| ??? 07/22/08 04:24 Read: times |
#156921 - Wear and tear Responding to: ???'s previous message |
10 million rewrites of the byte holding the least significant nibble. What is the specification of the EEPROM you have?
If you take 10 memory cells as "prescaler" and count the first 0..10, then the second 0..10, ... and first when the tenth byte reaches 10 (i.e. after 100 increments) then you reset all prescaler bytes and steps the normal counter, then you would get 11 writes for every 100 ticks of your counter. If you use 0..100 before switching to next prescaler cell, the prescaler cells would get 101 writes for every 1000 ticks. There are more complex schemes that rotates the writes for all involved EEPROM bytes, but the above suggestion is quite easy to implement and gives a big reduction in wear on the most modified EEPROM bytes. |
| Topic | Author | Date |
| incrementing a large number in assembly | 01/01/70 00:00 | |
| Wear and tear | 01/01/70 00:00 | |
| The EEPROM is | 01/01/70 00:00 | |
| F-RAM | 01/01/70 00:00 | |
| F-RAM problem | 01/01/70 00:00 | |
| what's the problem? | 01/01/70 00:00 | |
| found a substitute | 01/01/70 00:00 | |
| first do it in C, then | 01/01/70 00:00 | |
| Load/Save in loop | 01/01/70 00:00 | |
| To Answer Your Question ... | 01/01/70 00:00 | |
| Thanks Russ - slow event | 01/01/70 00:00 | |
| A sample code for your task | 01/01/70 00:00 | |
| Sample code irrelevant | 01/01/70 00:00 | |
| Socketed EEPROM? | 01/01/70 00:00 | |
| walking writes are dead simple | 01/01/70 00:00 | |
| Larger EEPROM = simple code | 01/01/70 00:00 | |
| the counter is | 01/01/70 00:00 | |
| good point | 01/01/70 00:00 | |
Blinkers | 01/01/70 00:00 |