Skip to main content

Butterworth order, poles and TF (1 kHz, 2.5 kHz)

Short answer

Order n ≥ log10((10^(0.1αmin)-1)/(10^(0.1αmax)-1)) / (2 log10(ωs/ωp)). For αmax=3dB, αmin=40dB, ωs/ωp=2.5, n=6. Poles at -1626.2±j6069.1, -4442.9±j4442.9, -6069.1±j1626.2 rad/s. T(s)=ωc^6/[(s^2+3252.4s+ωc^2)(s^2+8885.8s+ωc^2)(s^2+12138.2s+ωc^2)] with ωc=6283.2 rad/s.

Derive an expression to calculate the order of a Butterworth low pass filter. Use this expression to calculate the order of Butterworth low pass filter with the following specifications: Maximum passband attenuation: 3 dB at 1 kHz; Minimum stopband attenuation: 40 dB at 2.5 kHz. Also determine the pole locations and transfer function.

Order of a Butterworth low-pass filter

For the normalized Butterworth response (passband edge ωp\omega_p = 1 rad/s)

∣T(jω)∣2=11+ε2ω2n,α(ω)=10log⁡10(1+ε2ω2n) dB|T(j\omega)|^2 = \frac{1}{1+\varepsilon^2\omega^{2n}}, \qquad \alpha(\omega) = 10\log_{10}(1 + \varepsilon^2\omega^{2n})\ \text{dB}

Passband condition (ω=1\omega = 1, α=αmax\alpha = \alpha_{max}):

αmax=10log⁡10(1+ε2)  ⇒  ε2=100.1αmax−1\alpha_{max} = 10\log_{10}(1+\varepsilon^2) \;\Rightarrow\; \varepsilon^2 = 10^{0.1\alpha_{max}} - 1

Stopband condition (at ω=ωs/ωp\omega = \omega_s/\omega_p the attenuation must be at least αmin\alpha_{min}):

10log⁡10[1+ε2(ωsωp)2n]≥αmin(ωsωp)2n≥100.1αmin−1100.1αmax−1\begin{aligned} 10\log_{10}\left[1 + \varepsilon^2\left(\frac{\omega_s}{\omega_p}\right)^{2n}\right] &\ge \alpha_{min} \\ \left(\frac{\omega_s}{\omega_p}\right)^{2n} &\ge \frac{10^{0.1\alpha_{min}} - 1}{10^{0.1\alpha_{max}} - 1} \end{aligned}

Taking logarithms:

n≥log⁡10[100.1αmin−1100.1αmax−1]2log⁡10(ωs/ωp)n \ge \frac{\log_{10}\left[\dfrac{10^{0.1\alpha_{min}} - 1}{10^{0.1\alpha_{max}} - 1}\right]}{2\log_{10}(\omega_s/\omega_p)}

nn is rounded up to the next integer.

Order for the given specifications

αmax\alpha_{max} = 3 dB at fpf_p = 1 kHz, αmin\alpha_{min} = 40 dB at fsf_s = 2.5 kHz. Since only the ratio matters, ωs/ωp=fs/fp=2.5\omega_s/\omega_p = f_s/f_p = 2.5.

100.1αmin−1=104−1=9999.0000100.1αmax−1=100.3−1=0.99526ωsωp=2π(2500)2π(1000)=2.5n≥log⁡10(9999.0000/0.99526)2log⁡10(2.5)=4.00200.7959=5.028\begin{aligned} 10^{0.1\alpha_{min}} - 1 &= 10^{4} - 1 = 9999.0000 \\ 10^{0.1\alpha_{max}} - 1 &= 10^{0.3} - 1 = 0.99526 \\ \frac{\omega_s}{\omega_p} = \frac{2\pi(2500)}{2\pi(1000)} = 2.5 \\ n &\ge \frac{\log_{10}(9999.0000/0.99526)}{2\log_{10}(2.5)} = \frac{4.0020}{0.7959} = 5.028 \end{aligned}

The order must be an integer, so n = 6.

Check: with n=6n = 6, ε2=0.99526\varepsilon^2 = 0.99526, attenuation at ωs\omega_s is 10log⁡10(1+0.99526×2.512)=47.7310\log_{10}(1 + 0.99526 \times 2.5^{12}) = 47.73 dB ≥40\ge 40 dB.

Pole locations

Since αmax\alpha_{max} = 3 dB, ε2=0.995≈1\varepsilon^2 = 0.995 \approx 1, so 1 kHz is the half-power frequency and the poles lie on a circle of radius

ωc=2π×1000=6283.2 rad/s\omega_c = 2\pi \times 1000 = 6283.2\ \text{rad/s}

For nn = 6, θk=(2k−1)π/12\theta_k = (2k-1)\pi/12 (15°, 45°, 75°), pk=−ωcsin⁡θk±jωccos⁡θkp_k = -\omega_c\sin\theta_k \pm j\omega_c\cos\theta_k:

Polesθk\theta_kLocation (rad/s)
p1,p6p_{1}, p_{6}15°15°−1626.2±j6069.1-1626.2 \pm j6069.1
p2,p5p_{2}, p_{5}45°45°−4442.9±j4442.9-4442.9 \pm j4442.9
p3,p4p_{3}, p_{4}75°75°−6069.1±j1626.2-6069.1 \pm j1626.2

Normalized poles: −0.2588±j0.9659-0.2588 \pm j0.9659, −0.7071±j0.7071-0.7071 \pm j0.7071, −0.9659±j0.2588-0.9659 \pm j0.2588.

Transfer function

Normalized: TN(s)=1(s2+0.5176s+1)(s2+1.4142s+1)(s2+1.9319s+1)T_N(s) = \dfrac{1}{(s^2 + 0.5176s + 1)(s^2 + 1.4142s + 1)(s^2 + 1.9319s + 1)}. Replacing ss by s/ωcs/\omega_c:

T(s)=ωc6(s2+3252.4s+ωc2)(s2+8885.8s+ωc2)(s2+12138.2s+ωc2)T(s) = \frac{\omega_c^6}{(s^2 + 3252.4s + \omega_c^2)(s^2 + 8885.8s + \omega_c^2)(s^2 + 12138.2s + \omega_c^2)}

with ωc2=3.9478×107\omega_c^2 = 3.9478\times10^7 and ωc6=6.153×1022\omega_c^6 = 6.153\times10^{22}.

✓ Answer checked by Claude Sonnet 5.5

All Filter Design chapter 2 questions →