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

3D Spiral Please help me out

  1. Apr 29, 2005 #1
    I am having trouble getting started with this problem.. I guess I'll explain it first.

    Find one possible matrix A for which the solution to Y'=AY with initial condition Y(0)=(6,13,9) has the following property. As time progresses, the solution Y(t) spirals toward the plane 2x+3y+4z=0 where it continues to circultate about a radius five circle.

    I know I have to find some vectors Vr(real) and Vi(imaginary) that span the plane and are orthogonal to it. The eigen vectors if I remember correctly should come in a pair Vr + iVi and Vr-iVi that I can normalize with PDP-1(inverse) but I am having a problem with how to get these eigenvalues that span the plane.. will someone PLEASE (I'm begging you) help me.. this is driving me insane :bugeye:
    PS the 3rd (in z direction should just be a real lambda value I think)
  2. jcsd
  3. Apr 30, 2005 #2
    A way to attack the problem could be :

    Take a vector in the plane : a=(2,0,-1)/Sqrt(5), the normal vector to the plane : n=(2,3,4) and a vector in the plane, orthogonal to the first : b=(3,-10,6)/Sqrt(145).

    Then you notice that you're LUCKY : you can easily compute the distance of Y(0) to the direction given by n : it is 5.

    Hence the curve is not a 3d spiral but just a screw type one !!

    Hence you know that in the (a,b,n) basis, your vector expresses as :


    hence in the (a,b,n) basis : [tex] x'(t)=\left(\begin{array}{ccc} 0& -w & 0\\ w & 0 & 0\\0&0&-D\end{array}\right)x(t) [/tex]

    then you just have to change the basis to the canonical one and fit the initial condition.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook