RF & Communications — Complete Study Guide

A radio wave is electromagnetic energy — the same fundamental phenomenon as light, just at a much lower frequency. When alternating current flows in a conductor, it radiates an electromagnetic field that travels outward at the speed of light. Radio frequency (RF) is the slice of that spectrum useful for communication, roughly 3 kHz up to 300 GHz.

The spectrum is divided into named bands. You don't need to memorize every number, but you should recognize the order and what lives where:

The anatomy of a wave. Before going further it's worth naming the four properties every radio wave has, because the rest of RF is just manipulating them:

Modulation (Section 3) works by varying one of these — amplitude for AM, frequency for FM, phase for PM. So this little table is secretly the key to the whole guide.

How RF is made and radiated. Any time current changes, it produces a changing magnetic field, and a changing magnetic field produces a changing electric field — the two regenerate each other and the disturbance propagates outward as an electromagnetic wave. At low frequencies this radiation is negligible; as frequency climbs, a conductor of the right length becomes an efficient antenna, deliberately launching that energy into space. The electric and magnetic fields are perpendicular to each other and to the direction of travel — which is why "polarization" (Section 7) is defined by the electric field's orientation.

Why the band names matter in practice. The behavior shift across bands is gradual but real: at MF and below, waves follow the earth's curve (ground wave) and travel far at night. Through HF, the ionosphere becomes a mirror for long-distance "skip." From VHF up, waves punch through the ionosphere and travel essentially line-of-sight, trading range for abundant bandwidth — which is why high-data services (TV, Wi-Fi, cellular, GPS) live up high where there's room, while long-haul and low-data services sit down low.

Every wave has a frequency (how many cycles pass per second, in hertz) and a wavelength (the physical length of one cycle, in meters). They're locked together by the speed of light, so if you know one you can always find the other.

Velocity factor — waves slow down in cable. The speed of light applies in free space (and air, near enough). Inside a coax or wire, a wave travels slower — typically 0.66 to 0.95 of light speed — set by the cable's insulation. That ratio is the velocity factor (VF). It matters whenever a length of cable is part of a tuned circuit (a matching stub, a phasing line, a quarter-wave transformer):

Harmonics. A harmonic is an integer multiple of a frequency: the 2nd harmonic of 7 MHz is 14 MHz, the 3rd is 21 MHz. Harmonics matter two ways — antennas and circuits often work on harmonically related bands, and transmitters must suppress unwanted harmonics so they don't radiate interference. A wave's harmonics are always shorter in wavelength by the same integer factor.

Electrical vs physical length. An antenna's electrical length (what the wave "sees") is slightly longer than its physical length because of end effects — the wave fringes off the wire ends. That's why the practical half-wave dipole formula uses 468/f(MHz) in feet rather than the 492/f you'd get from pure free-space math (Section 7). The ~5% shortening is the end effect.

A bare radio wave at one frequency — the carrier — carries no information; it's just a steady tone. Modulation is how we ride information on top of it by varying one of its properties:

When you modulate a carrier, you don't get a single frequency anymore — you create sidebands, mirror-image copies of the carrier spaced above and below it by the modulating frequency. In AM, both sidebands carry the same information, and the carrier itself carries none.

That redundancy is wasteful, which is the whole insight behind SSB: throw away the carrier and one sideband, and you send the identical information in half the bandwidth using far less power. It's why long-distance voice on HF — and ham DX — is almost always SSB.

Modulation percentage (AM). How deeply you modulate matters. Modulation index for AM is the ratio of how far the amplitude swings to the carrier's level, usually given as a percentage:

At 100% the carrier amplitude swings from double down to zero. Push past 100% (overmodulation) and the signal clips at the zero crossings, generating spurious sidebands ("splatter") that interfere with adjacent channels and distort the audio. Undermodulation (say 30%) is clean but weak — you're not using the carrier efficiently. Broadcasters aim high but stay below 100%.

FM deviation and Carson's rule. FM doesn't have a modulation percentage; instead it has deviation — how far the carrier frequency swings from center. Wide deviation = louder/cleaner but more bandwidth. The occupied bandwidth of an FM signal is estimated by Carson's rule:

Why FM resists noise. Most natural and man-made noise shows up as amplitude spikes. Since FM carries its information in frequency, the receiver can clip off amplitude variations (with a limiter stage) and lose nothing — the noise is literally sliced away. AM has no such luxury, because amplitude is the signal.

Digital modulation, briefly. The same three knobs drive digital modes: ASK (amplitude-shift keying) varies amplitude, FSK (frequency-shift keying) shifts between tones, and PSK (phase-shift keying) flips phase. Combining amplitude and phase gives QAM, which packs many bits per symbol and underlies Wi-Fi, cable, and digital TV. They're not separate magic — just AM/FM/PM applied to discrete states instead of continuous audio.

RF signals span an enormous range — from a transmitter's hundreds of watts down to a received signal of a millionth of a microwatt. Linear numbers become unwieldy, so RF lives in decibels: a logarithmic way to express ratios.

When the reference is fixed, dB becomes an absolute unit. The big one in RF is dBm — decibels relative to 1 milliwatt:

The real superpower: dB just add. Because decibels are logarithmic, the multiplications of a signal chain become simple additions. A gain is +dB, a loss is −dB, and you walk the path adding them up:

The reference-unit family. Append a letter to "dB" and you've fixed a reference, turning a ratio into an absolute level. The common ones:

Signal-to-noise and why faint signals still work. What lets a receiver pull in a vanishingly small signal isn't raw level — it's the signal-to-noise ratio (SNR), the dB difference between the wanted signal and the noise floor. A −120 dBm signal is perfectly usable if the noise floor is −130 dBm (10 dB SNR). This is why "sensitivity" is quoted as a level for a given SNR, and why low-noise front-end design (Section 6) matters more than brute gain.

A transmitter's job is to generate a clean carrier, put information on it, and boost it to enough power to reach the receiver. Every transmitter, simple or complex, is built from the same chain of stages:

The oscillator sets the frequency — often crystal-controlled for stability. The buffer isolates the oscillator so later stages don't pull it off frequency. The modulator adds the information, and the power amplifier (PA) raises the signal to its final output level.

RF power amplifiers often run in Class C — highly efficient but distorting — which is fine for a constant-amplitude signal like FM or CW because a tuned circuit reconstructs the clean waveform. AM and SSB, whose amplitude carries the information, need more linear classes (A, AB) to avoid mangling the signal.

Setting and holding the frequency. An oscillator alone drifts with temperature and voltage. Three approaches keep a transmitter on frequency: a crystal oscillator (a quartz crystal's mechanical resonance is extremely stable — the piezoelectric effect), a VFO (variable-frequency oscillator, tunable but driftier), and the modern standard, a PLL frequency synthesizer, which locks a tunable oscillator to a crystal reference using a phase-locked loop — giving both crystal stability and the ability to step across a band.

Frequency multiplication. It's often easier to generate a stable signal at a lower frequency and then multiply it up. A multiplier stage deliberately distorts the signal to generate harmonics, then a tuned circuit selects the wanted multiple (×2, ×3, etc.). A 6 MHz crystal feeding a tripler yields a stable 18 MHz output. The trade-off: any frequency or phase change in the oscillator is multiplied too.

Keying and CW. The simplest "modulation" is just switching the carrier on and off — CW (continuous wave / Morse). It's still in use because it concentrates all power into a razor-thin bandwidth, punching through noise when voice can't. Keying must be shaped (soft on/off) to avoid "key clicks" — broadband splatter from switching too abruptly, the CW cousin of overmodulation.

Keeping the output clean. A transmitter must radiate only its intended signal. Two concerns: harmonics (multiples of the carrier) are suppressed by a low-pass filter at the output; and in tube/some high-power PAs, neutralization cancels internal feedback that would otherwise make the amplifier self-oscillate. A spurious or harmonic emission can interfere with services far from your own frequency — hence the legal requirement to filter them down.

A receiver has to pluck one faint signal out of the air from among thousands, amplify it enormously, and recover the information — all while rejecting everything nearby. The design that made this practical, and still dominates today, is the superheterodyne.

The key trick is heterodyning — mixing. The mixer combines the incoming signal with a tunable local oscillator (LO). When you mix two frequencies, you get their sum and difference. The receiver keeps the difference, which is always the same fixed intermediate frequency (IF) no matter what station you tune to.

Doing the amplification and filtering at a fixed IF is what gives a superhet its excellent selectivity (rejecting adjacent stations) and sensitivity (pulling in weak ones). After the IF stage, the detector recovers the audio and the audio amplifier drives the speaker.

Working the image frequency. The image sits two IFs away from your wanted signal (on the other side of the LO). With a 455 kHz IF and the LO above the signal, the image is 2 × 455 = 910 kHz away:

Sensitivity vs selectivity — the two receiver virtues. Sensitivity is the ability to hear weak signals (set by front-end gain and, crucially, low noise). Selectivity is the ability to separate one signal from its neighbors (set by the IF filter's bandwidth). They're different jobs: a sensitive-but-broad receiver hears everything at once; a selective-but-deaf one is quiet but clean. Good design needs both.

The detector depends on the mode. Recovering the information requires a detector matched to how the signal was modulated:

AGC — automatic gain control. Signals vary enormously as stations fade or you tune across strong and weak ones. AGC samples the output level and feeds it back to reduce gain on strong signals and raise it on weak ones, keeping the volume roughly constant and preventing strong-signal overload. It's why you can tune from a local blowtorch station to a distant whisper without lunging for the volume knob.

An antenna converts the transmitter's guided RF energy into a radiated wave (and the reverse on receive). Its size is tied directly to wavelength — which is why Section 2 matters so much here.

Antennas have gain — they focus energy in some directions at the expense of others, like a flashlight reflector. Gain is measured in decibels relative to a reference:

Polarization is the orientation of the radiated electric field — vertical or horizontal. Transmit and receive antennas should match; a vertical and a horizontal antenna can lose much of the signal between them. The feedline (usually coax) carries energy between radio and antenna, and it has loss that rises with frequency — a cable that's fine at HF can waste half your power at UHF.

Radiation patterns — antennas aren't equal in all directions. An ideal isotropic antenna radiates equally everywhere (a sphere) — it's only a math reference. Real antennas have shape:

• A half-wave dipole radiates broadside to the wire (a doughnut pattern), with nulls off the ends. Its feedpoint impedance is about 73 Ω — conveniently close to 50 Ω coax.
• A quarter-wave vertical (against a ground plane) is omnidirectional in the horizontal plane — good for mobile/repeater use. Feedpoint ~36 Ω.
• A Yagi adds parasitic director and reflector elements to focus energy into a beam — high gain in one direction, like a flashlight.

Gain always comes from focusing, never from creating energy — a high-gain antenna is just robbing the directions you don't care about to strengthen the one you do.

Feedlines and their loss. The transmission line between radio and antenna matters more than beginners expect:

Loss rises with frequency and length — and it's doubly costly under high SWR (Section 8), because reflected power makes extra trips through the lossy line. This is why a marginal cable at HF becomes a power-wasting liability at VHF/UHF.

Baluns and matching. A balun (balanced-to-unbalanced) connects a balanced antenna like a dipole to unbalanced coax, preventing feedline radiation and current imbalance. An antenna tuner doesn't change the antenna — it transforms the impedance the transmitter sees so the PA is happy, even when the antenna isn't a perfect match.

Ground systems. A vertical antenna needs a ground plane or radials to work against — the missing "other half" of the antenna. Poor grounding shows up as high SWR and lost efficiency. (This is distinct from safety grounding for lightning and shock, which every installation also needs.)

ERP — putting gain and loss together. Effective radiated power combines transmitter power, feedline loss, and antenna gain (all in dB — see Section 4). A 50 W transmitter into 3 dB of cable loss feeding a 6 dBd antenna nets roughly 100 W ERP in the favored direction. Antenna gain can beat raw transmitter power — and feedline loss can quietly eat it.

For maximum power to flow from transmitter to antenna, their impedances must match — and match the feedline's characteristic impedance (commonly 50 Ω for RF, 75 Ω for video). Characteristic impedance comes from the cable's geometry (its inductance and capacitance), not a DC resistance you can read with an ohmmeter.

When impedances don't match, some of the forward power reflects back down the line. Forward and reflected waves combine into a stationary interference pattern — standing waves — measured as the Standing Wave Ratio (SWR).

High SWR means power that never reaches the antenna, returning to heat the transmitter's final stage — which is why a bad antenna or feedline can damage a radio. Many transmitters fold back their power when they sense high SWR to protect themselves.

Reflection coefficient and return loss. SWR is one way to express a mismatch; there are two related measures you'll meet:

A perfect match means Γ = 0 (nothing reflected), SWR = 1:1, and infinite return loss. A total mismatch (open or short) means |Γ| = 1, SWR = ∞, and 0 dB return loss — everything bounces back. These are three views of the same thing: how much of your power the load refuses to accept.

How much power do you actually lose? Reflected power isn't all "lost" — in a lossless line it re-reflects and much eventually radiates — but it stresses the transmitter and exaggerates feedline loss. Rough reflected-power figures:

Fixing a mismatch. You don't always rebuild the antenna. A matching network — an L-network, pi-network, or antenna tuner — transforms the load impedance to what the source wants, using reactive components (inductors and capacitors) that add no loss in principle. A quarter-wave transformer (a precise length of line with a specific impedance) does the same trick at a single frequency. The tuner doesn't lower SWR at the antenna — it presents a happy match to the transmitter, which is what protects the PA.

Once a wave leaves the antenna, how it travels to the receiver depends mostly on its frequency. There are three main paths:

The ionosphere — charged layers high in the atmosphere — is the key to long-distance HF. It refracts HF sky waves back to earth, and its condition changes with the sun: time of day, season, and the 11-year solar cycle all shift which frequencies "open" for long-haul contacts. This is why a shortwave band alive at night may be dead at noon.

The ionospheric layers. The ionosphere isn't uniform — it has layers that behave differently and shift between day and night:

MUF and critical frequency. The ionosphere only refracts a wave back if the frequency is low enough for the current ionization. Two key numbers:

• Critical frequency — the highest frequency that reflects straight back when sent straight up. Above it, signals punch through to space.
• MUF (Maximum Usable Frequency) — the highest frequency that completes a given path (lower-angle signals reflect at higher frequencies than vertical ones). Operators work just below the MUF for best results; go above it and the signal escapes to space, below it and absorption increases.

A related idea is the skip zone — a dead ring between where the ground wave fades out and where the first sky-wave "hop" lands. Stations inside the skip zone can't hear a signal that's perfectly readable farther away.

Fading and multipath. Signals rarely arrive by a single clean path. When copies arrive via slightly different paths (different hops, or direct + reflected), they can add or cancel — producing fading. Rapid, frequency-selective fading distorts; slow fading just rises and falls (AGC handles that). Multipath is also why a moving car radio flutters under bridges and near buildings.

Above HF: it's not all line-of-sight. VHF/UHF are mostly line-of-sight, but a few modes extend them: tropospheric ducting (weather layers bending signals far beyond the horizon), sporadic-E (intense E-layer patches reflecting VHF), and knife-edge diffraction over ridges. Even "line-of-sight" reaches slightly past the visual horizon because the atmosphere bends RF a little — the radio horizon is about 15% farther than the optical one.