Microwave engineering basics

In short

Microwave engineering deals with electromagnetic signals from roughly 300 MHz to 300 GHz (free-space wavelength 1 m down to 1 mm), where circuit dimensions become comparable to the wavelength, so lumped R-L-C thinking fails and energy must be handled as guided waves in waveguides, coaxial lines and planar structures. The subject covers guided modes and cutoff, resonant cavities, scattering-parameter description of multiport junctions, passive components such as tees, couplers and ferrite devices, and the special sources needed at these frequencies (klystron, magnetron, TWT, Gunn and IMPATT diodes). It matters because radar, satellite communication, 5G, navigation, microwave links, RF heating and test instrumentation all live in this band. It is a reliable GATE/ESE scoring area (cutoff and mode calculations, S-matrix properties, device identification) and is directly tested in interviews at ISRO/VSSC/SAC, DRDO labs (LRDE, DEAL, RCI), BEL, Astra Microwave, HAL, Keysight/Anritsu/Rohde & Schwarz India, Reliance Jio and Bharti Airtel transmission teams.

Explained

Why microwaves need their own engineering

• Definition by wavelength, not by application: at 10 GHz, λ₀ = 30 mm, so a 15 mm lead is a half-wavelength radiator, not a wire. Voltage and current are no longer uniquely defined across a cross-section; power and travelling waves are the natural variables.
• Lumped elements fail: stray inductance of leads, interelectrode capacitance and transit time of carriers all become comparable to the signal period, so conventional tubes and low-frequency transistors stop amplifying.
• Low-loss guiding needed: skin effect confines current to a few micrometres, so ordinary wires are lossy; hollow metal waveguide or well-designed planar lines are used instead.
• Advantages of the band: very large absolute bandwidth (1% of 10 GHz = 100 MHz), highly directive beams from compact apertures, line-of-sight behaviour that allows frequency reuse, ability to pass through the ionosphere for satellite links, low atmospheric and galactic noise in 1–10 GHz (the "microwave window"), and fine resolution for radar and imaging.
• Limitations: line-of-sight only (horizon-limited hops), rain and atmospheric absorption above ~10 GHz (strong O₂ line near 60 GHz, H₂O near 22 GHz), expensive precision machining and connectors, and tight tolerances everywhere.

Rectangular waveguide: modes, cutoff and propagation

• A hollow single-conductor guide cannot support a TEM wave (no second conductor for the return path), so propagation is in TE modes (Ez = 0) and TM modes (Hz = 0), designated TEmn/TMmn with m half-cycle variations along the broad wall a and n along b.
• Cutoff: each mode propagates only above its cutoff frequency f_c; below it the fields decay exponentially (evanescent), so a waveguide is inherently a high-pass structure and passes no DC.
• Dominant mode TE₁₀ has the lowest cutoff, f_c = c/2a, and is used in practice; TM₁₁ is the lowest TM mode (TM₁₀ and TM₀₁ do not exist).
• Single-mode bandwidth: between TE₁₀ and the next mode. Choosing b ≈ a/2 pushes TE₀₁ above TE₂₀, giving a usable octave; commercial bands are kept well inside it (WR-90: cutoff 6.56 GHz, specified band 8.2–12.4 GHz).
• Dispersion: phase velocity v_p > c and group velocity v_g < c, with v_p·v_g = c². Energy travels at v_g; the guide wavelength λ_g is always longer than λ₀. This is not a violation of relativity — no information travels at v_p.
• Wave impedance is frequency dependent: Z_TE rises above 377 Ω and tends to infinity at cutoff; Z_TM falls towards zero at cutoff.
• Attenuation comes from finite wall conductivity (≈ 0.1 dB/m for copper WR-90 at 10 GHz) and any dielectric filling; power handling is limited by air breakdown (~30 kV/cm at STP, reduced by humidity, dust and altitude), so high-power guides are pressurised with dry air or SF₆.
• Dielectric filling of permittivity εr scales both f_c and λ by 1/√εr — the basis of compact dielectric-loaded components.
• Circular waveguide: modes come from Bessel-function roots; TE₁₁ is dominant, TM₀₁ next; TE₀₁ has the lowest attenuation and was used for long runs; circular guides suffer from polarisation rotation, so rectangular guide is preferred for most links. Degenerate modes (same f_c, e.g. TE₀₁ and TM₁₁ in circular guide, TEmn/TMmn pairs in rectangular guide) are a common exam point.

Cavity resonators and Q

• A length of waveguide shorted at both ends traps a standing wave and resonates at discrete frequencies — the microwave equivalent of a tuned L-C circuit, but with far higher Q (10³–10⁵ against ~100 for lumped circuits).
• Rectangular cavity: TEmnp/TMmnp, with p half-waves along the length d; the lowest practical mode is TE₁₀₁. Cylindrical cavity: TE₁₁₁ and TM₀₁₀ are common; TM₀₁₀ frequency depends only on radius.
• Unloaded Q₀ = 2π × (energy stored)/(energy lost per cycle) = ω₀W/P_loss; losses are in the walls (so silver/copper plating and polishing raise Q) and in any dielectric.
• Loaded Q_L includes the external circuit: 1/Q_L = 1/Q₀ + 1/Q_ext. The 3 dB bandwidth is f₀/Q_L. Coupling to a cavity is by a probe (electric), loop (magnetic) or iris/aperture, and may be under-coupled, critically coupled or over-coupled.
• Tuning is by a plunger or screw that changes volume; a frequency meter (wavemeter) is a calibrated cylindrical cavity whose absorption dip marks the frequency.
• Uses: klystron and magnetron tanks, filters, frequency standards, dielectric-constant measurement by perturbation, and dielectric-resonator oscillators.

Scattering parameters and multiport junctions

• Why S-parameters: at microwave frequencies true open and short terminations are hard to realise over a band and total voltage/current cannot be measured directly, but incident and reflected travelling waves can be sampled with couplers. S-parameters relate them: [b] = [S][a].
• Meaning: S₁₁ is input reflection coefficient with all other ports matched; S₂₁ is forward transmission (gain or insertion loss); S₁₂ is reverse transmission (isolation); S₂₂ output reflection. Each is complex — magnitude and phase.
• Properties: an n-port has an n×n matrix; reciprocal (passive, no ferrite/non-linearity) ⇒ [S] = [S]ᵗ, i.e. Sij = Sji; lossless ⇒ [S] is unitary (squared magnitudes in each column sum to 1, columns mutually orthogonal); a matched port has Sii = 0; shifting the reference plane multiplies entries by phase factors e^−jβℓ.
• Key theorems: a lossless, reciprocal three-port cannot have all three ports simultaneously matched — hence a matched reciprocal 3-port must be lossy (resistive power divider) or non-reciprocal (circulator). A lossless, reciprocal, matched four-port is necessarily a directional coupler.
• Measured with a vector network analyser (VNA) after calibration (SOLT / TRL) to move the reference plane to the device pins.

Passive components and ferrite devices

• E-plane tee (series tee): side arm in the plane of the E field; outputs in the collinear arms are out of phase when fed from the E arm.
• H-plane tee (shunt tee): side arm in the plane of the H field; collinear outputs are in phase.
• Magic tee (hybrid T): E and H tees combined. Fed at the H arm, power splits equally and in phase into arms 1 and 2; fed at the E arm, it splits equally and anti-phase; E and H arms are isolated from each other, and collinear arms are isolated. Needs internal matching posts/cones. Used in balanced mixers, duplexers, impedance bridges and as a power combiner.
• Hybrid ring (rat-race): a 1.5λ ring with four ports λ/4 apart — a printable 3 dB hybrid.
• Directional coupler: samples a known fraction of the forward wave only. Characterised by coupling C = 10 log(P₁/P₃), directivity D = 10 log(P₃/P₄), isolation I = C + D, and insertion loss. Two-hole, multi-hole, Bethe-hole and branch-line types; essential for power monitoring, reflectometers and VNAs.
• Ferrite devices (non-reciprocal, exploit gyromagnetic precession under a DC bias):
  • Isolator — low loss forward (0.3–1 dB), high loss reverse (20–30 dB); protects sources from reflected power. Built using Faraday rotation or resonance absorption (field-displacement type also common).
  • Circulator — 3- or 4-port device that passes 1→2→3→1; used as a transmit/receive duplexer in radar and base stations, and to turn a negative-resistance diode into an amplifier.
• Other hardware: fixed and variable attenuators (resistive card, PIN diode), phase shifters (dielectric slab, ferrite, switched-line PIN), matched loads, rotary joints, bends, twists, tapers, irises and tuning posts for reactive loading, flanges and waveguide-to-coax transitions (probe/loop), and planar/coaxial slotted lines.

Microwave sources and active devices

• Why conventional tubes/transistors fail: electron transit time across the gap becomes comparable to the RF period and interelectrode capacitance shunts the signal. Microwave tubes turn this transit time into an advantage.
• Two-cavity klystron (amplifier): the buncher cavity velocity-modulates the beam; drifting in the field-free space converts velocity modulation to density (current) modulation; the catcher cavity extracts power. Bunching parameter X peaks at X = 1.841 where J₁(X) = 0.582, so the fundamental RF current is at most 1.164 I₀. Gain ≈30 dB (up to 60 dB in multicavity designs), narrow band, high power (kW–MW) — used in radar transmitters and linear accelerators.
• Reflex klystron (oscillator): single cavity plus a negative repeller that returns the beam; oscillation requires the round-trip transit time to be N = n + 3/4 cycles (n = 0, 1, 2…), giving discrete repeller modes. Low power (10 mW–several W), electronically tunable; maximum theoretical electronic efficiency about 22.7%, practically far less. Mostly a laboratory and teaching source now.
• Magnetron (crossed-field oscillator): radial DC electric field with an axial magnetic field makes electrons follow cycloidal paths; a rotating space-charge spoke pattern interacts with the resonant anode cavities (π-mode, stabilised by strapping). Efficiency 40–70%, high power. The 2.45 GHz, ~700–1000 W oven magnetron and pulsed radar magnetrons are the standard examples.
• Travelling-wave tube (TWT): the beam interacts continuously with a slow-wave helix, so it is wideband (an octave or more) with 30–50 dB gain; needs attenuators/severs to prevent oscillation. The TWTA is still the workhorse of satellite transponders (and is built/qualified in India for ISRO payloads). BWO (voltage-tunable oscillator) and gyrotron (very high power, mm-wave, used in fusion heating) belong to the same family.
• Solid-state negative-resistance devices:
  • Gunn diode (TED): bulk GaAs/InP effect — electrons transfer from a light-mass to a heavy-mass valley above a threshold field of about 3.2 kV/cm, producing negative differential mobility and travelling domains; transit-time frequency f ≈ v_d/L. No junction. Milliwatt–watt CW, low noise; used in Doppler modules and intruder alarms.
  • IMPATT: avalanche + transit-time delay gives ~180° phase lag; high power and efficiency (10–20%) but noisy. TRAPATT gives higher efficiency at lower frequency; BARITT is low-noise, low-power.
  • Tunnel diode: negative resistance from band-to-band tunnelling, very low noise, very low power.
• Control and small-signal devices: PIN diode (RF switch, attenuator, phase shifter — acts as a variable resistance), varactor (voltage-variable capacitance for VCOs, frequency multipliers, parametric amplifiers), Schottky barrier diode (majority-carrier, no storage delay — the standard mixer and detector), GaAs MESFET/pHEMT and GaN HEMT (low-noise amplifiers and high-power amplifiers), realised as MMICs. GaN-on-SiC has displaced tubes in many radar transmitters; India has a gallium nitride/MMIC line-up through GAETEC and SITAR with DRDO and SCL Mohali in the chain.

Measurements, safety and Indian spectrum practice

• Microwave bench: source (klystron/Gunn with modulator) → isolator → variable attenuator → frequency meter → slotted line → detector/VSWR meter, with the device under test inserted. Modern labs replace most of this with a VNA and a spectrum analyser.
• Frequency: cavity wavemeter, or direct reading on a spectrum analyser/counter. Guide wavelength from the spacing of adjacent standing-wave minima (= λ_g/2) in a slotted line.
• Power: low power by bolometer (barretter — positive temperature coefficient; thermistor — negative), thermocouple or diode sensor heads with a power meter; high power by calorimetric (flow) methods or by measuring a known coupler tap. Readings are reported in dBm.
• Impedance/reflection: slotted line (VSWR and minimum position) or VNA S₁₁; attenuation by the power-ratio or substitution method; noise figure by the Y-factor method.
• Safety: microwave power causes tissue heating with the eye and testes most vulnerable; never look into an open waveguide or energise an unterminated high-power source. India's exposure limits for telecom base stations were tightened by the Department of Telecommunications in 2012 to one-tenth of the ICNIRP general-public limits, with TERM cells doing EMF audits and self-certification by operators; limits and test procedures are revised from time to time, so the current DoT/TEC circular should be checked.
• Spectrum governance: the WPC Wing (Ministry of Communications) issues licences and publishes the National Frequency Allocation Plan (latest major edition NFAP-2022), aligned to ITU Radio Regulations; TRAI recommends pricing; auctions since 2022 have assigned 3.3–3.67 GHz (n77/n78) and 24.25–27.5 GHz (n258) for 5G. Delicensed/ISM use includes 2.4–2.4835 GHz and parts of 5 GHz (Wi-Fi, under the 2018/2021 delicensing rules) and 865–867 MHz for RFID; 6 GHz delicensing rules were notified in 2025 and detailed conditions continue to evolve. Equipment must meet TEC certification under the MTCTE regime and BIS/EMC requirements; the governing Act is the Telecommunications Act, 2023, which replaced the Indian Telegraph Act, 1885 and the Wireless Telegraphy Act, 1933.

Key points

• Microwave range - 300 MHz to 300 GHz; λ₀ = 1 m to 1 mm; ITU calls 3–30 GHz SHF and 30–300 GHz EHF.
• Distributed behaviour - dimensions comparable to λ, so waves and S-parameters replace lumped V-I analysis.
• No TEM in hollow waveguide - only TE and TM modes; a second conductor (coax, microstrip) is needed for TEM.
• Dominant mode - TE₁₀ in rectangular guide (f_c = c/2a), TE₁₁ in circular guide; TM₁₁ is the lowest TM mode.
• Cutoff - below f_c the mode is evanescent; waveguide is a high-pass device and blocks DC.
• v_p·v_g = c² - phase velocity exceeds c, group velocity (energy) is less than c; λ_g > λ₀ always.
• Wave impedance - Z_TE = η₀/√(1−(f_c/f)²) rises at cutoff; Z_TM = η₀√(1−(f_c/f)²) falls to zero.
• b ≈ a/2 - standard aspect ratio giving the widest single-mode band with good power handling.
• Cavity Q - Q₀ = ω₀ × stored energy / power loss; 1/Q_L = 1/Q₀ + 1/Q_ext; BW = f₀/Q_L.
• S-matrix rules - reciprocal ⇒ symmetric; lossless ⇒ unitary; matched port ⇒ Sii = 0.
• Three-port theorem - a lossless, reciprocal, fully matched 3-port is impossible; hence circulators (non-reciprocal) and resistive dividers (lossy).
• Four-port theorem - a lossless, reciprocal, matched 4-port must be a directional coupler.
• Magic tee - E and H arms isolated; H-arm feed gives in-phase split, E-arm feed gives anti-phase split.
• Coupler figures - coupling = P_in/P_coupled, directivity = P_coupled/P_isolated, isolation = coupling + directivity (in dB).
• Ferrite devices - isolator and circulator are non-reciprocal; they rely on a biased ferrite and break Sij = Sji.
• Klystron vs TWT vs magnetron - klystron: narrowband high-gain amplifier; TWT: wideband amplifier; magnetron: efficient crossed-field oscillator (ovens, radar).
• Transit-time devices - Gunn (bulk GaAs, no junction, low noise), IMPATT (high power, high noise), TRAPATT, BARITT.
• Control devices - PIN for switching/attenuation, varactor for tuning and multiplication, Schottky for mixing and detection.

Formulas and examples

Example 1 — X-band waveguide WR-90 at 10 GHz. a = 22.86 mm, b = 10.16 mm, air-filled.

• f_c(TE₁₀) = c/2a = 3×10⁸/(2 × 0.02286) = 6.562 GHz; f_c(TE₂₀) = 13.12 GHz; f_c(TE₀₁) = 3×10⁸/(2 × 0.01016) = 14.76 GHz. Single-mode band = 6.56–13.12 GHz (specified operating band 8.2–12.4 GHz).
• At f = 10 GHz: λ₀ = 30 mm, f_c/f = 0.6562, (f_c/f)² = 0.4306, √(1 − 0.4306) = 0.7546.
• λ_g = 30/0.7546 = 39.76 mm; v_p = 3×10⁸/0.7546 = 3.98×10⁸ m/s; v_g = 3×10⁸ × 0.7546 = 2.26×10⁸ m/s (check: 3.98 × 2.26 ×10¹⁶ ≈ 9×10¹⁶ = c²).
• Z_TE = 377/0.7546 = 499.6 Ω, which is why waveguide-to-50 Ω coax transitions need a transformer.

Example 2 — rectangular cavity. a = 4 cm, b = 2 cm, d = 5 cm, TE₁₀₁ mode.

• f_r = (3×10⁸/2)·√[(1/0.04)² + (1/0.05)²] = 1.5×10⁸ × √(625 + 400) = 1.5×10⁸ × 32.02 = 4.80 GHz.

Example 3 — cavity Q and bandwidth. f₀ = 9 GHz, Q₀ = 8000, Q_ext = 4000.

• 1/Q_L = 1/8000 + 1/4000 = 1.25×10⁻⁴ + 2.5×10⁻⁴ = 3.75×10⁻⁴ ⇒ Q_L = 2667.
• BW = 9×10⁹/2667 = 3.38 MHz. (With the cavity uncoupled, BW would be 9 GHz/8000 = 1.13 MHz — loading always widens and de-Qs the response.)

Example 4 — S-parameters and a power budget. A two-port measures |S₁₁| = 0.10, |S₂₁| = 0.95, with 100 mW incident at port 1 and port 2 matched.

• Return loss = −20 log(0.10) = 20 dB; VSWR = 1.10/0.90 = 1.22.
• Insertion loss = −20 log(0.95) = 0.45 dB.
• Reflected power = |S₁₁|² × 100 = 1 mW; delivered = |S₂₁|² × 100 = 0.9025 × 100 = 90.25 mW.
• Missing power = 100 − 1 − 90.25 = 8.75 mW, so |S₁₁|² + |S₂₁|² = 0.9125 < 1 and the device is not lossless (a unitary matrix would sum to exactly 1).

Example 5 — directional coupler, and VSWR/frequency from a slotted line.

• A 20 dB coupler with 45 dB isolation fed with 1 W (30 dBm): coupled port = 1/100 = 10 mW (10 dBm); isolated port = 1/10^4.5 = 31.6 µW (−15 dBm); directivity = 45 − 20 = 25 dB; ideal through loss = −10 log(1 − 0.01) = 0.04 dB.
• On a WR-90 slotted line the detector reads V_max/V_min = 2.0 and adjacent minima are 24 mm apart.
  • VSWR = 2.0 ⇒ |Γ| = (2 − 1)/(2 + 1) = 0.333; return loss = −20 log(0.333) = 9.54 dB; reflected power = 0.333² = 11.1% of incident.
  • λ_g = 2 × 24 = 48 mm; λ_c = 2a = 45.72 mm ⇒ 1/λ₀² = 1/48² + 1/45.72² = 4.340×10⁻⁴ + 4.784×10⁻⁴ = 9.124×10⁻⁴ ⇒ λ₀ = 33.1 mm ⇒ f = 3×10⁸/0.0331 = 9.06 GHz.

Where it is used

• Space and satellite: ISRO (SAC Ahmedabad, VSSC, URSC) — C/Ku/Ka-band transponders, TWTAs, waveguide feeds and filters for INSAT/GSAT, RISAT SAR, NavIC payloads; deep-space network antennas at Byalalu.
• Defence electronics: DRDO labs (LRDE, DEAL, RCI, DLRL) and BEL, Astra Microwave, Data Patterns, Bharat Forge/Kalyani — radar transmit-receive modules, GaN HPAs, circulators, phase shifters, EW and jamming subsystems for Akash, LRDE AESA radars and naval sets.
• Telecom: microwave backhaul (6–42 GHz licensed links) and 5G mmWave for Jio, Airtel and Vodafone Idea; base-station duplexers, cavity filters and circulators made by Indian suppliers; TEC/MTCTE certification and EMF compliance work.
• Industrial, medical and scientific: 2.45 GHz ovens and industrial RF heating/drying, microwave-assisted chemistry, hyperthermia and ablation equipment, plasma sources; 915 MHz/2.45 GHz heating lines.
• Instrumentation and test: VNA/spectrum-analyser applications engineers at Keysight, Anritsu, Rohde & Schwarz and Tektronix India; calibration labs under NPL India/NABL for microwave power and attenuation standards.
• Job roles: RF/microwave design engineer, antenna and feed engineer, radar systems engineer, MMIC/GaN device engineer, EMI-EMC and test engineer, satellite payload integration engineer.

Interview questions

• Q: Why can a hollow rectangular waveguide not support a TEM wave? A: A TEM wave needs two separate conductors to carry equal and opposite currents and support a transverse-only field; a single closed conductor forces a longitudinal E or H component, giving TM or TE modes.
• Q: Why is the phase velocity in a waveguide greater than the speed of light? A: It is the apparent speed of the constant-phase pattern formed by waves bouncing obliquely off the walls; energy and information travel at the group velocity, which is less than c, and v_p·v_g = c².
• Q: What decides the operating band of a standard waveguide? A: The lower edge sits above the TE₁₀ cutoff (to avoid high attenuation and high wave impedance near cutoff) and the upper edge below the next mode's cutoff, so only the dominant mode propagates.
• Q: Why is b kept roughly half of a? A: It places TE₀₁ cutoff above TE₂₀, maximising the single-mode band, while keeping enough height for power handling and low loss.
• Q: Why are S-parameters preferred over Z or Y parameters at microwave frequencies? A: Ideal opens and shorts are not realisable over a band and total voltage/current cannot be measured directly, whereas incident and reflected waves can be sampled with directional couplers in a matched 50 Ω system.
• Q: How can you tell from an S-matrix that a component is lossless and reciprocal? A: Reciprocal if the matrix is symmetric (Sij = Sji); lossless if it is unitary — the squared magnitudes in each column sum to 1 and distinct columns are orthogonal.
• Q: Why can't a lossless reciprocal three-port be matched at all three ports? A: The unitary and symmetry conditions become mutually inconsistent; so a matched 3-port is either lossy (resistive divider) or non-reciprocal (circulator).
• Q: What is the difference between an isolator and a circulator? A: Both are ferrite non-reciprocal devices; an isolator is a two-port that passes one direction and absorbs the reverse, while a circulator routes power cyclically among three or four ports and can act as a duplexer.
• Q: Which arm of a magic tee is isolated from which, and why is that useful? A: The E and H arms are isolated from each other, and the two collinear arms are isolated; this allows simultaneous combining and splitting without interaction, as in balanced mixers and impedance bridges.
• Q: What is directivity of a coupler and why does it matter more than coupling? A: It is the ratio of coupled to isolated port power, showing how well the coupler distinguishes forward from reverse waves; poor directivity limits the accuracy of any reflection or VSWR measurement.
• Q: How does a reflex klystron oscillate with only one cavity? A: The repeller returns the velocity-modulated beam through the same gap after a transit time of n + 3/4 cycles, so the returning bunches deliver energy to the cavity in the correct phase.
• Q: Why is a Gunn diode called a transferred-electron device and not a junction device? A: It has no p-n junction; above about 3.2 kV/cm in GaAs, electrons transfer to a higher-energy, heavier-mass valley, giving negative differential mobility in the bulk material.
• Q: IMPATT or Gunn for a low-noise local oscillator? A: Gunn — IMPATT delivers more power and efficiency but its avalanche process makes it far noisier.

Common mistakes

• Using λ₀ instead of λ_g for guide lengths, stub and cavity dimensions, or forgetting that minima in a slotted line are λ_g/2 apart.
• Writing the cutoff formula without the dielectric factor when the guide is filled, or swapping a and b so that TE₀₁ and TE₁₀ are interchanged.
• Claiming v_p > c violates relativity, or using v_p to compute signal delay instead of v_g.
• Treating Z_TE as a constant 377 Ω, or applying the TM expression to a TE mode.
• Assuming |S₁₁|² + |S₂₁|² = 1 for a real device, and so under-reporting loss; or quoting S-parameter magnitudes while ignoring phase.
• Saying a ferrite isolator is reciprocal, or trying to satisfy the symmetry condition for a circulator's S-matrix.
• Mixing up the tubes: calling the magnetron an amplifier, the TWT narrowband, or expecting high power from a reflex klystron.
• Unit slips — mm with m, GHz with Hz, dB with dBm — and treating dB gains as multiplying factors.

Related topics

Electromagnetic waves and transmission lines · Smith chart and impedance matching · Antennas and wave propagation · RF circuit design basics · Radar systems and satellite communication · Microwave measurements and VNA calibration