The continuous-time signal x(t) = cos(20πt + π/4) is ideally sampled at a rate of 30 samples per second. The resulting sampled signal is then passed through a filter whose impulse response is h(t) = [sin(πt)/(πt)] · cos(80πt − π/2). Find the expression for the output signal of the filter.
15 marks · directive: Find
The core of a good answer
The filter is a narrow band pass filter centred on 40 Hz (passband 39.5 to 40.5 Hz, gain ½, phase −π/2 at +40 Hz); the impulse sampled signal has a 40 Hz image (10 + 30 Hz) of amplitude f_s = 30, so the output is y(t) = 15 cos(80πt − π/4) (= 15 sin(80πt + π/4)).
An outline written by upsc.help, not an official UPSC answer; every fact checked by a separate review. How we work
What the question asks
A sampling and filtering problem: ideal sampling replicates the 10 Hz spectrum around every multiple of 30 Hz, and the modulated sinc filter selects only the spectral line that falls in its narrow passband. The phase of the image and the phase of the filter must both be tracked.
Answer outline
Introduction
x(t) has frequency 10 Hz (ω = 20π). Ideal sampling: x_s(t) = Σ x(nT)δ(t − nT), T = 1/30 s, so X_s(f) = f_s Σ_k X(f − 30k).
Step 1: Frequency response of the filter
- sin(πt)/(πt) ↔ rect(f): 1 for |f| < 0.5 Hz, 0 otherwise (ideal low pass, cutoff 0.5 Hz, i.e. π rad/s)
- cos(80πt − π/2) = sin(80πt) = [e^{j(80πt − π/2)} + e^{−j(80πt − π/2)}]/2
- H(f) = ½e^{−jπ/2} rect(f − 40) + ½e^{+jπ/2} rect(f + 40)
- Passband 39.5 Hz < |f| < 40.5 Hz; gain ½, phase −π/2 at +40 Hz and +π/2 at −40 Hz
Step 2: Spectrum of the sampled signal
- x(t) = ½e^{jπ/4}e^{j2π(10)t} + ½e^{−jπ/4}e^{−j2π(10)t}
- Spectral lines of x_s(t) at ±10 + 30k Hz: 10, 20, 40, 50, 70, ... and their negatives, each line carrying weight f_s × ½ = 15 with the phase of its parent
- The only lines inside 39.5 to 40.5 Hz are +40 Hz (= +10 + 30, phase +π/4) and −40 Hz (= −10 − 30, phase −π/4); the 20 Hz and 50 Hz lines (from −10 + 30 and −10 + 60) fall outside
Step 3: Filter the selected lines
- At +40 Hz: 15e^{jπ/4} × ½e^{−jπ/2} = 7.5e^{−jπ/4}
- At −40 Hz: 15e^{−jπ/4} × ½e^{jπ/2} = 7.5e^{+jπ/4}
- y(t) = 7.5e^{−jπ/4}e^{j80πt} + 7.5e^{jπ/4}e^{−j80πt} = 15 cos(80πt − π/4)
Step 4: Checks and convention
- Output frequency 40 Hz is an image of the 10 Hz input created by sampling, not the original tone; this is why the result is not at 10 Hz.
- Numerical check: the complex amplitude of the 40 Hz component of the sampled sequence is 0.5∠45°; multiplying by f_s = 30 and by H(40) = 0.5∠−90° gives 15 cos(80πt − 45°).
- If sampling were defined with the factor T (x_s = T Σ x(nT)δ(t − nT)), every amplitude would be divided by 30, giving 0.5 cos(80πt − π/4); state the convention used.
Conclusion
y(t) = 15 cos(80πt − π/4): a 40 Hz sinusoid, the first upper image of the 10 Hz input, scaled by f_s × ½ and shifted by −π/2 by the filter.
Examples and data to cite
- Undersampling, also known as bandpass sampling or IF sampling, samples a band pass signal such as a radio intermediate frequency (IF) signal below twice its highest frequency and relies on sampling images, as this problem does.
- Multiplying a low pass impulse response by a cosine shifts its passband to plus and minus the cosine frequency (frequency shift property); that is how this sinc filter becomes a band pass filter at 40 Hz.
Facts to remember
- Ideal impulse sampling: X_s(f) = f_s Σ X(f − k f_s).
- sin(πt)/(πt) = sinc(t) ↔ rect(f) of unit height and width 1 Hz.
- Modulation: h(t)cos(2πf₀t + θ) ↔ ½[e^{jθ}H(f − f₀) + e^{−jθ}H(f + f₀)].
- An LTI system multiplies each complex exponential e^{j2πft} by H(f).