Transmission Lines and Waveguides
CSIR NET guided waves: telegrapher's equations, impedance, reflection, VSWR, quarter-wave matching, TE/TM modes, cut-off and cavities.
- 19 min
On this page · 11 sections
Prerequisites
- EM Waves, Reflection, Refraction and Polarization: plane waves, wave impedance \( \eta = \sqrt{\mu/\varepsilon} \), reflection at normal incidence.
- Maxwell’s Equations, Potentials and Gauge Invariance: boundary conditions at a perfect conductor (\( E_\parallel = 0 \), \( B_\perp = 0 \)).
- Linear ODEs and Special Functions: separation of variables and Bessel-function zeros for circular guides.
- Complex impedance and phasors from AC circuit theory, as in Semiconductor Devices and other electronics topics.
The Hook: Why Your Cable Has a Number Printed on It
Part 1: Transmission Lines
The distributed model
A line is described per unit length by series resistance \( R \) and inductance \( L \), shunt conductance \( G \) and capacitance \( C \). Kirchhoff’s laws on a short element \( dz \) give the telegrapher’s equations:
$$\frac{\partial V}{\partial z} = -RI - L\frac{\partial I}{\partial t}, \qquad \frac{\partial I}{\partial z} = -GV - C\frac{\partial V}{\partial t}$$With phasors \( V(z)e^{j\omega t} \) (the engineering convention used on this page; physics texts write \( e^{-i\omega t} \), which replaces \( j \) by \( -i \)):
$$\frac{d^2V}{dz^2} = \gamma^2V, \qquad \boxed{\gamma = \alpha + j\beta = \sqrt{(R + j\omega L)(G + j\omega C)}}, \qquad \boxed{Z_0 = \sqrt{\frac{R + j\omega L}{G + j\omega C}}}$$The general solution is \( V(z) = V^+e^{-\gamma z} + V^-e^{\gamma z} \), with currents \( I = (V^+e^{-\gamma z} - V^-e^{\gamma z})/Z_0 \). \( Z_0 \) is the voltage-to-current ratio of a single travelling wave.
Special cases
| Line | Condition | \( Z_0 \) | Propagation |
|---|---|---|---|
| Lossless | \( R = G = 0 \) | \( \sqrt{L/C} \), real | \( \alpha = 0 \), \( \beta = \omega\sqrt{LC} \), \( v = 1/\sqrt{LC} \) |
| Low loss | \( R \ll \omega L \), \( G \ll \omega C \) | \( \approx\sqrt{L/C} \) | \( \alpha \approx \dfrac{R}{2Z_0} + \dfrac{GZ_0}{2} \) |
| Distortionless (Heaviside) | \( R/L = G/C \) | \( \sqrt{L/C} \), real | \( \alpha = \sqrt{RG} \), \( v = 1/\sqrt{LC} \), both frequency-independent |
On a distortionless line every frequency travels at the same speed and decays at the same rate, so a pulse shrinks without changing shape. Real telephone cables have \( G/C \ll R/L \); historically, inductance was added to approach the condition.
Standard geometries
For any TEM line in a uniform dielectric, \( LC = \mu\varepsilon \), so \( v = 1/\sqrt{\mu\varepsilon} = c/\sqrt{\varepsilon_r} \) (non-magnetic).
| Line | \( L \) per length | \( C \) per length | \( Z_0 \) (non-magnetic) |
|---|---|---|---|
| Coaxial (radii \( a < b \)) | \( \dfrac{\mu}{2\pi}\ln\dfrac{b}{a} \) | \( \dfrac{2\pi\varepsilon}{\ln(b/a)} \) | \( \dfrac{60}{\sqrt{\varepsilon_r}}\ln\dfrac{b}{a} \) Ω |
| Two-wire (radius \( a \), centre spacing \( d \)) | \( \dfrac{\mu}{\pi}\cosh^{-1}\dfrac{d}{2a} \) | \( \dfrac{\pi\varepsilon}{\cosh^{-1}(d/2a)} \) | \( \dfrac{120}{\sqrt{\varepsilon_r}}\cosh^{-1}\dfrac{d}{2a} \approx \dfrac{120}{\sqrt{\varepsilon_r}}\ln\dfrac{d}{a} \) Ω (\( d \gg a \)) |
| Parallel plate (width \( w \gg \) gap \( h \)) | \( \mu h/w \) | \( \varepsilon w/h \) | \( \eta\,h/w \) |
The 60 and 120 come from \( \eta_0/2\pi \approx 60 \) Ω and \( \eta_0/\pi \approx 120 \) Ω, with \( \eta_0 = \sqrt{\mu_0/\varepsilon_0} \approx 376.7 \) Ω.
Reflection at a load
At the load (\( z = 0 \)), \( V/I = Z_L \). The reflection coefficient is
$$\boxed{\Gamma_L = \frac{V^-}{V^+} = \frac{Z_L - Z_0}{Z_L + Z_0}}$$| Load | \( \Gamma_L \) | What comes back |
|---|---|---|
| Matched, \( Z_L = Z_0 \) | 0 | nothing |
| Short, \( Z_L = 0 \) | \( -1 \) | inverted pulse (voltage node at load) |
| Open, \( Z_L = \infty \) | \( +1 \) | upright pulse (voltage antinode at load) |
| Pure reactance \( jX \) | \( \vert \Gamma\vert = 1 \), phase shift | everything, delayed in phase |
The voltage transmitted into the load is \( (1 + \Gamma_L)V^+ \). Power reflected is \( |\Gamma|^2 \) of the incident power; power delivered is \( 1 - |\Gamma|^2 \) (lossless line, real \( Z_0 \)).
Standing waves and VSWR
Incident and reflected waves form a partial standing wave with
$$|V|_{\max} = |V^+|(1 + |\Gamma|), \qquad |V|_{\min} = |V^+|(1 - |\Gamma|), \qquad \boxed{S = \text{VSWR} = \frac{1 + |\Gamma|}{1 - |\Gamma|}}$$Adjacent maxima (or minima) are \( \lambda/2 \) apart; a maximum and the next minimum are \( \lambda/4 \) apart. Inverting, \( |\Gamma| = (S - 1)/(S + 1) \). For a real load, \( S = Z_L/Z_0 \) if \( Z_L > Z_0 \) and \( Z_0/Z_L \) otherwise.
Input impedance
Looking into a lossless line of length \( l \) terminated in \( Z_L \):
$$\boxed{Z_{\text{in}} = Z_0\,\frac{Z_L + jZ_0\tan\beta l}{Z_0 + jZ_L\tan\beta l}}$$| Case | \( Z_{\text{in}} \) | Use |
|---|---|---|
| Short-circuited stub | \( jZ_0\tan\beta l \) | inductive for \( l < \lambda/4 \): a tunable reactance |
| Open-circuited stub | \( -jZ_0\cot\beta l \) | capacitive for \( l < \lambda/4 \) |
| \( l = \lambda/2 \) | \( Z_L \) | half-wave line repeats the load |
| \( l = \lambda/4 \) | \( Z_0^2/Z_L \) | impedance inverter: short \( \to \) open |
| Matched load | \( Z_0 \) for any \( l \) |
Quarter-wave transformer. To match a real load \( R_L \) to a line \( Z_0 \), insert a \( \lambda/4 \) section of impedance \( Z_T = \sqrt{Z_0R_L} \). It works exactly at one frequency (and its odd multiples).
Pulses and step response
A source of open-circuit voltage \( V_s \) and internal resistance \( R_s \) launches \( V_s Z_0/(R_s + Z_0) \) onto the line: at the instant of launch the line looks like a resistor \( Z_0 \). The wave then bounces between the ends with reflection coefficients \( \Gamma_L \) and \( \Gamma_s = (R_s - Z_0)/(R_s + Z_0) \); the voltage at any point settles to the DC value after many round trips. Time-domain reflectometry uses the delay of the first reflection to locate faults: distance \( = v\,t/2 \).
Part 2: Waveguides
Why a hollow pipe has no TEM mode
A TEM wave (\( E_z = B_z = 0 \)) inside a guide would need transverse fields satisfying the two-dimensional Laplace equation with the conductor as a single equipotential boundary, which forces the field to vanish. A TEM wave needs two separate conductors (coax, two-wire line). A hollow guide carries only
- TE modes: \( E_z = 0 \), \( B_z \neq 0 \);
- TM modes: \( B_z = 0 \), \( E_z \neq 0 \).
Rectangular guide (\( a \times b \), \( a > b \), filled with \( \mu, \varepsilon \))
Write fields as \( f(x, y)e^{j(\omega t - k_zz)} \). The longitudinal component obeys
$$\left[\partial_x^2 + \partial_y^2 + k_c^2\right]f = 0, \qquad k_c^2 = \omega^2\mu\varepsilon - k_z^2$$Boundary conditions give \( B_z = B_0\cos(m\pi x/a)\cos(n\pi y/b) \) for TE and \( E_z = E_0\sin(m\pi x/a)\sin(n\pi y/b) \) for TM, with
$$k_c^2 = \left(\frac{m\pi}{a}\right)^2 + \left(\frac{n\pi}{b}\right)^2, \qquad \boxed{f_{c,mn} = \frac{v}{2}\sqrt{\left(\frac{m}{a}\right)^2 + \left(\frac{n}{b}\right)^2}}, \quad v = \frac{1}{\sqrt{\mu\varepsilon}}$$| Mode family | Allowed \( (m, n) \) | Lowest |
|---|---|---|
| TE\( _{mn} \) | \( m, n \ge 0 \), not both zero | TE\( _{10} \): \( f_c = v/2a \), \( \lambda_c = 2a \) |
| TM\( _{mn} \) | \( m, n \ge 1 \) (a sine with zero index vanishes) | TM\( _{11} \) |
TE\( _{10} \) is the dominant mode for \( a > b \). TE\( _{mn} \) and TM\( _{mn} \) with the same indices (both \( \ge 1 \)) share a cut-off: they are degenerate.
Dispersion: the plasma hyperbola again
$$\boxed{\omega^2 = \omega_c^2 + v^2k_z^2}$$identical in form to the plasma dispersion relation. With \( F = \sqrt{1 - (f_c/f)^2} \):
| Quantity | Formula |
|---|---|
| Guide wavelength | \( \lambda_g = \lambda/F \) (longer than in free space) |
| Phase velocity | \( v_p = v/F > v \) |
| Group velocity | \( v_g = vF < v \); \( v_pv_g = v^2 \) |
| TE wave impedance | \( Z_{\text{TE}} = \eta/F \) (above \( \eta \)) |
| TM wave impedance | \( Z_{\text{TM}} = \eta F \) (below \( \eta \)) |
| Below cut-off | \( k_z = -j\alpha \), \( \alpha = \sqrt{\omega_c^2 - \omega^2}/v \): evanescent, no power flow |
Useful identity: \( \dfrac{1}{\lambda^2} = \dfrac{1}{\lambda_g^2} + \dfrac{1}{\lambda_c^2} \).
The bouncing-plane-wave picture of TE\( _{10} \)
\( E_y = E_0\sin(\pi x/a)\cos(\omega t - k_zz) \), with \( H_x \) and \( H_z \) completing the mode. Write \( \sin(\pi x/a) \) as the difference of two exponentials: the mode is two plane waves zig-zagging between the side walls at angle \( \theta \) to the axis, with \( \cos\theta = F \). Energy moves along the guide at \( v\cos\theta = v_g \); the crest pattern on the wall sweeps along at \( v/\cos\theta = v_p \). At cut-off \( \theta = 90^\circ \): the waves bounce straight across and nothing advances.
graph TD
A["Frequency f in guide a × b"] --> B{"f vs f_c,mn
for each mode"}
B -->|"f > f_c"| C["Mode propagates
λ_g = λ/√(1 − f_c²/f²)"]
B -->|"f < f_c"| D["Mode evanescent
decays as e^(−αz)"]
C --> E["Single-mode band:
f_c,10 < f < next cut-off"]
C --> F["v_g = v√(1 − f_c²/f²)
v_p v_g = v²"]Circular guide (radius \( a \))
The transverse dependence is a Bessel function \( J_m(k_c\rho) \). For TM modes \( E_z \) must vanish on the wall, so \( k_ca = x_{mn} \) (zeros of \( J_m \)); for TE modes \( \partial B_z/\partial\rho = 0 \), so \( k_ca = x'_{mn} \) (zeros of \( J_m' \)).
| Mode | Root | \( \lambda_c \) |
|---|---|---|
| TE\( _{11} \) (dominant) | \( x'_{11} = 1.841 \) | \( 2\pi a/1.841 = 3.41a \) |
| TM\( _{01} \) | \( x_{01} = 2.405 \) | \( 2.61a \) |
Coaxial line as a waveguide
Coax carries a TEM mode with no cut-off at all, so it works down to DC. Higher TE and TM modes appear once the circumference is comparable to the wavelength; the first (TE\( _{11} \)) has \( \lambda_c \approx \pi(a + b) \), a commonly used approximation. Keeping below that frequency keeps the coax single-mode.
Cavity resonators
Close a rectangular guide with conducting walls at \( z = 0 \) and \( z = d \). Then \( k_z = p\pi/d \) and the allowed frequencies are discrete:
$$\boxed{f_{mnp} = \frac{v}{2}\sqrt{\left(\frac{m}{a}\right)^2 + \left(\frac{n}{b}\right)^2 + \left(\frac{p}{d}\right)^2}}$$For TE modes \( p \ge 1 \) (the transverse \( E \) must vanish at the end walls); for TM modes \( p \ge 0 \) is allowed. A 10 cm air-filled cube has lowest frequency \( f_{101} = (c/2)\sqrt{2}/0.1 = 2.12 \) GHz. Wall losses give a finite quality factor \( Q = \omega\times\text{(energy stored)}/\text{(power lost)} \); a high-\( Q \) cavity resonates over a bandwidth \( f/Q \). Counting cavity modes below a frequency is the classic route to the Rayleigh–Jeans density of states.
Memory Tricks
Don’t Confuse
| This | vs | That |
|---|---|---|
| \( Z_0 \): ratio for one travelling wave | \( Z_{\text{in}} \): ratio of total \( V \) and \( I \) at a point | |
| Reflection coefficient \( \Gamma \) (amplitude) | Reflected power fraction \( \vert \Gamma\vert ^2 \) | |
| Coax: TEM, no cut-off | Hollow guide: TE/TM only, cut-off \( f_c \) | |
| Phase velocity \( v_p > c \) in a guide | Group velocity \( v_g < c \), carries energy | |
| Guide wavelength \( \lambda_g > \lambda_0 \) | Cut-off wavelength \( \lambda_c = 2a \) (TE\( _{10} \)) | |
| Distortionless: \( R/L = G/C \) | Lossless: \( R = G = 0 \) |
Common Mistakes
Practice
All questions below are original practice questions written for this page. They are not from any CSIR NET paper. Take \( c = 3.0\times10^8 \) m/s and \( \eta_0 = 377 \) Ω.
Level 1: Foundation
- Write \( Z_0 \) and the phase velocity of a lossless line.
- What is the lowest TM mode of a rectangular guide?
- What is the input impedance of a short-circuited \( \lambda/4 \) lossless line?
Answers
- \( Z_0 = \sqrt{L/C} \), \( v = 1/\sqrt{LC} \).
- TM\( _{11} \).
- Infinite (an open circuit): \( Z_0^2/0 \).
Level 2: Part B Style
Practice Q1. A coaxial line has outer-to-inner radius ratio 3.5 and is filled with a dielectric of \( \varepsilon_r = 2.25 \). Its characteristic impedance is about:
(a) 33 Ω (b) 50 Ω (c) 75 Ω (d) 113 Ω
Answer: (b)
\( Z_0 = (60/\sqrt{2.25})\ln3.5 = 40\times1.253 = 50.1 \) Ω. Option (c) forgets \( \sqrt{\varepsilon_r} \); option (d) multiplies by \( \sqrt{\varepsilon_r} \) instead of dividing.
Practice Q2. A 100 Ω resistive load terminates a lossless 50 Ω line. The VSWR and the fraction of incident power reflected are:
(a) 2 and 1/9 (b) 2 and 1/3 (c) 3 and 1/4 (d) 1.5 and 1/9
Answer: (a)
\( \Gamma = (100 - 50)/(100 + 50) = 1/3 \). \( S = (1 + 1/3)/(1 - 1/3) = 2 \). Reflected power \( |\Gamma|^2 = 1/9 \approx 11\% \). Option (b) gives the amplitude ratio as the power ratio.
Practice Q3. A quarter-wave section is to match a 200 Ω resistive load to a 50 Ω line. Its characteristic impedance should be:
(a) 125 Ω (b) 150 Ω (c) 100 Ω (d) 250 Ω
Answer: (c)
\( Z_T = \sqrt{50\times200} = 100 \) Ω. Option (a) is the arithmetic mean.
Practice Q4. An air-filled rectangular guide has \( a = 2.286 \) cm and \( b = 1.016 \) cm. At 10 GHz the propagating modes are:
(a) TE\( _{10} \) only (b) TE\( _{10} \) and TE\( _{01} \) (c) TE\( _{10} \), TE\( _{20} \) and TE\( _{01} \) (d) TE\( _{10} \) and TM\( _{10} \)
Answer: (a)
Cut-offs: TE\( _{10} \) 6.56 GHz, TE\( _{20} \) 13.1 GHz, TE\( _{01} \) 14.8 GHz. Only TE\( _{10} \) is below 10 GHz. TM\( _{10} \) (option d) does not exist.
Level 3: Part C Style
Practice Q5. For the guide of Q4 operating at 10 GHz in the TE\( _{10} \) mode, consider:
- The guide wavelength is about 3.97 cm.
- The group velocity is about \( 2.26\times10^8 \) m/s.
- The wave impedance is about 285 Ω.
- Filling the guide with a dielectric of \( \varepsilon_r = 2.25 \) lowers the TE\( _{10} \) cut-off to about 4.37 GHz.
Which are correct?
(a) 1, 2 and 4 only (b) 1 and 2 only (c) 1, 3 and 4 only (d) 2, 3 and 4 only
Answer: (a) 1, 2 and 4 only
\( f_c = 6.56 \) GHz, \( f_c/f = 0.656 \), \( F = \sqrt{1 - 0.430} = 0.755 \).
Statement 1: \( \lambda_g = 3.00/0.755 = 3.97 \) cm. True.
Statement 2: \( v_g = cF = 2.26\times10^8 \) m/s. True.
Statement 3: \( Z_{\text{TE}} = \eta_0/F = 499 \) Ω; 285 Ω is \( \eta_0F \), the TM formula. False.
Statement 4: \( f_c = c/(2a\sqrt{\varepsilon_r}) = 6.56/1.5 = 4.37 \) GHz. True.
Practice Q6. A lossless 50 Ω line is one-eighth of a wavelength long. Consider its input impedance for different terminations:
- Short-circuited: \( Z_{\text{in}} = +j50 \) Ω (inductive).
- Open-circuited: \( Z_{\text{in}} = -j50 \) Ω (capacitive).
- Terminated in 100 Ω: \( Z_{\text{in}} = (40 - j30) \) Ω.
- Terminated in 100 Ω: \( |\Gamma| \) at the input is smaller than at the load, because the line has transformed the impedance.
Which are correct?
(a) 1, 2 and 3 only (b) 1 and 2 only (c) 3 and 4 only (d) 1, 2, 3 and 4
Answer: (a) 1, 2 and 3 only
\( \beta l = (2\pi/\lambda)(\lambda/8) = \pi/4 \), \( \tan\beta l = 1 \).
Statement 1: \( jZ_0\tan\beta l = j50 \) Ω. True. Statement 2: \( -jZ_0\cot\beta l = -j50 \) Ω. True.
Statement 3: \( Z_{\text{in}} = 50\dfrac{100 + j50}{50 + j100} = 50\dfrac{(100 + j50)(50 - j100)}{50^2 + 100^2} = 50\dfrac{10000 - j7500}{12500} = 40 - j30 \) Ω. True.
Statement 4: on a lossless line \( |\Gamma| \) is the same everywhere; only its phase rotates. Check: \( \left|\dfrac{40 - j30 - 50}{40 - j30 + 50}\right| = \dfrac{\sqrt{1000}}{\sqrt{9000}} = \dfrac13 \). False.
Practice Q7. A line has \( L = 250 \) nH/m, \( C = 100 \) pF/m and series resistance \( R = 0.05 \) Ω/m. Its shunt conductance \( G \) is chosen to make it distortionless. Consider:
- \( Z_0 = 50 \) Ω and the signal speed is \( 2.0\times10^8 \) m/s.
- \( G = 2.0\times10^{-5} \) S/m.
- The attenuation constant is \( 1.0\times10^{-3} \) Np/m, so after 1 km the power has fallen to about 13.5% (about 8.7 dB loss).
- Pulses of different widths travel at different speeds on this line.
Which are correct?
(a) 1 and 2 only (b) 1, 2 and 3 only (c) 2, 3 and 4 only (d) 1 and 4 only
Answer: (b) 1, 2 and 3 only
Statement 1: \( Z_0 = \sqrt{250\times10^{-9}/100\times10^{-12}} = \sqrt{2500} = 50 \) Ω; \( v = 1/\sqrt{2.5\times10^{-17}} = 2.0\times10^8 \) m/s. True.
Statement 2: \( G = RC/L = 0.05\times100\times10^{-12}/250\times10^{-9} = 2.0\times10^{-5} \) S/m. True.
Statement 3: \( \alpha = \sqrt{RG} = \sqrt{10^{-6}} = 10^{-3} \) Np/m. Over 1000 m the voltage falls by \( e^{-1} \) and the power by \( e^{-2} = 0.135 \); \( 10\log_{10}(e^2) = 8.69 \) dB. True.
Statement 4: on a distortionless line \( v \) and \( \alpha \) are independent of frequency, so all Fourier components, and hence all pulse shapes, travel together. False.
Quick Revision Table
| Quantity | Formula |
|---|---|
| Propagation constant | \( \gamma = \sqrt{(R + j\omega L)(G + j\omega C)} \) |
| Characteristic impedance | \( Z_0 = \sqrt{(R + j\omega L)/(G + j\omega C)} \); lossless \( \sqrt{L/C} \) |
| Distortionless | \( R/L = G/C \); \( \alpha = \sqrt{RG} \) |
| Coax / two-wire | \( (60/\sqrt{\varepsilon_r})\ln(b/a) \); \( (120/\sqrt{\varepsilon_r})\cosh^{-1}(d/2a) \) |
| Reflection | \( \Gamma = (Z_L - Z_0)/(Z_L + Z_0) \) |
| VSWR | \( S = (1 + \vert \Gamma\vert )/(1 - \vert \Gamma\vert ) \) |
| Input impedance | \( Z_0(Z_L + jZ_0\tan\beta l)/(Z_0 + jZ_L\tan\beta l) \) |
| Quarter-wave | \( Z_{\text{in}} = Z_0^2/Z_L \); matching \( Z_T = \sqrt{Z_0R_L} \) |
| Rectangular cut-off | \( f_{c,mn} = (v/2)\sqrt{(m/a)^2 + (n/b)^2} \); TE\( _{10} \): \( v/2a \) |
| Guide relations | \( \lambda_g = \lambda/F \), \( v_g = vF \), \( v_p = v/F \), \( F = \sqrt{1 - f_c^2/f^2} \) |
| Wave impedance | \( Z_{\text{TE}} = \eta/F \), \( Z_{\text{TM}} = \eta F \) |
| Circular | TE\( _{11} \) \( \lambda_c = 3.41a \); TM\( _{01} \) \( \lambda_c = 2.61a \) |
| Cavity | \( f_{mnp} = (v/2)\sqrt{(m/a)^2 + (n/b)^2 + (p/d)^2} \) |
Key Takeaways
What’s Next
- Radiation from Moving Charges and Dipoles: what happens when a line ends in an antenna and the wave leaves the guide.
- Dispersion Relations in Plasma: the same cut-off physics produced by free electrons.
- Electromagnetic Theory (Advanced): Quick Revision.
- The boundary-value technique behind mode shapes: Electrostatics, Gauss’s Law, Laplace and Poisson.