Shop OBEX P1 Docs P2 Docs Learn Events
Some EEPROM lifespan questions/thoughts — Parallax Forums

Some EEPROM lifespan questions/thoughts

varnonvarnon Posts: 184
edited 2012-09-15 19:29 in Propeller 1
I was just pondering EEPROM lifespan. I think we can estimate 100,000 write/erase cycles. Is that right? Does that mean we can program to EEPROM ~ 100k times before it starts to wear out, or does programing to the EEPROM take multiple write/erase cycles.

If it is just one cycle per programming, it looks like you would have to program the EEPROM once a day for 274 years before it wears out. Seems pretty unlikely to wear out an EEPROM unless your application is constantly writing to it.

Has anybody had one that has actually reached the end of it's life?

Comments

  • Dave HeinDave Hein Posts: 6,347
    edited 2012-09-15 18:33
    If you wrote to the same location in a tight loop you could hit 100,000 writes in less than 10 minutes. So it's certainly possible to wear out a page in an EEPROM fairly quickly if you do it intentionaly. However, as you said, if you just write to it once a day it would take 100,000 days to reach the specified limit.
  • Peter JakackiPeter Jakacki Posts: 10,193
    edited 2012-09-15 18:58
    varnon wrote: »
    I was just pondering EEPROM lifespan. I think we can estimate 100,000 write/erase cycles. Is that right? Does that mean we can program to EEPROM ~ 100k times before it starts to wear out, or does programing to the EEPROM take multiple write/erase cycles.

    If it is just one cycle per programming, it looks like you would have to program the EEPROM once a day for 274 years before it wears out. Seems pretty unlikely to wear out an EEPROM unless your application is constantly writing to it.

    Has anybody had one that has actually reached the end of it's life?

    I don't know why the 100,000 is quoted so often and almost as a maximum too. The figure is 1,000,000 guaranteed minimum. To write to the same locations repetitively would involve having to readdress the EEPROM as the address counter increments automatically. This would slow down access to 100us per write at highest throughput possible but not including the programming cycle busy time which could be 5ms maximum. So you could wear it out if you were really trying in less than 83 minutes but the more typical endurance would be at least twice that. My general advice with EEPROMs is to not worry about endurance in normal operation such as reprogramming and backing up data etc. Counting is fine as long as it's not "crazy", that is, not practical. If it were a vending machine counting money and stock and errors etc for instance you would not worry at all.
  • Mike GreenMike Green Posts: 23,101
    edited 2012-09-15 19:09
    Keep in mind that some EEPROMs are specified for well under 10^6 writes. You have to look at the datasheet to make sure.
  • Peter JakackiPeter Jakacki Posts: 10,193
    edited 2012-09-15 19:29
    Mike Green wrote: »
    Keep in mind that some EEPROMs are specified for well under 10^6 writes. You have to look at the datasheet to make sure.
    True, but I only find that figure in really old 20th century datasheets but I haven't seen such low figures in anything current and contemporary with the Prop chip.
Sign In or Register to comment.