UNIT 4: Advanced Network Synthesis, Filter Design & Two-Port Networks
I. Advanced Two-Port Network Analysis
Symmetrical vs. Asymmetrical Networks
-
Symmetrical Network:
-
Image impedances equal: $$\displaystyle Z_{i1} = Z_{i2} $$
-
Parameters: $$\displaystyle A = D $$, $$\displaystyle AD - BC = 1 $$
-
-
Asymmetrical Network:
- Image impedances differ: $$\displaystyle Z_{i1} \neq Z_{i2} $$
-
Image Impedance: Impedance that, when terminating a port, makes the network appear infinite at the other port.
| Network Type | Image Impedance Formula |
|---|---|
| L-network (series $$\displaystyle Z_s $$, shunt $$\displaystyle Z_p $$) | $$\displaystyle Z_{i1} = \dfrac{Z_s + \sqrt{Z_s^2 + 4 Z_s Z_p}}{2} $$<br>$$\displaystyle Z_{i2} = \dfrac{Z_p + \sqrt{Z_s^2 + 4 Z_s Z_p}}{2} $$ |
| Symmetrical T-network (each series $Z$, shunt $Z'$) | $$\displaystyle Z_i = \sqrt{Z^2 + 2 Z Z'} $$ |
| Symmetrical π-network (series $$\displaystyle Z_s $$, each shunt $$\displaystyle Z_p $$) | $$\displaystyle Z_i = \sqrt{\dfrac{Z_s^2}{4} + Z_s Z_p} $$ |
| Asymmetrical T-network (series $$\displaystyle Z_1 $$, $$\displaystyle Z_2 $$, shunt $$\displaystyle Z_3 $$) | Solve simultaneously:<br>$$\displaystyle Z_{i1} = Z_1 + \dfrac{Z_3 (Z_2 + Z_{i2})}{Z_3 + Z_2 + Z_{i2}} $$<br>$$\displaystyle Z_{i2} = Z_2 + \dfrac{Z_3 (Z_1 + Z_{i1})}{Z_3 + Z_1 + Z_{i1}} $$ |
[!TIP] For asymmetrical T-networks, use the two equations above; substitute one into the other to solve.
Lattice and Bridged-T Networks
-
Lattice Network:
-
Structure: Crossed series impedances $$\displaystyle Z_a $$, $$\displaystyle Z_b $$.
-
Symmetrical, $$\displaystyle Z_i = \sqrt{Z_a Z_b} $$.
-
Applications: Equalizers, filters.
-
-
Bridged-T Network:
-
T-network with a bridge resistor across series arms.
-
Used for attenuation and impedance matching.
-
T and π Equivalent Circuits
- T → π (T: $$\displaystyle Z_1 $$, $$\displaystyle Z_2 $$ series, $$\displaystyle Z_3 $$ shunt):
$$ \begin{aligned} Z_{\pi1} &= \frac{Z_1 Z_3}{Z_1+Z_2+Z_3} \\ Z_{\pi2} &= \frac{Z_2 Z_3}{Z_1+Z_2+Z_3} \\ Z_{\pi3} &= \frac{Z_1 Z_2 + Z_2 Z_3 + Z_3 Z_1}{Z_3} \end{aligned} $$
- π → T (π: $$\displaystyle Z_{\pi1} $$, $$\displaystyle Z_{\pi2} $$ shunt, $$\displaystyle Z_{\pi3} $$ series):
$$ \begin{aligned} Z_1 &= \frac{Z_{\pi1} Z_{\pi3}}{Z_{\pi1} + Z_{\pi2} + Z_{\pi3}} \\ Z_2 &= \frac{Z_{\pi2} Z_{\pi3}}{Z_{\pi1} + Z_{\pi2} + Z_{\pi3}} \\ Z_3 &= \frac{Z_{\pi1} Z_{\pi2}}{Z_{\pi3}} \end{aligned} $$
Characteristic Impedance of Symmetric Networks
-
T-section: $$\displaystyle Z_0 = \sqrt{Z^2 + 2 Z Z'} $$
($Z$ = each series arm, $Z'$ = shunt)
-
π-section: $$\displaystyle Z_0 = \sqrt{\dfrac{Z_s^2}{4} + Z_s Z_p} $$
($$\displaystyle Z_s $$ = series impedance, $$\displaystyle Z_p $$ = each shunt impedance)
II. Filter Design & Approximation Techniques
Constant-K Filters
-
Design: Half-section with $$\displaystyle Z Z' = K^2 $$, cut-off $$\displaystyle \omega_c = 1/\sqrt{LC} $$.
-
Limitations: Stopband attenuation only 6 dB/octave; poor roll-off.
m-Derived Filters
-
T-section design (low-pass example):
-
Series: $$\displaystyle Z_s = m Z $$
-
Shunt: $$\displaystyle Z_p = \dfrac{Z'}{1-m^2} $$
-
Infinite attenuation frequency: $$\displaystyle \omega_{in} = \dfrac{\omega_c}{\sqrt{1-m^2}} $$ ($$\displaystyle m < 1 $$)
-
-
π-section: Similar transformation.
-
Role: Sharper cutoff than constant-K; used in composite filters.
Chebyshev Approximation
-
Passband: Equiripple (equal amplitude variations).
-
Stopband: Monotonic.
-
Roll-off: Steeper than Butterworth.
-
Design: Use Chebyshev polynomials; apply frequency transformation for HPF/BPF/BSF.
Composite Filters
-
Combine constant-K and m-derived sections.
-
Use different $m$ values to balance stopband attenuation and cutoff sharpness.
Frequency Transformation (from LPF prototype)
-
LPF → HPF: $$\displaystyle s \rightarrow \dfrac{\omega_c}{s} $$
-
LPF → BPF: $$\displaystyle s \rightarrow \dfrac{s^2 + \omega_0^2}{B s} $$
($$\displaystyle \omega_0 $$ = center freq, $B$ = bandwidth)
-
LPF → BSF: $$\displaystyle s \rightarrow \dfrac{B s}{s^2 + \omega_0^2} $$
Reactance Curves
-
LPF: $$\displaystyle X_L = \omega L \uparrow $$, $$\displaystyle X_C = 1/(\omega C) \downarrow $$; intersect at $$\displaystyle \omega_c $$.
-
HPF: $$\displaystyle X_L \downarrow $$, $$\displaystyle X_C \uparrow $$; intersect at $$\displaystyle \omega_c $$.
[!TIP] Reactance curves indicate passband (where $$\displaystyle |X_L| \neq |X_C| $$) and stopband (where $$\displaystyle |X_L| = |X_C| $$).
III. Network Synthesis Methods
Fundamentals
-
Positive Real (PR) Function: $\text{Re}[F(s)] \geq 0$ for $$\displaystyle \text{Re}(s) > 0 $$.
-
Conditions:
-
No poles in $$\displaystyle \text{Re}(s) > 0 $$.
-
Poles on $j\omega$ axis are simple with positive residues.
-
At $$\displaystyle s=\infty $$, if pole exists, residue positive.
-
-
-
Hurwitz Polynomial: All roots in $\text{Re}(s) \leq 0$.
Check: coefficients positive, all Hurwitz determinants positive.
Foster Synthesis
- Foster I (for $Z(s)$ PR, no pole at $\infty$):
$$ Z(s) = R_0 + sL_\infty + \sum_{k=1}^m \frac{R_k}{s + a_k} + \sum_{k=1}^n \frac{2 R_k' s}{s^2 + \omega_k^2} $$
Realization: Series connection of $R$, $L$, and parallel $RC$ circuits.
- Foster II (for $Y(s)$ PR):
$$ Y(s) = G_0 + sC_\infty + \sum \frac{G_k}{s + a_k} + \sum \frac{2 G_k' s}{s^2 + \omega_k^2} $$
Realization: Parallel connection of $G$, $C$, and series $RL$ circuits.
Cauer Synthesis
-
Cauer I (ladder):
Start with $Z(s)$ → remove pole at $\infty$ (subtract $$\displaystyle sL_\infty $$) → take reciprocal → remove pole at $\infty$ (subtract $$\displaystyle G_0 $$) → repeat.
Yields: series $L$, shunt $C$.
-
Cauer II:
Start with $Y(s)$ → remove pole at $\infty$ (subtract $$\displaystyle sC_\infty $$) → reciprocal → etc.
Yields: shunt $C$, series $L$.
Brune's Method
-
For functions with complex poles.
-
Remove a pair of complex conjugate poles using a Brune section (parallel $RLC$).
-
Brune coefficient $k$ determines component values.
-
Continue with Foster/Cauer for remaining PR function.
Bott-Duffin Method
-
Synthesizes minimum positive real functions.
-
Principle: Remove one pole at a time using a specific network configuration.
Realization of Network Functions
-
Given $F(s)$, ensure it is PR/Hurwitz.
-
Expand in partial fractions.
-
Implement using Foster or Cauer forms.
- Example: $$\displaystyle F(s) = \dfrac{(s+2)(s+6)}{(s+3)(s+9)} $$ → partial fractions → realize.
IV. Attenuators
Types and Design
-
Symmetrical: Input impedance = output impedance = $$\displaystyle Z_0 $$.
-
Asymmetrical: Input impedance ≠ output impedance.
-
T-Attenuator (symmetrical):
$$ \begin{aligned} R_1 &= R_2 = Z_0 \frac{k-1}{k+1} \\ R_3 &= \frac{2 Z_0 k}{k^2-1} \end{aligned} $$
where $$\displaystyle k = 10^{A/20} $$, $A$ = attenuation in dB.
- π-Attenuator (symmetrical):
$$ \begin{aligned} R_1 &= R_2 = Z_0 \frac{k+1}{k-1} \\ R_3 &= \frac{2 Z_0 k}{k^2-1} \end{aligned} $$
-
Bridged-T Attenuator:
T-attenuator with a bridge resistor across the series arms.
Used for impedance matching networks.
[!TIP] For asymmetrical attenuators, design equations depend on given input/output impedances and loss.
V. Advanced Transmission Line Topics
Microstrip Lines
-
Structure: Dielectric substrate ($$\displaystyle \varepsilon_r $$), ground plane, conducting strip (width $W$, thickness $t$).
-
Effective dielectric constant ($W/h \geq 1$):
$$ \varepsilon_{\text{eff}} = \frac{\varepsilon_r+1}{2} + \frac{\varepsilon_r-1}{2} \frac{1}{\sqrt{1+12h/W}} $$
- Characteristic impedance ($W/h \geq 1$):
$$ Z_0 = \frac{120\pi}{\sqrt{\varepsilon_{\text{eff}}} \left[ \frac{W}{h} + 1.393 + 0.667 \ln\left(\frac{W}{h}+1.444\right) \right]} $$
-
Losses:
-
Conductor loss: Increases with $\sqrt{f}$ due to skin effect.
-
Dielectric loss: $G \propto f \tan\delta$.
-
High-Frequency Line Parameters
-
Skin effect: $R \propto \sqrt{f}$.
-
Dielectric losses: $G \propto f \tan\delta$.
-
At high $f$, both $R$ and $G$ become significant.
Equalizers
-
Compensate for frequency-dependent attenuation on lines.
-
Full Series Equalizer: Series inductor $L$ and shunt capacitor $C$.
Design for flat attenuation over band $$\displaystyle [f_1, f_2] $$:
$$ \begin{aligned} L &= \frac{Z_0}{\pi (f_2 - f_1)} \\ C &= \frac{\pi (f_2 - f_1)}{Z_0 \cdot (2\pi f_c)^2} \quad \text{(where $$\displaystyle f_c $$ is center frequency)} \end{aligned} $$
Exam Focus & Quick Reference
| Topic | Key Formula / Concept | Past Paper Frequency |
|---|---|---|
| Image Impedance (T-network) | $$\displaystyle Z_i = \sqrt{Z^2 + 2 Z Z'} $$ | High |
| m-Derived Filter Design | $$\displaystyle Z_s = mZ $$, $$\displaystyle Z_p = Z'/(1-m^2) $$, $$\displaystyle \omega_{in} = \omega_c/\sqrt{1-m^2} $$ | High |
| Foster I Synthesis | $$\displaystyle Z(s) = R_0 + sL_\infty + \sum \frac{R_k}{s+a_k} + \sum \frac{2R_k' s}{s^2+\omega_k^2} $$ | High |
| Cauer I Synthesis | Ladder: series $L$, shunt $C$ | High |
| Attenuator Design | T: $$\displaystyle R_1=Z_0\frac{k-1}{k+1} $$, $$\displaystyle R_3=\frac{2Z_0 k}{k^2-1} $$ | High |
| Microstrip $$\displaystyle Z_0 $$ | $$\displaystyle Z_0 = \dfrac{120\pi}{\sqrt{\varepsilon_{\text{eff}}} \left[ \frac{W}{h} + 1.393 + 0.667 \ln\left(\frac{W}{h}+1.444\right) ]} $$ | Medium |
| Chebyshev vs Butterworth | Equiripple passband vs maximally flat | Medium |
| Frequency Transformation | LPF→HPF: $$\displaystyle s \rightarrow \omega_c/s $$ | Medium |
[!TIP] Always check realizability: PR function for synthesis, Hurwitz for denominator of stable systems. For filter design, remember $$\displaystyle m < 1 $$ for m-derived LPF to have $$\displaystyle \omega_{in} > \omega_c $$.