S. Ranzini, V. E. Parahyba, T. Vilela, E. Schneider, Joao B. Tardelli, Julio Oliveira, Eudes P. Lopes, Tiago C. Lima, J. Reis, J. Oliveira, G. Fraidenraich
{"title":"高速相干光学系统的数字反向传播ASIC设计","authors":"S. Ranzini, V. E. Parahyba, T. Vilela, E. Schneider, Joao B. Tardelli, Julio Oliveira, Eudes P. Lopes, Tiago C. Lima, J. Reis, J. Oliveira, G. Fraidenraich","doi":"10.1109/IMOC.2015.7369081","DOIUrl":null,"url":null,"abstract":"We analyze the performance limits of the digital back-propagation (DBP) algorithm implemented in application specific integrated circuits (ASIC). The proposed hardware implementation, jointly with a set of other digital signal processing (DSP) blocks, was experimentally evaluated by processing a 32 GBd PDM-16QAM signal. In the transmitter side, it is discussed the linear filters complexity for Nyquist spectral shaping and the trade-off between number of taps and filter roll-off. Then, at the receiver, linear filters and nonlinear compensators are used to equalize chromatic dispersion (CD) and nonlinear distortions, respectively. These stages characterize the DBP block, which is able to enhance the transmission reach and the optimum launch power of the optical system. Transmission performance is optimized for different DBP implementations, considering two CMOS technology nodes: 16 nm and 28 nm. The DBP complexity is evaluated in terms of the required die size and in terms of number of the DBP stages and fiber span-lengths. For a target gain of 35% (in comparison with the CD-only compensation) on the transmission reach, the required DBP chip size is around 1.6 mm x 1.6 mm using 16 nm CMOS technology. The chip area of the nonlinear stage is only 16% larger than the linear stage.","PeriodicalId":431462,"journal":{"name":"2015 SBMO/IEEE MTT-S International Microwave and Optoelectronics Conference (IMOC)","volume":"8 6","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2015-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"4","resultStr":"{\"title\":\"Digital back-propagation ASIC design for high-speed coherent optical system\",\"authors\":\"S. Ranzini, V. E. Parahyba, T. Vilela, E. Schneider, Joao B. Tardelli, Julio Oliveira, Eudes P. Lopes, Tiago C. Lima, J. Reis, J. Oliveira, G. Fraidenraich\",\"doi\":\"10.1109/IMOC.2015.7369081\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"We analyze the performance limits of the digital back-propagation (DBP) algorithm implemented in application specific integrated circuits (ASIC). The proposed hardware implementation, jointly with a set of other digital signal processing (DSP) blocks, was experimentally evaluated by processing a 32 GBd PDM-16QAM signal. In the transmitter side, it is discussed the linear filters complexity for Nyquist spectral shaping and the trade-off between number of taps and filter roll-off. Then, at the receiver, linear filters and nonlinear compensators are used to equalize chromatic dispersion (CD) and nonlinear distortions, respectively. These stages characterize the DBP block, which is able to enhance the transmission reach and the optimum launch power of the optical system. Transmission performance is optimized for different DBP implementations, considering two CMOS technology nodes: 16 nm and 28 nm. The DBP complexity is evaluated in terms of the required die size and in terms of number of the DBP stages and fiber span-lengths. For a target gain of 35% (in comparison with the CD-only compensation) on the transmission reach, the required DBP chip size is around 1.6 mm x 1.6 mm using 16 nm CMOS technology. The chip area of the nonlinear stage is only 16% larger than the linear stage.\",\"PeriodicalId\":431462,\"journal\":{\"name\":\"2015 SBMO/IEEE MTT-S International Microwave and Optoelectronics Conference (IMOC)\",\"volume\":\"8 6\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2015-11-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"4\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2015 SBMO/IEEE MTT-S International Microwave and Optoelectronics Conference (IMOC)\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/IMOC.2015.7369081\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2015 SBMO/IEEE MTT-S International Microwave and Optoelectronics Conference (IMOC)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/IMOC.2015.7369081","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 4
摘要
我们分析了在专用集成电路(ASIC)中实现的数字反向传播(DBP)算法的性能限制。通过对32gbd PDM-16QAM信号的处理,对所提出的硬件实现以及其他数字信号处理(DSP)模块进行了实验评估。在发射端,讨论了奈奎斯特频谱整形的线性滤波器复杂度以及抽头数和滤波器滚降之间的权衡。然后,在接收端,使用线性滤波器和非线性补偿器分别平衡色散和非线性畸变。这些阶段是DBP块的特征,它能够提高光学系统的传输距离和最佳发射功率。考虑到16纳米和28纳米两种CMOS技术节点,针对不同的DBP实现优化了传输性能。DBP复杂性是根据所需的模具尺寸和DBP阶段的数量和光纤的长度来评估的。为了在传输距离上获得35%的目标增益(与纯cd补偿相比),使用16nm CMOS技术所需的DBP芯片尺寸约为1.6 mm x 1.6 mm。非线性级的芯片面积仅比线性级大16%。
Digital back-propagation ASIC design for high-speed coherent optical system
We analyze the performance limits of the digital back-propagation (DBP) algorithm implemented in application specific integrated circuits (ASIC). The proposed hardware implementation, jointly with a set of other digital signal processing (DSP) blocks, was experimentally evaluated by processing a 32 GBd PDM-16QAM signal. In the transmitter side, it is discussed the linear filters complexity for Nyquist spectral shaping and the trade-off between number of taps and filter roll-off. Then, at the receiver, linear filters and nonlinear compensators are used to equalize chromatic dispersion (CD) and nonlinear distortions, respectively. These stages characterize the DBP block, which is able to enhance the transmission reach and the optimum launch power of the optical system. Transmission performance is optimized for different DBP implementations, considering two CMOS technology nodes: 16 nm and 28 nm. The DBP complexity is evaluated in terms of the required die size and in terms of number of the DBP stages and fiber span-lengths. For a target gain of 35% (in comparison with the CD-only compensation) on the transmission reach, the required DBP chip size is around 1.6 mm x 1.6 mm using 16 nm CMOS technology. The chip area of the nonlinear stage is only 16% larger than the linear stage.