WM2D

[Gregory]

Custom point properties:
X: body[?].radius*cos(body[?].p.r)
Y: -body[?].radius*sin(body[?].p.r)

:-))

Great! Thank you very much!
I have also figured out myself just right the moment and wrote on a slip of paper just before I saw your post. :) I realized that Wm2d calculate it this way... I'm improving slowly.

And Rainer, If some day I will have something workable, then you will be one of the firsts to hear about it.

[Gregory]

Now, here is another pulley example for everyone.
I think it works similarly to a real pulley.

There are 2 + 2 slightly different formulas for the coordinates of the attachment points, and two other formulas to control the lenghts of the ropes, and finally I attached a motor to it which also contains a special formula for the driving torque.

The formulas for the ropes are only approximate ones.
Have a look at it if you like.

Greg

[Tinhead]

Nice one,

Welcome to the advanced WM2D 'conf'user club ;)

Cheers,
Rainer

[jim_mich]

The model is cute to watch but its not accurate. The variation in rope length is defined as rotation (degrees) / 1.65 which equals 53.5454 for 90 degrees of rotation. One quarter of the pulley circuference is 78.5398. This means that the weights do not move far enough to equal 1/4 pulley rotation in the simulation.

Also the motor torque driving the pulley is based on the pulley's rotation rather than on rope forces. In reality the weights will balance, but once they are moved they would continue to move at a constant speed and would unwrap the rope from the pulley.

[Gregory]

Welcome to the advanced WM2D 'conf'user club ;)

Thanks... Yeah, confuser. :) It's great!


And yes, Jim... you're right.

Now, it's corrected. Except that the motor remained the same. I don't know how to measure rope forces, do you?
Sure, there is no such motor exist in reality, but it's only to show the motion.

I also added a "constant speed after a push" version next to it.

[jim_mich]

Gregory,
You used 125 + (body[2].p.r) * 0.8726 to define the length of the rope. This is good until/if you decide to change the radius of the pulley, in which case you need to remember to go back and change the 0.8726 values. The 0.8726 value is the circumference of the pulley divided by 360. If you make the formula as follows then it will be right for any size pulley.

125 + body[2].p.r * body[2].radius * 2 * Pi / 360

Rather than messing around with dropping a weight onto the pulley to start it moving you can add a momentary torque to the pulley and set it active when t < 0.5 so that it's active while time is less than 0.5 seconds, or whatever time looks good.

To find a rope force formula, I selected the rope, then clicked measure, then clicked tension, then looked at the properties of the tension output graph. It showed me that the formula is -constraintforce(17).x

You could use the contraint force of the two ropes to define the motor torque on the pulley. You'll need to figure in the pulley diameter.

One more comment; when the weights contact the pulley on the right they swing outward. This should cause the rope point connection on the pulley to change, which it doesn't do. It is little things like this that can keep a model from being exactly accurate.

[Gregory]

Thanks for the advice Jim mich, I appreciate it.
I like the general formula... How could I miss that...? 8)

Yes, I know the momentary torque trick, just sometimes like watching a ball drop & hit something.

Now I see the measures of the rope... Correct.

And great, you have a good sense to find any inaccuracy.
I think I will ask for the help of your sense when it needed.

Good luck with your CF wheel design!

Greg

[Gregory]

I made a few new experiment with the pulley and found a problem.

I simulated an experiment where the ropes lose their tension and become loose at some points. When they are periodically become loose Wm2d fail to calculate the rope lenghts correctly, and the ropes become longer and longer over time which seems really stupid.

Can this problem be fixed without replacing the ropes with rods?
Why does Wm2d fail to simulate it correctly?

[jim_mich]

Tighten up the simulation accuracy error parameters (integrator error, overlap error, assembly error). If you tell it to be accurate within .080 then it could gain or loose that much each step, which might be good enough to simulate a train wreck but not for a small wheel.

I usually use 0.001 or even smaller.

[Gregory]

OK, really. It seems to solve the problem.

[Gregory]

Is it possible in Wm2d to switch on Gravity for some objects, and switch off for others?

I would need it to test a few three dimensional ideas with a CF design concept...

If not, is there any way to imitate Gravity for some objects in the sim, and not for others?

I thought about Electrostatics. I think I can use it as a constant by introducing some formulas, and then I can switch it on and off as I choose for every object...

[Fletcher]

Don't know if this will help - I've never used it & can't find the menu item for it at the moment.

Sometimes Rainer uses a plan view i.e. looking down from above a model. That takes all gravity out of it [eg. a horizontal wheel] so you can look at CF alone. Perhaps if he is reading this or someone else knows how to change the setup they'll respond for you.

[Tinhead]

Hi Gregory,

hmmmm, not sure if there is an easy way to do this. One way to go would be to create a counter force attached to some objects to cancel out gravity.
In the 'Y' field I would use a formula like this: Body[x].mass*9.80665
For sure, replace 'x' with the number of the object in question.

Or the other way around, turn gravity off in the World menu and attach your own 'gravity force' to some objects.

There are other 'work arounds', it just depends on what you try to do. Without knowing this it is hard to make suggestions.

Cheers,
Rainer

[wikiwheel]

As I get into this. I think that the only thing that will work is two or more independent systems that balance individually at different angles so that one or the other forces the other one to shift. Over then over then over..oto...oto..oto............................................infinity


Mik

[Gregory]

Thanks Fletcher,

If you simply switch off gravity in the World menu, then you can consider that you're working in plan view, say your wheel is placed horizontally and gravity has 'no effect' on it. Can be good in some cases. Most likely for CF, or magnets, etc.

Hi Gregory,

hmmmm, not sure if there is an easy way to do this. One way to go would be to create a counter force attached to some objects to cancel out gravity.
In the 'Y' field I would use a formula like this: Body[x].mass*9.80665
For sure, replace 'x' with the number of the object in question.

Or the other way around, turn gravity off in the World menu and attach your own 'gravity force' to some objects.

There are other 'work arounds', it just depends on what you try to do. Without knowing this it is hard to make suggestions.

Cheers,
Rainer

Hey Rainer, thanks!

You win again... That is a very simple way to do. What would I do without you! I can imagine myself messing around with electrostatics and such when I can easily attach a counter force... that's funny.

I have an idea based on Gravesande's CF experiment, and I want to use a "horizontal unit", but Wm2d is just 2D.

Now ok, I can equal the force of gravity on the CF builder weights, while the wheel can turn by normal overbalancing weights & gravity. At least this is the concept.

Sometimes I use a kind of backward designing in wm2d... Including some formulas, tricks, cheats etc. So I can make the wheel turn this way, and after I usually think about how can I modify and replace the unrealistic aspects with real ones. Now I thought why not try a CF design... Maybe I can finally make it running without a cheat. It would be good, but in the worst case I will still end up with another cheated runner in my collection.

Sure, sometimes it's good to see one revolving just for inspiration!