inequality for IM products? plz

Sources of inequality for IM products? plz
as u know, testing of a practical RF amplifier design over a range of
tone-spacings will usually result in some diffrences emerging in the
levels of the upper and lower IMD products of a given order or a chang
in the level of both,with changes in tone spacing.
the origin of the differences is what? maybe from AM-AM and AM-PM,
maybe supply modulation effects, maybe bias modulation effects.
give me details with the above origins.....or upload the relative IEEE papers.....
and, show me the methods of Measurement of AM/Am and AM/PM Characteristics,especially pulsed versions.....urls and pdf are welcome.
thanx alt

The most common explanation for the inequality in the IM products is due to memory effects. If you plot Upper/Lower IM's versus tone spacing and you find that the rfelative difference between the upper/lower is a function of the spacing then you know that you have memory effects. If not, then most likely it is due to the difference in the envelope impedence; in other words the impedence at the envelope frequencies (2w1+w2 or 2w2-w1..etc) are different for the Upper/Lower IM bands

Regards,
Brett