Suppose we have a DSBSC signal given by :

$$s(t) = 10 \cos (2\pi f t + \phi ) \cdot m(t) $$

where carrier is random signal with phase ($\phi$) uniformly distributed in $[0,2\pi]$.

If the power spectral density of the message signal $m(t)$ is :

enter image description here

Q.How can I express the power spectral density of $s(t)$ in terms of given PSD of $m(t)$?

My attempt:

By Wiener–Khinchin relations,

Autocorrelation of $m(t) = 3 \mathrm{sinc}^2(3000t)$

Now to get the autocorrelation of $s(t)$,

$$R_{ss} (T) = E[ x(t)x^*(t+T)\cdot m(t) m^*(t+T) ]$$

where $x(t)$ is the carrier.

If I assume they are independent, then :

$$R_{ss}(T) = R_{mm}(T) R_{xx}(T)$$

which gives me

$$R_{ss}(T) = 150\cos(2\pi T)\cdot\mathrm{sinc}^2(3000T)$$

  • 1
    $\begingroup$ Also check this answer to a related question. $\endgroup$ – Matt L. Dec 1 '15 at 13:05

The random phase $\phi$ makes the carrier signal $c(t)=A\cos(2\pi f_0+\phi)$ wide-sense stationary with the autocorrelation function

$$R_c(\tau)=\frac{A^2}{2}\cos(2\pi f_0\tau)\tag{1}$$

The Fourier transform of $(1)$ is the power spectrum of the carrier:


where $\delta(f)$ is the Dirac impulse.

Since the carrier signal $c(t)$ and the message signal $m(t)$ are assumed to be uncorrelated, the autocorrelation function of the modulated signal $s(t)=c(t)m(t)$ equals the product of the autocorrelation functions of $c(t)$ and $m(t)$:


From $(3)$ it follows that the power spectrum $S_s(f)$ of $s(t)$ is the convolution of the power spectra of $c(t)$ and $m(t)$:

$$S_s(f)=S_c(f)\star S_m(f)=\frac{A^2}{4}\left[S_m(f-f_0)+S_m(f+f_0)\right]\tag{4}$$

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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