Time and Frequency Domains
- Time Domain: Representation of signals as they vary over time.
- Frequency Domain: Representation of signals in terms of frequency components (Fourier Transform).
-Integration in the time domain is equal to 1/s in the Laplace domain
-cos(ωt)=(e^(jωt)+e^(-jωt))/2
-A 1st order low pass filter in the Laplace domain has an amplification of 0Db for frequencies much lower than cutoff
-The real part of the poles of a transfer function H(s) should be negative for stability
-The reverse coefficients of 𝐻(𝑧) are all zero for a FIR filter
-The DFT creates a spectrum that is periodic with 2𝜋in the Ω-domain. The DFT is inherently periodic with a period of 2𝜋
-The analog Fourier series decomposes a function into a sum of sines and cosines, not rectified ones
-Multiplying by 𝑧 in the z-domain shifts the signal one sample time 𝑇𝑠 to the left in the time domain.
-Applying a low-pass filter twice reduces the gain at the cutoff frequency by 6 dB. True. Each application reduces it by 3 dB, so twice leads to 6 dB
-The difference equation 𝑦[𝑛]=𝑥[𝑛−1]−𝑥[𝑛−2]−2𝑦[𝑛−1] is an example of a IIR filter.
Key Transformations
- Fourier Transform (FT): Converts time-domain to frequency-domain.
X(jω) = ∫[−∞,∞] x(t)e^(−jωt) dt
- X(jω): Frequency domain representation of the signal.
- x(t): Time-domain signal.
- ω: Angular frequency (rad/s).
- Laplace Transform: For analyzing continuous-time systems.
X(s) = ∫[0,∞] x(t)e^(−st) dt
- s: Complex frequency (s = σ + jω).
- X(s): Laplace domain representation of the signal.
- Z-Transform: For discrete-time systems.
X(z) = Σ[−∞,∞] x[n]z^(−n)
- X(z): Z-domain representation of the signal.
- z: Complex variable for discrete systems.
Sampling and Nyquist Theorem
- Sampling Frequency: Fs (Hz), rate at which the signal is sampled.
- Nyquist Criterion: Fs ≥ 2F_signal_max, ensuring no aliasing.
- Aliasing occurs when Fs < 2F_signal_max, causing overlap in frequency.
Stability Criteria
- s-domain: Poles must lie in the left half of the complex plane.
- z-domain: Poles must lie inside the unit circle in the z-plane.
Types of Filters
- Low-Pass Filter (LPF): Allows frequencies below cutoff.
- High-Pass Filter (HPF): Allows frequencies above cutoff.
- Band-Pass Filter (BPF): Allows frequencies within a range.
- Band-Stop Filter (BSF): Blocks frequencies within a range.
Filter Design (IIR & FIR)
- IIR Filter: Feedback; uses both current and past outputs (infinite impulse response).
Example Transfer Function:
H(z) = (b0 + b1z^−1 + ... + bMz^−M) / (1 + a1z^−1 + ... + aNz^−N)
- bk: Feedforward coefficients.
- ak: Feedback coefficients.
- z: Complex variable for discrete systems.
- FIR Filter: No feedback; only uses current and past inputs (finite impulse response).
Example Transfer Function:
H(z) = b0 + b1z^−1 + ... + bMz^−M
Filter Characteristics
- Cutoff Frequency: ωc, the threshold frequency.
- Warped Frequency: ωa = 2tan(πFc/Fs), accounts for non-linearities.
- Bode Plot: Visual representation of gain and phase vs. frequency.
- Gain: 20 dB/dec per zero, -20 dB/dec per pole.
- Phase: +90° per zero, -90° per pole.
Difference Equations
- General Form:
y[n] = Σ[0,M] bk x[n−k] − Σ[1,N] aj y[n−j]
- y[n]: Output signal.
- x[n]: Input signal.
- bk: Feedforward coefficients.
- aj: Feedback coefficients.
- Use Case: Relates input-output relationships for digital filters.
Discrete Fourier Transform (DFT)
- Forward DFT:
X[k] = Σ[0,N−1] x[n]e^(−j(2πkn/N))
- X[k]: Frequency domain representation.
- x[n]: Time domain samples.
- N: Total number of samples.
- Inverse DFT:
x[n] = (1/N) Σ[0,N−1] X[k]e^(j(2πkn/N))
- Use Case: Converts between time and frequency domains for sampled signals.
Fourier Series Coefficients
- DC Component:
a0 = (1/T) ∫[0,T] f(t) dt
- a0: Average value over one period.
- Cosine Terms:
an = (2/T) ∫[0,T] f(t)cos(nωt) dt
- an: Amplitude of cosine components.
- Sine Terms:
bn = (2/T) ∫[0,T] f(t)sin(nωt) dt
- bn: Amplitude of sine components.
- Use Case: Decomposes periodic signals into harmonic components.
s to z Transformations
- Euler:
s ≈ (1−z^−1)/T
- Approximates derivative in the z-domain.
- Bilinear (Tustin):
s ≈ (2/T)(1−z^−1)/(1+z^−1)
- Maps the entire s-plane into the z-plane.
- Use Case: Converts continuous-time systems into discrete-time systems.
Frequency Warping
- Relation:
ωa = tan(ωd T / 2)
- ωa: Analog frequency.
- ωd: Desired digital frequency.
- Important: Adjusts for non-linearities during transformation.
1. High-Pass Filter Design
- Question: Design a high-pass filter with Fs = 1kHz and Fc = 50Hz.
- Solution:
1. Highest frequency without aliasing:
Fs/2 = 500 Hz.
2. Warped frequency:
ωa = 2tan(πFc/Fs).
3. Transfer Function in s-domain:
H(s) = s / (s + ωc).
4. Discrete Transfer Function:
Use Bilinear Transform to convert H(s) to H(z):
H(z) = (1−z^−1)/(1+z^−1).
5. Bode Plot: Sketch amplitude and phase vs. frequency, noting the cutoff behavior.
2. Fourier Series Example
- Question: Find the Fourier series representation of:
f(t) = {1, 0 ≤ t < 0.01; 0, 0.01 ≤ t < 0.02}.
- Solution:
1. Fundamental frequency: f = 50 Hz, ω = 2πf.
2. Coefficients:
- a0 = (1/T) ∫[0,T] f(t) dt.
- an and bn: Integrate f(t)cos(nωt) and f(t)sin(nωt).
3. Stability Check
- Question: Determine if a system is stable.
- Solution:
- For s-domain: Ensure Re(pole) < 0.
- For z-domain: Ensure |pole| < 1.
Question 11: High-pass filter design
a-Highest signal frequency without aliasing (𝐹𝑠=1kHz) 𝐹max=𝐹𝑠/2=500Hz
b-Calculate warped frequency (𝜔𝑎 ): Use 𝜔𝑎=2tan(𝜋𝐹𝑐/𝐹𝑠), where 𝐹𝑐=50Hz
c-Transfer function 𝐻(𝑠): A first-order high-pass filter has 𝐻(𝑠)=𝑠/(s+)𝑐 , where =2
𝑐 𝐹𝑐
d-Discrete 𝐻(𝑧) using Euler: Replace 𝑠 with (2/𝑇)*(𝑧-1)/(z+1) in 𝐻(𝑠)
e-Forward/reverse coefficients: Use the derived 𝐻(𝑧) to calculate coefficients for difference equations.
f-Bode plot: Sketch amplitude (gain) and phase shift as a function of frequency. Highlight 50 Hz cutoff behavior.
Difference between IIR and FIR:
IIR: Feedback, infinite impulse response.
FIR: No feedback, finite impulse response.
Question 12: FIR filter design
Calculate Ω:Ω𝑐=(2𝐹𝑐)/𝐹𝑠
Coefficients ℎ(𝑛): Use the window method for FIR filter design.
𝐻(𝑧): Convert coefficients to transfer function 𝐻(𝑧)
Difference equation: Derive using coefficients 𝑏.
Question 13: Fourier series
Identify zero coefficients: Determine which terms vanish based on symmetry.
Radial frequency :𝑓 =2𝜋𝑓
⋅ =2𝜋⋅ 50rad/s
Calculate 𝑎0,𝑎𝑛,𝑏𝑛: Integrate over one period using: 𝑎𝑛=2/𝑇∫0𝑇𝑓(𝑡)cos(𝑛𝑡)𝑑𝑡, 𝑏𝑛=2/𝑇∫0𝑇𝑓(𝑡)sin(𝑛𝑡) 𝑑𝑡.
