When you hit the Run button, neither model should move at all.  The one on the left has a free pivot and is balanced.  The model on the right is also balanced and has a velocity motor with a value of 0 (locking it).  The torque should be at or near 0.

Quote 2:  "I am also not fussed about his rather aggressive demeanor. I mean the guy was virtually calling us a bunch of idiots for querying his programme, when all we tried to do was help."

Response:  This is not true.  To me it appeared that because everyone had given up on the fact it was accelerating, uncontrollably spinning, or whatever, that the software program basically did not work.  If people are going to use these tools, they need to be used correctly with an understanding of at least some of the basic theory behind how they solve the equations of motion and when and where the User might need to intervene with adjustments.

Now, using the corrected model, here is something that might be of interest.  Apply the pin joint AND a motor at the same location.  For the motor, set the value at something like 60 deg/sec.  For the "Active When" field, set it to TIME<3.  When you run the simulation, the motor will be deactivated and the device will freewheel (there is no friction).  See my second attached model (BS7-corrected_w_free wheel.wm2d).

[hansvonlieven]

G'day Scott,

My apologies for being a bit hard on you.

The fact remains that I still think there is a problem with the programme. When I first tested the simulation I could get the programme to work as expected by changing a single value, hence my suspicion it was the value entered in the connecting bar that caused the programmme to behave as it did. See my original post below.

Now, if I can do this, it cannot be a problem with the animation steps. The imbalance is still there and the programme caters for it in the way one would expect. The animation steps are also unchanged. So both of these factors are NOT the cause. It is interesting to note however that by balancing the elements the programme works as expected. Maybe a combination of factors is responsible here for the erratic behaviour.

I am a mechanical engineer,not a software engineer; I do have a fair amount of experience with programming though. These are just my impressions.

I like the programme, it is useful as it is and I wish you the best of luck with it.

Greetings from Australia

Hans von Lieven

@niente

The simulation is invalid. Whoever did this put an illegal value in the field that specifies the mass of the polygon. Change it to 1.0 kg and it will behave as expected. See screenhot below. The programme should have rejected the value entered, but that is an editing problem not a real fault in the simulation.

Hans von Lieven

[Bessler007]

Hello Hans,

If you make two compasses with rods pivioting on precisely placed bodies and adjust the rod lengths (x,y in a unit circle) you can connect the loose ends to find a precise point.

Another way is to paste a calculated value into the length of a rectangular body pivoting from one end;  then set the angle appropriately (relative polar coordinates).

No matter which way you precisely place bodies, wm2d has a serious flaw in it's calculus.  You're right not to trust it.  You're also very correct to think the more mass in motion in wm2d the less it's ability to track it.

I have more confidence in a piece of paper and a pencil..

@steve,

Yes, polygons are a problem. Fletcher found this out too.

@erick,

You have run into the same trouble as I did. I think the moment you have 8 kinetic bodies in the simulation the thing goes haywire. It cannot cope with that many moving parts.

@ all,

One of the problems I have found is that it is extremely difficult to place parts accurately. You really have to get into high magnification to see it. This throws the whole thing into disarray. I had one simulation today that behaved irrationally until I found the culprit to be an imbalance in the system caused by poor placement of components.

At any rate it's been fun, sort of anyway. Or has it?      

Hans von Lieven

[Low-Q]

Hi All,
almost falling asleep at my PC I finally changed a few settings
with a rod only version and only 2 weights and suddenly I came up
with an endless rotating wheel !
Which also speeds up from time to time.

It uses the centrifugal forces.

It is very easy to setup.
A disc with 200Kg weight, being 2.54 Meters in diameter,
the 2 weights hang each at 1.20 Meter from the axis.
The upper weight has 100 Kg
the lower weight has 101 kg !
the upper rod is 0.600 Meters long
the lowerrod is 0.601 Meters long
The connection rod between the 2 weights is 3.056 Meters long.

It never stops, although I did set air resistance already very strong and
also enabled the friction model.

I will also now post amovie of it  in the next message.
Here is the WM2D file:

http://www.overunity.com/bessler

Enjoy !

Regards, Stefan.

I havent read all the replies, but what happens if you scale down the size and use 100g weights in stead of 100kg, and a 200g disc instead of 200kg. I bet a lot of the work is done by the enormous inertia you have in such weight.