Circular Polarization in Square Slot Antenna

Hi ,

I am designing circular slot antenna ( fr = 475 MHz) on a cube of size 10cm(length) by 10cm (width) by 30 cm (height) cube. My square slot 'wraps around' the cube. I have also a parasitic slot at a distance of 170 cm from the first slot. I want to create the CP using this slot antenna. What is the best way to generate the CP using this design. I want to keep the design as simple as possible. I have excited the two adjacent sides of my slot antenna with two feedlines that are of magnitude 1 and 1<90degree. I am getting very nice CP using these two separate feedlines. But practically I will have to use power divider to make it work . I want to use some technique using striplines that has only one port and that may excite CP in this square slot antenna.

I will appreciate the quick response. Thanks.

If you have boardspace for striplines: from the 50 Ohms source/receiver, make a quarter wave line to go down from 50 Ohms to 25 Ohms (35 Ohms), split into two 50 Ohms lines so that you get two feeds for the two-feed slot antenna. Make one of the paths 90 degrees longer to get the 90 degrees phase shift between the feeds.

A real combiner/splitter (with resistors) is only required of the load impedance can go crazy, but in your case both ports will have good VSWR. You may even design the antenna for 100 Ohms input impedance, then you don't need the quarter wave line as 2 lines of 100 Ohms in parallel gives you 50 Ohms.

You can get circular polarization with a single feed by using cuts (same idea as in circular pol. single feed patch antennas). The disadvantage is that the bandwidth where you have good circularity of the field reduces, and the design takes more time.

I am thinking to design the Wilkinson power divider to provide the 90 degree phase shift. Will it be a good idea to do that. ? Secondly can I send you my geometry so that u may better visualize that what is my design ?

You can make a microstrip Wilkinson divider with a 90 degrees delay in one of the outputs. You can also use a simple T-splitter with 90 degrees delay as shown in figure 8 of:
**broken link removed**.

If you don't get problems with your contractor, just post your geometry here so that other people can comment also.

Here is my geometry and I am describing my problem again. There are two substrates. The lower slot is main slot and upper slot is parasitic slot. Rest of substrate is covered with perfect copper as is required for slot antennas. The two striplines are visible and are feeding the two adjacent sides of lower slots. I have set the magnitudes for both of them as 1<0 and 1<90 and I am getting the CP. But I want to achieve the CP using only one feed and some simple phase shifting network that may excite both sides of slot with 90 degree phase difference.

To be honest, I don't think I understand your geometry. Besides this, as discussed before, you can use a stripline T power divider or stripline Wilkinson divider with 90 degrees delay to enable single feed.

How is the isolation (S12) between the two feeds?

The geometry is complementary to the patch antenna. There are two substrates. The lower surface of first substrate is ground and is all covered with perfect E sheets.
The Upper surface of second substrate is also copper except for the two slots that are meant to provide radiation. The top and bottom rectangular ring is also copper and I have subtracted some copper to make space for the 2 ports on both the sides, where there are strip lines. Is the geometry is clear now ? Kindly let me know.

I have used simplified method of creating CP and I have divided the stripline coming from the 50 ohm lumped port to two 100 ohm lines. Length of one line is longer than other by the factor of lambda by 4. Also what are important things to take care of , if I want to get good CP using this method.

ISloation between two feeds in separate feed configuration is around -41 dB at resonant frequency. It is showing resonance at 496 MHz and I have yet to bring it down to 475 MHz.

I still can't get your structure (sorry for that).

If you don't want to use two additional quarter wave matching sections, you should redesign/relocate the antenna feed so that its impedance is 100 Ohms (for both feeds). When done, you can directly connect your simple T-divider to it (with an electrically quarter wave length differerence in one arm to get the CP). The isolation is very good!

Make sure that you have the right physical length to get 0.25 lambda electrical length. Also check whether your CP has the right direction. If not you need to put the additional quarter wave section in the other arm.

Let me try once again. The structure is all copper except the slots that have been cut and the rest of material are two substrates of equal thickness and feed lines have been sandwiched between them.

My antenna is resonant at 497MHz , while my solution frequency in HFSS is 475MHz. Should I evaluate 0.25 lambda electrical length according to 497MHz or 475MHz ?

I did not understand this part of your answer (If not you need to put the additional quarter wave section in the other arm.) . can you please explain. By other arm , what do you mean..?

Your 50 Ohms line (coming from the source) splits into two 100 Ohms lines/arms. You can't connect them to your two feed points directly when you designed the antenna feed for 50 Ohms impedance (that would result in VSWR=2).

So you need to redesign your antenna feed to have 100 Ohms impedance, or you need to add a quarter wave transformer in each arm (quarter wave line of 71 Ohms). This line transforms 50 Ohms into 100 Ohms so you will have traveling waves in your two 100 Ohms arms.

To get the CP, you need to extend one of the 100 Ohms arms with a quarter wave line

Just to test the principle (that is antenna + simple divider) I would design the divider+phase shifter for the frequency where S11 for each antenna is best (496 MHz from posting #7). When the result at 496 MHz is fine, you can put your focus on changing the geometry to get it resonant at 475 MHz. to redesign the splitter, you just need to change the quarter wave phasing line as all other lines carry traveling waves only. When you use two quarter wave transformers, you need to change their length also.

Here is what I did finally. I have included the details and labels of everything.

Kindly let me know , if there is any mistake.

- - - Updated - - -

Secondly what are some fundamental and critical points to check to make sure that HFSS simulation is perfect. Because even though the HFSS converges to the solution frequency and gives good s11 curve , it remain uncertain that measured results will come out the same or not. I have tried to ask this question from many persons but nobody gave satisfactory answer.

Now it starts to make sense

I don't know HFSS, but i would recommend you to seperately simulate the microstrips and check their simulated characteristic impedance against results based on standard formulas. I assume that the vertical metal walls with slots are structures that will be meshed also. Drawing microstrips that use a meshed metallic structure as ground may result in large inacurracy. I don't how HFSS handles this, so you may check the manual.

Electrically half wave slot antennas (narrow slots) tend to have high current density at the edges (because the edges are close together). To enable the simulator to simulate the current density gradient well, you may use a scheme that allows a denser mesh at the edges. It may not be a concern for you as the slot goes around completely (your wrapped around label). I know IE3D has a feature for that, but I don't know for HFSS.

When you change the meshing density (Cells/lambda) and you get different results, you can be sure you need to increase the cells/lambda, especially at positions where there is a high current density gradient.

You need to figure out many things yourself as vendor documentation mostly focuses on the good things of their simulator only.

Hi ,

I am trying a lot to generate CP in a planar slot antenna at around 500 MHz. Can you please give me some hints as how can I do that. I will be really thankful.