[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Transwarp Card problems



Henry wrote:
Hello Michael.

Actually it's not that some EPROM's are better then others (except for speed), it's a factor of time. Most data sheets specify that a programmed EPROM (and most other devices too) is only good for about 10 years. Now we all know that the programming can last longer then that, but there's guarantee that they will not loose bits from time to time. Really that's the greatest risk.

I understand that they are only guaranteed for a certain period of
time, but in my experience I have not found any that were losing bits.
(I know it can happen, but with a sample of hundreds, I haven't seen a
single one.)

It is important to distinguish between guaranteed life and practical
life.  We are now doing the experiment to determine practical life.

I think it's a great idea to back up these EPROMs, but I haven't yet
seen the beginning of a rising failure rate.  Of course, I could just
be very lucky, but with hundreds of samples, that seems unlikely.

I would expect the problem to show up on EPROMs that were marginally
programmed, or at least not "heavily" programmed.  The older adaptive
algorithms used to program a bit until it seemed stable, then program
it more to double the stored charge.  That's pretty heavy programming,
unless there is high leakage in the die--which could result from high
temperature, stray UV light, or just poor manufacturing quality.  So
far, none of those problems seems to afflict the EPROMs I have.

When you program an EPROM you're 'trapping' a small charge in a cell. The computer in return can tell how the cell is charged - i.e. 1 or 0. It still amazes me how even over a few weeks that the cells in these type of devices don't loose their charge. Most if not all complex devices are programmed like this too these days. Makes one wonder what will happen in the future to old computers. Will a 20 year old CPLD or FPGA still hold its program? Cracking and cloning those devices is monumentally harder then any PAL or GAL - so don't ask! ;-) Although I have read about services that brag how then can crack an FPGA. Must be expensive though. It can be around $1000 just to crack a GAL!

I understand the storage mechanism of EPROMs, and I also understand the
extremely low leakage of intrinsic silicon.  Hundreds of years is a
possible life, and we're already on our way to several decades.

-michael

Music synthesis for 8-bit Apple II's!
Home page:  http://members.aol.com/MJMahon/

"The wastebasket is our most important design
tool--and it is seriously underused."