I'm just curious about this because a lot of my PIC chips say in the datasheet that they hold their programming for 40 years. Don't get me wrong, thats long enough, I'm sure anything I make will be forgoten or broken in another way by 2050 anyhow. But I can't find anything that goes into more detail about this and im curious. Do they really loose their programming? What is the main cause of it? Are their chips that boast extended program retention?

 

They probably won't. But they can't guarentee it. I think the old CMOS OTP chips could hold it longer.

__________________
Bill
Smart Kits build Smart People
http://www.blueroomelectronics.com/

 

In 40 years they will not go blank. Likely one bit will change.

 

Yeah, I kind of figured it was just their way of saying that these chips don't last forever. I still wonder if there are special chips that are supposed to last longer. Not that I need it myself.

 

If that, it's a pessimistic specification - but EEPROM's, and EPROM's as well, don't hold their memory for ever.

PROM's should do, as they are physically 'damaged' during programming, so can't 'recover'.

__________________
PIC programmer software, and PIC Tutorials at:
http://www.winpicprog.co.uk

 

Sure. The old 16Cxxx series. You blew tiny fuses, hard to imagine that self repairing.

__________________
Bill
Smart Kits build Smart People
http://www.blueroomelectronics.com/

 

That was only the configuration 'fuses' (on VERY old PIC's) which is why they are still called that today - the program memory was EPROM, and UV eraseable if you had a windowed part.

__________________
PIC programmer software, and PIC Tutorials at:
http://www.winpicprog.co.uk

 

I actually have one of those, a chip with a window, I got it out of an old power box. I dont know exactly what the box did but it was a big grey metal box with a really high capacity lever switch on the side and a circuit board in the back. I've never tried to program it cause I have no idea how, but it has a little window where you can see the chip inside, initially there was a foil sticker covering it.

 

Most probalby it's just an EPROM, and not a processor - the foil sticker was to keep UV out and prevent it getting erased.

__________________
PIC programmer software, and PIC Tutorials at:
http://www.winpicprog.co.uk

 

I did look it up using the code number, I just cant remember what it was now. It probably was just memory like you said. I read the datasheet because I was so curious as to why it had a window, it mentioned about sheilding it because a few hours of UV would erase it. I got a lot of parts from those because they were all in sockets where I could just pull them out. It was a big pile of them too. But almost all the chips were single and quad comparators, I pulled that stuff before Id really learned anything about electronics, so now Ive got a large jarfull of iffy quad comparators, and 2 or 3 larger chips.

 

If I remember correctly?, they were numbers like 2708, 2716 etc.

Commonly they were used to store the operating system for microprocessors, and for larger systems enough to start the system and boot from floppy disk.

__________________
PIC programmer software, and PIC Tutorials at:
http://www.winpicprog.co.uk

 

eeproms and flash memory are basically the same thing just slight variations on a theme. The information is actually stored inside a capacitor in the chip. With chip sizes getting ever lower the insulation rating around these capacitors gets lower so leakage current increases which causes a shorter lifespan because even not connected to anything, just the tiniest little bit of current is leaking on. I doubt it's ever really an issue, but many modern micros have the ability to write to their own program memory so in theory if you had an RTC clock you could write code that could make the chip refresh it's own contents every so many years if you're on the paranoid side.

__________________

 

In the same vein, If you really needed an electronic device to last 100 years, you could put in a couple chips with the same programming, with some kind of query they could run on eachother to make sure it was still programmed right, and if not it could reprogram that chip. But I cant think of many reasons you would need to do that.

 

RAID 5 eproms? Thats an interesting idea.

 

And virtually totally impractical, the cost/benefit ratio is almost 0

__________________