Life expectancy of EEPROM

Julie in Texas

How many "write" cycles is the EEPROM on a Stamp good for? Millions? Billions? Forever?

I've tried finding the information, but can't find that key parameter.

Mike Green

If you look at the schematic given on Parallax's website for the BS2 for example, it shows that the BS2 uses a 24LC16 EEPROM. Microchip's datasheet for this part shows that it's specified for at least 1,000,000 write/erase cycles. Other manufacturers of this part have similar "wear" characteristics and the larger EEPROMs used in other Stamp models also have similar "wear" characteristics.

Julie in Texas

Mike Green said...
If you look at the schematic given on Parallax's website for the BS2 for example, it shows that the BS2 uses a 24LC16 EEPROM. Microchip's datasheet for this part shows that it's specified for at least 1,000,000 write/erase cycles. Other manufacturers of this part have similar "wear" characteristics and the larger EEPROMs used in other Stamp models also have similar "wear" characteristics.

Thanks -- I'm fairly new to this particular collection of hardware.

Is there an easy (meaning, on the order of "READ" and "WRITE" easy) PBASIC way to access a serial SRAM that's connected to some collection of I/O pins? My application updates several words of EEPROM memory once a second and I'd prefer for the devices I'm prototyping to be good for more than 11 days before reaching that "at least 1,000,000" limit. A small serial SRAM would do the trick.

I intend to eventually do a full-on custom PIC, but the Stamps will get me through hardware prototyping (and with serial SRAM, early hardware release) in the meantime.

Mike Green

The simplest way to do this would be to use a RTC chip that includes a small amount of SRAM. These are designed to use a backup battery to retain the memory contents when the board's power is off and there's already code available. The easiest one to use with a Stamp is the DS1302 which includes 31 bytes of RAM (www.parallax.com/StoreSearchResults/tabid/768/txtSearch/rtc/List/0/SortField/4/ProductID/233/Default.aspx). The Nuts and Volts Columns #33 and #34 also discuss this chip and include examples of its use (www.parallax.com/Resources/NutsVoltsColumns/NutsVoltsVolume1/tabid/444/Default.aspx). See page 375 of #33 near the bottom of the page for the statements required to access any location in the DS1302.

Post Edited (Mike Green) : 12/7/2009 4:58:00 PM GMT

Julie in Texas

Mike Green said...
The simplest way to do this would be to use a RTC chip that includes a small amount of SRAM. These are designed to use a backup battery to retain the memory contents when the board's power is off and there's already code available. The easiest one to use with a Stamp is the DS1302 which includes 31 bytes of RAM (www.parallax.com/StoreSearchResults/tabid/768/txtSearch/rtc/List/0/SortField/4/ProductID/233/Default.aspx). The Nuts and Volts Columns #33 and #34 also discuss this chip and include examples of its use (www.parallax.com/Resources/NutsVoltsColumns/NutsVoltsVolume1/tabid/444/Default.aspx). See page 375 of #33 near the bottom of the page for the statements required to access any location in the DS1302.

RTC chips aren't RAM enough, although having a clock on board might be handy. Something else I can add to my growing list of "Things I'm going to MUX/DEMUX along the way."

(And battery backup is something I hope to avoid -- these devices are going to be monitoring power systems. If they don't have electricity, they have far bigger problems.)

Mike Green

The next best option is the Microchip 23K256 SPI SRAM. These are 32K SRAMs. There's no ready-to-go code for the Stamps, but you'd use the same SHIFTIN/SHIFTOUT statements to access it. It's a bit more complex to use than the DS1302 since you have to send a command byte, then an address. The datasheet is pretty complete in terms of what to send to the chip and what you get back.

Julie in Texas

Thanks!

32K is probably more than I could ever hope to use, but my brain is already thinking of uses!

allanlane5

I think they're physically small, and you don't have to use the whole thing.

Julie in Texas

allanlane5 said...
I think they're physically small, and you don't have to use the whole thing.

I don't HAVE to ... but that doesn't mean I won't [noparse];)[/noparse]

My first computer had 1KB of RAM. I've used "small" before, and I can code in "small" fairly well. But having owned PCs with as much as 16GB of RAM, I'm tired of "small".