# Homework Help: Laplace transform dirac delta

1. Nov 8, 2014

### redundant6939

1. The problem statement, all variables and given/known data
$x(t) = cos(3πt)$
$h(t) = e-2tu(t)$

Find $y(t) = x(t) * h(t)$(ie convolution)

2. Relevant equations
Y(s) = X(s)H(s) and then take inverse laplace tranform of Y(s)

3. The attempt at a solution
$L(x(t)) = [πδ(ω - 3π) + πδ(ω + 3π)]$
$L(h(t)) = \frac{1}{s+2}$

Laplace Transform inverse :
$y(t) = \frac{1}{2} \int \frac{e^{st}}{s+2} * δ(w-3π)$ + $\frac{1}{2} \int \frac{e^{st}}{s+2} * δ(w+3π)dt = \frac{1}{2} \frac {e^{3sπ} + e^{-3sπ}}{s+2}$

Im not sure about direc delta integration..

Last edited: Nov 8, 2014
2. Nov 8, 2014

### vela

Staff Emeritus
The Laplace transform of $\cos 2\pi t$ isn't what you said it is. (It should be a function of $s$, not $\omega$.) You're thinking of the Fourier transform.

3. Nov 8, 2014

### redundant6939

Is it $\frac {s}{s^2 + (3π)^2}$ ?

4. Nov 8, 2014

### vela

Staff Emeritus
Yes.

5. Nov 8, 2014

### redundant6939

But isn't that the LT of cos(3pi*t)u(t)?

6. Nov 8, 2014

### vela

Staff Emeritus
Yes. Usually, it's assumed you're working with the unilateral Laplace transform, and you essentially ignore what's going on before t=0. Are you supposed to be using the bilateral Laplace transform?

7. Nov 8, 2014

### Ray Vickson

The "bilateral" transform of $\cos(3 \pi t)$ does not exist, because the integral
$$\int_{-\infty}^{\infty} e^{-st} \cos(3 \pi t) \, dt$$
is not convergent. However, the "unilateral" LT of $u(t) \cos(3 \pi t)$ certainly does exist.