TwinPop: A Resource-efficient and Highly Linear FPGA-based Time-to-Digital Converter  

Project Description:

Time to Digital Converter (TDC) is a frequently used high-precision time interval measurement unit in various electronic and industrial scenarios. The notorious nonlinearity problems are the leading cause that affect the performance of a TDC. To deal with these issues, conventional approaches often demand excessive redundant resources or tedious calibrations, or alternatively sacrifice conversion rates. To this end, in this paper, we propose TwinPop, a new highly Linear, resource-efficient and temperature-invariant TDC. ToPo utilizes only 1/4 chain resources than conventional Tapped Delay Line (TDL) TDC, without introducing any increased dead time and even being less vulnerable to temperature fluctuations. The key technique of TwinPop is a novel two-phase interpolation structure. In this structure, we utilize few temperature-invariant delay units to imprint timestamps along a single short chain. These delay units, acting like a ring oscillator, switch respective states along the short chain if a predefined time interval is triggered, leaving decodable landmarks. Meanwhile, since only 1/4 chain resources is needed, we can constrain all logic resources within a single logic array block, significantly mitigating the nonlinearity problem, especially the ultra-wide bins. We have implemented our TDC structure in popular FPGA platforms including Intel and Xilinx FPGAs. Results demonstrate that our TDC can readily achieve a less than 0.3~LSB INL even without any calibration tricks.

Acknowledgement:

  • This project is filed on December 2021 and is patented. For more information please visit our git repository

    People:

    • Dr. Chao Cai (Associate Proffessor, College of Life Science & Engineering) - Huazhong University of Science and Technology
    • Fei Wang(Ph.D student, College of Life Science & Engineering) - HUST
    • Jiawen Zhou (Master student, College of Life Science & Engineering) - HUST
    • Ziyi Wong (Master student, College of Life Science & Engineering) - HUST

    Related Publications:

  • [1] Chao Cai. "TwinPop: A Resource-efficient and Highly Linear FPGA-based Time-to-Digital Converter", in IEEE Transactions on Instrumentation and Measurement, 2024