This harmonic oscillator is driven and damped, with the form:

$$\ddot{x} + \lambda \dot{x} + \omega_0^2 x = A \cos(\omega_d t)$$

Now, I have used the ansatz (guess): $x(t) = B \cos(\omega_d t + \phi)$, and have written B in the form:

$$B = \frac{A} {\sqrt{(\omega_o^2-\omega_d^2)^2+\lambda\omega_d^2}}$$

Next, I am required to "approximate B using the Lorentzian form"

$$B = \frac{C}{(\omega_d - \Omega)^2+\biggl(\frac{\Gamma}{2}\biggr)^2}$$

However, this is where I am stuck. I know that because it says "approximate" I will somehow have to drop out terms from my first expression to B, but I don't know where to start. How can I write B in this form?

EDIT: I have found a wikipedia article on resonance which show a form very similiar to what I seek, however, I can't seem to find a derivation http://en.wikipedia.org/wiki/Resonance

  • 1
    $\begingroup$ I gess that you could start expanding the square root in a Taylor series. $\endgroup$
    – Ana S. H.
    Jan 3, 2013 at 18:06
  • $\begingroup$ ...and which term is smaller? $(\omega_0^2 - \omega_d^2)^2$ or $\lambda \omega_d^2$? $\endgroup$
    – Vibert
    Jan 3, 2013 at 20:51
  • 1
    $\begingroup$ @Vibert I am not given any information in the question that would allow me to deduce which is smaller. In fact, I have provided all the information that was given to me above. $\endgroup$
    – Mel
    Jan 3, 2013 at 21:28
  • $\begingroup$ If you take the limit $\omega_d \rightarrow \infty$ you find $A=C$. Then $\omega_d=0$ and $\omega_d=\omega_0$ give two other equations which should be able to determine $\Omega$ and $\Gamma$. Not sure this will work properly though. Probably a better bet is to Taylor expand both forms around the maximum point and match terms. $\endgroup$
    – Michael
    Jan 4, 2013 at 2:46
  • $\begingroup$ how did you guess x(t)=Bcos(ωdt+ϕ)? $\endgroup$
    – user22564
    Mar 29, 2013 at 19:48

1 Answer 1


Since the maximum is the most important point of the curve, I suggest matching the derivatives 0-2 of the two curves at $\omega_0$. This is equivalent to doing a Taylor/power expansion on both functions and matching the first three coefficients. Since there are three constants, we can match three criteria (=equations).

First derivative: set $\Omega=\omega_0$, and by differentiating both curves you can show that the first derivative of both curves at $\omega_0$ is zero.

Zeroth derivative: By setting the two curves equal and solving for $C$, you find that $C=\frac{A (\frac{\Gamma}{2})^2}{\sqrt{\lambda}\omega_d}$.

Second derivative: We still haven't set $\Gamma$. The necessary equation comes from setting the second derivative at $\omega_0$ equal. This gives $\frac{\Gamma}{2}=\sqrt{\frac{\lambda}{2}}$.

The plot (all parameters in the original resonance curve are 2; blue is original, red is Lorentzian) looks pretty good to me:

enter image description here

  • $\begingroup$ Thanks. One thing though, how come your blue curve doesn't (appear) to have two peaks? When I plotted both curves, the original curve had two peaks, at $\omega_d = \omega_0$ and $\omega_d = -\omega_0$ $\endgroup$
    – Mel
    Jan 20, 2013 at 13:47
  • $\begingroup$ You're absolutely right - I forgot a square root. I redid the calculation, the $C$ hasn't changed, but the $\Gamma$ did. $\endgroup$
    – zonksoft
    Jan 20, 2013 at 14:22

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.