nobody has ever found pi to be equal to anything but pi, and if they used a numerical representation in any rational form they were just estimating. it wasnt even until like 17th century until it was found to a higher degree of accuracy.
rational approximations were used (22/7, 256/81, 25/8, 355/113...) for a long time, id say about 3500 years minimum. in fact, we still use rational approximations, but we generally have accepted the value of pi as an irrational number.
the paper definition of a ratio between the circumference and diameter has been dated to before 1650 BC, and although archimedies made the first decimal approximation the definition was already on paper. so the exact value had already been defined. in 1650, thus, archimedes could not have found the exact value of pi. rational estimations had also already been made for use in egypt. beginning with pi=3, then to 22/7. that was before archimedes, and 22/7 is as close to pi as 3.14.
there is also some evidence that babylonians used the 25/8 figure for the ratio, but had never defined the ratio to be its own value. that was like 2000 BC! so the first approximation was like almost 2 millenia before archimedes.
the origin of the use of the letter "pi" to represent this ratio didnt start until nearly two millenia after archimedes. so i don't see what he has to do with it, excluding his excellent geometric estimation that teacher still tell us to use if our calculators don't have a pi button, 22/7.
the proof that irrational numbers existed was what pythagoras objected to allowing to become public. and indeed, the student who proved it during class was killed some months after in a suspicious fishing accident. here's a little proof on irrationality
http://www.homeschoolmath.net/other_topics/proof_square_root_2_irrational.php
heres a good one
http://www.symynet.com/fb/pi_hoax.htm#
more use between 1700-1800 gained pi the rep that it has today, one of the most, if not the most, usefull trancendental number.