An ADC samples a continuously varying voltage at regular intervals. If the sampling rate is too slow, those samples can resemble a completely different, slower signal. This is aliasing.
A graph showing aliasing as signal and sampling frequencies change
1. Choose an example or adjust the sliders
The graph covers 0.5 seconds. Raise the signal frequency above half the sampling frequency to see the red apparent signal diverge from the gray original.
Nyquist frequency
Half the sampling frequency is the Nyquist frequency. Higher-frequency components can fold into lower frequencies. The apparent frequency is the absolute difference between the signal frequency and its nearest integer multiple of the sampling frequency. For example, sampling 45 Hz at 50 Hz makes it appear as |45 − 50| = 5 Hz.
To reconstruct the signal: sampling frequency > 2 × highest signal frequencyIn practice
Once sampled, an aliased component cannot be distinguished from a true low-frequency signal using those samples alone. An analog low-pass anti-aliasing filter removes unwanted components before the ADC. Real filters need a transition band, so choose a sampling rate with enough margin; 5–10 times the frequency of interest is a useful illustrative choice, not a universal requirement.