Patch Antenna Design and it's challanges (low cx-pol, robust design) 1

[DenVo]

Dear Experts,

I am designing a patch antenna in X-Band. This patch is a multilayer aperture coupled (H-polarization) and pin coupled (V-polarization) patch antenna which uses distances between the patches and the reflector plane. The patch is designed for good cx-pol suppression in both polarizations.

The distances between the layer are my problem. Inside my simulation I use a ROHACELL material to establish the distances of arround 3-5 mm.

The problem is that ROHACELL is not easy to combine stable with substates. The advantage is the low E_r=1.005, which results in a high antenna gain in my design.

Has somebody similar problems or an idea to separate the patches and refelctor on another way? Or a hint to use other material for separation or frame structures? If frame structures are an option, the frame should not siginficatly influence the antenna radiation pattern. The frame thickness should not exceed 1.5mm, because the frame separates subarrays from each other and the max. element distance is arround lambda/2 for phasing the beam from broadside to endfire direction.

I have high requirements on reproduceablity of the antenna pattern and therefore high requirements on fabricating such an element. The antenna parameter are used in the radar equation. The beamwdith and gain are from extrem importance.

Regards DenVo

 

[Higgler]

Rohacell is also a pain to get good thickness tolerances. It has burrs on it when you cut it. Do some test coupons for glueing your rohacell to other substrates, you'll find the concern.

Your writeup needs a picture, your words are not clear, your question is not clear.

 

[DenVo]

Hey Higgler,

thank you for your response. I try to give you a better picture of the design.

This X-Band antenna is a multilayer structure with around 4 substrates. On the separate substrates are the radiating patches which are pin coupled for V-Polarisation and EM- coupled for the H-Polarization, the feeding network, shielding GND to reduce unwanted radiation generated by the feeding network and a GND reflector.

To establish now the distances between GND reflector and patch, I use a foam-layer with 5mm height. What I noticed by discussing with PBC manufacturer is, that they are not able to handle distances of 5mm height between substrates in a multilayer process.

My main question is, if somebody knows other manufacturing ways/ methods to establish these distances. Are ceramic frame structures or plastic cast technologies an option?

Is it not state of the art to use foam layer in a multilayer microstrip patch antenna design? I know from NASA and ESA satellite projects, where these kind of antenna structures are in use. How they handle this, especially for this temperature range?

Regards DenVo

 

[Higgler]

I understand the rohacell for a larger patch and hence higher gain. Are you analyzing this with expensive computer software? HFSS or CST? Or empirically building it?

Aperture coupled/fed patches radiate backwards I know, and some effort has been made to reduce that backlobe. What are you doing for that backwards energy? feeding it with stripline? I'm used to seeing a microstrip aperture fed patch.

The ground plane size has a large effect on gain. Can change it by 4 dBi going from one Lambda to 4 Lambda square (per Pozar's yellow book).
If this is a single dual polarized patch and not an array, and you want higher gain than normally attainable from a patch, there are 3 things you can do for that.
1) make ground plane as large as possible, or add a cup shape. i.e. put patch inside a metal box. Control gain and pattern shape by depth into box.
2) Add a single director out in front of the patch, dual pole director would be a + or X shape
3) Patch floating 1/4 wave above a ground plane gives +10 dBi gain. Read that one in the Antennas APS monthly articles. Essentially making a patch on stilts and let the ground plane become an image antenna to increase your gain.

 

[DenVo]

Hey Higgler,

nice that you are open for discussion. Indeed I simulate/optimize the design by using CST.

The single patch is part of an array with arround 32 patches in one subarray.

What should I do with the backwards energy? I optimized the design for low back radiation and good cx-pol results. Do you have special methods to handle this?

It seems to me, that you don't like aperture coupled patch antennas. Why?

One point you mentioned is interesting. You expect more gain, by implemeting a patch antenna in a metal box. Is the same true for a subarray of 32 patches or would you expect problems with mutual coupling between the aperture coupled patches?

Could you please post a reference paper on your point 3. I can not imagine how a patch is floating above the gnd.

I have the yellow book too ;-).

Greeds DenVo

 

[Higgler]

Array designs are different. What is your bandwidth and scan angle? What is the end use of this array?

One problem with having "too much gain" in a single element in an array is that excess gain can be the cause of a scan blindness.

I can see having a single element (no array) be two different types of antennas for convenience, but arrays usually want both elements to be the same type and have the same phase so that you can create circular polarization.

A simple patch with probes in two locations or 4 locations would work.