Analysis of COEX WIFI and Bluetooth on NXP SOC

Abstract

Wireless local and personal-area networks provide complimentary services in the same unlicensed (UL) radio frequency band of operation. As the mutual benefits of utilizing these services become increasingly apparent, the likelihood of mutual interference may also increase. NXP Semiconductor is a market leader in the making of wifi and Bluetooth combine SOCs since past many years. The job role of an SQA in NXP is to validate the Bluetooth and Wifi profiles for different firmware and drivers on NXP SOC. This thesis presents a performance evaluation of coexisting wireless networks in the 2.4 GHz Industrial, Scientific, and Medical (ISM) frequency band. Specifically, the study focuses on the impact of Bluetooth on Wi-Fi throughput in different COEX scenarios. The study is categorized into three different sections, each evaluating the impact of different factors on Wi-Fi performance. The first scenario evaluates the throughput of different, Wi-Fi standards (IEEE 802.11n, 802.11ac, and 802.11ax) with Advance Audio Distribution Profile(A2DP) Sink COEX scenario. The second scenario investigates the effect of different Bluetooth profiles Advance Audio Distribution Profile(A2DP), Human Interface Devices (HID), Inquiry, and Bluetooth Low Energy Core Configuration (LE_CC) on Wi-Fi throughput when used with the 802.11n standard. Finally, the third scenario evaluates the effect of different combinations of Bluetooth profiles on Wi-Fi throughput. The results show significant degradation in Wi-Fi throughput in the 2.4 GHz band compared to the 5 GHz band, with degradation levels of 38% and 2%, respectively. Moreover, the presence of different Bluetooth profiles has a significant impact on Wi-Fi performance, with an average throughput degradation of 41% occurring when the Inquiry profile is on, and a minimum degradation of 10% achievable with the LE_CC profile. In multi-profile Bluetooth scenarios, the average throughput degradation is 60%, which is significantly higher than in single-profile scenarios (25%).