quantization error, inherent (of an ideal analog-to-digital converter)

Within a step, the maximum (positive or negative) possible deviation of the actual analog input value from the nominal midstep value.

NOTE 1 This error follows necessarily from the quantization procedure. For a linear analog-to-digital converter, its value equals ±½ LSB.

NOTE 2 Use of the term "resolution error" in place of "inherent quantization error" is deprecated, because "resolution" as a design parameter has a nominal value only.

References