How to characterize an ADC?
An analog-to-digital converter (ADC) converts a continuous analog input signal (e.g. voice from a microphone or measurement values from a sensor) into discrete digital values. There are two main characteristics that are important:
- The vertical resolution is given by the so-called bit width of the converter: e.g. a 4-bit ADC divides the input signal range into 2^4 = 16 digital values. As of today, common bit widths are in the range of 18 to 24. Hence, the smallest representable value is way smaller than the interfering signals that are normally present, as for xample, signal noise.
- The sampling rate, sometimes also called sampling frequency, equals the time between two samples and is thus the horizontal resolution on the time axis. The more values per time unit the analog-to-digital converter captures (usually stated as samples per second Sps) the faster signals can be recorded. But this implies that more values are to be stored; this again increases the memory requirement. As a conclusion: the sampling rate must be chosen carefully. Common sampling rates are in the range of just a few samples per second (e.g. for temperatures) up to several million samples per second for very dynamic signals (e.g. from explosions).