CST simulate a free space (2-Port) measurement

[niccolinifr]

Hi, I'm new in CST, I would like to know how to setup a simulation to measure reflection(S11 and S22) and trasmission (S12 and S21) with 2 horn antenna and a DUT places in the middle. I have made some measures with VNA but the result doesn't match with the simulation in CST Studio
I attach a picture of the setup I'd like to model:

 

[Amr Wael]

Maybe this video provides something to start from.

You can mirror the antenna to get the second one.

 

[niccolinifr]

This is what I've done, but the S-parameter are too far from what I expected.
There is something wrong with the model i draw but I don't know where is the error.
Boundary conditions are open(add space), I set the reference plane to the DUT interface, the background is air (eps =1 mu = 1) and frequency is E-Band.
Another strange thing is if i remove the DUT and leave the antennas I obtain S12 and S21 near -7dB but I expect to measure 0dB and that's really strange.
F

 

[volker@muehlhaus]

Why do you expect 0 dB insertion loss? The antennas will not direct all power completely to the other antenna.

For far field, you could estimate path loss using De Friis equation, but it seems that you are in the near field where things get complicated.

 

[niccolinifr]

Why do you expect 0 dB insertion loss? The antennas will not direct all power completely to the other antenna.

Because I would like to see the influence of my DUT only, as if you were taking a VNA measurement. In a VNA measurement, before you insert DUT you make a calibration and what you see in S12 and S21 is 0dB, I would like to simulate this condition in CST, or to do something similare so that i can measure the influence in trasmission of my DUT.

 

[volker@muehlhaus]

I understand that you wish to remove the test setup effect, and have some "DUT only" numbers, but it is not that easy.

VNA calibration methods are designed for TEM propagation with a single mode of propagation. Your test setup in the near field, not TEM, with signal leaking around the DUT. It does not have that clean single mode propagation, so I doubt that VNA calibration meth is valid for this setup.

 

[niccolinifr]

I understand that you wish to remove the test setup effect, and have some "DUT only" numbers, but it is not that easy.

VNA calibration methods are designed for TEM propagation with a single mode of propagation. Your test setup in the near field, not TEM, with signal leaking around the DUT. It does not have that clean single mode propagation, so I doubt that VNA calibration meth is valid for this setup.

Ok I understand, so what could I do?
If I model my setup to work in far field, how could I implement VNA calibration method?

 

[volker@muehlhaus]

Long ago during my university days, we created a similar transmission setup, but in a waveguide to characterize thin superconducting sheets. One important aspect in measurement was to reduce leakage, in your case I would place absorbing material around the DUT. Check with known DUT like solid metal plate.

In our project, we did the evaluation "manually" after solving equations for transmission trough the DUT by hand. That took my student a few weeks, but it was convenient to have it in closed form.

 

[niccolinifr]

Long ago during my university days, we created a similar transmission setup, but in a waveguide to characterize thin superconducting sheets. One important aspect in measurement was to reduce leakage, in your case I would place absorbing material around the DUT. Check with known DUT like solid metal plate.

Yes my next step is to simulate the DUT into a perfect waveguide.

In our project, we did the evaluation "manually" after solving equations for transmission trough the DUT by hand. That took my student a few weeks, but it was convenient to have it in closed form.

I think manually is really complicated because of the geometry of the DUT.
I was hoping for the existence of algorithms that would do the calibration in post processing, or similar to the macro for material
characterization (like the picture), where I load THRU , LINE and DUT s2p file and CST MWS give me the result.

 

[jesuschrist]

Hello there, Your device is basically a dielectric so you can use the periodic structures feature of CST MWS. You can define Floquet ports and send plane wave from 1 port to other and then measure reflection and transmission. That is the way to design frequency selective surfaces and your application is similar to that.

 

[rekekurafremo]

Hi,Did you solve this problem ? I tried to simulation to masure reflection parameters in CST but the result is so for like you. Error rate is over the 30 percent

Hi, I'm new in CST, I would like to know how to setup a simulation to measure reflection(S11 and S22) and trasmission (S12 and S21) with 2 horn antenna and a DUT places in the middle. I have made some measures with VNA but the result doesn't match with the simulation in CST Studio
I attach a picture of the setup I'd like to model:
View attachment 167894