RADAR & DFS - WHY IT MATTERS TO 802.11#
The 5 GHz UNII-2A (5250-5350 MHz, ch 52-64) and UNII-2C/2-Extended (5470-5725 MHz, ch 100-144) bands are shared with radar as the PRIMARY user. Wi-Fi is SECONDARY and must yield. IEEE 802.11h + regional rules (FCC 47 CFR 15.407, ETSI EN 301 893) require every device on these channels to run DFS: detect radar, stop transmitting, and move - protecting weather, air-traffic, military and maritime radar.
Escalation relevance: an unexplained mass client disconnect, a channel that silently changes, or an AP that "goes quiet for a minute at boot" on ch 52-144 is very often DFS - not a Wi-Fi bug. This sheet lets you tell a real DFS event from a false trigger, and map a detected waveform back to the radar that caused it. 6 GHz has NO DFS, which is a key reason to move high-density traffic there.
DFS CHANNELS (5 GHz) - WHERE RADAR RULES APPLY6 rows#
| Channel | Center (MHz) | UNII band | DFS? | CAC required | Typical radar here | Notes |
|---|---|---|---|---|---|---|
| 36-48 | 5180-5240 | UNII-1 | No | None | - | Non-DFS. Indoor/low-power. Safe from radar evacuation. |
| 52-64 | 5260-5320 | UNII-2A | Yes | 60 s standard | Military, some ASR | Standard DFS. 4 channels (52/56/60/64). |
| 100-144 | 5500-5720 | UNII-2C (2-Ext) | Yes | 60 s / 600 s | TDWR, ASR, ARSR, NEXRAD | Largest DFS block (12 ch). Contains weather-radar sub-band. |
| 120, 124, 128 | 5600/5620/5640 | UNII-2C | Yes (weather) | 600 s (10 min) | TDWR (terminal Doppler) | WEATHER-RADAR band 5600-5650 MHz. Extended 10-min CAC. Many vendors disable these by default. |
| 149-165 | 5745-5825 | UNII-3 | No | None | - | Non-DFS. Highest power (US). Preferred for reliability. |
| 169-177 | 5845-5885 | UNII-4 | No | None | - | FCC 2020 indoor low-power. Non-DFS. |
REAL-WORLD RADAR SYSTEMS ON DFS CHANNELS8 rows#
| Channel | Center (MHz) | UNII band | DFS? | CAC required | Typical radar here | Notes |
|---|---|---|---|---|---|---|
| Radar type | Channels (5 GHz) | Pulse width | PRF / PRI | Burst / scan pattern | Rotation / scan period | Threat to Wi-Fi |
| TDWR - Terminal Doppler Weather Radar | 116, 120, 124 (5600-5650 MHz) | ~1 µs | ~1000-2000 Hz (0.5-1 ms PRI) | Bursts of ~18-25 pulses, gap, repeat; can gate-average | ~1 min surveillance / longer volume scan | Critical |
| ASR - Airport Surveillance Radar | 108-128 (5500-5640 MHz) | 1-5 µs | 700-1200 Hz (~0.8-1.4 ms PRI) | Continuous pulse train as beam sweeps past | ~4.5-5 s per rotation | Critical |
| ARSR - Air Route Surveillance (en-route) | 100-136 (5470-5725 MHz) | 2-6 µs | ~250-500 Hz (2-4 ms PRI) | Long pulse bursts, widely spaced | ~10-12 s per rotation | High |
| NEXRAD / WSR-88D - NWS weather radar | 100-128 (5500-5640 MHz) | 1.57-4.5 µs | 318-1300 Hz (0.77-3.1 ms PRI) | Variable - switches VCP scan modes | ~4.5-10 min per volume scan | High |
| Military / defence radar (various) | Any 5 GHz DFS ch (5250-5350, 5470-5725) | 0.5-100 µs (variable) | Classified - staggered/agile | Staggered PRF, frequency-agile, hard to predict | Often non-rotating / electronically scanned | Critical |
| Ship-borne / marine radar | 100-124 (5470-5600 MHz) | 0.05-1 µs | 500-3000 Hz | Short dense burst as antenna sweeps | ~2-3 s per rotation | Moderate (coastal/port) |
| Doppler wind profiler | 100-128 (5470-5640 MHz) | 0.2-16 µs | 10-50 Hz (very slow) | Very long pulses, low repetition, points upward | Fixed beam - no rotation | Low |
FCC DFS TEST WAVEFORMS (FCC 06-96 / 13-22) - WHAT DEVICES MUST DETECT7 rows#
| Type | Category | Pulse width (µs) | PRI (µs) | Pulses / bursts | Min. detection prob. | Notes |
|---|---|---|---|---|---|---|
| Type 0 | Short pulse (fixed) | 1 | 1428 | 18 pulses | 60% (per trial) | The baseline test pulse. Used for channel-move-time and detection-bandwidth tests. (Was "Type 1" in FCC 06-96, renamed in 13-22.) |
| Type 1 | Short pulse (variable) | 1-5 | 518-3066 (test-B set) | ≥ per roll | 60% | Added in FCC 13-22. Wide PRI range; ≥30 unique waveforms. |
| Type 2 | Short pulse (variable) | 1-5 | 150-230 | 23-29 | 60% | ≥30 unique waveforms per trial set. |
| Type 3 | Short pulse (variable) | 6-10 | 200-500 | 16-18 | 60% | ≥30 unique waveforms. |
| Type 4 | Short pulse (variable) | 11-20 | 200-500 | 12-16 | 60% | ≥30 unique waveforms. Aggregate types 1-4 ≥ 80% (96/120). |
| Type 5 | Long pulse (frequency-modulated) | 50-100 (chirp) | 1000-2000 | 8-20 bursts × 1-3 pulses | 80% | Chirped long pulse. Radar sweeps frequency during the pulse. 12 s total duration. |
| Type 6 | Frequency hopping | 1 | 333 | 9 pulses/burst × 100 bursts | 70% | 100 hops across 5250-5724 MHz (475 freqs, no repeat). Hardest to catch. |
ETSI DFS TEST WAVEFORMS (EN 301 893) - EUROPE7 rows#
| Type | Category | Pulse width (µs) | PRI (µs) | Pulses / bursts | Min. detection prob. | Notes |
|---|---|---|---|---|---|---|
| Type | Pulse width (µs) | PRF (Hz) | Pulses / burst | Character | Min. detection prob. | Notes |
| Type 1 | 0.5-5 | 200-1000 | 10 | Constant PRF | 60% | Loosest test - 10 pulses. |
| Type 2 | 0.5-15 | 200-1600 | 15 | Constant PRF | 60% | |
| Type 3 | 0.5-15 | 2300-4000 | 25 | Constant PRF | 60% | |
| Type 4 | 20-30 | 2000-4000 | 20 | Constant PRF + chirp (±2.5 MHz) | 60% | Chirped. |
| Type 5 | 0.5-2 | 300-400 (×2-3 staggered) | 10 | Staggered PRF (2-3 values, Δ 20-50 Hz) | 60% | Staggered - mimics weather radar. |
| Type 6 | 0.5-2 | 400-1200 (×2-3 staggered) | 15 | Staggered PRF (Δ 80-400 Hz) | 60% | Wider stagger. |
DFS TIMING CONSTANTS - FCC vs ETSI7 rows#
| Type | Category | Pulse width (µs) | PRI (µs) | Pulses / bursts | Min. detection prob. | Notes |
|---|---|---|---|---|---|---|
| Parameter | What it is | FCC (US) | ETSI (EU) | Escalation meaning | ||
| CAC - Channel Availability Check | Passive listen before first use of a DFS channel; no TX, no clients. | 60 s (standard) | 60 s (standard) | The "AP goes quiet for a minute at boot / on channel change" - normal, not a fault. | ||
| Extended CAC (weather) | Longer CAC on weather-radar channels. | 600 s (10 min) on ch 120/124/128 | 600 s on 5600-5650 MHz | Ten-minute silent period. Why many disable ch 120-128. | ||
| Channel Move Time | Max time to stop TX + leave channel after detection. | ≤ 10 s | ≤ 10 s | Clients drop within 10 s of a real hit. | ||
| Channel Closing Transmission Time | Aggregate TX allowed during the move (to send CSA). | ≤ 200 ms + 60 ms control | ≤ 1 s | Time to announce the switch (CSA) to clients. | ||
| NOP - Non-Occupancy Period | How long a channel stays banned after a hit. | 30 min | 30 min | The channel "disappears" for 30 min. Repeated hits = channel unusable. | ||
| In-Service Monitoring | Continuous radar watch while serving clients. | Required | Required | Detection can fire any time, not just at boot. |
REAL RADAR vs FALSE DFS TRIGGER - HOW TO TELL#
FALSE-POSITIVE sources (common): poorly-shielded microwave ovens (2.4 GHz leak rarely, but some 5 GHz industrial heaters), certain FHSS/wireless-video/cordless bridges, non-compliant outdoor links, perimeter/security radar, and adjacent AP self-detection. A false trigger looks identical to a real one to the regulator - the AP must evacuate either way.
Signs it is a REAL radar: (1) correlates with a physical source - within ~35 km of an airport (ASR/TDWR), coast/port (marine), or a weather-radar site (NEXRAD); (2) repeats on the SAME channels that map to that radar in the catalogue above; (3) periodicity matches a rotation (e.g. hits every ~4.5 s = ASR sweep, or bursts consistent with TDWR).
Signs it may be FALSE: hits on random DFS channels with no nearby radar source; strongly correlated with a specific piece of on-site equipment powering on; only one AP in a cluster ever triggers (localized RF source). Enhanced/Enterprise DFS (EDFS) and vendor "radar signatures" reduce but do not eliminate these.
Data to collect for an escalation: AP model + regulatory domain, exact channel(s) and time of each DFS event, whether it was CAC-time or in-service, the vendor radar-type classification if logged (maps to the FCC/ETSI type tables above), site geography (airport/coast/weather-radar proximity), and whether one AP or many detected it.
DFS OPERATIONS & MITIGATIONS7 rows#
| Type | Category | Pulse width (µs) | PRI (µs) | Pulses / bursts | Min. detection prob. | Notes |
|---|---|---|---|---|---|---|
| Feature / practice | What it does | Trade-off / note | ||||
| Zero-Wait DFS (ZW-DFS) | A spare radio pre-clears (pre-CACs) DFS channels so the move is instant. | Needs a dedicated/scanning radio; not all APs have it. Removes the 60 s outage on move. | ||||
| Disable weather channels (120/124/128) | Exclude the 600 s-CAC channels from the auto pool. | Loses 3 channels but avoids 10-min outages. Very common default. | ||||
| Exclude all DFS (UNII-1/3/4 + 6 GHz only) | Run only non-DFS channels. | Zero radar risk; fewer/narrower channels in 5 GHz. Good for real-time/voice or radar-dense sites. | ||||
| Prefer 6 GHz for capacity | Move high-density/wide-channel traffic to 6 GHz (no DFS). | Requires Wi-Fi 6E/7 clients. Best long-term answer to DFS pain. | ||||
| DFS-event monitoring | Track DFS hits per AP over time (e.g. syslog / dashboard API). | High event count at one location = move that AP off DFS. Key operational signal. | ||||
| Band-steer DFS-incapable clients | Keep legacy/IoT clients (no DFS support) on non-DFS or 2.4 GHz. | Some older clients cannot join DFS channels at all - appears as "can see SSID, cannot connect". |
RELATED SHEETS#
See "Channel Plans" for the full per-region channel/DFS map, "FCC Part 15 UNII Rules" for the US power/DFS rules per band, "Regulatory Domains" for FCC/ETSI/ARIB differences, "Advanced Regulatory" for corner cases, and "802.11 Frame Types" for the Channel Switch Announcement (CSA) action frame the AP sends when it evacuates.
Source basis: FCC 47 CFR 15.407, FCC 06-96 / FCC 13-22 & KDB 905462 (test waveforms), ETSI EN 301 893, IEEE 802.11h. Real-world radar parameters are representative/typical (some military values are classified); treat them as field guidance, not exact per-installation specs.