| ??? 06/17/05 15:28 Read: times |
#95181 - endurance - the 5/3V puzzle Responding to: ???'s previous message |
They can simply cheat. They can simply implement the "walking write" in harware, hidden for the user, having multiple memory cells per "public" bit. Based on knowledge of the technology (which I don't have) this may imply less than linear increase of area. This could also explain the relatively low densities of available memories (while AFAIK they are ,anufactured using a quite state-of-the-art 0.35um technology, allowing significantly bigger SRAMs...) But more importantly this explains the difference in endurance between the 5V and 3V part - the lower voltage allows smaller geometry hence higher density... But I don't know how to prove if this is true or not... (having not enough funds for a decent electron microscope... :-) Something similar happens with modern superhigh dnsity hard disks - the increased error rate due to approaching physical limits is hidden using run-time remapping of faulty sectors.. Jan Waclawek |
| Topic | Author | Date |
| 89s52 | 01/01/70 00:00 | |
| similar to SRAM | 01/01/70 00:00 | |
| IAP | 01/01/70 00:00 | |
| Problem with IAP and 89c669 | 01/01/70 00:00 | |
| think | 01/01/70 00:00 | |
| non-volatile ram | 01/01/70 00:00 | |
| eeprom or nvram | 01/01/70 00:00 | |
| ramtron | 01/01/70 00:00 | |
| FRAM | 01/01/70 00:00 | |
| 5V FRAM/3V FRAM | 01/01/70 00:00 | |
| how much data and how fast | 01/01/70 00:00 | |
| Another simple possibility.... | 01/01/70 00:00 | |
| RAMTRON FRAMs + SPI, parallel | 01/01/70 00:00 | |
| SRAM, NVRAM, FLASH, EEPROM | 01/01/70 00:00 | |
| FRAM | 01/01/70 00:00 | |
| FRAM Life | 01/01/70 00:00 | |
| FRAM life II. | 01/01/70 00:00 | |
| Ramtron, Fujitsu, and others | 01/01/70 00:00 | |
| others? | 01/01/70 00:00 | |
| FRAM life | 01/01/70 00:00 | |
| cycles | 01/01/70 00:00 | |
endurance - the 5/3V puzzle | 01/01/70 00:00 |