Ultra-wideband (UWB) wireless communication is prospering as a powerful partner of the Internet-of-things (IoT). Due to the ongoing development of UWB wireless communications, the demand for high-speed and medium resolution analog-to-digital converters (ADCs) continues to grow. The successive approximation register (SAR) ADCs are the most powerful candidate to meet these demands, attracting both industries and academia. In particular, recent time-interleaved SAR ADCs show that multi-giga sample per second (GS/s) can be achieved by overcoming the challenges of high-speed implementation of existing SAR ADCs. However, there are still critical issues that need to be addressed before the time-interleaved SAR ADCs can be applied in real commercial applications. The most well-known problem is that the time-interleaved SAR ADC architecture requires multiple sub-ADCs, and the mismatches between these sub-ADCs can significantly degrade overall ADC performance. And one of the most difficult mismatches to solve is the sampling timing skew. Recently, research to solve this timing-skew problem has been intensively studied. In this paper, we focus on the cutting-edge timing-skew calibration technique using a window detector. Based on the pros and cons analysis of the existing techniques, we come up with an idea that increases the benefits of the window detector-based timing-skew calibration techniques and minimizes the power and area overheads. Finally, through the continuous development of this idea, we propose a timing-skew calibration technique using a comparator offset-based window detector. To demonstrate the effectiveness of the proposed technique, intensive works were performed, including the design of a 7-bit, 2.5 GS/s 5-channel time-interleaved SAR ADC and various simulations, and the results prove excellent efficacy of signal-to-noise and distortion ratio (SNDR) and spurious-free dynamic range (SFDR) of 40.79 dB and 48.97 dB at Nyquist frequency, respectively, while the proposed window detector occupies only 6.5% of the total active area, and consumes 11% of the total power.