Critical Cells problem in CST

I'm trying to simulate a log-periodic antenna in CST 2006. I've placed the port in the center at the feed point.
Transient solver keeps reporting error "port is inside of a perfect conductor" (not the exact words) and terminates.
Since I'm a beginner with CST, please help me solve this problem. Check my structure for errors...
Thanks!
EDIT:
it seems i have a problem with the meshing. Some critical cells are expanded with PEC and this causes a short circuit to my port...

hi: I simulated a log-peroidic antenna with CST in these days. Maybe I can help you. However I only has CST 5.0. I cannot open your file. Can you convert it into a lower version. In CST5.0, one antenna project includes a lot of files, you need to upload all of them.
Do you compare this log-periodic antenna simulation with CST and with other sover, such as ADS momentum, HFSS. I found momentum has the similiar result. However, I could have a wide band simulation. It consumed a lot of memory.

i dont understand

yzriver, thanks for your input!
I had somehow found a solution for my "expanded critical cells" problem. I had modified one parameter in the meshing options menu and critical cells no longer short circuits my port.
Surprisingly, CST and HFSS results actually do compare very well!
However CST still needs a lot of time (its not as memory hungry as HFSS). For 2M cells it required 14 hours for 1st -pass, second pass had 3.5M cells and I terminated it, cos I wouldnt wait a couple of daysfor results...

rfmw,could pls upload ur sat and mod file of ur logperiodic,i could help u in my cst 5.1
regards
note ur substrate thick and epsr=?

rfmw: cong.
could you tell me the band width of your antenna. Mine is from 200G-2T. The mesh number in HFSS will be huge. Could you please take a look at my case in h**p://www.edaboard.com/ftopic234779.html. I couldn't have the similiar result as in CST and momentum. Could you please try 100G~200GHz, it should be done faster.
Thank you.

rfmw,try to increase ur frequency band in simulator

I'll try to save files for your version 5 and 5.1. Will appreciate it very much if you can take a look at it. The main problem I have are these damned "expanded critical cells", which short circuits my port!
@abdoeng
"note ur substrate thick and epsr=?"
Cos the application is microwave and mmwave planar antenna, my substrate is 150um-thick InP (eps=12.4) and HR Si hemispherical lens (eps=11.9).
@yzriver
"could you tell me the band width of your antenna. Mine is from 200G-2T. The mesh number in HFSS will be huge."
Mine is from 30-300GHz. Indeed I could not go far with HFSS on a 32-bit CPU/32-bit WinXP with 2GB of RAM. 120GHz is the highest I could reach for my structure. I can never imagine to reach 300GHz in HFSS, at least on a 32-bit maschine that is...
@abdoeng
"try to increase ur frequency band in simulator"
Currently in CST I have 30-150GHz and the cells hit 2.2M and the simulation for the first pass took approx. 14 hours. I feel that if I enlarge frequency band, I'll get into deeper trouble.
Thanks a lot for your help and tips. I'll try to prepare the files asap...

hi rfmw:
In my case, I would spend more than 24 hours to simulate from 100G to 500G before I stopped it.
The question is which one is better, CST or HFSS. There were hundreds of post to discuss this one. I am consider to install one of these in cluster.

@yzriver
sorry, i will not simulate your structure before you update your lumped port. I think you will have to get the correct results after you place the port correctly. If you again have trouble, then i'll give it a go...