- 2074 Chaitra · 12 marks
Impulse invariance Butterworth, 0.7 dB and 14 dB
Short answer
Order N=3, Ωc=0.6301 rad/s. H(z)=0.63015/(1−0.53251z⁻¹) + (−0.63015+0.53085z⁻¹)/(1−1.24748z⁻¹+0.53251z⁻²); parallel realization.
Design a digital low-pass filter with the following specification: i) Pass-band magnitude constant to 0.7 dB below the frequency of 0.15π. ii) Stop-band attenuation at least 14 dB for the frequencies between 0.6π to π. Use Butterworth approximation as a prototype and use impulse invariance method to obtain the digital filter.
Step 1: specifications
Pass band within 0.7 dB of 0 dB up to 0.15π (|H| ≥ 10^(−0.7/20) = 0.9226); stop-band attenuation at least 14 dB (|H| ≤ 0.1995) from 0.6π to π. T is not given; take T = 1 s.
Step 2: analog edge frequencies (impulse invariance: Ω = ω/T)
Ωp = 0.4712 rad/s, Ωs = 1.8850 rad/s (T = 1 s)
Step 3: order
N ≥ log[(1/0.1995² − 1)/(1/0.9226² − 1)] / (2 log(Ωs/Ωp)) = log[24.1189/0.1749] / (2 log 4.0000) = 1.7769
The analog formula gives N = 2. However, impulse invariance causes aliasing: the usual N = 2 design (Ωc = 0.7287 rad/s, H(z) = 0.3033z⁻¹/(1 − 1.0396z⁻¹ + 0.3568z⁻²)) gives only 0.899 at 0.15π, below the required 0.9226, and a numerical search shows that no choice of Ωc lets a second-order design meet both edges. So take N = 3.
Step 4: cut-off
Ωc = Ωp / (0.1749)^(1/6) = 0.6301 rad/s
Step 5: analog poles and H(s)
Poles: s_k = Ωc exp[jπ(2k + N + 1)/(2N)], k = 0 … 2 (left half plane):
- s0 = −0.3151 + j0.5457
- s1 = −0.6301
- s2 = −0.3151 − j0.5457
H(s) = Ωc³ / [(s + Ωc)(s² + Ωc s + Ωc²)] = 0.2502 / [(s + 0.6301)(s² + 0.6301s + 0.3971)] = Σ A_k/(s − s_k), with residues
- A0 = −0.3151 − j0.1819
- A1 = 0.6301
- A2 = −0.3151 + j0.1819
Step 6: impulse invariant transformation
Each term A/(s − s_k) becomes T·A/(1 − e^(s_k T) z⁻¹), so h[n] = T·h_a(nT). Combining the complex-conjugate pair into a real second-order section:
- 0.63015 / (1 − 0.53251z⁻¹)
- (−0.63015 + 0.53085z⁻¹) / (1 − 1.24748z⁻¹ + 0.53251z⁻²)
H(z) is the sum of these two terms (parallel form).
Step 7: check
|H(e^jωp)| = 0.9231 (required ≥ 0.9226); |H(e^jωs)| = 0.0360 (required ≤ 0.1995); |H(1)| = 0.9996. Both conditions are satisfied.
Answer: N = 3, Ωc = 0.6301 rad/s, H(s) = 0.2502/[(s + 0.6301)(s² + 0.6301s + 0.3971)], H(z) = 0.63015/(1 − 0.53251z⁻¹) + (−0.63015 + 0.53085z⁻¹)/(1 − 1.24748z⁻¹ + 0.53251z⁻²).
✓ Answer checked by Claude Opus 5.5 (high effort)
Related questions
- Impulse invariance Butterworth, 0.2π, 0.5 dB · Digital Signal Analysis and Processing, chapter 6 · asked 2×
- Impulse invariance Butterworth, 200 Hz and 500 Hz · Digital Signal Analysis and Processing, chapter 6
- Low-pass FIR window design, αs = 41 dB · Digital Signal Analysis and Processing, chapter 5
- Low-pass FIR window design, αs = 40 dB · Digital Signal Analysis and Processing, chapter 5 · asked 2×
- Bilinear Butterworth, 0.2π passband, −15 dB at 0.4π · Digital Signal Analysis and Processing, chapter 6 · asked 2×
- Kaiser low-pass FIR, passband 0.9π, stop 0.14π · Digital Signal Analysis and Processing, chapter 5 · asked 2×
- Low pass to band stop transformation · Filter Design, chapter 3 · asked 10×
- Band stop filter (2000 rad/s, 400 rad/s) from 3rd order · Filter Design, chapter 3 · asked 4×
- Matched filter approximation using RC low pass filter · Communication Systems, chapter 9 · asked 2×
- Impulse response of ideal low pass filter · Signals and Systems, chapter 5 · asked 5×
All Digital Signal Analysis and Processing chapter 6 questions →