Minglei Zhang

Dr. Minglei Zhang


Research Focus & Metrics

Academic Career Length

13Years (2014-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 February 2020 in IEEE Journal of Solid-State Circuits
Verified
Journal Article
Closed access
WJWenning Jiang
Wenning Jiang
YZYan Zhu
Yan Zhu
MZMinglei Zhang
Minglei Zhang
CCChi-Hang Chan
Chi-Hang Chan
RMRui Paulo Martins
Rui Paulo Martins
A temperature-stabilized 12-bit single-channel successive approximation register (SAR)-assisted pipelined analog-to-digital converter (ADC) running at 1 GS/s with Nyquist signal to noise and distortion ratio (SNDR) above 60 dB is presented. The ADC uses a three-stage (4 b-4 b-6 b) SAR-assisted pipeline hybrid architecture to achieve an attractive energy efficiency along with an extended sampling rate. A high-linearity open-loop Gm-R-based residue amplifier (RA) with both complete-settled and dynamic features improves the residue amplification efficiency and speed, while reducing the gain variation over a temperature drift. The inter-stage gain variation over the temperature is compensated through complementary temperature coefficients (TCs) from the inner devices of the RA. Furthermore, a cascade amplification topology in the backend RA alleviates the effect of the input parasitic capacitance to its front-end capacitor DAC (CDAC), thus leading to a small CDAC size to accelerate amplification and conversion. The prototype ADC was fabricated in a 28-nm CMOS process and consumes 7.6 mW from a 1-V power supply at 1 GS/s. The measured inter-stage gain variation is less than 2.3% with a temperature range from 0 °C to 80 °C. The SNDR and SFDR are 60 and 74.6 dB with a Nyquist input, respectively, achieving a Walden figure-of-merit (FoM) of 9.3 fJ/conversion-step and a Schreier FoM of 168.2 dB.

25 September 2019 in IEEE Journal of Solid-State Circuits
Verified
Journal Article
Closed access
MZMinglei Zhang
Minglei Zhang
CCChi-Hang Chan
Chi-Hang Chan
YZYan Zhu
Yan Zhu
RMRui P. Martins
Rui P. Martins
This article presents a low power-supplied 13-bit 20-MS/s time-to-digital converter (TDC)-assisted successive approximation register (SAR) analog-to-digital converter (ADC). In this hybrid architecture, the voltage-to-time converter (VTC) and TDC realize an inherent process, voltage, and temperature (PVT) robustness by inner tracking, thus inducing no extra power and circuit overheads. The voltage-domain and time-domain speed-enhanced techniques accelerate the first- and second-stage ADC conversions under a low power supply, respectively. Furthermore, in cooperation with a detect-and-skip switching scheme in the SAR ADC and an offset bit-shifting scheme in the two-step TDC, the ADC achieves linearity of 13 bit. The prototype ADC was fabricated in a 65-nm CMOS process with a 0.6-V power supply, achieving a 71.0-dB signal-to-noise and distortion ratio (SNDR) and an 89.5-dB spurious-free dynamic range with a Nyquist input at 20 MS/s, while exhibiting a Schreier figure-of-merit (FoM) of 181.9 dB. The ADC presents a sub-1-dB SNDR drop across a temperature range of -50 to 90 °C and ±5% power-supply variation.

11 August 2020 in IEEE Journal of Solid-State Circuits
Verified
Journal Article
Closed access
MZMinglei Zhang
Minglei Zhang
YZYan Zhu
Yan Zhu
CCChi-Hang Chan
Chi-Hang Chan
RMRui P. Martins
Rui P. Martins
This article presents an 8-bit time-domain analog-to-digital converter (ADC) that achieves 10 GS/s by aggregating only four time-interleaved channels. It also experiences less than 3.0-dB signal-to-noise and distortion ratio (SNDR) drop at an 18-GHz input frequency from a dc input due to its small input capacitance and inherent voltage-to-time converter (VTC)-based sub-channel buffer. A 16x time-interpolation-based time-to-digital converter (TDC) resolves in two steps while allowing both the inter-stage gain and the quantization step to be free from calibration over process, supply voltage, and temperature (PVT) variations. Furthermore, through a timing-extended residue transfer scheme, the metastability error rate is suppressed to <10⁻⁸. Fabricated in a 65-nm CMOS process, the prototype ADC achieves a 40.1-dB SNDR for a Nyquist input signal at 10 GS/s while consuming 50.8 mW from a 1.0-V power supply, yielding a Walden figure-of-merit of 61.5 fJ/conversion-step.

7 September 2017 in IEEE Journal of Solid-State Circuits
Verified
Journal Article
Closed access
MZMinglei Zhang
Minglei Zhang
KNKyoohyun Noh
Kyoohyun Noh
XFXiaohua Fan
Xiaohua Fan
ESEdgar Sanchez-Sinencio
Edgar Sanchez-Sinencio
This paper presents an energy-efficient 13-bit 10-50 MS/s subranging pipelined-successive approximation register (SAR) analog-to-digital converter (ADC) with power supply scaling. In the presented ADC, an SAR-assisted subranging floating capacitive DAC switching algorithm reduces switching energy along with enhanced linearity and speed in the first-stage SAR ADC. A following temperature-insensitive time-based residue amplifier realizes open-loop residual amplification without background calibration, while maintaining the benefits of dynamic operation and noise filtering. Furthermore, asynchronous SAR control logic employs a pre-window technique to accelerate SAR logic operations. The prototype ADC was fabricated in a 130-nm CMOS process with an active area of 0.22 mm². With a 1.2-V power supply and a Nyquist frequency input, the ADC consumes 1.32 mW at 50 MS/s and achieves signal-to-noise and distortion ratio and spurious-free dynamic range of 69.1 and 80.7 dB, respectively. The operating speed is scalable from 10 to 50 MS/s with a scalable power supply range of 0.8-1.2 V. Walden FoMs of 4-11.3 fJ/conversion-step are achieved.

22 June 2018 in IEEE Transactions on Very Large Scale Integration (VLSI) Systems
Verified
Journal Article
Closed access
KNKyoohyun Noh
Kyoohyun Noh
JAJudy Amanor-Boadu
Judy Amanor-Boadu
MZMinglei Zhang
Minglei Zhang
ESEdgar Sanchez-Sinencio
Edgar Sanchez-Sinencio
This paper presents low-power circuit techniques for comparators and controllers to reduce increasing power loss of an active diode with offset/delay calibration in a 13.56-MHz CMOS active rectifier. Unlike a conventional switched-offset comparator, the proposed voltage-mode switched-offset comparator reuses its bias current to generate continuously adjustable offset voltage, thereby reducing power loss. Offset calibration for accurate switching is adopted to improve power conversion efficiency (PCE) and voltage conversion ratio (VCR). The proposed switch controller for offset calibration uses dynamic logic circuits and reduces power overhead that NOR-based set-reset latches and edge detectors of conventional switch controllers cause. Under 200-Ω loading and 5-V output, the proposed comparator and controller reduce power consumption by up to 73% and 24% in simulations compared with the conventional comparator and controller, respectively. This paper is implemented with 5-V devices of a 0.35-μm CMOS process to support high voltage. A peak PCE of 89.0%, a peak VCR of 90.1%, and a maximum output power of 126.7 mW are measured for 200-Ω loading.

1 February 2019
Verified
Conference Paper
Closed access
WJWenning Jiang
Wenning Jiang
YZYan Zhu
Yan Zhu
MZMinglei Zhang
Minglei Zhang
CCChi-Hang Chan
Chi-Hang Chan
RMRui P. Martins
Rui P. Martins
Continuous technology scaling has allowed unceasing growth of the sampling rate of a single channel ADC in past decades. Such development not only helps reduce the number of channels in massively time interleaved ADCs, but also contributes to lower their overall jitter and input capacitance, thus enabling a further push on the ADC performance boundary. Being limited by metastability, the conventional SAR architecture is not suitable when both high resolution and speed are essential. While the pipeline SAR offers a higher speed alternative, it also keeps the low and dynamic power nature of the SAR architecture through adopting dynamic or integrating residue amplifiers (RAs) in recent works [1, 2]. With an integrating characteristic, the amplification time can be short but in contrast, the linearity, extra reset time of the load, and PVT sensitivity pose significant design challenges. In this work, rather than adopting the integrating-type amplifier, we present a Gm-R based RA which has a complete-settled amplification characteristic, thus allowing us to compensate the gain variation over temperature easily with a tracking bias technique. Besides this, we use a two-stage amplification to alleviate the RA input parasitic capacitance that enables a small DAC size in all stages. The single channel prototype reaches 1GS/s with 60.02dB SNDR at a Nyquist input consuming 7.6mW from a 1V supply.