Wouter van Ooijen wrote:

>> You and Harold suggested  to use range rather then single value.
>> How wide range should I use? suppose I m looking for 951
>> is 950-952 acceptable ?
> 
> depends on what *you* want.

Which includes how stable your readings are. 

Basically, you want to minimize false positives (readings that fall into
your selected range without being really the ones you want to catch) and at
the same time minimize false negatives (readings that fall outside your
selected range when they are really the ones you want to catch).

Sometimes there is a range that has almost 0% of both, sometimes you need
to live with a few percent of false results of either or both.

The degenerated range of checking for a single value usually minimizes
false positives but gives you a lot of false negatives. By broadening the
range, you include more readings and decrease the number of false
negatives. But there probably will be a point where false positives start
to appear. 

Gerhard

-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist