in my opinion , the circuit is determined, the ip3 is determined too.
but when i simulate rfic by ads , the ip3 is changed with input power.
what the matter?
Yes. IP3 does change with input power. The closer you are
to the -1dB compression point of your device the worse your
IP3 will be. The Intemod-products grows 3 time faster that the
fundamental and as you go into compression the fundamental
slows down but the Intermods don’t (or not as fast).
I'm not agree with Element_115, because the IIP3 of circuit is depends on the circuit characteristic not the input power. but for the right measurement of IIP3.
If you apply a large signal to circuit then the measured IIP3 is not right.
so you should apply a small signal to circuit for IIP3 measurement and measure the IIP3 in linear region of circuit not the nonlinear region or near 1-dB.