Frequency shift in simulated S-Parameters in HFSS

[mav3rik]

Hi there!

I just started working with HFSS a while ago, so i´m still not totally aware of all relevant settings in my simulations.

The following problem is nagging me right now:

I am designing patch antennas that should work for a frequency range from 76-81GHz and transmit into a WR12 Waveguide. I got a few promising designs out of it and did some prototype boards based on my simulations.

The good news: So basically the S-parameters measured with the VNWA look quite like how they should, according to my simulations.

The bad news: Compared to my simulations, all measured parameters are shifted by 1.5GHz to the right hand side.

I used a solution frequency of 83GHz, since i was told i should choose a solution frequency a little higher than my frequency of interest. I´m not sure if this was the best hint...
Maximum number of passes 15 at a maximum Delta S of 0.03.
Mixed order solution and direct solver method.

I also am using the Huray surface roughness model, because otherwise i can´t simulate the losses correctly.

Any sugestions at where i am doing something wrong?

BR

 

[PlanarMetamaterials]

Which direction is "the right hand side"? Upwards in frequency?
What dielectric are you using?
Are you simulating and measuring the antenna and waveguide together?

 

[mav3rik]

Yes, by right hand side i mean upwards in frequency. So i have to shift my simulated S-parameters 1.5GHz to a higher frequency to make them fit my measurements.

As dielectric i am using Rogers 3003.

I am simulating a short microstripline that feeds my antenna, the antenna itself and the waveguidetransition, that connects the PCB and the waveguide to eachother.
So basically i am simulating everything that i am then measuring on the VNWA.

 

[PlanarMetamaterials]

Try a convergence of delta s = 0.01 (-20dB). As well, you might want to do some more adaptive passes after convergence to make sure nothing changes.