Measuring IIP3 and the Conversion Gain of a Mixer in Spice Eldo

[solrcarlos]

Hello guys,

I am designing a differential mixer and I am facing some problems with measuring its conversion gain and IIP3. Trying to solve this problem I am using a block in verilogA of an ideal mixer, but the IIP3 and ConvGain set in the block are not the one I am measuring. Hereby is the code used.

****** INPUT SPECTRUM
.EXTRACT FSST LABEL=VDB_RF	   YVAL(VDB(RF+,RF-), F_RF)
.EXTRACT FSST LABEL=PDBM_RF     YVAL(PDBM(VIN), F_RF)

****** OUTPUT SPECTRUM
.EXTRACT FSST LABEL=VDB_IF	YVALl(VDB(IF+,IF-), F_IF)
.EXTRACT FSST LABEL=PDBM_IF	YVAL(PDBM(Rload), F_IF)

****** CONVERSION GAIN
.EXTRACT FSST LABEL=AV_DB	MEAS(VDB_IF) - MEAS(VDB_RF)
.EXTRACT FSST LABEL=AP_DB	MEAS(PDBM_IF) - MEAS(PDBM_RF)

****** IIP3 CALCULATIONS ****** 
.EXTRACT FSST LABEL=O1	 	yval(VDB(IF+,IF-), F_IF)
.EXTRACT FSST LABEL=O3_A	        yval(VDB(IF+,IF-), F_LO - (2*F_RF - F_FUND3) )
.EXTRACT FSST LABEL=O3_B	        yval(VDB(IF+,IF-), F_LO - (2*F_FUND3 - F_RF) )
.EXTRACT FSST LABEL=IM3_1 	{ meas(O1) - ( meas(O3_A) + meas(O3_B) )/2 }
.EXTRACT FSST LABEL=IIP3_A3 	{ meas(IM3_1)/2 + Pin }

What is wrong with all these ? Thank you so much guys!

- - - Updated - - -

Just to add information to the problem.

View attachment IIP3_A1_A3_v2.png

Should I just consider that IIP3 and the gain as the OIP-IIP3 given by the graphic?