Scholarlink

<p>As high-bandwidth arbitrary waveform generators (AWGs) place higher demands on digital-to-analog converters (DACs), ultra-high sampling rates and high precision have become the main research directions. To meet these demands, time-interleaved DAC (TI-DAC) technology has emerged. It employs a multichannel parallel structure to effectively increase the sampling rate by several times. However, the TI-DAC’s zero-order hold (ZOH) characteristic manifests a sinc-type frequency response, which significantly degrades amplitude–frequency flatness, and impairs signal accuracy. The existing pre-calibration algorithms increase system complexity, limit application scenarios, and struggle to meet the requirements for low power consumption, low latency, and high precision. To obtain better compensation performance, this article proposes digital pre-calibration network (DPCN) based on pruning-enhanced covariance matrix adaptation evolution strategy (Pe-CMA-ES). The algorithm can effectively compensate for the frequency response distortion of the TI-DAC. Additionally, this research introduces a normalized signal-to-noise ratio (SNR) loss factor. It is used to quantify the influence of ZOH characteristics on the signal frequency response in the TI-DAC. Simulation experiments and hardware evaluations are carried out on a two-channel TI-DAC. The simulation results show that the proposed method improves the passband amplitude-frequency flatness by 45.3% relative to the uncompensated case. The hardware-based tests show that an improvement of 27.8% in passband flatness relative to the uncompensated case, which is consistent with the simulation results. The results show that the proposed method achieves high accuracy while effectively reducing hardware resources.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

TI-DAC ZOH Frequency Response Attenuation Compensation Method Based on Pe-CMA-ES Algorithm

  • 期刊论文

  • 发表日期：

    2026年4月1日

查看全文
  • Guangyu Gu,
  • Yu Peng,
  • Xinyue Yu,
  • Liansheng Liu

摘要

As high-bandwidth arbitrary waveform generators (AWGs) place higher demands on digital-to-analog converters (DACs), ultra-high sampling rates and high precision have become the main research directions. To meet these demands, time-interleaved DAC (TI-DAC) technology has emerged. It employs a multichannel parallel structure to effectively increase the sampling rate by several times. However, the TI-DAC’s zero-order hold (ZOH) characteristic manifests a sinc-type frequency response, which significantly degrades amplitude–frequency flatness, and impairs signal accuracy. The existing pre-calibration algorithms increase system complexity, limit application scenarios, and struggle to meet the requirements for low power consumption, low latency, and high precision. To obtain better compensation performance, this article proposes digital pre-calibration network (DPCN) based on pruning-enhanced covariance matrix adaptation evolution strategy (Pe-CMA-ES). The algorithm can effectively compensate for the frequency response distortion of the TI-DAC. Additionally, this research introduces a normalized signal-to-noise ratio (SNR) loss factor. It is used to quantify the influence of ZOH characteristics on the signal frequency response in the TI-DAC. Simulation experiments and hardware evaluations are carried out on a two-channel TI-DAC. The simulation results show that the proposed method improves the passband amplitude-frequency flatness by 45.3% relative to the uncompensated case. The hardware-based tests show that an improvement of 27.8% in passband flatness relative to the uncompensated case, which is consistent with the simulation results. The results show that the proposed method achieves high accuracy while effectively reducing hardware resources.