Digital Communication
Digital communication questions cover sampling, quantisation, PCM, line coding, multiplexing, error performance and synchronization. You should be able to compute bit rates and explain why companding exists.
Core rules
- Sampling theorem: fs ≥ 2fm for a band-limited signal; practical telephony uses fs = 8 kHz for 4 kHz voice.
- PCM steps: sampling, quantisation, encoding. Quantisation introduces quantisation noise, reduced by more levels.
- PCM bit rate = fs × n, where n is bits per sample.
- Companding (A-law or μ-law) gives small signals more quantisation levels, improving perceived SNR.
- TDM interleaves samples from multiple channels into one frame; E1 carries 32 timeslots at 2.048 Mbit/s.
- Line codes such as NRZ, RZ, AMI and Manchester differ in DC content, bandwidth and clock recovery.
- BER depends on Eb/N0; eye diagrams show margin and ISI.
Worked example
A voice channel is sampled at 8 kHz and encoded with 8 bits per sample. Bit rate = 8000 × 8 = 64,000 bit/s = 64 kbit/s. Thirty such channels plus two signalling/framing timeslots give 32 × 64 = 2048 kbit/s, the E1 rate.
Traps that cost marks
- Using fs = 4 kHz instead of 8 kHz for telephony.
- Adding bits per sample instead of multiplying by sample rate.
- Confusing quantisation noise with channel noise.
- Assuming Manchester coding is bandlimited; it needs more bandwidth but is self-clocking.
- Forgetting that PCM output is a bit stream, not an analogue waveform.
Exam approach
Identify fs and n first, then multiply. For line-code questions, compare DC component, bandwidth and clock-recovery ability in a table.