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

Homework Help: Help with a vector proof about perpedndicular unit vectors

  1. Sep 13, 2010 #1
    Prove that if a(t) is a unit vector whose components are functions of t, then the vector b(t) = da(t)/dt is perpendicular to a(t)

    here i have the parameters that make a(t) a unit vector:

    1 = f(t)^2 + g(t)^2

    i know of the example sinxi + cosxj = a(t)
    but that is the only one i can get to work. I need a general proof though so:

    a(t) = f(t)^2 i + g(t)^2 j

    b(t) = f '(t) i + g '(t) j

    (dot) product of a(t) and b(t) is:

    f(t)*f '(t) + g(t)*g '(t)

    i tried to take the derivative of a(t) and solve for both f'(t) and g'(t)

    1 = f(t)^2 + g(t)^2

    0 = 2*f(t)*f'(t) + 2*g(t)*g'(t)

    i treated this like one big equation and solved each for g'(t) and f'(t)

    f ' (t) = [-g(t)*g ' (t)]/ f(t)

    g ' (t) = [-f(t)*f ' (t)]/ g(t)

    when i carry out the dot product i still dont get zero, i need some advice, am i even on the right track.

    the only other way i could think of was to setting each component equal to zero to solve for the derivative of the functions, but i did not know if this was mathematically legal


    2*g(t)*g'(t) = 0

    g'(t) = 0
    Last edited: Sep 13, 2010
  2. jcsd
  3. Sep 13, 2010 #2


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    Use this expression to express g(t) in terms of f(t). That will help greatly.

    I don't understand the part in red. Why have you squared the functions? They should not be squared if your definition of a is to be consistent with your definition of b. The rest of what you have doesn't make sense either. See if following my advice above helps.
  4. Sep 13, 2010 #3
    did you mean this one: 1 = f(t)^2 + g(t)^2

    so f(t) = (1 - g(t)^2)^1/2

    and g(t) = (1- f(t)^2)^1/2

    then the derivative of the above:

    f ' (t) = -[g(t)*g ' (t)] / [1-g(t)^2]
    g ' (t) = -[f(t)*f ' (t)] / [1-f(t)^2]

    so then if i take the dot product

    [f(t) * f ' (t)] + [ g(t) * g ' (t) ]

    (1 - g(t)^2)^1/2 * -[g(t)*g ' (t)] / [1-g(t)^2] + (1- f(t)^2)^1/2 * -[f(t)*f ' (t)] / [1-f(t)^2]

    simplified :

    -[g(t)*g ' (t)] -[f(t)*f ' (t)]

    still dont understand it, where is my error?? :confused:
  5. Sep 13, 2010 #4


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member


    You just need one of these. The whole point is to eliminate one by expressing it in terms of the other. For example, if you express g in terms of f, then henceforth everything is in terms of f and you don't need g anymore. Do you understand?

    Here is the problem. This differentiation is wrong. You need to use the chain rule since:

    dg/dt = dg/df * df/dt

    What is g(f), and what is its derivative with respect to f? That is the first step. The second step is to multiply that by df/dt.

    Also, again, you only have to compute ONE of the above derivatives.
  6. Sep 13, 2010 #5
    got it finished, thanks. i wasnt thinking of eliminating terms correctly, like you said
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook