Hi
I'm a designing a fully differential opamp for CT Filter. The circuit has different gain and phase margin values according to the analysis I chose.
Case 1:
XF analysis. 1 Port element connected between diff input and biasing voltages are applied thru a AC blocking resistors (50K). Gain and phase curve drawn at the output.

Case 2:
AC analysis. Common mode (vdc) and diff mode (vsin) sources connected thru an ideal_balun (from analog lib). Gain and phase curve of (vop-von)/(vip-vin) are drawn.

There is 6 dB gain difference and 80 degree phase difference between 2 simulations. (XF has better results)

I wonder if there is any trick that I'm missing in the setup and which case is more proper for and OpAmp simulation?

Any helps will be highly appreciated.