A class of electronic systems containing both analog and digital subsystems whose most common application is to the conversion of analog signals to digital form and vice versa. The device is also known as a delta-sigma converter. The main advantage of the sigma-delta approach to signal conversion is its minimal reliance on the quality of the analog components required. To achieve this end, the system uses pulse density modulation to create a high-rate stream of single-amplitude pulses. For analog-to-digital conversion, the rate at which the pulses are generated depends on the amplitude of the analog voltage being sensed. For digital-to-analog conversion, the pulse density depends on the numeric digital quantity applied at the converter input. See also Pulse modulation.
The simplest implementation of a sigma-delta analog-to-digital converter uses an analog circuit to generate the single-valued pulse stream from an analog source, and a digital system to repeatedly sum the number of these pulses over a fixed number of pulse intervals. The summing operation converts the pulses to a numeric value, achieving analog-to-digital conversion. Conversely, in a digital-to-analog converter, a digital circuit is used to convert numeric values from a digital processor to a pulse stream, and these pulses then are low-pass-filtered by a relatively simple analog system to produce an analog waveform. This low-pass filtering effectively sums the uniform analog pulse amplitudes over a fixed interval. The circuit—analog or digital—used to generate the pulse stream is called a sigma-delta (or delta-sigma) modulator. See also Electric filter.
In each case, analog information is contained not in the pulse amplitude but in the number of pulses that occur during the conversion interval. This distribution of the analog information makes the conversion process essentially independent of the amplitude of the pulses and greatly simplifies the design and fabrication of the analog portion of the converter. It does, however, require that the sampling process be rapid, since the resolution of the conversion depends on the number of pulses that can exist in the conversion interval.