FEKO Model of Exponentially Flared Horn - Question

Hello everyone,

I'm attempting to accurately model the Narda 642 standard gain horn. The aim is the match up this model with measurements taken in an antenna chamber, specifically pattern and gain. As far as I can tell, this is an exponentially flared horn, with the exponential flare being both in E and H. Can someone verify this? I can't seem to get any good information from Narda's support.

I've generated a model in FEKO based on the above assumptions with a waveguide feed. Unfortunately, results aren't matching up as well as I'd like with measurements in chamber. So a few questions regarding horn antenna modeling:

1. I've used PEC conductor sheets. Should I actually account for metal thickness (about 3.2mm)? Will this affect the pattern significantly? I added finite conductivity as a material instead of PEC and the difference was negligible.
2. The actual measurement featured a waveguide (WR137) to n-type adapter. The simulation lacks this. Should it be included?

Any general guidance or links to literature would be appreciated. I'm used to modeling planar structures, which is a different game.

Thanks!