Analogue-to-digital conversion: sampling, quantisation and resolution, the sampling theorem and aliasing, quantisation error, and the trade-off between resolution and data rate.

Levels (up to 2 marks): an $n$-bit converter has $2^n$ levels. For 8 bits, $2^8 = 256$ levels.

Resolution (up to 2 marks): the resolution is the full range divided by the number of steps: $\dfrac{5.0}{256} = 1.95 \times 10^{-2}\ \text{V} \approx 19.5\ \text{mV}$ per step (using $2^n - 1 = 255$ steps gives $19.6\ \text{mV}$, either is accepted).

Output code (up to 2 marks): the code is the input divided by the resolution, rounded: $\dfrac{2.0}{19.5 \times 10^{-3}} = 102.6$, so code $103$ (binary $0110\,0111$).

Markers reward $256$ levels, a resolution near $19.5\ \text{mV}$, and an output code of about $102$ or $103$.

Sampling theorem (up to 2 marks): to reconstruct a signal faithfully it must be sampled at a rate at least twice the highest frequency present in the signal (the Nyquist rate), $f_\text{sample} \ge 2 f_\text{max}$.

Aliasing (up to 2 marks): if the signal is sampled too slowly (below the Nyquist rate), high-frequency components are misread as lower frequencies that were not present, appearing as false (alias) signals that distort the reconstruction.

Prevention (up to 1 mark): pass the signal through a low-pass anti-aliasing filter before sampling to remove any frequencies above half the sampling rate, and/or sample faster.

Markers reward the at-least-twice sampling rate, aliasing as high frequencies appearing as false low frequencies, and the anti-aliasing low-pass filter (or faster sampling) as the cure.