Greetings. I'm measuring IIP3 of my designed LNA in 3 different ways, but achieving different results. The question is: which result is correct? All simulations are provided with 2-tone test.

Case 1a. I run HB-analysis with two sinusoids: Frf1= 1.75G, Frf2 = 1.755G, Pin1 = -43dBm, Pin2 = -43dBm. I find IIP3 as IIP3 = Pin + IM3/2, where Pin - total input power: Pin1+Pin2 = -43dBm + -43dBm = -40dBm. IM3 - difference between power level of one of the 1st order tones, and power level of one of the 3rd order tones. In this case, the IIP3 equals -9.5dBm.

Case 1b. I run HB-analysis with two sinusoids: Frf1= 1.75G, Frf2 = 1.755G, Pin1 = Pin2 = Pin(variable) - 3 (in dBm). So, total input power equals Pin variable value. I find IIP3 using IPN curves by choosing 1st order freq 1.75G, and 3rd order freq 1.745G (2Frf1-Frf2). In this case, the IIP3 also equals -9.5dBm.

Case 2. I run AC Rapid IP3 analysis. Input source 1 - input port. Freq1 = 1.75G. Input source 2 - input port. Freq2 = 1.755G. Input power = -40dBm. Boxes named "Power 2 and 3" remain empty. 1st order and 3rd order frequencies are set same as in previous analysis. In this case, the IIP3 equals -13.5dBm.

Please, tell me which result is correct? Maybe in equation IIP3 = Pin - IM3/2 I should use the the power of only one of 1st order tones as Pin? Or Pin is total input power, equal to power sum of the tones?

Thank you.