Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Correct equations of motion for object?

  1. Sep 28, 2007 #1

    I am currently trying to simulate an object that can rotate around an axis bouncing on a piece of yarn. I have as equations of motion for the object:

    I d²Theta/dt² = Gravity_Torque + Yarn_Torque

    With the gravity torque alone this system is perfectly conserving energy: the rotating lever will rotate down from horizontal to horizontal at the other side and back. However, if I also put in the yarn torque, this is only directed upwards, so in my simulations the lever will bounce back from the yarn higher than the position it left from and hence energy is continuously added and my object will bounce back higher and higher. How to solve this issue, what are the correct equations of motion? I'm not so good at physics.

    In attachment the situation is displayed.

    Attached Files:

    • Test.JPG
      File size:
      38.4 KB
  2. jcsd
  3. Sep 29, 2007 #2


    User Avatar
    Gold Member

    The object can't bounce off with more energy it arrived with. I don't know what you mean by yarn torque. Is this the resistance caused by stretching the yarn ?

    You should try analysing the system as if there were no gravity, in which case it is like a SHO with a period of free fall in the middle.

    It's not a trivial problem. Do you know how to work out EOMs with Lagrange's method ?
  4. Oct 1, 2007 #3
    Sorry, I'm being a bit overwhelmed. No I don't know those things. I do suspect the equations of motion used for the object are alright but the interaction with the yarn is at fault. The yarn is discretized in a number of 0-dimensional nodes. I have discovered a flaw in my modeling though. I have modeled the effect of collision of the nodes with the object: they can bounce off of the object but then the object should get a momentum change in the opposite direction, which I have not modeled yet. That I have not modeled yet. I don't know though whether that will suffice to make the system conserve energy. I'll get back here when it doesn't.
  5. Oct 1, 2007 #4


    User Avatar
    Gold Member

    Yes, it is vital that objects bouncing off each other should conserve momentum or your simulation is bound to fail.
  6. Oct 1, 2007 #5


    User Avatar
    Staff Emeritus
    Science Advisor

    I don't understand your example well enough to give you the equations of motion for it, but the general equations of motion for a rigid body can be described by Euler's equations .

    Euler's equations use body-fixed coordinates. It looks to me like you might be using absolute space coordinates, in which case you can't use the above directly as written.

    The topic is discussed in Goldstein's "Classical mechanics", but since it's a graduate level textbook and you say your math is not that great, it might be a bit advanced.

    Unfortunately I don't understand the exact coordinate choice you've made to be able to give you any specifics.

    If your body is constrained to rotate only in a plane, I.e. if you assume the yarn and the body are always coplanar, you might be able to generate the equations of motions by using Lagrangian mechanics.

    Using a Lagrangian approach, if the yarn doesn't stretch and has some constant length L, the state of a planar system should be able to be described by two angles, the angle between the yarn and the vertical, and the angle between the end of your rod and the yarn, along with characteristics of the problem such as the length of the yarn, and the moment of inertia (i.e. mass, length, width) of the rod.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook