QAM

QAM works by combining two amplitude-modulated signals 90° out of phase (in-phase "I" and quadrature "Q" components). Each combination of I and Q amplitude represents a unique symbol on a constellation diagram — a grid of points where the x-axis is the I amplitude and the y-axis is the Q amplitude. The number before "QAM" tells you how many distinct constellation points the scheme uses: 64-QAM has 64 points arranged in an 8×8 grid (6 bits each), 256-QAM has 256 points in a 16×16 grid (8 bits), 1024-QAM has 1024 points (10 bits). Higher QAM means more bits per symbol — more throughput for the same channel width. The trade-off: denser constellations require better signal quality to decode reliably, so higher QAM is only used when SNR is high.

How the constellation diagram works

In a QAM constellation diagram, each point represents one symbol — one transmission state that the radio can send. The receiver must determine which of the N points was sent based on the received signal. When the channel adds noise, the received signal is displaced from the true constellation point. If the displacement is small enough, the receiver correctly identifies the nearest point. If noise displaces the signal past the decision boundary between two adjacent points, a symbol error occurs, corrupting the bits that symbol encoded. The spacing between constellation points — the noise margin — decreases as the number of points increases for a fixed signal power. This is why 4096-QAM requires approximately 24 dB more SNR than 4-QAM to achieve the same error rate.

QAM orders and their bit efficiency

SNR requirement and why higher-order QAM is demanding

SNR (Signal-to-Noise Ratio) measures how much stronger the signal is than the background noise, expressed in decibels. Each doubling of the QAM order (from 256 to 1024, for example) adds 10 more constellation points per row and column, reducing the spacing between adjacent points. To maintain the same bit error rate, the SNR must increase by roughly 6 dB for each doubling. 4096-QAM, used in Wi-Fi 7 and DOCSIS 3.1, requires approximately 40 dB SNR — meaning the signal must be 10,000 times stronger than the noise floor. This is only achievable at very close range over a clean, interference-free channel, or over a controlled coaxial cable plant. Even minor interference from a neighbour's Wi-Fi or a microwave oven can push SNR below the threshold, causing the radio to drop back to a lower order automatically.

QAM in cable internet (DOCSIS)

Cable internet uses QAM on its downstream channels. DOCSIS 3.0 uses 256-QAM (8 bits/symbol) on individual 6 MHz or 8 MHz channels. DOCSIS 3.1 introduced OFDM channels up to 192 MHz wide and supports up to 4096-QAM (12 bits/symbol) on those channels. The coaxial cable plant is a controlled medium with predictable noise characteristics, making high-order QAM more reliable than over-the-air Wi-Fi. A DOCSIS 3.1 modem negotiating 4096-QAM on a 192 MHz OFDM channel can carry several gigabits per second downstream on a single channel — enabling multi-gigabit cable internet plans.

QAM in Wi-Fi generations

Each Wi-Fi generation has increased the maximum QAM order. Wi-Fi 4 (802.11n) introduced 64-QAM. Wi-Fi 5 (802.11ac) added 256-QAM, increasing peak throughput by 33% over Wi-Fi 4 for the same channel width and stream count. Wi-Fi 6 (802.11ax) added 1024-QAM, another 25% gain. Wi-Fi 7 (802.11be) adds 4096-QAM, a further 20% increase in bits per symbol. In practice, 4096-QAM in Wi-Fi 7 is only achievable at very close range — typically within 3–5 metres of the access point with excellent RF conditions. Its real value is in short-range high-density deployments like conference rooms and enterprise offices.

QAM in LTE and 5G

4G LTE initially used 64-QAM on the downlink and QPSK/16-QAM on the uplink. LTE-Advanced (LTE-A, 3GPP Release 10+) added 256-QAM downlink capability, contributing to peak throughputs exceeding 1 Gbps in ideal conditions. 5G NR (New Radio) supports 256-QAM on both downlink and uplink in FR1 (sub-6 GHz) bands. In mmWave (FR2) bands, the very short range and line-of-sight requirement mean high-order QAM is more consistently achievable. 5G's combination of 256-QAM, wider channels (up to 100 MHz in sub-6 GHz, 400 MHz in mmWave), and massive MIMO produces the multi-gigabit peak rates cited in 5G specifications.

Adaptive modulation

Wi-Fi and cellular radios use adaptive modulation and coding (AMC) to automatically select the highest QAM order the current signal quality supports. Move close to your router with a clear line of sight and the radio negotiates 1024-QAM. Walk to another room and it drops to 256-QAM or 64-QAM to maintain a reliable link. This is why Wi-Fi "speed" shown in your OS settings changes as you move around — it reflects the PHY rate at the current QAM order and channel width, not your internet plan speed.

Frequently Asked Questions

What does 1024-QAM mean in Wi-Fi 6?

1024 distinct signal states, each encoding 10 bits. Wi-Fi 5 used 256-QAM (8 bits/symbol), so Wi-Fi 6's 1024-QAM encodes 25% more bits per symbol. It requires very strong SNR — only achievable at close range with a clean signal.

Why does higher QAM require better signal quality?

More QAM states means the constellation points are packed closer together. Any noise can push a received signal point to an adjacent state, causing bit errors. Higher SNR is needed to keep the error rate acceptable — which is why radios fall back to lower QAM automatically in noisy or distant conditions.

Where is QAM used besides Wi-Fi?

DOCSIS cable internet (256-QAM on DOCSIS 3.0, up to 4096-QAM on DOCSIS 3.1), 4G LTE (64–256-QAM), 5G NR (256-QAM), digital cable TV (DVB-C), and DSL (QAM-based DMT). Essentially any high-speed digital radio or cable system.