Digital Baseband Architecture and Simulated Low-IF FSK Receiver Performance

As the Internet of Things continues to grow in popularity, the implementation of low-power, low-noise RF demodulation for these types of devices is becoming increasingly necessary. One of the most prevalent modulation schemes used in these types of devices is Frequency-Shift Keying (FSK), which is utilized in IEEE 802.15.4 and Bluetooth Low-Energy. FSK demodulation is usually performed using matched filtering as a way to determine the received frequency after mixing the received RF signal down to some intermediate frequency (IF). In this paper, we demonstrate how the digital baseband architecture can influence the expected signal-to-noise ratio of the receiver. Through simulation, we explore the degree to which various parameters, such as IF and ADC sampling rate and bit depth, affect the expected signal-to-noise ratio of the matched filter implementation. We show that with specific parameter choices, the relative phase between the incoming signal and local oscillator has negligible impact on performance, obviating the need for phase alignment via a Phase-Locked Loop. Understanding these parametric spaces is critical for efficiently scoping future AIassisted design exploration.