一种用于JESD204C接口的低延迟同步头检测算法与电路

殷鹏; 张超; 雷长安; 侯卫周; 舒洲; 刘术彬; 朱樟明

doi:10.11999/JEIT260163

一种用于JESD204C接口的低延迟同步头检测算法与电路

doi: 10.11999/JEIT260163 cstr: 32379.14.JEIT260163

1.
河南大学物理与电子学院开封 475004
2.
西安电子科技大学集成电路学院西安 710071

基金项目: 中国教育部基础与交叉学科突破计划(YB2025XDXM105)，陕西省创新能力支撑计划(2024RS-CXTD-06)，国家自然科学基金(62434007)，河南大学本科教学改革研究与实践项目(DHXJJG2026-146)

详细信息

作者简介:
殷鹏：男，讲师，研究方向为高速数据传输接口、信号处理等数模混合集成电路设计

张超：男，硕士生，研究方向为高速数据传输接口、图像处理等数字电路设计

雷长安：男，硕士生，研究方向为高速数据传输接口、图像处理等数字电路设计

侯卫周：男，教授，研究方向为电子电路分析、设计及仿真

舒洲：男，教授，研究方向为低功耗模拟/混合信号IC设计，主攻高速通信接口设计

刘术彬：男，教授，研究方向为混合信号集成电路、射频集成电路、高速数据接口等

朱樟明：男，教授，研究方向为高精度ADC/DAC、芯片化雷达、模拟前端集成电路与模拟信号链芯片、三维集成电路与微系统等

通讯作者:
舒洲　shuzhou@xidian.edu.cn

中图分类号: TN402; TN492
计量
- 文章访问数: 11
- HTML全文浏览量: 6
- PDF下载量: 0
- 被引次数: 0
出版历程
- 收稿日期: 2026-02-09
- 修回日期: 2026-05-29
- 录用日期: 2026-05-29
- 网络出版日期: 2026-06-08

A Low-Latency Synchronization Header Detection Algorithm and Circuit for JESD204C Interface

1.
School of Physics and Electronics, Henan University, Kaifeng 475004, China
2.
School of Intergrated Circuits, Xidian University, Xian 710071, China

Funds: Fundamental and Interdisciplinary Disciplines Breakthrough Plan of the Ministry of Education of China (YB2025XDXM105), Innovation CapabilitySupport Program of Shaanxi (2024RS-CXTD-06), The National Natural Science Foundation of China under Grant 62434007, Research and Practice Project of Undergraduate Teaching Reform at Henan University (HDXJJG2026-146)

摘要

摘要: JESD204C串行(SerDes)接口作为高速芯片互连和边缘AI计算网络的核心器件，其链路初始化延迟长、同步过程功耗高等问题制约了系统实时性与能效。为此，该文提出一种基于迭代式集合筛查(Iterative Set Screening, ISS)的同步头(Synchronization Header, SH)检测方法，旨在提升检测效率。该方法利用同步头序列固有的极性翻转特性，构建并行筛查机制，动态压缩搜索空间，高效剔除无效候选位置，实现同步头的快速精准定位。所提电路已集成至JESD204C接收链路，并在FPGA平台完成验证。实验结果表明：对比典型方案，该设计将同步头锁定时间缩短70 %，且检测时延对同步头的位置不敏感，同步过程稳定性与能效同步提升。
- JESD204C接口 /
- SerDes /
- 同步头(SH) /
- FPGA
Abstract: Objective With rapid technological development, front-end high-speed data converters (ADC/DAC) see rising sampling rates and resolutions, while back-end processing units (FPGA/ASIC) gain stronger computing power. This imposes strict requirements on their high-speed data interface: high bandwidth, low latency, low power consumption and high synchronization reliability. As a mainstream high-speed serial (SerDes) interface in high-speed analog-to-digital converters (ADCs), JESD204C interface are plagued by problems such as long link initialization latency and high power consumption during synchronization, which restrict the real-time performance and energy efficiency of the system. Therefore, this study aims to optimize the link-layer design of the JESD204C receiver, propose an efficient synchronization header (SH) detection method, and realize the core mechanism of exponential compression of the search set through global observation and iterative convergence. The efficiency of synchronization header detection is improved, fast and accurate positioning of the synchronization header is achieved, link synchronization latency is shortened, and the stability and energy efficiency of the synchronization process are enhanced, thereby addressing the performance bottlenecks of typical schemes. Methods The typical JESD204C interface employs a serial sliding detection method for synchronization header detection, which results in issues such as high link initialization latency and large delay jitter. To solve these problems, an iterative set screening (ISS)-based synchronization header detection algorithm is proposed in this paper. The synchronization header detection problem is abstracted as the core scientific issue of “rapidly locating a deterministic pattern in a binary random sequence”, and a theoretical model based on information theory and stochastic processes is constructed. Parameters such as expected space utilization and bit error rate are introduced to provide a quantitative evaluation basis for algorithm performance. In this model, synchronization header candidate positions are defined as a dynamic set. Leveraging the inherent polarity inversion characteristics of the synchronization header and per-cycle global observation information, Multi-level XOR logic is utilized to perform parallel verification on all candidate hypotheses, eliminate non-inverted bits, and dynamically compress the search space. This significantly improves the synchronization establishment speed and position robustness, and provides a low-latency and high-reliability initialization solution for high-speed SerDes links. Results and Discussions The proposed synchronization header detection algorithm based on ISS has been comprehensively validated under harsh scenarios such as synchronization header crossing block boundaries and link loss due to burst errors, demonstrating excellent robustness: rapid completion of synchronization header locking and link resynchronization under all test conditions (Figures 11, 12, 13, 14, 15, 16). To comprehensively evaluate performance, this paper replicates four representative existing schemes: single-bit serial locking circuit, 66-bit serial locking architecture, register-intensive block synchronization method, and parallel search circuit. Then, a systematic comparative analysis is conducted between these schemes and the proposed scheme. Test results show that the normalized locking time of the single-bit serial locking circuit, 66-bit serial locking architecture, and register-intensive block synchronization method fluctuates significantly with synchronization header position (Figure 17(a)), especially at block boundaries (Figure 17(b)): when the synchronization header is at the MSB, typical sliding detection consumes approximately 1.8 times the time at the LSB, reflecting sensitivity to starting position and search path dependence. In contrast, the proposed ISS scheme maintains stable normalized locking time within 1.0 ± 0.05 at all positions, with a standard deviation reduced by over 70%. This scheme eliminates position dependence by equally evaluating all candidate positions via a parallel filtering mechanism. Whether the synchronization header is at the LSB, MSB, or any block position, synchronization can be completed within tens of clock cycles. Experimental results verify that the ISS algorithm greatly improves synchronization robustness and predictability while accelerating link initialization. Table 1 summarizes the performance metrics of the proposed scheme. The average locking time is only 24.4 clock cycles, with an overall improvement of more than 70% compared to the single-bit serial locking circuit, 66-bit serial locking architecture, and register-intensive block synchronization method. At the same time, its locking time standard deviation is minimal 4.7, indicating a more stable synchronization process. In resource utilization, the design consumes 509 LUTs and only 2.0 mW, far lower than the 3503 LUTs and 94.1 mW of the register-intensive scheme; its energy efficiency reaches 0.03 milliwatts per bit, superior to the three traditional methods. In addition, compared to the parallel search circuit, the average lock time is reduced by 6.11%, power consumption is reduced by 50.3%, and energy efficiency is improved by 53.8%. Therefore, the proposed JESD204C receive link has comprehensive advantages in synchronization header detection speed, stability, power consumption, and energy efficiency. Conclusions This study proposes an ISS-based synchronization header detection algorithm for the JESD204C receiver. By performing parallel screening of the data stream through multi-level XOR logic, dynamically compressing the search space, and efficiently eliminating non-inverted candidate bits, the algorithm dynamically screens the candidate positions of the synchronization header to the only real position, innovating the traditional serial detection mechanism. The design is verified on the Xilinx KC705 FPGA platform. A 31-bit pseudo-random binary sequence (PRBS31) simulates inverted data random distribution, and a high-speed SMA cable completes data loopback transmission tests. The test results showed that the algorithm achieved an average lock time of only 24.4 clock cycles with a standard deviation as low as 4.7, maintaining high robustness for the synchronization header at any position within the 66-bit block; the energy efficiency reaches 0.03  mW/bit. The algorithm outperforms existing typical schemes in three core indicators: lock speed, delay stability, and energy efficiency. It provides a new paradigm for low-latency, high-reliability, and high-energy-efficiency synchronous initialization of high-speed SerDes links.
- JESD204C Interface /
- SerDes /
- Synchronization Header /
- FPGA

HTML全文

图 1 JESD204C典型架构

下载: 全尺寸图片幻灯片

图 2 JESD204C链路层数据结构

下载: 全尺寸图片幻灯片

图 3 典型同步头检测电路的结构

下载: 全尺寸图片幻灯片

图 4 ISS同步头检测算法的核心原理

下载: 全尺寸图片幻灯片

图 5 期望空间大小

下载: 全尺寸图片幻灯片

图 6 SNR与BER的量化映射关系

下载: 全尺寸图片幻灯片

图 7 搜索集合期望衰减曲线(L=66)

下载: 全尺寸图片幻灯片

图 8 算法1的关键处理过程

下载: 全尺寸图片幻灯片

图 9 data132的格式

下载: 全尺寸图片幻灯片

图 10 验证平台示意图

下载: 全尺寸图片幻灯片

图 11 同步头锁定

下载: 全尺寸图片幻灯片

图 12 同步头和有效数据输出

下载: 全尺寸图片幻灯片

图 13 锁定同步头后失锁

下载: 全尺寸图片幻灯片

图 14 同步头出现错误

下载: 全尺寸图片幻灯片

图 15 载荷表现翻转特征

下载: 全尺寸图片幻灯片

图 16 多块边界识别

下载: 全尺寸图片幻灯片

图 17 同步头锁定时间测试

下载: 全尺寸图片幻灯片

表 1 ISS算法与典型方案的对比

	典型方案	ISS算法
基本对象	单个候选位置	候选位置集合
搜索方式	按位滑动、逐位置试探	全位置并行筛查、集合迭代筛查
判断依据	每次滑动后重新判断当前窗口	使用寄存器跨周期累计保留信息
收敛原理	依赖搜索路径，逐步接近目标位置	错误候选随周期指数衰减
位置敏感性	高，锁定时间随同步头位置变化明显	低，对同步头所在位置不敏感
非理想信道适配性	未考虑干扰和噪声对结果的影响	引入误码率、期望空间利用率等参数修正模型，推导非理想信道下的搜索集合衰减公式

下载: 导出CSV

表 2 参数映射

参数	参数描述
周期t	时钟周期数
初始空间大小L	对应一个数据块内的初始候选位数量，决定了66b_reg的最小位宽
集合期望数E[N_t]	66b_reg中仍为“1”的位数，即当前仍被保留的候选位置数量
错误候选期望数量阈值δ	当E[N_t]<δ时，认为搜索空间已收敛至唯一位置。
理论收敛时间T_c	同步头锁定所需的理论时钟周期数，为接收链路初始化提供依据

下载: 导出CSV

1 同步头检测

输入：datain
输出：同步头位置j
1) 数据拼接与极性翻转检测
•通过将当前和之前的数据块合并，形成连续的数据流数据：
data_r={datain_t[65:0],datain_(t−1)[65]}
• 利用异或运算计算翻转特征值序列的异或数据：
data_x[i]=data_r[i]$\oplus $data_r[i+1], i=0,1,···,65
2) 同步头的迭代式集合筛查
•初始化一个66位寄存器，初始化为全“1”：
66b_reg₀=[1,1,···,1]
•在每个时钟周期t中使用按位与更新66b_reg：
66b_reg_t=66b_reg_(t−1)&data_x_t
3) 同步头锁定与复位机制
• 检查同步头是否被锁定：
sh_test=(66b_reg&(66b_reg−1))==0
•如果sh_test为“0”，则锁定同步头位置j，其中　　 66b_reg[j]=1，否则从步骤1重新开始。
4) 数据对齐与输出
•一旦同步头锁定，输出同步头位置j。

下载: 导出CSV

表 3 资源消耗与性能指标

方案	同步头平均锁定时间(时钟周期)	锁定时间标准差	LUT资源	寄存器资源	功耗(mW)	能效(mW/bit)
本算法	24.4	4.7	509	222	2.0	0.030
[11]	7438.0	1281.7	195	312	1.2	1.200
[17]	146.4	36.7	491	228	4.5	0.068
[18]	112.9	17.9	3503	4552	94.1	1.426
[19]	26.2	1.7	309	213	4.3	0.065

下载: 导出CSV

参考文献(19)

[1]	ZHOU Guoqing, ZHANG Haotian, XU Chao, et al. A real-time data acquisition system for single-band bathymetric LiDAR[J]. IEEE Transactions on Geoscience and Remote Sensing, 2023, 61: 5702721. doi: 10.1109/TGRS.2023.3282624.
[2]	王知非, 黄之闻, 叶天辰, 等. 面向芯粒互连的单端64 Gb/s全双工收发机设计[J]. 电子与信息学报, 2025, 47(9): 2979–2993. doi: 10.11999/JEIT250506. WANG Zhifei, HUANG Zhiwen, YE Tianchen, et al. A 64 Gb/s single-ended simultaneous bi-directional transceiver for die-to-die interfaces[J]. Journal of Electronics & Information Technology, 2025, 47(9): 2979–2993. doi: 10.11999/JEIT250506.
[3]	ZHAO Wenhao, WANG Houjun, LIU Ke, et al. Design of synchronous multichannel cascadable high-speed arbitrary waveform generator[J]. IEEE Transactions on Instrumentation and Measurement, 2025, 74: 2003617. doi: 10.1109/TIM.2025.3545161.
[4]	骆建军, 沈一凡, 周迪, 等. 一种高性能硬件加密引擎阵列架构[J]. 电子与信息学报, 2021, 43(12): 3743–3748. doi: 10.11999/JEIT200855. LUO Jianjun, SHEN Yifan, ZHOU Di, et al. High performance crypto module with array of hardware engines[J]. Journal of Electronics & Information Technology, 2021, 43(12): 3743–3748. doi: 10.11999/JEIT200855.
[5]	KOU Zhengchang, YOU Qi, KIM J, et al. High-level synthesis design of scalable ultrafast ultrasound beamformer with single FPGA[J]. IEEE Transactions on Biomedical Circuits and Systems, 2023, 17(3): 446–457. doi: 10.1109/TBCAS.2023.3267614.
[6]	文溢, 陈建军, 黄俊, 等. 采用自适应连续时间线性均衡器和判决反馈均衡器算法的一种16 Gbit/s 并转串/串转并接口[J]. 电子与信息学报, 2023, 45(11): 3984–3990. doi: 10.11999/JEIT230668. WEN Yi, CHEN Jianjun, HUANG Jun, et al. A 16 Gbit/s serializer/deserializer with adaptive continuous time linear equalizer and decision feedback equalizer equalization algorithm[J]. Journal of Electronics & Information Technology, 2023, 45(11): 3984–3990. doi: 10.11999/JEIT230668.
[7]	ZHAO Haoyu and JIANG Ziyun. Implementation of deterministic latency at the receiver of JESD204C[C]. Proceedings of 2024 4th International Conference on Computer Science, Electronic Information Engineering and Intelligent Control Technology (CEI), Guangzhou, China, 2024: 301–304. doi: 10.1109/CEI63587.2024.10871333.
[8]	LI Shijie, MA Ruichang, DENG Mingxing, et al. A 312.5 Mbps-32 Gbps JESD204C wireline transceiver back-compatible with JESD204B in 28 nm CMOS[J]. Integrated Circuits and Systems, 2024, 1(2): 109–118. doi: 10.23919/ICS.2024.3423852.
[9]	YIN Peng, SHU Zhou, XIA Yingjun, et al. A low-area and low-power comma detection and word alignment circuits for JESD204B/C controller[J]. IEEE Transactions on Circuits and Systems I: Regular Papers, 2021, 68(7): 2925–2935. doi: 10.1109/TCSI.2021.3072772.
[10]	WANG Qiushi, YIN Peng, WU N, et al. A high-logic-density, low-power control character detection and identification circuit for the JESD204B data link layer[J]. IEEE Transactions on Circuits and Systems II: Express Briefs, 2023, 70(4): 1600–1604. doi: 10.1109/TCSII.2022.3226520.
[11]	JEDEC. JESD204C. 1 Serial interface for data converters[S]. JEDEC, 2021. (查阅网上资料, 未找到本条文献出版地信息, 请确认).
[12]	代苏杰, 杨定坤, 阳江平, 等. JESD204C同步算法研究[J]. 自动化与信息工程, 2025, 46(2): 18–24. doi: 10.12475/aie.20250203. DAI Sujie, YANG Dingkun, YANG Jiangping, et al. Research on JESD204C synchronization algorithm[J]. Automation & Information Engineering, 2025, 46(2): 18–24. doi: 10.12475/aie.20250203.
[13]	PEREZ-RESA A, GARCIA-BOSQUE M, SANCHEZ-AZQUETA C, et al. Chaotic encryption for 10-Gb ethernet optical links[J]. IEEE Transactions on Circuits and Systems I: Regular Papers, 2019, 66(2): 859–868. doi: 10.1109/TCSI.2018.2867918.
[14]	ZENG Jinrui, YANG Junfeng, ZHANG Lei, et al. FPGA implementation of fixed-latency command distribution based on aurora 64B/66B[J]. IEEE Transactions on Nuclear Science, 2024, 71(6): 1348–1356. doi: 10.1109/TNS.2024.3400378.
[15]	IEEE. IEEE Standard 802.3-2022 IEEE standard for Ethernet[S]. IEEE, 2022. (查阅网上资料, 未找到本条文献出版地信息, 请确认).
[16]	ALI I, ASIF M, UR REHMAN M R, et al. A configurable, multi-mode comma detection and word alignment controller for high speed serial interface in 130 nm CMOS technology[C]. Proceedings of 2018 14th International Conference on Emerging Technologies (ICET), Islamabad, Pakistan, 2018: 1–5. doi: 10.1109/ICET.2018.8603553.
[17]	RANJULA W P, SENARATH R M A U, SENARATNA D P D, et al. Implementation techniques for IEEE 802.3ba 40Gbps Ethernet Physical Coding Sublayer (PCS)[C]. Proceedings of 2015 12th International Conference on Electrical Engineering/Electronics, Computer, Telecommunications and Information Technology (ECTI-CON), Hua Hin, Thailand, 2015: 1–5. doi: 10.1109/ECTICon.2015.7207093.
[18]	WANG Bin and HU Qingsheng. A high-speed 64b/66b decoder used in SerDes[J]. Applied Mechanics and Materials, 2014, 556/562: 1549–1552. doi: 10.4028/www.scientific.net/AMM.556-562.1549.
[19]	YIN Peng, CHEN Haoran, MA Rui, et al. A low-latency synchronization header detector and hardware-efficient correction decoder for JESD204C receiver[J]. IEEE Transactions on Circuits and Systems I: Regular Papers, 2026: 1–13. doi: 10.1109/TCSI.2026.3662461.