Linear polarized microstrip patch antenna - design help

[Tom48]

hi there,im pretty new to antenna design, still have a lot of theory to go through, so please be kind. ;)I have to design a linearly polarized microstrip patch antenna (rectangular) that is fed with 2 coaxial probe feeds.

The patch should be a square (L=W), so i can place 2 feeds orthogonally to each other.

That way i will hopefully be able to receive/separate the vertically and horizontally polarized signal "portion".

(thats why the antenna needs to be a square, as both signal "portions" (vertical and horizontal) are on the same frequency).

So far i have only modeled and simulated one coaxial feed, i would like to optimize the antenna first before adding the second orthogonal feed.

I want to design the patch on FR-4 substrate for 867 MHz.

I have done various calculations (found on the net and Balanis 3th edition) to get my patch width and length.

Also tried to calculate the inset position so i can match the feed to 50 Ohm.I made a first design in CST and i just can not get rid of the "zero point" in the middle of the radiation pattern.

Some hints on what else i could try to improve the radiation pattern of the antenna would be great! :)

(radiation pattern is attached)

Is a good/the common radiation pattern even possible with a square shape (W=L)?

Most patch antennas i saw so far with the common radiation (like e.g. this: http://forums.radioreference.com/attachments/scanner-receiver-antennas/32485d1305662856-radiation-pattern-dual-band-microstrip-patch-antenna-3.5-ghz.jpg) pattern are rectangular (W!=L).The return loss is quite good, about -31dB at resonance frequency of 867 MHz.I suspect the coaxial feed to be modeled so it does not result in a 50 Ohm impedance at the port, i used the following

parameters in order to modulate my coaxial probe in CST:inner radius (pin): 0.5 mm

outer radius (coating): 2.5 mm

teflon: 2.1

What dimensions do you usually use for a coaxial feed to match to 50 Ohm?Some help/hints are highly appreciated!

Thank you so much in advance! :)Kind regards, tom.

 

[Paul Colby]

It looks to me like your patch is two large for the desired frequency. To get a peak along z you need 1/2 a wavelength in the patch mode when viewed as a transmission line. This will place a maximum of the patch current on the origin. The pattern looks like a full wavelength under the patch yielding two current peaks 180 out of phase.

 

[Baluncore]

Is there a ground plane on the bottom of the FR-4 ?
How do the two coaxial feeds connect in quadrature to the patch ?
Is there a connection between the ground plane and the patch ?

 

[Tom48]

hi,

thank you so much for your answers! :)
I made a mistake causing the patch to be a full wavelength patch, just like Paul already suspected!

I have another question, what level of cross-polarization is acceptable in general for low profile patch antennas?
What ways to reduce cross-polarization do you know?

thank you!

 

[Paul Colby]

Be careful how polarization is defined. Two feed will yield two polarizations neither of which need be linear at any particular point in the far field. So, a lot depends on what you're trying to achieve. To get maximal independence of the feeds you need to reduce the cross coupling between them in the band of operation. In your modeling software monitor the power out of one feed while driving the other. Be sure to properly load the monitored feed.