Your description confuses me a bit.
Is vertical and Horizontal radiation patterns the polarization of the antenna or the physical location.

The location on the roof of the car (center vs. edge) will affect results a bit more than other items.

Changing dielectric material thickness a little should primarily affect antenna bandwidth, with no change in the antenna pattern. Unless your dielectric thickness increases to between 1/4 and 1/2 wavelength.

kchiggins

Hi,
The signal polarization needs to be mainly horizontal, aalthough I accept that the vertical influence will be there too because of the proximity to the car roof/groundplane.

Johnor,
that's unusual use of a patch antenna, the primary polarization is vertical polarization for a flat patch mounted flat attached to a car roof.  The H plane polarization is usually horizontal for angles far off broadside, but placing the antenna on a large groundplane/car roof reduces the horizontal polarization gain at low angles towards the horizon.

Why do you need Horizontal polarization?

If you really need horizontal, you need to model the ground plane effects accurately.
You'd need to mount a patch on a (?2"diameter) vertical tube with the long length of the patch wrapped around the tube horizontally. The antenna pattern shape as you look upwards in elevation will be set based on the height of the antenna off your car roof. You can get good horizon coverage if you don't mind large dips in antenna gain upwards in elevation.

kchiggins

Hi,
I have simulated a simple patch structure with other software (mmana)and there does seems to be a reasonable horizontal pattern.I am into amateur radio as a hobby and use the works cst software during my lunch hour, to model antenna for home use.

I believe you for a single antenna simulation, The current flows along the length of the patch which would look horizontal polarization on the sides of the car and vertical polarization in front and behind the car (if you oriented the long length of the patch with the long length of the car). But adding a large ground plane to simulate the car roof in the analysis may not be possible. Horizontal polarization can't really travel across a flat car roof but vertical polarization can. If you can add a small circular or square ground plane, try it and you should see the horizontal polarization die out at the horizon. You may get in the -20 dBi gain range.

I've seen horizontally polarized omni antenna papers for patches, but they bent them horizontally around a vertical pipe.
kchiggins

Hi,
Ok..will try this and see what results I get.I also am looking into placing two patch antennas vertically mounted &  back to back with a gnd plane inbetween and then feeding both at the same time.
Any ideas on the phasing possibilities ???

Thanks for your input.

Johnor,
that sounds ok. Will they be Horizontal polarization patches? I'll assume so.

There are four options;
In phase combiner (two patches point in same or opposite directions).
Out of phase combiner (two patches point in same or opposite directions).

 
In phase combiner, antennas facing the same direction gives deep antenna pattern nulls frontward and backwards (assumes the ground plane in direction of the car front to back).

Similar but opposite results as described above, if you use the out of phase combiner.

In phase combiner with antennas facing opposite directions gives peaks front and back, and bumps in the 10-40 degree zone either side of front and back if you have a large ground plane. If the ground plane is very short, you minimize those 10-40 degree bumps. A short ground plane idea mimics the  patch on a cylinder idea flattened. It may be best to keep the ground plane small then, face the antennas opposite and use an in-phase combiner. If you don't want to buy a combiner, you can take one coax. and solder it to two coaxes, i.e. one 50 ohm to two 50 ohmers (= 25 ohms). With short lengths of line you can match into your patch pair by moving the location of the feed. I've seen that done before to save space and cost.

Plus, your car hood is going to make some odd antenna patterns if you look upward due to the roof energy reradiating.

kchiggins.  

Hi,
Yes, they will be horizontal, and I am expecting  the car roof to play a big part in the radiation pattern.I will hook up a network analyzer and see whats the antennas are doing when connected in/out of phase.