基于两级时间插值的FPGA-TDC设计与测试

Design and testing of FPGA-TDC based on two-stage time interpolation

  • 摘要: 提出一种结合多相时钟与延迟链插值的多通道FPGA-TDC结构,以降低工作频率、提升线性度并减少资源消耗,同时保持高分辨率。设计采用两级插值结构,利用多相时钟与延迟链构建细时间单元,从而减小延迟非线性积累并缩小编码器规模。系统在Xilinx ZYNQ-7035平台实现,并在0~16000 ps范围内进行测试。实验结果表明,所设计的TDC系统分辨率优于4 ps,微分非线性在−1~+7 LSB之间,积分非线性在−2 LSB至+14 LSB之间。与传统结构相比,该方案在同频率下延迟链长度成倍缩短,在相同链长下频率更低。所提两级插值结构在提升分辨率和线性度的同时显著节省逻辑资源,具备良好的应用潜力。

     

    Abstract:
    Background
    Field-programmable gate array (FPGA)-based time-to-digital converters (TDCs) have been extensively employed for high-precision time interval measurements, where picosecond-level resolution is often required. Among existing approaches, the delay-line method remains widely used, while the system clock frequency and the delay chain design are recognized as the primary factors affecting resolution and linearity.
    Purpose
    The objective of this study is to develop a multi-channel FPGA-TDC architecture that integrates multiphase clocking with delay-line interpolation, thereby lowering the operating frequency, improving linearity, and reducing hardware resource utilization, while maintaining high measurement resolution.
    Methods
    A two-stage interpolation scheme was introduced, where fine time measurement cells were implemented through the combination of multiphase clocks and shortened delay chains. This configuration mitigates the accumulation of nonlinearity in the delay elements and reduces the scale of thermometer-to-binary encoders, resulting in decreased logic overhead. The proposed TDC was implemented on a Xilinx ZYNQ-7035 device, and its performance was evaluated within a measurement range of 0–16000 ps.
    Results
    The experimental evaluation demonstrated that a time resolution better than 4 ps was achieved. The measured differential nonlinearity (DNL) was in the range of −1 least significant bit (LSB) to +7 LSB, while the integral nonlinearity (INL) ranged from −2 LSB to +14 LSB. Compared with conventional architectures, the proposed scheme shortens the delay chain length by several times at the same operating frequency, and achieves a lower frequency with the same chain length.
    Conclusions
    The proposed two-stage interpolation architecture not only enhances resolution and linearity but also significantly reduces logic resource consumption, demonstrating strong application potential.

     

/

返回文章
返回