由于现场可编程逻辑门阵列(FPGA)功能实现的多元化,往往会出现不同时钟域的信号.不同时钟域的信号进行交互,若不进行同步处理,经常会产生数据丢失、时序错误等问题,所以跨时钟域检查对FPGA功能实现特别重要.本文主要阐述了在开展“龙鳞”平台通信模块FPGA软件验证与确认工作中跨时钟域检查的测试流程和方法,对跨时钟异常进行分类,分析通信模块FPGA软件的跨时钟异常并提供解决方案,为FPGA测试工程师提供一种测试思路.
Different clock domain signals often appear due to the diversification of field programmable gate array (FPGA) functions. Signals from different clock domains interact with each other. Data loss, timing errors and other problems often occur without synchronous processing. Therefore, cross-clock domain checking is especially important for FPGA function realization. This paper mainly describes the testing process and method of cross-clock domain checking in the verification and validation of FPGA software for nuclear advanced safety platform of I&C (NASPIC) communication module.It classifies the cross-clock domain exceptions, analyzes the exceptions of FPGA software for communication module and presents solutions, and provides a test idea for FPGA test engineers.
[1]邹晨.FPGA设计中跨时钟域信号同步方法[J].航空计算技术, 2014, 44(4): 131-134.
ZOU Chen. Method of signal synchronization of cross-clock domain in design of FPGA[J].Aeronautical Computing Technique, 2014, 44(4): 131-134.
[2]俞帆, 张伟欣. FPGA设计中的跨时钟域问题[J].现代电子技术, 2014, 37(7): 151-156.
YU Fan, ZHANG Weixin. Problem of clock domain crossing in FPGA design[J]. Modern Electronics Technique, 2014, 37(7): 151-156.
[3]周伟, 杜玉晓, 杨其宇, 等. FPGA跨时钟域亚稳态研究[J]. 电子世界, 2012(3): 87-89.
ZHOU Wei, DU Yuxiao, YANG Qiyu, et al. Research on the metastable state of FPGA across clock domains[J]. Electronics World, 2012(3): 87-89.
[4]胡昌顺, 高嵩, 吴春瑜, 等. SOC设计中多bits数据跨时钟域的解决方法[J]. 辽宁大学学报(自然科学版), 2011, 38(1): 11-15.
HU Changshun, GAO Song, WU Chunyu, et al.Solution for multi-bit data across clock domains in SOC design[J]. Journal of Liaoning University (Natural Sciences Edition), 2011, 38(1): 11-15.
[5]黄忠朝, 赵于前. 一种实现高速异步FIFO的FPGA方法[J]. 计算机工程与应用, 2010, 46(3): 13-15.
HUANG Zhongchao, ZHAO Yuqian. Implementation method of high-speed asynchronous FIFO using FPGA[J]. Computer Engineering and Applications, 2010, 46(3): 13-15.