C. Calligaro, A. Manstretta, A. Pierin, G. Torelli
{"title":"多电平非易失性存储器传感方案的比较分析","authors":"C. Calligaro, A. Manstretta, A. Pierin, G. Torelli","doi":"10.1109/ICISS.1997.630269","DOIUrl":null,"url":null,"abstract":"Reading multilevel non-volatile memories is a very demanding task. Three sensing techniques (the parallel scheme, the binary-serial scheme and the mixed parallel-serial scheme) are considered here. Their operation principles are described and a comparative evaluation in terms of both access time and circuit complexity is carried out. The parallel approach is the most suitable for 4-level-cell memories (2 bit per cell). The mixed approach seems to be the most attractive for a large number of programmable levels (at least 16 levels per cell). In this case, the sensing area overhead is limited to less than 1% of the memory array while access time penalty is less than 50%.","PeriodicalId":357602,"journal":{"name":"1997 Proceedings Second Annual IEEE International Conference on Innovative Systems in Silicon","volume":"16 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"1997-10-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"7","resultStr":"{\"title\":\"Comparative analysis of sensing schemes for multilevel non-volatile memories\",\"authors\":\"C. Calligaro, A. Manstretta, A. Pierin, G. Torelli\",\"doi\":\"10.1109/ICISS.1997.630269\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Reading multilevel non-volatile memories is a very demanding task. Three sensing techniques (the parallel scheme, the binary-serial scheme and the mixed parallel-serial scheme) are considered here. Their operation principles are described and a comparative evaluation in terms of both access time and circuit complexity is carried out. The parallel approach is the most suitable for 4-level-cell memories (2 bit per cell). The mixed approach seems to be the most attractive for a large number of programmable levels (at least 16 levels per cell). In this case, the sensing area overhead is limited to less than 1% of the memory array while access time penalty is less than 50%.\",\"PeriodicalId\":357602,\"journal\":{\"name\":\"1997 Proceedings Second Annual IEEE International Conference on Innovative Systems in Silicon\",\"volume\":\"16 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"1997-10-08\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"7\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"1997 Proceedings Second Annual IEEE International Conference on Innovative Systems in Silicon\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/ICISS.1997.630269\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"1997 Proceedings Second Annual IEEE International Conference on Innovative Systems in Silicon","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ICISS.1997.630269","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Comparative analysis of sensing schemes for multilevel non-volatile memories
Reading multilevel non-volatile memories is a very demanding task. Three sensing techniques (the parallel scheme, the binary-serial scheme and the mixed parallel-serial scheme) are considered here. Their operation principles are described and a comparative evaluation in terms of both access time and circuit complexity is carried out. The parallel approach is the most suitable for 4-level-cell memories (2 bit per cell). The mixed approach seems to be the most attractive for a large number of programmable levels (at least 16 levels per cell). In this case, the sensing area overhead is limited to less than 1% of the memory array while access time penalty is less than 50%.