Modulation is the process of writing data onto an electric current or EM wave. The process can be done by varying the amplitude, the frequency, or the phase of the current or wave. Another method is to transmit a series of pulses, whose duration, amplitude, or spacing is made to vary.

The Carrier

The heart of a wireless signal is a sine wave known as the carrier. The lowest carrier frequency used for radio communications is a few kilohertz (kHz). The highest frequency is in the hundreds of gigahertz (GHz). For efficient data transfer, the carrier frequency must be at least 10 times the highest frequency of the modulating signal.

On/Off Keying

The simplest form of modulation is on/off keying. This can be done in the oscillator of a radio transmitter to send Morse code, which is a binary digital mode. The duration of a Morse-code dot is one bit (binary digit). A dash is 3 bits long. The space between dots and dashes within a character is 1 bit. The space between characters in a word is 3 bits. The space between words is 7 bits. The key-down (full-carrier) condition is called mark, and the key-up (no-signal) condition is called space.
Morse code is slow. Human operators use speeds ranging from about 5 words per minute (wpm) to 40 or 50 wpm. A few human operators can work at 60 to 70 wpm. These people usually copy the signals in their heads.

Frequency-Shift Keying

Simplified block diagram of an AFSK transmitter.
Digital data can be sent over wireless links by means of frequency-shift keying (FSK). In some FSK systems, the carrier frequency is shifted between mark and space conditions, usually by a few hundred hertz or less. In other systems, a two-tone audio-frequency (AF) sine wave modulates the carrier.
This is known as audio-frequency-shift keying (AFSK). The two most common codes used with FSK and AFSK are Baudot (pronounced “baw-DOE”) and ASCII (pronounced “ASK-ee”). The acronym ASCII stands for American Standard Code for Information Interchange. In radioteletype (RTTY) FSK and AFSK systems, a terminal unit (TU) converts the digital signals into electrical impulses that operate a teleprinter or display characters on a computer screen. The TU also generates the signals necessary to send RTTY as an operator types on a keyboard. A device that sends and receives AFSK is sometimes called a modem.
This acronym stands for modulator/ demodulator. A modem is basically the same as a TU. above figure is a block diagram of an AFSK transmitter. The main advantage of FSK or AFSK over on/off keying is the fact that there are fewer errors or misprints, because the space part of the signal is identified as such, rather than existing as a gap or pause in the data. A sudden noise burst in an on/off keyed signal can confuse a receiver into reading the space as a mark signal, but when the space is positively represented by its own signal, this is less likely to happen.

Amplitude Modulation

An amplitude modulator using an NPN bipolar transistor
An AF voice signal has frequencies mostly in the range between 300 Hz and 3 kHz. Some characteristic of an RF carrier can be varied, or modulated, by these waveforms, thereby transmitting voice information. Above figure shows a simple circuit for obtaining amplitude modulation (AM). This circuit can be imagined as an RF amplifier for the carrier, with the instantaneous gain dependent on the instantaneous audio input amplitude. Another way to think of this circuit is as a mixer that combines the RF carrier and audio signals to produce sum and difference signals at frequencies just above and below that of the carrier.
The circuit shown in above figure works well, provided the AF input amplitude is not too great. If the AF input is excessive, then distortion occurs, intelligibility is degraded, system efficiency is reduced, and the bandwidth of the signal is increased unnecessarily. The extent of AM is expressed as a percentage, from 0 percent (an unmodulated carrier) to 100 percent (full modulation). Increasing the modulation past 100 percent causes the same problems as excessive AF input. In an AM signal modulated 100 percent, 1⁄3 of the power is used to convey the data, and the other 2⁄ 3 is consumed by the carrier wave.
Spectral display of a typical AM voice communications signal.
Above figure shows a spectral display of an AM voice radio signal. The horizontal scale is calibrated in increments of 1 kHz per division. Each vertical division represents 3 dB of change in signal strength. The maximum (reference) amplitude is 0 dB relative to 1 mW (abbreviated as 0 dBm). The data exists in sidebands above and below the carrier frequency. These sidebands resemble the sum and difference signals produced by a mixer. In this case the mixing occurs between the AF input signal and the RF carrier. The RF between −3 kHz and the carrier frequency constitutes the lower sideband (LSB); the RF from the carrier frequency to +3 kHz represents the upper sideband (USB). The bandwidth is the difference between the maximum and minimum sideband frequencies, in this case 6 kHz.
In an AM signal, the bandwidth is twice the highest audio modulating frequency. In the example of above figure, all the voice energy is at or below 3 kHz, so the signal bandwidth is 6 kHz. This is typical of a communications signal. In standard AM broadcasting, the AF energy is spread over a wider bandwidth, nominally 10 kHz.

Single Sideband

Spectral display of a typical SSB voice communications signal. In this example, the carrier and the USB energy are eliminated, leaving only the LSB energy.
In AM, most of the RF signal power is consumed by the carrier alone; the two sidebands are mirrorimage duplicates. This is inefficient, and is also unnecessarily redundant! If the carrier and one of the sidebands is eliminated, these shortcomings can be overcome. That makes the signal stronger for a given amount of RF power, or allows the use of lower RF power in a given communications scenario. Another bonus is the fact that the bandwidth is reduced to less than half that of an AM signal modulated with the same data, so more than twice as many signals can fit into a specific range, or band, of frequencies.
When the carrier is removed from an AM signal along with one of the sidebands, the remaining RF energy has a spectral display resembling above figure. This is single sideband (SSB) transmission. Either the LSB or the USB alone can be used, with equally good results.
A balanced modulator using two NPN bipolar transistors. The inputs are in push-pull, but the outputs are in parallel.
An SSB signal can be obtained with a balanced modulator, which is an amplitude modulator/ amplifier using two transistors with the inputs in push-pull and the outputs in parallel (above figure). This cancels the carrier wave in the output, leaving only LSB and USB energy. The result is a double sideband suppressed carrier (DSBSC) signal, often called simply double sideband (DSB). At some stage following the balanced modulator, one of the sidebands is removed from theDSB signal by a bandpass filter to obtain an SSB signal.
Block diagram of a basic SSB transmitter.
Above figure is a block diagram of a simple SSB transmitter. The balanced modulator is placed in a low-power section of the transmitter. The RF amplifiers that follow any type of amplitude modulator, including a balanced modulator, must all be linear to avoid distortion and unnecessary spreading of signal bandwidth (“splatter”). They generally work in class A, except for the PA, which works in class AB or class B.

Frequency and Phase Modulation

Generation of FM by reactance modulation of a Colpitts oscillator. Other oscillator types can be similarly modified.
In frequency modulation (FM), the instantaneous amplitude of a signal remains constant, and the instantaneous frequency is varied instead. A nonlinear PA such as a class C amplifier can be used in an FM transmitter without causing signal distortion, because the amplitude does not fluctuate.
Frequency modulation can be obtained by applying the audio signal to a varactor in a tuned oscillator. An example of this scheme, known as reactance modulation, is shown in above figure. The varying voltage across the varactor causes its capacitance to change in accordance with the audio waveform. The changing capacitance results in variation of the resonant frequency of the inductance-capacitance (LC) tuned circuit, causing small fluctuations in the frequency generated by the oscillator.
Another way to get FM is to modulate the phase of the oscillator signal. This causes small variations in the frequency, because any instantaneous phase change shows up as an instantaneous frequency change (and vice versa). When phase modulation is used, the audio signal must be processed, adjusting the frequency response of the audio amplifiers. Otherwise the signal will sound unnatural when it is received.
Deviation is the maximum extent to which the instantaneous-carrier frequency differs from the unmodulated-carrier frequency. For most FM voice communications transmitters, the deviation is standardized at 5 kHz. This is known as narrowband FM (NBFM). The bandwidth of an NBFM signal is comparable to that of an AM signal containing the same modulating information. In FM hi-fi music broadcasting, and in some other applications, the deviation is much greater than 5 kHz. This is called wideband FM (WBFM).
The deviation obtainable by means of direct FM is greater, for a given oscillator frequency, than the deviation that can be obtained by means of phase modulation. The deviation of a signal can be increased by a frequency multiplier. When an FM signal is passed through a frequency multiplier, the deviation is multiplied along with the carrier frequency. The deviation in an FM signal should be equal to the highest modulating audio frequency if optimum fidelity is to be obtained. Thus, 5 kHz is more than enough for voice. For music, a deviation of 15 kHz or even 20 kHz is required for good reproduction when the signal is received.
In any FM signal, the ratio of the frequency deviation to the highest modulating audio frequency is called the modulation index. Ideally, this figure is between 1:1 and 2:1. If it is less than 1:1, the signal sounds muffled or distorted, and efficiency is sacrificed. Increasing it beyond 2:1 broadens the bandwidth without providing significant improvement in intelligibility or fidelity.