2014IEEE Transactions on Nuclear ScienceOpen access

Picosecond Resolution Time-to-Digital Converter Using ${{\rm G}_{\rm m}} \hbox{-C}$ Integrator and SAR-ADC

Zule Xu, Masaya Miyahara, Akira Matsuzawa

Open full text 15 citations

Abstract

A picosecond resolution time-to-digital converter (TDC) is presented. The resolution of a conventional delay chain TDC is limited by the delay of a logic buffer. Various types of recent TDCs are successful in breaking this limitation, but they require a significant calibration effort to achieve picosecond resolution with a sufficient linear range. To address these issues, we propose a simple method to break the resolution limitation without any calibration: a Gm-C integrator followed by a successive approximation register analog-to-digital converter (SAR-ADC). This translates the time interval into charge, and then the charge is quantized. A prototype chip was fabricated in 90 nm CMOS. The measurement results reveal a 1 ps resolution, a -0.6/0.7 LSB differential nonlinearity (DNL), a -1.1/2.3 LSB integral nonlinearity (INL), and a 9-bit range. The measured 11.74 ps single-shot precision is caused by the noise of the integrator. We analyze the noise of the integrator and propose an improved front-end circuit to reduce this noise. The proposal is verified by simulations showing the maximum single-shot precision is less than 1 ps. The proposed front-end circuit can also diminish the mismatch effects.

Open-access reader

About this research paper

What this paper is about

A picosecond resolution time-to-digital converter (TDC) is presented. The resolution of a conventional delay chain TDC is limited by the delay of a logic buffer. Various types of recent TDCs are successful in breaking this limitation, but they require a significant calibration effort to achieve picosecond resolution with a sufficient linear range. To address these issues, we propose a simple method to break the resolution limitation without any calibration: a Gm-C integrator followed by a successive approximation register analog-to-digital converter (SAR-ADC). This translates the time interval into charge, and then the charge is quantized. A prototype chip was fabricated in 90 nm CMOS. The measurement results reveal a 1 ps resolution, a -0.6/0.7 LSB differential nonlinearity (DNL), a -1.1/2.3 LSB integral nonlinearity (INL), and a 9-bit range. The measured 11.74 ps single-shot precision is caused by the noise of the integrator. We analyze the noise of the integrator and propose an improved front-end circuit to reduce this noise. The proposal is verified by simulations showing the maximum single-shot precision is less than 1 ps. The proposed front-end circuit can also diminish the mismatch effects.

Why it matters

OpenAlex reports 15 citations for this work. Citation counts describe recorded attention and do not establish research quality.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

A picosecond resolution time-to-digital converter (TDC) is presented. The resolution of a conventional delay chain TDC is limited by the delay of a logic buffer. Various types of recent TDCs are successful in breaking this limitation, but they require a significant calibration effort to achieve picosecond resolution with a sufficient linear range. To address these issues, we propose a simple method to break the resolution limitation without any calibration: a Gm-C integrator followed by a successive approximation register analog-to-digital converter (SAR-ADC). This translates the time interval into charge, and then the charge is quantized. A prototype chip was fabricated in 90 nm CMOS. The measurement results reveal a 1 ps resolution, a -0.6/0.7 LSB differential nonlinearity (DNL), a -1.1/2.3 LSB integral nonlinearity (INL), and a 9-bit range. The measured 11.74 ps single-shot precision is caused by the noise of the integrator. We analyze the noise of the integrator and propose an improved front-end circuit to reduce this noise. The proposal is verified by simulations showing the maximum single-shot precision is less than 1 ps. The proposed front-end circuit can also diminish the mismatch effects.

Key concepts: Integrator, Integral nonlinearity, Time-to-digital converter, Differential nonlinearity, Physics, CMOS, Successive approximation ADC, Comparator

Related papers

Back to paper searchBrowse research topicsOriginal source
Picosecond Resolution Time-to-Digital Converter Using ${{\rm G}_{\rm m}} \hbox{-C}$ Integrator and SAR-ADC — Research Paper | ScholarLens