How to combine results for an antenna that uses multiple ports?

Hi there, I have a quick query about how to combine results for an antenna that uses multiple ports, with a different phase excitation for each port. You see I've designed an open-ended waveguide antenna fed using balanced probes, which are excited with 1|0° and 1|180° signals, well in CST it's straighforward to excite each using a coaxial feed with a discrete port, and then use the 'results > combine results' function to provide the phase shift for port 2. However, I have experimental results from farfield measurements of the fabricated structure and I'm just wondering if I'm combining the results correctly. I have the gain in dBi and phase of the received signal at certain frequencies and I've been combining the results in the following way:

mag1 = 10^(gain1(dBi)/20)
real1 = mag1 * cos(phase1)
img1 = mag1 * sin(phase1)

mag2 = 10^(gain2(dBi)/20)
real2 = mag1 * cos(phase2+180°)
img2 = mag1 * sin(phase2+180°)

realsum = real1 + real2
imgsum = img1 + img2
magsum = √(realsum^2+imgsum^2)

gainsum(dBi) = 20*log(magsum)
phasesum = argument(realsum + j*imgsum)

I'm not sure if this is the method CST uses to combine results, it's hard to compare as CST doesn't give phase information for the farfield results :( I'm especially uncertain about whether the logs have a factor of 10 or 20, aside from that it's just a straightforward vector summation I know, could anyone shed any light on this? Thanks!

I guess you are going the right way as we need to o vector sum of the fields for both the ports.

Do you think it's correct to use a factor of 20 in the calculations though? I think since it's the electric field intensities of the farfield that are being summed (i.e. in V/m) it is correct to do so but I wasn't entirely sure, I know that when you apply 'linear scaling' to the farfield results in CST they use 10^(gain(dBi)/10), which I presume is because these values are related to the power of the farfield. Can anyone clarify this?