On Sat, 27 Feb 1999 14:39:14 +0100, you wrote:

>I asked this question already to the list, but I did'nt exactely understand
>the answer !!
>
>So, I'll ask my question again !! Is it normal that when your program uses
>1/5 of the total amount of memory, the UV erase time is about 20 minutes ???
>Also, is it normal that the UV-erase time depends on the amount of memory
>used by a program ??
20 mins sounds OK.
>And, if it IS normal, can someone give me a the full explanation ??
There isn't any dependence on the _amount_ of memory used, but there
could be on _which_ locations are used, which may appear to give
similar results - different parts of the chip may have different
sensitivities, or (more likely) the edges of the chip are being
shadowed by the edge of, or dirt/defects in the erase window 
It's also slightly conceivable that a poor programmer could change the
programming voltage progressively, e.g. due to converter heating up
etc., so not all cells are programmed with the same threshold.