HFSS: "Correct" Port Options for Microstrip lines based RF Circuits

Hi

I have recently shifted from Agilent's Momentum/EMDS to Ansoft's HFSS to simulate microstrip lines based RF Circuits (like, WPD, BLC, etc.). Need suggestions for different port options for correct simulation.

Currently, I am trying Waveport option with dimensions being calculated from WavePort Calculator @ http://www.emtalk.com/waveport_calc.htm
Is this sufficient for simulating a port with port impedance of 50-Ohm?
OR does it indicate approximate dimensions and need to be calibrated manually using "ports only" simulation(s) to reach the 50-Ohm target?
What should be done in case required port impedance is different, as in a Wilkinson Power Divider with different port impedances?

Will simulating the SMA connector (which I use in actual fabricated circuit to give RF signal to the circuit) in HFSS, as port for RF circuit, be more accurate?
Can the above SMA port be modeled through a simple co-axial cable, with waveport at its ends OR there are other accurate ways?

Thanks in Advance!
Vivs

My confusion is further highlighted as below:
Through discussions with my colleagues and reading certain online comments, I observed following "thumb rules" (TR) have been suggested to decide waveport dimension for a microstrip line based RF circuit simulation:
(a) in HFSS (TR1): Wp = 5 x Wt; Hp = 5 x Hs
(b) in HFSS earlier version Help (TR2): Wp = 20 x Wt; Hp = 15 x Hs
(c) in HFSS v15 Help Document (TR3): Wp = 10 x Wt; Hp = 8 x Hs
(d) in ADS-2008 Momentum/EMDS: Wp = 07 x Wt; Hp = 04 x Hs
(e) in ADS-2009 Momentum/EMDS: Wp = 21 x Wt; Hp = 11 x Hs
(f) in CST Help: Wp = 10 x Wt; Hp = 5 x Hs
(g) a suggestion in EdaBoard.com: Wp = 10 x Wt; Hp = 10 x Hs
(h) some else suggestion to use Waveport Calculator in em Talk (mentioned earlier in the problem)
where, Wp and Hp are the width and the height of a waveport, while Wt and Ht are the microstrip line's width and substrate height respectively.
I wish to know the concrete method (rather than any ambiguous "thumb rule") to ensure correct power transfer to my RF circuit, without any mismatch at the input end.
Thanks