H. Smith, S. Kuppinger, P. Venkatachalam, W. Becker
{"title":"对IBM G5/G6大型机进行跨3个包装层次的噪声验证","authors":"H. Smith, S. Kuppinger, P. Venkatachalam, W. Becker","doi":"10.1109/EPEP.1999.819185","DOIUrl":null,"url":null,"abstract":"For IBM's G5/G6 mainframes, both signal rise times and machine cycle times have reduced to the point where signal integrity issues such as noise containment at the system level represent a significant challenge for comprehensive verification of the off-chip nets. The total noise is composed of coupling noise and the switching or delta-i noise. These noise sources are evaluated for all MCM and board nets to ensure coverage is not compromised. The noise verification process has been developed within IBM's S/390 division over several generations of technology and machine designs (Rude, 1994; Venkatachalam et al, 1993; Smith and Katopis, 1996). It is intended to provide a bounding calculation of total noise and identify nets which exceed their design limits for subsequent rerouting. Bounding calculation accuracy is a function of the resolution of deterministic parameters such as physical layouts and statistical variations such as switching time uncertainty. There are several packaging components which must be characterized before system level noise analysis can be performed on the interconnects. The packaging components are the MCM, its board, the memory cards, and the complex connector structures used at the package interfaces. The glass-ceramic MCM contains 20 plane pairs of wiring in ceramic and one plane pair of wiring in thin film. The board has six plane pairs of wiring. The memory cards have five plane pairs of wiring. This paper addresses noise checking across the three levels of packaging which include the on-MCM nets as well as signals from chips on the MCM to chips on the memory card through the board.","PeriodicalId":299335,"journal":{"name":"IEEE 8th Topical Meeting on Electrical Performance of Electronic Packaging (Cat. No.99TH8412)","volume":"134 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"1999-10-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"4","resultStr":"{\"title\":\"Noise verification across 3 levels of packaging hierarchy for the IBM G5/G6 mainframes\",\"authors\":\"H. Smith, S. Kuppinger, P. Venkatachalam, W. Becker\",\"doi\":\"10.1109/EPEP.1999.819185\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"For IBM's G5/G6 mainframes, both signal rise times and machine cycle times have reduced to the point where signal integrity issues such as noise containment at the system level represent a significant challenge for comprehensive verification of the off-chip nets. The total noise is composed of coupling noise and the switching or delta-i noise. These noise sources are evaluated for all MCM and board nets to ensure coverage is not compromised. The noise verification process has been developed within IBM's S/390 division over several generations of technology and machine designs (Rude, 1994; Venkatachalam et al, 1993; Smith and Katopis, 1996). It is intended to provide a bounding calculation of total noise and identify nets which exceed their design limits for subsequent rerouting. Bounding calculation accuracy is a function of the resolution of deterministic parameters such as physical layouts and statistical variations such as switching time uncertainty. There are several packaging components which must be characterized before system level noise analysis can be performed on the interconnects. The packaging components are the MCM, its board, the memory cards, and the complex connector structures used at the package interfaces. The glass-ceramic MCM contains 20 plane pairs of wiring in ceramic and one plane pair of wiring in thin film. The board has six plane pairs of wiring. The memory cards have five plane pairs of wiring. This paper addresses noise checking across the three levels of packaging which include the on-MCM nets as well as signals from chips on the MCM to chips on the memory card through the board.\",\"PeriodicalId\":299335,\"journal\":{\"name\":\"IEEE 8th Topical Meeting on Electrical Performance of Electronic Packaging (Cat. No.99TH8412)\",\"volume\":\"134 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"1999-10-25\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"4\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"IEEE 8th Topical Meeting on Electrical Performance of Electronic Packaging (Cat. No.99TH8412)\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/EPEP.1999.819185\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"IEEE 8th Topical Meeting on Electrical Performance of Electronic Packaging (Cat. No.99TH8412)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/EPEP.1999.819185","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
对于IBM的G5/G6大型机,信号上升时间和机器周期时间都减少到了这样的程度:信号完整性问题(如系统级的噪声控制)对片外网络的全面验证构成了重大挑战。总噪声由耦合噪声和开关噪声或δ - 1噪声组成。所有MCM和板网都对这些噪声源进行了评估,以确保覆盖范围不受影响。IBM的S/390部门在几代技术和机器设计中开发了噪声验证过程(Rude, 1994;Venkatachalam等,1993;Smith and Katopis, 1996)。它的目的是提供总噪声的边界计算,并确定超过其设计限制的网,以便随后改道。边界计算精度是确定性参数(如物理布局)和统计变化(如切换时间不确定性)分辨率的函数。在对互连进行系统级噪声分析之前,必须对几个封装组件进行特性分析。封装组件包括MCM、MCM板、内存卡和封装接口上使用的复杂连接器结构。该微晶玻璃MCM包含20对陶瓷平面布线和1对薄膜平面布线。这块电路板有六对平面布线。存储卡有五对平面布线。本文讨论了三个封装级别的噪声检查,包括on-MCM网络以及从MCM上的芯片到通过板的存储卡上的芯片的信号。
Noise verification across 3 levels of packaging hierarchy for the IBM G5/G6 mainframes
For IBM's G5/G6 mainframes, both signal rise times and machine cycle times have reduced to the point where signal integrity issues such as noise containment at the system level represent a significant challenge for comprehensive verification of the off-chip nets. The total noise is composed of coupling noise and the switching or delta-i noise. These noise sources are evaluated for all MCM and board nets to ensure coverage is not compromised. The noise verification process has been developed within IBM's S/390 division over several generations of technology and machine designs (Rude, 1994; Venkatachalam et al, 1993; Smith and Katopis, 1996). It is intended to provide a bounding calculation of total noise and identify nets which exceed their design limits for subsequent rerouting. Bounding calculation accuracy is a function of the resolution of deterministic parameters such as physical layouts and statistical variations such as switching time uncertainty. There are several packaging components which must be characterized before system level noise analysis can be performed on the interconnects. The packaging components are the MCM, its board, the memory cards, and the complex connector structures used at the package interfaces. The glass-ceramic MCM contains 20 plane pairs of wiring in ceramic and one plane pair of wiring in thin film. The board has six plane pairs of wiring. The memory cards have five plane pairs of wiring. This paper addresses noise checking across the three levels of packaging which include the on-MCM nets as well as signals from chips on the MCM to chips on the memory card through the board.