ADC

An ADC is an electronic circuit that transforms continuous analog signals, like voltage or current, into discrete digital values that a computer or microcontroller can process. The device samples the analog input at regular intervals, measures its amplitude, and encodes each measurement as a binary number. Without ADCs, digital systems would have no way to interface with the physical world: temperature sensors, microphones, strain gauges, and pressure transducers all feed their analog outputs into ADCs before that information can be stored, computed on, or transmitted.

The core performance metrics of an ADC are its resolution and sampling rate. Resolution, measured in bits, determines how finely the converter can divide the input voltage range: an 8-bit ADC produces 256 discrete levels, while a 12-bit ADC produces 4,096. The sampling rate, given in samples per second or hertz, sets how often the input is measured. The Nyquist theorem dictates that the sampling rate must be at least twice the highest frequency present in the input signal, or aliasing will distort the digital output. A 10 kHz audio signal requires sampling at 20 kHz minimum; high-speed instrumentation may demand rates in the megahertz or gigahertz range.

Common ADC architectures and their trade-offs

Successive approximation register (SAR) ADCs dominate industrial and automotive applications because they balance speed, power consumption, and cost. Flash ADCs achieve the highest conversion rates, converting in a single clock cycle, but consume significant power and are limited to lower resolutions. Sigma-delta ADCs excel at very high resolution and noise immunity, making them popular in precision measurement, though they require oversampling and digital filtering. Pipeline ADCs handle the gap between SAR and flash speeds, converting multiple bits per cycle at moderate power. The choice depends entirely on the application: a temperature controller may tolerate 12-bit SAR performance at 100 kHz, while an oscilloscope front end needs 8-bit or 10-bit flash conversion in the gigahertz range.

Errors and limitations are practical concerns. Quantization error, the irreducible rounding that occurs when an analog value is forced into a digital bin, equals plus-or-minus half the smallest bit value. Linearity errors, offset, and gain drift degrade accuracy over temperature and time. Input impedance must be considered: many ADCs present a high-impedance input, but sampling also draws transient current that can load weak signal sources. Protection circuits, input filtering, and careful grounding are essential in noisy industrial environments to prevent false counts and runaway data.

ADCs appear in nearly every modern instrument and control system. Programmable logic controllers read analog sensor inputs through ADC modules. Data acquisition systems sample multiple channels in parallel or sequence them through a multiplexer to a single ADC. Mixed-signal microcontrollers integrate ADCs on-chip with CPU and memory, enabling standalone measurement without external chips. Standalone ADCs range from simple 8-bit converters in basic instrumentation to precision 16-bit and 24-bit devices in test equipment and medical devices. Understanding the ADC's place in your signal chain, its sampling behavior, and its error sources is essential to building systems that measure correctly.