very serious CST 2006 BUG?

I think there is a serious bug in CST 2006. Here is the problem. I have a simple project: just two waveguides facing each other. I ran the simulation 4 times with the same exact settings. To my suprise, the results of these 4 runs were quite different than each other. I have the latest service pack installed and I am using a licensed version (I mean legal). I have attached the results and the project file to my post. Please let me know whether you can produce the same symptoms.

Hello Irfan1,

I'm wondering, being that it's a square waveguide, maybe you need to specify a mode polarization angle?

I did that, ran it twice, and got the same result twice. Attached is the S-parameter result.

thank you very much for goingh through the trouble of running the simulation. But I wonder whether you can get the same result at successive runs without any modification to the original file.

Added after 2 hours 32 minutes:

Problem solved. I uninstalled all the previous and new versions. And then re-installed only the new version. Now, it is working.

Did your results look like the one I posted?

I first ran it twice with your exact settings, and the S-Parameters looked the same each time (like your first run example).

Then I looked at the port modes and saw that one of them was TE10 and the other apparently both TE10 and TE01.

Then I set the mode polarization angle to 0.0, and got the result that I posted, which I think is correct.

Out of curiosity I later went back and ran it a few more times using your settings again, and this time the results were different (some were like yours and some not).