SV: Re: [sdiy] DROID3, Another AVR Synth?

[ASSI]

On Sonntag, 12. Februar 2006 01:04, karl dalen wrote:
> A pecularity, for example, Ericsson uses/ed the MC68HC11 MCU's
> in every GSM base station for years, and i know that most of these
> systems are still in use today, so lets take one station made in
> 1990 the data retentention is stated  to 10 years over selected temp
> range, don't remember wich, but i have a bounch of these MCU's
> and they are still going strong no failures, lets see,  they are on
> their 16'th year of working!

That doesn't surprise me very much. Even in a GSM station (which is a 
harsh environment) you are not at worst case conditions all the time 
and to be able to specify 10 year data retention you need to have 
sufficient margin - and that margin has to be some multiple of the 
specification for a temperature activated process with relatively low 
activation energy like retention.

> What if you exceed the maximum number of
> writes, shouldent that fuzz up the memory?

Yes. If you _exceed_ the maximum number of writes anything can 
happen. :-)

More seriously, cycling a Flash memory creates tiny defects that will 
eventually ruin retention and the ability to program and erase the 
memory cells. The details are complex and different from one memory 
type to the other. The first cycles or so don't affect the retention 
very much, depending on memory type this is in the 100 to 1000 cycles 
ballpark. 


Achim.
-- 
+<[Q+ Matrix-12 WAVE#46 Neuron microQkb Andromeda XTk sonic heaven]>+

SD adaptation for Waldorf rackAttack V1.04R1:
http://Stromeko.Synth.net/Downloads.html#WaldorfSDada