Gold gear vs steel gear

[ftworks]

Steel vs gold gears

noiseplayer

Rating:
Filesize: 310 kB
Comments: 2
Ratings: 3

Hi all. I want you to try this:
Get two identical gears.
Make one of gold and the other of steel.
Put two identical motors on them of 100 Nm torque.
Turn on the two motors.
What you expect to happen?
Two identical torques one against each other... then nothing must move... or not?
See what Algodoo says. Very interesting bug... if it is a bug...
Bye

ft

steel_and_gold_gears.jpg

(116.68 KiB) Not downloaded yet

[RicH]

Your screenshot is questionable. Why do they weigh the same? It shouldn't as the preset materials(gold and steel) have different densities.

[ftworks]

Thank you, there was an error, I corrected it in the post.

[RicH]

I believe this happens because the hinge strength is relative to the weight of the object.

[RaRaMalum]

I believe this happens because the hinge strength is relative to the weight of the object.

This has been a "bug" in phun/algodoo since the beginning of time.

[ftworks]

I believe this happens because the hinge strength is relative to the weight of the object.

What do you mean by hinge strenght?
Do you mean motor strenght?
If so, what do you mean by "relative to the weight of the object".
Does the heavier gear motor strenght is 100 Nm or not?

[RicH]

Yes, I do mean motor strength.
For example:
Object A weighs 100 kg
Object B weighs 50 kg
When you give both objects equal motor strength, Object A turns out to be more powerful than Object B. Get it?

[TimTheOne]

yea RicH is right.

A heavier object will be able to produce more 'force' when it has the same strenght

[izacque]

Rich is totally right here

[ftworks]

Object A weighs 100 kg
Object B weighs 50 kg
When you give both objects equal motor strength, Object A turns out to be more powerful than Object B. Get it?

Yes, I saw. But does it correspond to a real situation?
I wonder how motor torque is measured?
Power = torque * angular velocity.
If torques are the same, and angular velocities are the same, we have two objects with the same power, but the heavier one is powerful than the other... Is it realistic?

[Frank]

The NM is a lie!

[ftworks]

The NM is a lie!

An algorithm is an approximation, not a lie...
It helps to know and question...
Let's try to answer please...

[Mystery]

No it is a lie.

[davidz40]

Nm is a total lie. I made one metre long 1 kg stick and attached hinge to one end. The "force" needed to keep it horizontal was about 35 Nm (should be about 0.5 Nm), and didn't depend on stick's mass

[KarateBrot]

Nm is a total lie. I made one metre long 1 kg stick and attached hinge to one end. The "force" needed to keep it horizontal was about 35 Nm (should be about 0.5 Nm), and didn't depend on stick's mass

Sure you're talking about the force? 35Nm is no force. It's torque.

[Antotabo]

Nm is a total lie. I made one metre long 1 kg stick and attached hinge to one end. The "force" needed to keep it horizontal was about 35 Nm (should be about 0.5 Nm), and didn't depend on stick's mass

Sure you're talking about the force? 35Nm is no force. It's torque.

He wanted to say N and not NM. This test totally proves that it depend on weight and it should not be called NM but NM/KG. Let me explain: The gravity force on the object is proportional to the weight because gravity is an acceleration. So if you increase density, the force needed to keep it there IS actually increasing but with no appearing movement. So yes it depend on mass. Try it with a mass attached you will see now that the force will then change.

So yes, Phun /Algodoo is wrong on that.

I think this happen because the parameter is directly linked to the acceleration of the body, and it should be fixed in some way. But it would maybe cause problem to old scene.

Edit: shorter way to explain: A torque proportional to mass = angular Acceleration= acceleration= gravity (explain the above test)= not a torque.

[KarateBrot]

I tested it with a 100Nm torque motor and with higher density the force also increases. And there's no Nm somewhere. There's just the unit N like there should be for forces.

Or may I see the experimental set-up, please? Maybe I'm misunderstanding something

[davidz40]

torque experiment

davidz

Rating:
Filesize: 12.65 kB
Comments: 1
Ratings: 1

Here's my experiment scene
In previous post, I wanted to say torque of course.

Anyway:
-Acting force depends on torque and length of the arm. More torque-more force. Longer arm-less force.
In algodoo it seems to be reversed somewhat
You can make the longer arm even longer (50m or so), and it will be capable of holding many tonnes instead of 10kg.
And torque definitely shouldn't depend on mass of object attached to hinge.

[KarateBrot]

Okay, I see. But I wouldn't jump to a conclusion because you only can make arms in algodoo that contain mass (so the force is different at any position of the geometry and you have to sum an infinite amount of forces) and they also have got dimensions. In school physics you only learn about torque with a point mass at the end of a virtual arm. So there really should be a difference between the real formulas and the easy-going school formulas.
But I have to admit it is a weird thing that the resultant force that comes from the motor's torque is dependent from the geometry's mass. It's just too weird to be real ( <= It's dangerous to say something like that but I hope no physics professor will slap me after reading it ). But yeah, seems like it's wrong.

A feedback from Algoryx would be really nice.

Or maybe we should make a "torque bug" topic in the bug section.


Edit:
I just got an idea. As you perhaps know to have force there also needs to be an interaction partner. Maybe algodoo calculates it correct if one geometry is attached to another and it gets wrong if the motor gets attached to the rigid/fixed/inflexible background. Of course there's no interaction partner if it is attached to the background so maybe the geometry interacts with itself.
Just a few suggestions. Maybe I'm totally wrong.

[Antotabo]

Killinich, in fact you can do calculation very easily with no nead of complicated infinite force. In the case of a rectangular arm, just consider the force resulting from the object mass being in the center of gravity (simply divide arm lenght by two) and pulling down. So yes your easy going shool formula are still good.

[KarateBrot]

Kilinich? My name is KarateBrot

But they're not good anymore if you also have got gravity. But yeah it's only a small difference.

[tatt61880]

Motor torque seems to depend on not only length of the arm but also on area of the geom.

I want to know how motor torque works exactly. :s
[scene]44965[/scene]

[Mystery]

I think it works of the mass of the object the hinge is on, i notice it all the time. This actually might deserve a bug report.

[tatt61880]

I've tried to get equation for describing motor works.
Algodoo/Phun's motors seem to give torque(=N) to the geom, with following equation.

N = I_center * motorTorque

which N, I_center and motorTorque are defined as below sentences.

N is torque which the motor gives to the geom. Unit: Nm (= kg * m^2/s^2)
I_center is moment of inertia for rotations around the center of mass with the geoms. Unit: kg*m^2
motorTorque is torque of the motor. Unit: 1/s^2

motorTorque seems to have 1/s^2 (= s^-2) as its unit, if my equation is correct...
However, motorTorque had better have Nm(=kg * m^2/s^2) as its unit.

I recommend to adopt the equation N = motorTorque for describing motors work, instead of N = I_center * motorTorque

Thanks for your time.

/Tatt

[emilk]

"Oops"

Yes, you are correct - the motor strength is multiplied with the (combined) moment of inertia of the two objects the hinge is attached to, to get the real motor torque. So as tatt pointed out, the correct unit for the motor strength should be 1/s^2 (or rad/s^2, to be precise), describing the rate of change of the angular velocity of the affected object.

The reasoning behind this choice of unit is that a machine works the same if you scale it or change the involved densities. But obviously I messed the unit up when adding it to the GUI (much later).

Now that that has been cleared out - the obvious question is, would Nm be a better choice of unit? For some things, yes. But I'm betting there will be people who are gonna be irritated that their machines no longer works the same when scaled.

What do you think? Keep the current strength (rad/s^2, i.e. angular acceleration), and only fix the unit displayed in the menus? Or change to Nm? And then when scaling, also scale the Nm strength as to keep constant angular acceleration (as now)?