How does code rate affect BER performance?

Lower code rates provide more parity bits per information bit, giving the decoder more constraints to correct errors. At a target BER of 10^-6, a rate 1/3 LDPC code requires approximately 1.5 dB Eb/No while a rate 3/4 code requires approximately 4.5 dB Eb/No, a 3 dB difference. However, comparing at the same bandwidth (spectral efficiency), the rate 1/3 code needs 3x the bandwidth, and converting from Eb/No to SNR: SNR = Eb/No + 10*log10(R*m) where m is bits per symbol. A rate 1/3 code with QPSK (m=2) gives spectral efficiency of 0.67 bps/Hz, while rate 3/4 with QPSK gives 1.5 bps/Hz. The system designer must balance the error correction capability (lower rate) against spectral efficiency (higher rate) based on the available bandwidth and required reliability. In 5G NR, this trade-off is automated through adaptive modulation and coding (AMC), where the base station selects the highest MCS (code rate and modulation order) that achieves the target BLER of 10% based on CQI reports.

What is the relationship between code rate and Shannon capacity?

Shannon's channel capacity theorem states that reliable communication is possible at rates up to C = B * log2(1 + SNR) bits per second, where B is bandwidth and SNR is signal-to-noise ratio. For a coded system, the spectral efficiency is R * log2(M) bps/Hz where R is code rate and M is modulation order. The maximum code rate for error-free communication at a given SNR and modulation order is R_max = C / (B * log2(M)) = log2(1+SNR) / log2(M). At SNR = 10 dB (linear 10): C/B = log2(11) = 3.46 bps/Hz. With 16-QAM (log2(16) = 4): R_max = 3.46/4 = 0.865. With QPSK: R_max = 3.46/2 = 1.73, meaning QPSK cannot achieve this capacity even at rate 1 since its maximum spectral efficiency is 2 bps/Hz and the capacity exceeds the modulation limit. Modern LDPC codes in 5G NR operate within 0.5 to 1.5 dB of the Shannon limit for code block sizes above 1,000 bits, meaning they achieve BER below 10^-6 at SNR values only 0.5 to 1.5 dB above the theoretical minimum.

Digital Communications

Code Rate Detail

Q: How does rate matching work in 5G NR?

The LDPC encoder always produces coded bits at the mother code rate (1/3 for BG1, 1/5 for BG2), generating all systematic and parity bits. Rate matching then selects a subset of these coded bits to achieve the desired effective code rate, which can be higher than the mother rate (by puncturing parity bits) or lower (by repeating bits, though this is rare). The rate matching operates on a circular buffer containing all coded bits. For a target rate of 3/4, only k/(3/4) = 4k/3 coded bits are selected from the n = 3k available (for BG1), sending all k systematic bits plus k/3 parity bits. For incremental redundancy HARQ, subsequent retransmissions read from different positions in the circular buffer, effectively lowering the combined rate. The redundancy version (RV) index controls the starting position: RV0 starts at the beginning (systematic bits), RV2 at approximately 1/3 of the buffer, providing mostly new parity bits for code combining. This allows the system to adaptively lower the effective rate through retransmissions without re-encoding.

/kohd rayt/

Code rate R = k/n is the ratio of information bits (k) to total coded bits (n) in FEC. Rate 1/2 adds 100% redundancy. Lower rates: stronger correction but less spectral efficiency. 5G NR LDPC: 1/5 (BG2 mother) to 948/1024 (~0.926), selected via AMC based on CQI. Rate matching reads from circular buffer; HARQ retransmissions use different redundancy versions. Modern LDPC operates 0.5 to 1.5 dB from Shannon limit.

Category: Digital Communications

5G NR range: 1/5 to 0.926

Gap to Shannon: 0.5 to 1.5 dB

Understanding Code Rate

Code rate is the single most important parameter describing the trade-off between reliability and throughput in a digital communication system. A rate 1/2 code doubles the number of transmitted bits relative to the information content, consuming twice the bandwidth (or half the time) but providing substantial error correction capability. A rate 7/8 code adds only 14% overhead, maximizing throughput but offering limited error protection. The art of system design is selecting the right code rate for the channel conditions: too high and the decoder cannot correct errors; too low and bandwidth is wasted.

Modern wireless systems do not use a fixed code rate. Instead, adaptive modulation and coding (AMC) continuously adjusts both the modulation order and code rate based on channel quality feedback. In 5G NR, the UE reports Channel Quality Indicator (CQI) values that map to specific MCS indices, each defining a modulation order (QPSK, 16-QAM, 64-QAM, 256-QAM) and a target code rate. The MCS table covers 29 entries from MCS 0 (QPSK, rate ~0.12, spectral efficiency 0.23 bps/Hz) to MCS 28 (256-QAM, rate ~0.93, spectral efficiency 7.41 bps/Hz), a 32x range in spectral efficiency. The base station scheduler selects the highest MCS that achieves a target BLER of 10% (first transmission), relying on HARQ retransmissions to recover the 10% of failed blocks.

Code Rate Formulas

Code Rate:
R = k/n (0 < R ≤ 1)

Shannon Capacity Limit:
R_max = log₂(1 + SNR) / log₂(M)

SNR from Eb/No:
SNR = E_b/N₀ + 10 log₁₀(R × m) (dB)

Where k = info bits, n = coded bits, M = modulation order, m = log₂(M). At SNR=10 dB, 16-QAM: R_max = 0.865. Rate 1/2 LDPC requires ~1.5 dB E_b/N₀ for BER 10^-6; rate 3/4 requires ~4.5 dB.

5G NR MCS and Code Rate Examples

MCS Index	Modulation	Code Rate	Spectral Eff.	Use Case
MCS 0	QPSK	~0.12	0.23 bps/Hz	Cell edge, deep fade
MCS 9	QPSK	~0.60	1.21 bps/Hz	Moderate coverage
MCS 16	64-QAM	~0.55	3.32 bps/Hz	Good channel
MCS 24	256-QAM	~0.70	5.55 bps/Hz	Near cell center
MCS 28	256-QAM	~0.93	7.41 bps/Hz	Excellent SNR

Common Questions

Frequently Asked Questions

How does code rate affect BER?

Lower rate = more parity = better correction. Rate 1/3 needs ~1.5 dB E_b/N₀; rate 3/4 needs ~4.5 dB (for BER 10^-6). But rate 1/3 uses 3x bandwidth. System trade-off: AMC selects highest rate achieving target BLER 10% based on CQI. HARQ handles residual errors.

How does rate matching work in 5G NR?

LDPC encodes at mother rate (1/3 BG1, 1/5 BG2). Rate matching selects subset from circular buffer for target rate. Rate 3/4: send all k systematic + k/3 parity bits. HARQ retransmissions read from different buffer positions (RV0/1/2/3), lowering combined rate via code combining without re-encoding.

Relationship to Shannon capacity?

Shannon: C/B = log₂(1+SNR). At 10 dB, 16-QAM: R_max = 0.865. Modern 5G LDPC operates within 0.5 to 1.5 dB of this limit for blocks >1,000 bits. Practical codes cannot exceed Shannon limit; operating closer requires larger block sizes and more decoder iterations.

Related Terms

← Code Domain Noma Codebook →