Transmission Lines and Impedance Matching

Introduction and Context

At low frequencies, interconnections are often treated as ideal wires whose voltage is the same at every point. At microwave frequencies this approximation fails. A conductor may be several wavelengths long, and voltage and current then vary with position because electromagnetic energy propagates as a wave. Such an interconnection is a transmission line.

Transmission-line theory provides the foundation for microwave circuits, antennas, radar systems, satellite links, filters, oscillators, and measurement systems. It explains propagation, reflection, standing waves, power transfer, losses, and impedance transformation. Impedance matching is the systematic process of making a source, line, and load compatible so that reflections are reduced and power transfer is controlled.

A useful engineering criterion is the ratio of physical length ℓ to wavelength λ. Lumped-circuit treatment is generally acceptable when ℓ is much smaller than λ, commonly ℓ ≤ λ/10. When this condition is not satisfied, distributed-parameter analysis is required.

Learning Objectives

After completing this chapter, the reader should be able to:

1. Explain the physical basis and distributed nature of microwave transmission lines.
2. Derive and interpret the telegrapher’s equations.
3. Analyze propagation constant, characteristic impedance, reflection coefficient, standing waves, and power flow.
4. Use the Smith chart and analytical equations to transform impedances.
5. Design quarter-wave, single-stub, and double-stub matching networks.
6. Evaluate practical losses, bandwidth, tolerances, and measurement limitations.
7. Select an appropriate matching method for a microwave application.

1. Distributed-Parameter Model

A uniform transmission line is represented by four primary parameters per unit length:

For a differential section of length Δz, the series impedance is (R + jωL)Δz and the shunt admittance is (G + jωC)Δz. Here ω = 2πf is angular frequency, f is frequency, and j = √−1. The model assumes a uniform line, linear materials, sinusoidal steady state, and a cross-section sufficiently small for a one-dimensional description.

Applying Kirchhoff’s laws to an infinitesimal section gives the telegrapher’s equations:

V(z) is voltage and I(z) is current at position z. Differentiating and substituting gives the wave equations:

where the propagation constant is

α, in nepers per metre, is the attenuation constant; β, in radians per metre, is the phase constant. The general voltage and current solutions are

V⁺ and V⁻ are forward- and reverse-travelling voltage amplitudes. The characteristic impedance is

The negative sign in the reverse-current term follows from the opposite direction of reverse-wave power flow.

Lossless and Low-Loss Lines

For a lossless line, R = 0 and G = 0. Therefore,

The phase velocity and wavelength are

For a low-loss line, R ≪ ωL and G ≪ ωC. Approximate expressions are

and

The first term in α represents conductor loss and the second represents dielectric loss. These approximations apply only when the stated inequalities hold.

2. Physical Transmission-Line Structures

Common microwave transmission lines include coaxial cable, two-wire line, parallel-plate line, microstrip, stripline, coplanar waveguide, and rectangular or circular waveguide. Coaxial and two-wire lines support a transverse electromagnetic (TEM) mode over their normal operating range. Stripline is also ideally TEM. Microstrip is technically a quasi-TEM line because fields partly occupy the dielectric and partly occupy air.

For a TEM line, voltage and current have familiar circuit meanings. In waveguides, fields must be described using transverse-electric or transverse-magnetic modes, and the concept of characteristic impedance is mode dependent. The present chapter concentrates primarily on TEM and quasi-TEM transmission lines.

3. Reflection at a Load

Consider a line with characteristic impedance Z₀ terminated by load impedance ZL. At the load, the incident and reflected waves must satisfy the boundary condition V = ZLI. The voltage reflection coefficient is

This expression assumes a single mode and a reference impedance Z₀. If ZL = Z₀, ΓL = 0 and the line is matched. An open circuit produces ΓL = +1, while a short circuit produces ΓL = −1.

At a distance l measured from the load toward the generator, the reflection coefficient on a lossless line is

Its magnitude is constant on a lossless line, while its phase changes with position. On a lossy line,

so the magnitude decreases as the observation point moves toward the generator.

The input impedance of a lossless line of length l is

This equation is periodic with period λ/2. Important special cases are:

• A line of length λ/2 repeats the load impedance: Zᵢₙ = ZL.
• A line of length λ/4 transforms a load according to Zᵢₙ = Z₀²/ZL.
• A short-circuited quarter-wave line appears as an open circuit.
• An open-circuited quarter-wave line appears as a short circuit.

For a lossy line, the general impedance expression is

The lossless expression is obtained by setting γ = jβ.

4. Standing Waves and Power

The total voltage on a lossless line is the sum of incident and reflected waves. Interference produces standing-wave maxima and minima. The voltage standing-wave ratio (VSWR) is

where 0 ≤ |Γ| < 1 for a passive load with finite delivered power. Conversely,

A matched line has VSWR = 1. An ideal open or short has VSWR → ∞.

For a real load, the time-average power delivered to it is

where P_inc is incident power. The reflected power is P_ref = P_inc|ΓL|². This relation assumes the same real reference impedance is used for incident and reflected power normalization.

The local voltage and current magnitudes on a lossless line are

Voltage maxima occur where the reflected and incident voltages are in phase; minima occur where they are out of phase. Adjacent maxima, or adjacent minima, are separated by λ/2. A maximum and the nearest minimum are separated by λ/4.

The normalized load impedance is zL = ZL/Z₀. In terms of normalized impedance,

Normalization makes results independent of the absolute line impedance and is central to Smith-chart analysis.

5. Worked Example: Reflection and VSWR

A 50 Ω lossless line is terminated by ZL = 100 − j50 Ω. Find the reflection coefficient, its magnitude, VSWR, and the fraction of incident power delivered to the load.

First normalize the load:

Then

Thus |ΓL| = √(0.4² + 0.2²) ≈ 0.447. Therefore,

The delivered fraction is

Hence 80% of the incident power is delivered to the load, while 20% is reflected.

6. The Smith Chart

The Smith chart is a graphical representation of the complex reflection coefficient and normalized impedance or admittance. It allows engineers to perform impedance transformation without repeated complex arithmetic.

For normalized impedance z = r + jx, constant-resistance circles and constant-reactance arcs are plotted in the reflection-coefficient plane. Moving toward the generator on a lossless line rotates the reflection coefficient clockwise by 2βl. A distance of λ/2 produces one complete rotation; λ/4 produces a half rotation.

The admittance form is y = g + jb = 1/z. On the Smith chart, converting from impedance to admittance corresponds to a 180° rotation. This is especially useful for shunt-stub design, because admittances in parallel add directly.

A practical Smith-chart procedure is:

1. Normalize ZL by Z₀.
2. Plot the normalized load.
3. Rotate toward generator by the required electrical length.
4. Read the transformed impedance or admittance.
5. Add the matching element graphically or analytically.
6. Convert the result back to dimensional units.

The chart is not limited to lossless lines, but lossy-line analysis requires a more general chart or numerical computation.

7. Impedance-Matching Principles

A matching network transforms a load impedance into the line impedance at the chosen frequency. Matching can maximize power transfer, reduce reflections, protect amplifiers, improve measurement accuracy, or control the voltage distribution.

For a source with Thevenin impedance ZS and load ZL, maximum average power transfer occurs when

where * denotes complex conjugation. This maximum-power condition should not be confused with the condition ZL = Z₀, which matches a load to a transmission line. In a practical system, both source-to-line and line-to-load matching may be required.

Matching is always a compromise among bandwidth, loss, physical size, power handling, tuning sensitivity, and fabrication complexity. A perfect match at one frequency may be narrowband, while a broadband match may require multiple reactive sections or resistive loss.

7.1 Quarter-Wave Transformer

A quarter-wave transformer is a line of characteristic impedance Zt and length λ/4 inserted between a real load RL and a main line of impedance Z₀. Since

the matching condition is

This simple formula applies to real, positive impedances at the design frequency. It does not directly match an arbitrary complex load unless the load is first transformed to a resistance by another reactive element or line section.

For Z₀ = 50 Ω and RL = 100 Ω, Zt = √(50 × 100) ≈ 70.7 Ω. The transformer is exactly matched at its design frequency but becomes mismatched as frequency changes because its electrical length departs from 90°.

7.2 Single-Stub Matching

A stub is a section of transmission line, either open-circuited or short-circuited, connected in series or shunt. A stub supplies a controlled reactance or susceptance. Shunt stubs are commonly designed using admittance.

For a normalized load admittance yL, move along the line until

where d is the distance from the load and b is the normalized susceptance. A stub with normalized susceptance −jb is then connected in parallel, producing ytotal = 1 and therefore a match.

For a lossless short-circuited stub of characteristic impedance Z₀,

and its admittance is

where ls is stub length. For an open-circuited stub,

and

The sign and periodicity must be handled carefully; physically equivalent solutions differ by integer multiples of λ/2.

7.3 Lumped L, T, and π Networks

At lower microwave frequencies, surface-mount inductors and capacitors may provide matching. An L-network uses one series and one shunt reactive element. T and π networks provide additional degrees of freedom and can realize filtering or harmonic suppression.

At higher frequencies, parasitic inductance, parasitic capacitance, finite quality factor, self-resonance, and package dimensions make lumped models inaccurate. Distributed stubs, stepped-impedance sections, and resonator-based networks are then preferable.

7.4 Multisection and Tapered Transformers

A multisection quarter-wave transformer uses several impedance sections to distribute the mismatch over frequency. Properly selected impedances can produce binomial or Chebyshev responses. A continuously tapered line changes impedance gradually and can achieve broadband matching, although its physical length may be substantial.

8. Worked Example: Quarter-Wave Matching

Design a quarter-wave transformer between a 50 Ω line and a 200 Ω resistive load at 2.4 GHz. Assume a microstrip effective relative permittivity εeff = 2.25.

The required transformer impedance is

The guided wavelength is

where c ≈ 3 × 10⁸ m/s. Thus

The physical quarter-wave length is

The design therefore requires an approximately 100 Ω microstrip section 20.8 mm long. In practice, electromagnetic simulation and fabrication calibration are required because εeff varies with frequency and conductor thickness, and discontinuities occur at the junctions.

9. Practical Losses, Dispersion, and Power Handling

Conductor loss arises from finite conductivity and skin effect. The skin depth is

where μ is permeability and σ is conductivity. As frequency increases, current is confined closer to conductor surfaces, increasing effective resistance. Surface roughness can further increase loss.

Dielectric loss is commonly represented by loss tangent tanδ. In a dielectric-dominated line, attenuation generally increases with frequency and is related to dielectric dissipation. Radiation and leakage may be significant in microstrip bends, connectors, transitions, and discontinuities.

Dispersion means that phase velocity depends on frequency. It is particularly important in quasi-TEM structures and waveguides. Dispersion alters electrical length, impedance, pulse shape, and matching bandwidth.

Power handling is limited by conductor heating, dielectric breakdown, corona, multipactor effects in vacuum systems, and local electric-field enhancement. A low VSWR is therefore important not only for efficiency but also for reliability.

10. Measurement and Calibration

Vector network analyzers (VNAs) measure scattering parameters, usually S11, S21, and related quantities. For a one-port measurement with reference impedance Z₀,

because S11 is the input reflection coefficient at the measurement reference plane.

The reference plane may not coincide with the device terminals. Calibration methods such as SOLT (short-open-load-through), TRL (thru-reflect-line), and LRM (line-reflect-match) remove systematic errors from cables, adapters, and fixtures. De-embedding mathematically shifts the measurement plane through a known fixture model.

Good practice includes connector inspection, torque control, appropriate calibration standards, verification with an independent check standard, cable stability, and avoiding excessive power that could compress or damage the device. An apparent impedance mismatch may originate from calibration error, connector repeatability, fixture radiation, or an incorrect dielectric model rather than from the device itself.

11. Illustrative Case: Matching a Microwave Power Amplifier

Background

Consider an illustrative 2.45 GHz power amplifier using a 50 Ω microstrip system. Electromagnetic and nonlinear device simulations indicate that the transistor presents a complex optimum load of 18 + j12 Ω at the fundamental frequency. The output network must transform this impedance to 50 Ω while maintaining acceptable efficiency and harmonic behaviour.

Problem

A direct connection would produce substantial reflection. The network must also tolerate substrate-permittivity variation and maintain safe voltage and current levels under a 1 W output condition.

Analysis

The normalized target load is

A Smith-chart analysis first uses a short series line or series capacitor to remove the reactive component. The resulting resistance is then transformed with a shunt open stub or a two-section distributed network. A single narrowband match may provide excellent S11 at 2.45 GHz but excessive mismatch over the required 2.4–2.5 GHz band.

The design team compares three alternatives:

Approach and Solution

A two-section microstrip network is selected. Initial dimensions are obtained from Smith-chart synthesis, then optimized with a full-wave electromagnetic model including bends, pads, vias, conductor thickness, and the transistor’s nonlinear embedding impedance. Harmonic terminations are included because the optimum fundamental-frequency load is not necessarily the optimum overall large-signal load.

Outcome

The final design achieves a low fundamental reflection near the centre frequency, acceptable bandwidth, and voltage below the substrate breakdown limit. Prototype measurements show a small frequency shift, corrected by shortening one section after calibration verification.

Learning Points

1. The impedance requiring transformation may be complex and power dependent.
2. A schematic transmission-line model is only the first stage of microwave design.
3. Bandwidth, loss, power handling, and manufacturability must be optimized together.
4. Calibration and de-embedding are essential for comparing simulation with measurement.
5. Matching networks can also act as harmonic filters and stability-control networks.

12. Design Methodology and Best Practices

A reliable matching design follows these steps:

1. Define frequency range, source and load impedances, power, allowable reflection, and physical constraints.
2. Establish the reference planes and obtain accurate S-parameters or large-signal load-pull data.
3. Choose an appropriate line model, including frequency-dependent loss and effective permittivity.
4. Normalize impedances and determine whether series or shunt realization is convenient.
5. Synthesize an initial network analytically or with a Smith chart.
6. Include discontinuities, connectors, packaging, bias networks, and thermal constraints in simulation.
7. Evaluate bandwidth, insertion loss, VSWR, group delay, power handling, and tolerance sensitivity.
8. Fabricate a test coupon where practical and verify dimensions and dielectric properties.
9. Calibrate the VNA and de-embed fixtures before measurement.
10. Tune only after distinguishing design error from measurement error.

13. Limitations and Contemporary Developments

Classical transmission-line theory assumes linear, homogeneous, and often quasi-static cross-sectional behaviour. It becomes less accurate for electrically large discontinuities, strongly dispersive materials, multimode propagation, nonlinear loads, and complex three-dimensional transitions. Full-wave numerical methods such as the finite-element method, method of moments, and finite-difference time-domain method are therefore used alongside circuit models.

Modern design increasingly combines symbolic synthesis, electromagnetic simulation, optimization, and measurement-based model extraction. Machine-learning methods can assist surrogate modelling and optimization, but they do not remove the need for physically valid training data, passivity checks, causality checks, uncertainty analysis, and engineering verification. Automated optimization should not be accepted solely because it produces a low simulated reflection coefficient; manufacturability, power stress, thermal reliability, and measurement correlation remain essential.

In regulated systems, designers must also consider electromagnetic compatibility, spectrum allocation, safety limits, and product standards. Deliberate impedance control reduces unwanted radiation and interference, while accurate documentation supports repeatability and certification.

Chapter Summary

Transmission lines are distributed electromagnetic structures described by per-unit-length resistance, inductance, conductance, and capacitance. The telegrapher’s equations lead to travelling-wave solutions characterized by propagation constant γ and characteristic impedance Z₀. Load mismatch produces reflected waves, quantified by the reflection coefficient Γ, and creates standing waves described by VSWR. Input impedance varies periodically with electrical length and can be transformed using line sections.

Impedance matching transforms a load into the desired line or source impedance. Important techniques include quarter-wave transformers, single-stub networks, lumped reactive networks, multisection transformers, and tapered lines. Each has trade-offs involving bandwidth, loss, size, power, and sensitivity. Smith charts provide a graphical method for impedance and admittance transformation. Practical microwave design must include conductor and dielectric losses, dispersion, discontinuities, calibration, de-embedding, manufacturing tolerances, and power limitations. Modern matching is therefore an integrated process combining analytical synthesis, full-wave simulation, measurement, and engineering judgment.

Key Terms

Transmission line: A structure that guides electromagnetic energy with voltage and current varying along its length.

Distributed parameter: A circuit parameter specified per unit length.

Telegrapher’s equations: Differential equations relating voltage and current along a transmission line.

Characteristic impedance: The impedance seen looking into an infinitely long line or by a single travelling wave.

Propagation constant: Complex quantity γ = α + jβ describing attenuation and phase variation.

Attenuation constant: Real part of propagation constant, expressed in nepers per metre.

Phase constant: Imaginary part of propagation constant, expressed in radians per metre.

Reflection coefficient: Ratio of reflected-wave amplitude to incident-wave amplitude.

VSWR: Ratio of maximum to minimum standing-wave voltage.

Standing wave: Spatial interference pattern produced by forward and reverse waves.

Electrical length: Phase length βl of a transmission-line section.

Smith chart: Graphical tool for transforming normalized impedance and admittance.

Quarter-wave transformer: A λ/4 line section used to transform a real load impedance.

Stub: Open- or short-circuited line section used as a reactive matching element.

Admittance: Reciprocal of impedance, Y = 1/Z.

Quasi-TEM: Mode with predominantly transverse electric and magnetic fields but some longitudinal components.

De-embedding: Removal of fixture and interconnect effects from measured data.

Dispersion: Frequency dependence of propagation velocity or impedance.

Insertion loss: Reduction in transmitted power caused by a network.

Reference plane: Defined physical or analytical location at which impedance or scattering data are specified.

Review Questions

Short Answer Questions

1. Why must a microwave interconnection be treated as a transmission line rather than an ideal wire?
2. Define characteristic impedance and propagation constant.
3. What are the reflection coefficients of an open circuit and a short circuit?
4. State the relationship between VSWR and the magnitude of the reflection coefficient.
5. Why is a quarter-wave transformer narrowband?

Long Answer Questions

1. Derive the telegrapher’s equations and obtain the expressions for γ and Z₀.
2. Explain standing-wave formation and derive the VSWR formula.
3. Discuss the operation and limitations of the Smith chart.
4. Compare quarter-wave, stub, lumped-element, multisection, and tapered-line matching networks.
5. Explain how VNA calibration and de-embedding affect impedance measurements.

Analytical / Application Questions

1. A 75 Ω lossless line is terminated by 25 Ω. Calculate ΓL, VSWR, and the fraction of incident power delivered to the load.
2. Design a quarter-wave transformer between a 50 Ω line and a 200 Ω resistive load at 3 GHz. If εeff = 4, calculate the transformer impedance and physical length.
3. A microwave amplifier presents a complex output impedance. Propose a measurement, modelling, and matching workflow that accounts for bandwidth, power, fabrication tolerance, and verification uncertainty.

Multiple Choice Questions with Answers

Question 1

For a lossless transmission line, the characteristic impedance is

A. √(L/C)

B. √(C/L)

C. L/C

D. C/L

Correct Answer: A. √(L/C)

Question 2

A load equal to the characteristic impedance produces which reflection coefficient?

A. −1

B. 0

C. +1

D. j

Correct Answer: B. 0

Question 3

If the magnitude of the reflection coefficient is 0.5, the VSWR is

A. 1.5

B. 2

C. 3

D. 4

Correct Answer: C. 3

Question 4

A quarter-wave transformer matching 50 Ω to 200 Ω requires a characteristic impedance of

A. 25 Ω

B. 75 Ω

C. 100 Ω

D. 250 Ω

Correct Answer: C. 100 Ω

Question 5

For a lossless line, a short-circuited quarter-wave section appears at its input as a

A. Short circuit

B. Matched load

C. Resistor equal to Z₀

D. Open circuit

Correct Answer: D. Open circuit

Question 6

The voltage reflection coefficient at a load is

A. (Z_L+Z₀)/(Z_L−Z₀)

B. (Z_L−Z₀)/(Z_L+Z₀)

C. Z_L/Z₀

D. Z₀/Z_L

Correct Answer: B. (Z_L−Z₀)/(Z_L+Z₀)

Question 7

For a passive load, the fraction of incident power delivered to the load is

A. |Γ|²

B. 1+|Γ|²

C. 1−|Γ|²

D. 1/|Γ|²

Correct Answer: C. 1−|Γ|²

Question 8

Which matching method is most naturally designed by adding normalized susceptances?

A. Shunt-stub matching

B. Series-only resistive matching

C. Quarter-wave matching of a real load only

D. Direct cable termination

Correct Answer: A. Shunt-stub matching

Question 9

For a line with Z₀ = 50 Ω and load ZL = 100 Ω, the magnitude of ΓL is

A. 0.20

B. 0.33

C. 0.50

D. 0.67

Correct Answer: C. 0.50

Question 10

The principal purpose of VNA de-embedding is to

A. Increase source power

B. Remove fixture and interconnect effects

C. Convert an open circuit into a short circuit

D. Eliminate dielectric loss physically

Correct Answer: B. Remove fixture and interconnect effects