Are these two antenna the same polarization?

Hi Tony,

The axial ratio is specifically defined in the HFSS documentation as Ey'/Ex', where the coordinate system x'-y' is rotated to align with the maximum of the polarization ellipse. So, it is impossible to determine based on two separate antennas that may have different performance.

However, owing to the discrepancies between the two, I would guess they do not have the same polarizations at all bands. You can look at the fields far from the antennas to see if they are similar.

Hi, Planar,
Nice to get your reply.
I checked the axis ratio on the 76--81GHz band, and found on some freq point, it may be circular polarization.
And I designed is the comb antenna, i.e. the 1st sheet in the attachment document, which is widely used in many famous radar vendor, such as Bosch and Continent.
The attachment is the axis ratio summary on the whole band.
Pls comment if I used the RX and TX antenna, is there any problem?
Due to the mmwave, the reflection should be diffused, does this help?
Best,
Tony Liu

I'm not familiar with the comb antenna, are they supposed to be linearly polarized?

I'm not sure what to make of your data; the left and right plots seem to be plotting the same parameters, but contain different data.

The Axial Ratio plots in linear format are not the most informative. The antenna will be circularly polarized when the axial ratio is 1, but linear when the value is large or small. It's difficult to tell when the values are near 0 or closer to 1. For that reason, I would plot these in Log format. Also, you could plot LHCP and RHCP gain to get a better sense of how the antenna is radiating.

Hi, Planar,
I ploted RHCP and LHCP @80GHz.
Phi=0 means the azimuth plane, and Phi=90 means elevation plane.
It seems the antennas are circular polarization.
Best,
Tony Liu

Comb array antennas used in automotive are linear polarized, even that sometime (to avoid cross-polarization) some manufacturers use elements inclination, usually of 45 degrees. But those antennas are still linear polarized.

Because the RHCP and LHCP gains are nearly the same at broadside [(theta,phi) = (0,0)], the antennas appear to be linearly polarized. There is some net circular polarization off-axis, but this is typical of most antennas.

Just a caution, It sounds like you may be using one antenna as a transmitter and the other as a receiver and have established a link between the pair. Now the relative positioning also plays into the evaluation.

Azulykit said:

Just a caution, It sounds like you may be using one antenna as a transmitter and the other as a receiver and have established a link between the pair. Now the relative positioning also plays into the evaluation.

Click to expand...

Hi, Azulykit,
The TX and RX antennas are about 3lamda apart.
I also defined it as 4D radar, almost based on the experience and simple calculation.
But I am curious about how to evaluate the TX and RX antennas together as a whole in HFSS or some other software?
Best,
Tony Liu

This is now a different question. That being said, this to be something that could be done with HFSS. It is essentially the evaluation of s21 (or s12) where one antenna is port one and the other is port 2. The 3 wavelength separation makes a straight forward evaluation possible without needing to using some of the newer modeling techniques where objects are separated by much greater distances. I suspect that you could treat this as a simple finite element model, particularly if you have a computer with plenty of RAM. Enclose your tx/rx antenna pair inside a cubic airbox with sides of 10 wavelengths or so and radiation boundary conditions to simulate free space and see how it works. You could compare the results to Friis to see if you can believe the results.

Are you still near field or are you into the far field?

Hi, Azulykit,
I guess your method is for testing near field. That is part of evaluation.
But for far field, how should I do? Do I need to build a corner reflector in far field?
Best,
Tony Liu

Tony,

The modeling I suggested does not depend on being in the near field. I was just trying to get a feeling for the dimensions when I asked about near/far field to see how it would impact computer RAM requirements. Overall it appears that you could model the tx/rx antenna system with HFSS. If you are thinking of system modeling (tx/rx antenna and some sort of radar target too) the dimensions and separation of the tx/rx antennas is at issue but also the distance to the radar target comes into play. With HFSS the antennas could probably be modeled. If you add in a radar target into the mix the modeled volume would increase and probably become unwieldy with just the FEM modeling features. It might still be possible using the integral equation solving features. Are you now thinking of modeling the two way radar range equation using HFSS?

Of course, it is always possible to build hardware and do actual tests.

I do not understand your question about a corner reflector. Are you thinking about using the corner reflector as a radar target? Would a conductive sphere be more appropriate as a target?

This thread started as a question about polarization and now it seems to have changed into something that might be described as radar system testing. Is this where you are at the moment? Am I understanding your intent correctly?



I just looked at your "77G_Comb...." document above. I am guessing that you have two different arrays of dipoles above some sort of ground plane and the images provided are of only one layer of a multi-layer PCB/antenna. Is this an antenna with three conductive layers and a mix of dielectrics? Am I understanding correctly? I would not expect the two antennas to have the same pattern as you stated above. An orthographic view of your antennas would be helpful to get a better understand of their structure. If I am understanding your antenna correctly I would expect primarily linear polarization in the broadside direction.

You also asked about other software. I am a HFSS user but not as familiar with other packages. I am hesitant to make recommendations particularly as it is not clear to me what you are trying to accomplish.

Azulykit

Thanks a lot, Azulykit.
Yes, a conductive sphere be more appropriate as a target.
I am desinging a automobile SRR MRR LRR radar. The antenna placement are all microstrip patch or comb antennas on the same PCB layers, as the typical BOSCH or Continential radar.