Design solutions that pursue the sound quality and experience you need
The global 2.4 GHz ISM band spans 2400–2483.5 MHz with 83.5 MHz total bandwidth, shared by Wi-Fi, Bluetooth, ZigBee, wireless keyboards, and even microwave ovens. Wi-Fi typically occupies three non-overlapping channels (1, 6, 11), each 22 MHz wide, while Bluetooth hops across 79 channels of 1 MHz. In dense office environments TWS earbuds collide with neighbouring Wi-Fi APs and routers, causing audio packet loss, control failures, and occasional short-range drops. On the Liwei dual-mode reference platform, in-house coexistence testing shows Bluetooth PER climbing from 1% baseline to over 8% when a neighbour AP saturates Wi-Fi channel 1. This is the failure mode Adaptive Frequency Hopping (AFH) is designed to mitigate.
AFH was introduced in the Bluetooth 1.2 specification and partitions the 79 hopping channels into three buckets: good, bad, and unknown. The piconet master and slave exchange an AFH channel map through LMP (Link Manager Protocol), marking interfered channels as "bad" and excluding them from the hop sequence. Channel classification is driven by Packet Error Rate (PER) and Received Signal Strength (RSSI), with the PER threshold typically set between 1% and 3%. The channel map is not static: the link periodically re-evaluates every few seconds, and if more than 30% of channels are flagged bad, an AFH remap triggers. Liwei's dual-mode solution sets the re-evaluation period at 4–5 seconds, balancing responsiveness against protocol-stack overhead. Compared with the early 79-hop equal-probability scheme, AFH stabilizes the audio stream in noisy RF environments without sacrificing hop randomness on clean channels.
A 2.4 GHz Wi-Fi deployment on channels 1/6/11 overlaps with Bluetooth channels by over 70%. Two coexistence strategies dominate: AFH for frequency-domain sharing and PTA (Packet Traffic Arbitration) for time-domain sharing. The Liwei reference design on the JieLi AC7006 dual-mode SoC uses a hardware PTA pin: when Bluetooth requests transmission, it asserts the priority line to request Wi-Fi to back off; the reverse path also exists. Time-sliced arbitration drops measured conflict from roughly 15% to under 2% in our lab. PTA is more aggressive than AFH alone and is preferred for dense Wi-Fi deployments. In a combined mode, AFH removes the ±10–20 channels occupied by the Wi-Fi centre frequency, and PTA arbitrates the remaining time slots; private 2.4G low-latency links benefit from the same PTA grant to achieve stable 20 ms timing.
AFH solves protocol-layer contention; physical-layer isolation depends on PCB layout. The antenna keep-out zone should be 3–5 mm and free of metal shells, with the feed line tuned to 50 Ω and S11 better than −10 dB. When Wi-Fi and Bluetooth coexist on the same board, mount their antennas at least 15 mm apart to maximize spatial diversity gain. On the Liwei dual-mode headset reference, with Wi-Fi RSSI around −50 dBm, enabling AFH reduced Bluetooth audio PER from 8% to below 1% in production testing. Power rail ripple is held under 50 mV; the antenna-matching LC uses 0402 high-precision components with a slightly damped Q to broaden the working bandwidth above 100 MHz, which helps AFH jump between channels without impedance drift. The bottom-layer ground plane must remain continuous—any split creates a return-path discontinuity that detunes both antennas.
Q: Does AFH noticeably increase Bluetooth audio latency?
A: No. AFH operates at the channel-management layer and never enters the audio data path. Hop-sequence switching is handled by the baseband in 250 µs slots. Perceived audio latency is dominated by codec buffering, not AFH. On the Liwei dual-mode platform the measured end-to-end audio latency stays under 80 ms, with under ±2 ms variation between AFH active and inactive—well within competitive-gaming-grade limits.
Q: How does a Bluetooth headset detect which channels are occupied by Wi-Fi?
A: Through LMP exchanges. The slave tallies CRC failures per channel and converts them to PER, reporting back to the master. Combined with RSSI, the master decides whether to flag a channel as bad, then pushes an updated AFH channel map to the slave. The mechanism is fully automatic with no user intervention required, and when the Wi-Fi AP switches channels, AFH typically converges within 5 seconds.
Q: How effective is AFH against microwave-oven or wireless-mouse interference?
A: Microwave ovens emit wideband noise across the 2.4 GHz band. AFH can almost entirely exclude the polluted channels, allowing the audio stream to fall back onto the clean portion of the band. Wireless mice typically hop in the 2476–2483 MHz region, partially overlapping Bluetooth's higher channels but rarely the centre band. Liwei's reference design concentrates Bluetooth's primary hop channels between index 2 and 60, deliberately away from the ISM band edge, leaving the cleanest mid-band spectrum reserved for audio traffic. Lab tests with an active microwave oven showed almost no audible artefacts on the audio stream after AFH convergence.
Shenzhen Liwei Electronics Technology Co., Ltd has specialized in audio headset electronic solution design for 12 years, serving 15+ brand customers and 300+ projects, with deep experience tuning AFH/PTA coexistence on Bluetooth and 2.4G dual-mode platforms. For production-line coexistence data or customised reference designs for Bluetooth and Wi-Fi coexistence, contact the Liwei engineering team.
Keywords: 2.4GHz coexistence, AFH adaptive frequency hopping, Bluetooth anti-interference, Wi-Fi coexistence, TWS headset RF design
