Ubiquitous personal healthcare inspection (uPHI) in a wireless body area network (WBAN) requires continuous signal monitoring up to several days or even longer. To enable cable-free body monitoring with µW biomedical acquisition devices [1,2], a submW wireless transceiver is required for long-term observation. However, existing systems [3,4] have large power dissipation with poor interference-rejection capability, limiting device coexistence and the use of thin-film flexible-batteries. For better operation reliability in an environment with interference, the wireless medical telemetry service (WMTS) constraints defined in [5] are collectively considered. As a result, a multi-tone CDMA (MT-CDMA) baseband processor with a 31-chip spreading code is implemented in a wireless sensor node (WSN) and a central processing node (CPN), respectively, as shown in Fig. 20.5.1, for interference-free and reliable communications in body monitoring. In the battery-capacity-limited WSN, both dynamic and leakage power are optimized, comprising µW-scale power dissipation for extended body-signal monitoring. This power optimization is achieved using a power island approach with the techniques of multi-V t CMOS (MVTCMOS), multiple supply voltage (MSV) domains, and distributed coarse-grain power-gating cells (DCG-PGC), improving the operating duration by 11×. For better system performance, the CPN applies both adaptive I/Q-mismatch calibration (A-IQMC) and dynamic sampling-timing control (DSTC) with phase-tunable clock generation (PTCG) [6] to further improve system SNR by 2.2dB and reduce power dissipation in the ADC by 43.75%. Figure 20.5.1 shows the proposed uPHI-WSN and uPHI-CPN in a coexistence scenario. This transceiver pair may connect with various possible sensor categories and achieves a maximum datarate of 143kb/s for a maximum 10-WSN coexistence, providing large interference-rejection capability and better system reliability. Because uPHI-WSN/CPN communicates with a datarate much higher than a human body's sensor signal rate in this work, a WSN stays asleep for most of time (sleep-phase) and transmits data in a burst mode (active-phase). Therefore, leakage power minimization is emphasized for this long-sleep operation. When an ECG is considered as a body signal source with 16b samples at a sampling rate of 500-samples/s, a 2048×16b SRAM is designed into the WSN. This SRAM size has an optimum leakage saving efficiency, resulting in a 6.96% active-sleep duty cycle. The sampled signals from the sensor are modulated by QPSK with a 16-point IFFT, where the subcarrier allocation follows the conjugate-symmetric rule to further reduce by 50% the DAC and frontend circuit area and power. Then, the baseband signal is transmitted at 5M chip/s. In the CPN on the receiver side, the A-IQMC circuit is provided to compensate for gain and phase signal distortion. Also, a DSTC is applied for best sample timing search with the aid of the PTCG circuit, which provides 5MHz 8-phase evenly spaced clock phases for selection. Each clock phase is sepa...