Fourier expansion of $x(t) = t$ on a given period

42 Views Asked by At

Suppose I have a periodic function defined on a period as:

$$x(t) = t$$ $$ 0 < t < T_0$$

I wish to find the fourier expansion of this. So:

$$x_n = \frac{1}{T_0}\int_{0}^{T_0}te^{-i\frac{2\pi}{T_0}nt}dt$$

If I allow $$-i\frac{2\pi}{T_0}n = \phi$$ then:

$$x_n = \frac{1}{T_0}\int_{0}^{T_0}te^{\phi t}dt$$ and the result is straightforward through integration by parts (work omitted):

$$x_n = \frac{1}{T_0}(\frac{te^{\phi t}}{\phi} - \frac{e^{\phi t}}{\phi^2})$$ $$ 0 < t < T_0$$

Replace phi:

$$x_n = \frac{1}{T_0}(\frac{-te^{-i\frac{2\pi}{T_0}nt}}{i\frac{2\pi}{T_0} n} + \frac{e^{-i\frac{2\pi}{T_0}nt}}{4\frac{\pi^2}{T_0^2} n^2})$$

On the given interval. So I plug in T nought and 0, and I get an expression that I cannot figure out how to simplify. Here is my final answer:

$$x_n = \frac{-T_0e^{-i2\pi n}}{i2\pi n} + \frac{e^{-i2\pi n}}{4\frac{\pi^2}{T_0} n^2} - \frac{1}{4\frac{\pi^2}{T_0} n^2}$$

But this does not seem correct to me. I considered ways to simplify the solution, but none were obvious to me. I could convert the exponentials to their sine/cosine form, but I don't see any likely results that are any more useful than what I already have.

I'm pretty sure I've made an error somewhere.

1

There are 1 best solutions below

0
On BEST ANSWER

Your work and the resulting answer for $x_n$ were correct.

Simply note that $e^{i2n\pi}=1$.

So, the second and third terms cancel each other and we have

$$x_n=-\frac{T_0}{i2n\pi}$$

which is the result you have after recognizing that $e^{i2n\pi}=1$.