My topology is a pi network with only passive lumped elements and I want to use the Y parameters to calculate the admittance of the three parts of the pi network.
As can be seen from the figure, my imagY(2,1)=imagY(1,2)=(imag(Y21) + imag(Y12))/2 (because the three curves in the lower left corner overlap), imag(Y11) = imag(Y22) (similarly, the two curves in the lower right corner also overlap).
So the question is, why is (imag(Y21) + imag(Y12))/2 + imag(Y11) ≠ (imag(Y21) + imag(Y12))/2 + imag(Y22)? (As can be seen from the figure), their curves are not equal, and it is not a simple "negative" relationship. I tried other circuits (even including converting the S parameters obtained by testing passive components such as capacitors into Y parameters and then calculating them), and found that they are not equal? Why?
Ports 3/4 are connected to my passive pi network, and port 12 is connected to the test data I imported through the DAC (testing a passive capacitor)
As can be seen from the figure, my imagY(2,1)=imagY(1,2)=(imag(Y21) + imag(Y12))/2 (because the three curves in the lower left corner overlap), imag(Y11) = imag(Y22) (similarly, the two curves in the lower right corner also overlap).
The difference that you search for is rather small, 0.0x
The visual overlap of the curves in the lower left graph at a scale of +/- 5 would not show such difference.
You better plot the difference to see how much difference you really have, instead of trusting your visual check for curve overlap.