Loading content...
Marketing materials advertise WiFi routers with impressive numbers: "3,000 Mbps!", "WiFi 6E AX11000!". But what do these numbers actually mean? Why does your file transfer achieve only a fraction of the advertised speed? And how can you predict the actual throughput a deployment will deliver?
WiFi data rates are not straightforward. Multiple factors interact: channel bandwidth, modulation scheme, coding rate, spatial streams, guard intervals, and protocol overhead. Understanding these components is essential for capacity planning, troubleshooting, and setting realistic performance expectations.
This page dissects WiFi data rates from first principles. We'll build up from individual OFDM symbol calculations to complete MCS (Modulation and Coding Scheme) tables, then examine the overhead factors that reduce PHY rates to actual application throughput.
By the end of this page, you will: (1) Calculate PHY data rates from fundamental parameters, (2) Navigate MCS tables for 802.11n/ac/ax, (3) Understand the gap between PHY rate and application throughput, and (4) Estimate real-world performance for deployment planning.
All modern WiFi (802.11a/g/n/ac/ax) uses OFDM (Orthogonal Frequency-Division Multiplexing). To understand data rates, we must understand OFDM's structure.
OFDM Symbol Structure:
An OFDM channel is divided into parallel subcarriers, each carrying an independent modulated signal. The key parameters:
802.11a/g/n/ac (20 MHz channel):
802.11ax (20 MHz channel):
Data Rate Formula:
The fundamental data rate calculation is:
Data Rate = (Data Subcarriers × Bits per Symbol × Coding Rate × Spatial Streams) / Symbol Duration
Example: 802.11ac, 80 MHz, 256-QAM, 5/6 coding, 2 spatial streams, short GI
Data Rate = (234 × 8 × 5/6 × 2) / 3.6 μs
= (234 × 8 × 0.833 × 2) / 0.0000036 s
= 3120 / 0.0000036
= 866.7 Mbps
This matches MCS 9 at 80 MHz with 2 spatial streams—a common configuration for mid-range 802.11ac devices.
| Channel Width | 802.11ac Data Subcarriers | 802.11ax Data Subcarriers | 802.11ax Improvement |
|---|---|---|---|
| 20 MHz | 52 | 234 | 4.5× |
| 40 MHz | 108 | 468 | 4.3× |
| 80 MHz | 234 | 980 | 4.2× |
| 160 MHz | 468 | 1960 | 4.2× |
Guard Interval (GI) Impact:
The guard interval is 'dead time' between OFDM symbols, preventing multipath-induced inter-symbol interference. Shorter GI increases throughput but requires cleaner RF environment:
Standard vs. Short Guard Interval:
| Standard | Normal GI | Short GI | Throughput Increase |
|---|---|---|---|
| 802.11n | 800 ns | 400 ns | +11% |
| 802.11ac | 800 ns | 400 ns | +11% |
| 802.11ax | 800 ns | 400 ns | +6% |
| 802.11ax | 1600 ns | 800 ns | +11% |
| 802.11ax | 3200 ns | 1600 ns | +6% |
802.11ax introduces multiple GI options (0.8, 1.6, 3.2 μs) for different deployment scenarios. The longer symbol duration already provides multipath resilience, so normal GI (0.8 μs) is often practical.
Short guard interval works reliably in low-multipath environments: open office spaces, outdoor deployments, and close-range connections. Enable it for maximum throughput but monitor for retransmissions—elevated retry rates may indicate multipath-related errors where normal GI would be more reliable.
Modulation determines how many bits each subcarrier carries per symbol. Higher-order modulation packs more bits but requires better signal-to-noise ratio (SNR).
WiFi Modulation Hierarchy:
BPSK (Binary Phase Shift Keying):
QPSK (Quadrature Phase Shift Keying):
16-QAM:
64-QAM:
256-QAM (802.11ac/ax):
1024-QAM (802.11ax):
4096-QAM (802.11be/WiFi 7):
| Modulation | Bits/Symbol | Constellation Points | Min SNR | Typical Range* |
|---|---|---|---|---|
| BPSK | 1 | 2 | ~6 dB | 100% |
| QPSK | 2 | 4 | ~9 dB | 85% |
| 16-QAM | 4 | 16 | ~15 dB | 65% |
| 64-QAM | 6 | 64 | ~21 dB | 45% |
| 256-QAM | 8 | 256 | ~27 dB | 25% |
| 1024-QAM | 10 | 1024 | ~33 dB | 15% |
| 4096-QAM | 12 | 4096 | ~39 dB | 10% |
*Range percentages are approximate, relative to BPSK range under identical conditions.
The SNR Reality:
Each 3 dB of additional SNR requirement approximately halves the usable range (due to inverse-square power law). Thus:
This explains why maximum data rates are only achievable at close range, while devices automatically fall back to lower modulation as distance increases.
WiFi devices continuously monitor signal quality and adjust modulation/coding dynamically. This 'rate adaptation' happens on a per-frame basis in many implementations. A client might transmit one frame at MCS 9 (256-QAM) and the next at MCS 4 (16-QAM) if conditions temporarily degrade. Rate adaptation algorithms are vendor-specific and can significantly impact real-world performance.
Coding Rate and Error Correction:
Forward Error Correction (FEC) adds redundancy to detect and correct transmission errors. The coding rate indicates the proportion of bits that carry actual data:
| Coding Rate | Data Bits | Redundancy | Error Resilience |
|---|---|---|---|
| 1/2 | 50% | 50% | Highest |
| 2/3 | 67% | 33% | High |
| 3/4 | 75% | 25% | Medium |
| 5/6 | 83% | 17% | Low |
Lower coding rates (more redundancy) are used with higher-order modulation to maintain reliability. For example:
The MCS index encodes the specific combination of modulation and coding rate.
The MCS (Modulation and Coding Scheme) index is WiFi's way of specifying the exact combination of modulation, coding rate, and spatial streams. Understanding MCS tables is essential for interpreting diagnostic data and planning capacity.
802.11n MCS Structure:
802.11n uses a linear MCS index (0-31) where:
| MCS | Modulation | Coding | 20 MHz (800ns GI) | 20 MHz (400ns GI) | 40 MHz (800ns GI) | 40 MHz (400ns GI) |
|---|---|---|---|---|---|---|
| 0 | BPSK | 1/2 | 6.5 Mbps | 7.2 Mbps | 13.5 Mbps | 15 Mbps |
| 1 | QPSK | 1/2 | 13 Mbps | 14.4 Mbps | 27 Mbps | 30 Mbps |
| 2 | QPSK | 3/4 | 19.5 Mbps | 21.7 Mbps | 40.5 Mbps | 45 Mbps |
| 3 | 16-QAM | 1/2 | 26 Mbps | 28.9 Mbps | 54 Mbps | 60 Mbps |
| 4 | 16-QAM | 3/4 | 39 Mbps | 43.3 Mbps | 81 Mbps | 90 Mbps |
| 5 | 64-QAM | 2/3 | 52 Mbps | 57.8 Mbps | 108 Mbps | 120 Mbps |
| 6 | 64-QAM | 3/4 | 58.5 Mbps | 65 Mbps | 121.5 Mbps | 135 Mbps |
| 7 | 64-QAM | 5/6 | 65 Mbps | 72.2 Mbps | 135 Mbps | 150 Mbps |
802.11ac/ax MCS Structure:
802.11ac introduced a cleaner MCS structure where index is independent of spatial streams:
| MCS | Modulation | Coding | 20 MHz | 40 MHz | 80 MHz | 160 MHz |
|---|---|---|---|---|---|---|
| 0 | BPSK | 1/2 | 6.5 Mbps | 13.5 Mbps | 29.3 Mbps | 58.5 Mbps |
| 1 | QPSK | 1/2 | 13 Mbps | 27 Mbps | 58.5 Mbps | 117 Mbps |
| 2 | QPSK | 3/4 | 19.5 Mbps | 40.5 Mbps | 87.8 Mbps | 175.5 Mbps |
| 3 | 16-QAM | 1/2 | 26 Mbps | 54 Mbps | 117 Mbps | 234 Mbps |
| 4 | 16-QAM | 3/4 | 39 Mbps | 81 Mbps | 175.5 Mbps | 351 Mbps |
| 5 | 64-QAM | 2/3 | 52 Mbps | 108 Mbps | 234 Mbps | 468 Mbps |
| 6 | 64-QAM | 3/4 | 58.5 Mbps | 121.5 Mbps | 263.3 Mbps | 526.5 Mbps |
| 7 | 64-QAM | 5/6 | 65 Mbps | 135 Mbps | 292.5 Mbps | 585 Mbps |
| 8 | 256-QAM | 3/4 | 78 Mbps | 162 Mbps | 351 Mbps | 702 Mbps |
| 9 | 256-QAM | 5/6 | 86.7 Mbps | 180 Mbps | 390 Mbps | 780 Mbps |
802.11ax (WiFi 6) MCS Additions:
802.11ax extends the MCS table with 1024-QAM options and modified symbol structure:
| MCS | Modulation | Coding | 20 MHz | 40 MHz | 80 MHz | 160 MHz |
|---|---|---|---|---|---|---|
| 0 | BPSK | 1/2 | 8.6 Mbps | 17.2 Mbps | 36 Mbps | 72.1 Mbps |
| 1 | QPSK | 1/2 | 17.2 Mbps | 34.4 Mbps | 72.1 Mbps | 144.1 Mbps |
| 2 | QPSK | 3/4 | 25.8 Mbps | 51.6 Mbps | 108.1 Mbps | 216.2 Mbps |
| 3 | 16-QAM | 1/2 | 34.4 Mbps | 68.8 Mbps | 144.1 Mbps | 288.2 Mbps |
| 4 | 16-QAM | 3/4 | 51.6 Mbps | 103.2 Mbps | 216.2 Mbps | 432.4 Mbps |
| 5 | 64-QAM | 2/3 | 68.8 Mbps | 137.6 Mbps | 288.2 Mbps | 576.5 Mbps |
| 6 | 64-QAM | 3/4 | 77.4 Mbps | 154.9 Mbps | 324.3 Mbps | 648.5 Mbps |
| 7 | 64-QAM | 5/6 | 86 Mbps | 172.1 Mbps | 360.3 Mbps | 720.6 Mbps |
| 8 | 256-QAM | 3/4 | 103.2 Mbps | 206.5 Mbps | 432.4 Mbps | 864.7 Mbps |
| 9 | 256-QAM | 5/6 | 114.7 Mbps | 229.4 Mbps | 480.4 Mbps | 960.8 Mbps |
| 10 | 1024-QAM | 3/4 | 129 Mbps | 258.1 Mbps | 540.4 Mbps | 1080.9 Mbps |
| 11 | 1024-QAM | 5/6 | 143.4 Mbps | 286.8 Mbps | 600.5 Mbps | 1201 Mbps |
Even at identical MCS and bandwidth, 802.11ax achieves higher rates than 802.11ac. This comes from the increased subcarrier density (4× more data subcarriers) despite the longer symbol duration. Compare MCS 9 at 80 MHz: 802.11ac = 390 Mbps, 802.11ax = 480.4 Mbps—a 23% increase before considering 1024-QAM.
MIMO (Multiple-Input Multiple-Output) enables multiple simultaneous data streams between AP and client. Each additional spatial stream multiplies the peak data rate.
Spatial Stream Requirements:
Both transmitter and receiver must support the same number of spatial streams:
Device Classifications:
| Client Type | Typical MIMO Config | Spatial Streams |
|---|---|---|
| Basic IoT/Sensors | 1×1 | 1 SS |
| Budget smartphones | 1×1 or 2×2 | 1-2 SS |
| Mid-range phones/laptops | 2×2 | 2 SS |
| Premium laptops | 2×2 or 3×3 | 2-3 SS |
| High-end workstations | 4×4 | 4 SS |
| Enterprise APs | 4×4 to 8×8 | 4-8 SS |
| High-density APs | 8×8 or 12×12 | 8-12 SS (MU-MIMO) |
| Spatial Streams | MCS 9 Rate | Typical Devices |
|---|---|---|
| 1 SS | 433.3 Mbps | Entry smartphones, IoT |
| 2 SS | 866.7 Mbps | Most smartphones, laptops |
| 3 SS | 1300 Mbps | Premium laptops |
| 4 SS | 1733.3 Mbps | Desktop, enterprise APs |
The MIMO Reality Check:
Marketing materials advertise maximum rates assuming maximum spatial streams. Real-world observations:
Most clients are 2×2: Enterprise surveys show 80-90% of devices are 2-stream capable. Planning for 2 SS is reasonable for capacity.
MIMO degrades with distance: As SNR decreases, MIMO becomes less effective. A 4×4 system at range may perform no better than 2×2 at close range.
Antenna diversity vs. MIMO: Extra antennas can provide diversity gain (improved reliability) without additional spatial streams. A 3×3:2 configuration has 3 antennas but only 2 streams.
MIMO requires multipath: Spatial multiplexing depends on distinct radio paths. In line-of-sight outdoor environments, MIMO gains may be minimal.
MU-MIMO (Multi-User MIMO) lets APs use spatial streams for different clients simultaneously. An 8×8 AP might serve four 2×2 clients concurrently, each getting 2 spatial streams. This increases aggregate capacity but doesn't increase individual client speeds. MU-MIMO is most effective when clients are in spatially distinct directions from the AP.
The data rates in MCS tables are PHY rates—the raw bit rate at the physical layer. Actual application throughput is significantly lower due to multiple overhead sources:
Major Overhead Categories:
Frame Headers and Trailers (~5-10%)
Medium Access Protocol (~15-25%)
Encryption Overhead (~2-5%)
IP/TCP Headers (~5-10%)
Contention and Collisions (~5-20%+)
Calculating Realistic Throughput:
A practical rule of thumb:
Actual TCP Throughput ≈ PHY Rate × 0.50 to 0.65 (ideal conditions)
Actual TCP Throughput ≈ PHY Rate × 0.30 to 0.45 (typical enterprise)
Actual TCP Throughput ≈ PHY Rate × 0.15 to 0.30 (congested/edge coverage)
Example:
| Scenario | PHY Rate | Expected Throughput | Efficiency |
|---|---|---|---|
| Close range, clear channel | 866.7 Mbps | 450-550 Mbps | 52-63% |
| Medium range, moderate traffic | 400-600 Mbps | 150-300 Mbps | 38-50% |
| Edge coverage, high density | 100-300 Mbps | 30-100 Mbps | 30-33% |
| Mixed 802.11n/ac network | Variable | 20-40% reduction | — |
Frame aggregation (A-MPDU/A-MSDU) dramatically improves efficiency by amortizing protocol overhead across multiple data frames. When aggregation is working well (large file transfers, video streaming), efficiency approaches 70%. With small packets (VoIP, gaming), aggregation provides little benefit, and efficiency may be only 30-40%.
802.11ax Efficiency Improvements:
802.11ax specifically targets overhead reduction:
802.11ax Efficiency Comparison:
| Traffic Type | 802.11ac Efficiency | 802.11ax Efficiency | Improvement |
|---|---|---|---|
| Large file transfer | ~60% | ~70% | +17% |
| Web browsing | ~40% | ~55% | +38% |
| VoIP | ~25% | ~45% | +80% |
| IoT sensors | ~15% | ~35% | +133% |
The gains are most dramatic for small-packet, bursty traffic where legacy WiFi overhead was most punishing.
Consumer router marketing uses aggregate numbers that can mislead buyers. Understanding these numbers prevents unrealistic expectations.
The "AX11000" Example:
A router advertised as "AX11000" might have:
Reality Check:
No single device uses all bands simultaneously. A client connects to ONE band at a time.
160 MHz channels are rare. Most deployments use 80 MHz due to interference.
1024-QAM requires perfect conditions. Most clients operate at 256-QAM or lower.
4×4 clients are uncommon. Most devices are 2×2, halving the maximum rate.
Actual throughput is 40-60% of PHY rate. Protocol overhead is unavoidable.
More Realistic Assessment:
A consumer connecting a 2×2 laptop to an "AX11000" router typically achieves:
Still excellent, but far from "11 Gbps".
| Marketing Claim | Theoretical Max | Realistic Single-Client | Typical User Experience |
|---|---|---|---|
| AC1200 | 1200 Mbps combined | 866 Mbps PHY (5 GHz) | 300-400 Mbps actual |
| AC1900 | 1900 Mbps combined | 1300 Mbps PHY (5 GHz) | 500-650 Mbps actual |
| AX3000 | 3000 Mbps combined | 2402 Mbps PHY (5 GHz) | 800-1100 Mbps actual |
| AX6000 | 6000 Mbps combined | 4804 Mbps PHY (5 GHz) | 1000-1500 Mbps actual* |
| AX11000 | 11000 Mbps combined | 4804 Mbps PHY (5 GHz) | 1000-1500 Mbps actual* |
*Requires 160 MHz channel, 4×4 client, and ideal conditions.
Enterprise vs. Consumer Specifications:
Enterprise specifications tend to be more honest:
You now understand WiFi data rates from first principles: how OFDM parameters, modulation schemes, coding rates, and spatial streams combine to determine PHY rates, and how protocol overhead reduces those to actual throughput. The next page examines range considerations—how distance, obstacles, and environmental factors affect delivered performance.