I can't understand meaning of your questions.
Please explain more clearly using correct terminologies.

IP2(IM2) have various type. On the other hand IP3(IM3) definition is unique.
What type of IP2(IM2) do you intend to evaluate ?

http://www.designers-guide.org/Forum/YaBB.pl?num=1193566207
http://www.designers-guide.org/Forum/YaBB.pl?num=1216663909/6#6

LeroyJenkins wrote on Dec 28^th, 2009, 8:24am:
You don't seem to understand IP2.
Being different from IP3, there are many definitions for IP2.
Again see http://www.designers-guide.org/Forum/YaBB.pl?num=1216663909/6#6

LeroyJenkins wrote on Dec 28^th, 2009, 8:24am:
For IP3, it is easy.
See http://www.designers-guide.org/Modeling/modeling-rf-systems.pdf

An explanation of intercept points is given in http://www.designers-guide.org/Analysis/rf-sim.pdf starting on page 35.

IP2 characterizes even order distortion and IP3 characterizes odd order distortion. IP2 and IP3 are conceptually similar, so I will only describe IP3. At low power levels, the third order distortion will dominate over higher level odd terms (5th, 7th, etc), thus at low power levels if you double the signal power the odd order distortion will increase by 8. If you keep increasing the signal power level then eventually the contributions of the higher order terms will dominate over the third order term, but IP3 is measured by extrapolating from low signal power levels, so this effect is ignored by the IP3 measurement. By definition the third-order intercept point (which is referred to as IP3) is the level where, based only on extrapolation, the power in the third order term is equal to the power in the fundamental (the first order term).

The actual number for IP3 does not represent a signal level you could hope to achieve from the circuit. Rather you only use it to compare the distortion of two designs. If all else is equal, the one with the higher IP3 is more linear.

-Ken

Okay, so let me see if i've got this right.

In general something like an RF amp is considered a static nonlinear system, that is it can be modeled with an indefinate-order polynomial equation.

If we take the first 3 order terms the first one is the one we are usually interested in, this is the linear gain times the input. Every other higher order term in the equation is essentially distortion contribution.

The constant before the second order term is related to IP2, and the constant before the third order term is related to IP3.

As input power is increased, the output power increases linearly (ideally indefinately, but practially reaches compression). The issue is that in 'intermodulation terms' the output power actually increases quicker (because the slopes of IP2 and IP3 are 2:1 3:1 accordingly) so a system can become sucepcitble to distortion before compression.

The purpose of extrapolating IP2 and IP3 to an actual intercept point (although its not a realistic power reading) is to put some sort of quantifiable weight on how suceptible a system is to intermodulation distortion?

Please correct me where i'm wrong... as you can see i'm simply trying to paint a picture of this in my head.