EASA Module 5 · 5.3 Data Conversion
An analogue-to-digital converter can only report one of 2ⁿ discrete levels, so every reading is rounded to the nearest step. Add bits to shrink the step; drop the sampling rate below twice the signal frequency and watch the reconstruction alias into a completely false waveform.
Full-scale range here is 0–10 V. Resolution is set by the number of bits: Δ = range ÷ 2ⁿ, and the worst-case quantisation error is half a step, ±Δ/2. Adding one bit halves the step; it never removes the error entirely. The sampling rate is a separate matter — it must exceed twice the highest signal frequency (the Nyquist criterion). Sample too slowly and the samples fit a lower-frequency wave that no downstream filtering can undo.