24 Apr
2001
24 Apr
'01
8:12 p.m.
On Tue, 24 Apr 2001, Eric Dittman wrote:
I've actually had an EPROM that would not take the new
programming until I erased it, and the programmer did
not require the device to be blank first.
An intelligent (modern) programmer will only burn the location until it
reads back correctly. This is done to avoid stressing the cells and
allows more erase cycles. What probably happened is the programmer wrote
it once, read the location was good, and moved on. For some facinating
reading, look at the programming algorithms for an EPROM some day....
clint
Maintaining original labels is going to be hard to
do,
as the EPROMs age. I've already had to refresh some
EPROMs from equipment as recent as 1982.
Unless your programmer insists on doing a blank-check before programming,
you can often refresh EPROMs without erasing them. And thus the label can
stay in place. You can often refresh OTP devices in the same way.
After all, bit-rot is caused by the charge on the floating gate leaking
away. Charge that was put there when the device was programmed. So it can
be put back by reprogramming _without erasing first_.