Problems when editing sources in HFSS

I have an antenna array that I simulated. I edit sources and I found that the results are accurate only for certain phase choices. My antenna array is symmetric so the results should be symmetric for many phase choices when editting sources. Has anyone had similar problems or know how to fix this error? Thanks.

ya man u r right. for this problem u have to increase no of solution setups or no of passes. that may be between 15 to 20.it will need much time and large memory...
u will have to wait as per yr processor speed and ram capacity.
secondly dnt get 2d solutions,make polar 3d plots...u will surely get all yr desired shapes.

try it.

regards
ali

strange reply by "Rf engineer"!
dear kevinj, do you use a registered version or a cracked one?
editing sources is a post-processing action and It has nothing to do with solution setups or no of passes. If the structure has been analyzed correctly, everything should be predictable.
I have heard that cracked HFSS has such problems.

have you tried using "master/slave" boundary condition?. I know this is not the question you have asked, but it may be another option for array analysis.

Added after 32 minutes:

kevinj, I just tried with two element antennas. I used edit source and my results are correct. I use V.11 of HFSS. I don't think your port definition will really matter, but just in case you are interested, I am using wave port for driven terminal.

eastern, you may be right. i have both the cracked and full version. I'm testing on the full version later today. If you are right then you get a helped. lol. regardless, I'll post the result of cracked vs. full version when editting sources. Thanks.

Hi. I found out what the problem was. I made a mistake when I created the lumped element. I created some of the elements directing from the top layer to the bottom layer and some from the bottom layer to the top. So they were out of phase by 180 degrees. HFSS 10 tells you this when you edit your sources. Version 9.2 doesn't. Hopefully this will help someone in the future. I found no difference from the full vs. cracked version. Also, I perform most simulations using 12 passes ( I think that's default). Thank you for your time.