Embedded Files
Patch Antenna
1. First Step: Linearly Polarized Antenna
• Calculate the physical length(L=W) of the patch antenna
Considering the substrate Rogers RT5880 lossy, we have found the length L at (1). Of course, we are going to tune it in order to have the proper resonance.
Constructing the proper layout
We have constructed the components properly. Note that we have placed the feeding line to the lower part, not the upper part as instructed. Therefore, we have defined the port in the positive(not negative) y-direction. We have found Wf as 1.96mm using the ”Calculate analytical line impedance” feature of CST. You may see some important parts of our design below figures.
Figure 1: feeding line placement
Figure 2: Short wire
Figure 3: Overall design with feeding line highlighted
Inserting the symmetry plane
In our case, the electrical field is in a positive y-direction because of our feeding line position. Thus, we cannot insert PEC(Et=0) plane for the yz plane because Et is in also the positive y-direction which contradicts our electrical field vector. So we have placed PMC(Ht=0) plane for the yz plane in order to halve the mesh dimension and thus the solving system dimension. Note that we have also checked this placement by trying both PMC, PEC, and no symmetry plane solutions(We have the actual version of CST Studio Suite).
Performing transient analysis
As instructed, we have used a variable ”yc” which is the distance of short wire position from the lowest part of the antenna patch. At first, we had to guess this value and we have put it as 5mm. This will be adjusted according to the matching procedure later. Moreover, the instruction says that a mesh with 10 lines per wavelength but we have to simulate with 15 lines per wavelength since we do not have a limit(non-student version).
Figure 4: S11 behavior of our first attempt
As it can be seen in Figure 4, S11 is the lowest at 2.355Ghz. So it needs to be tuned to 2.45GHz. Notice also the dB values, they are pretty low. We thought that it is because of the matching issue and will be handled later.
Tuning the length L
In order to tune the frequency to 2.45GHz from 2.355Ghz, we need to lower the length L from the expression (1). Thus we have used the ”Parametric Sweep” feature and found proper L to be 39.6mm. This actually makes sense because we have said in the lecture that electrical length is bigger than physical length and used 0.48 instead of 0.5 in the expression (1). By applying this, the length can be calculated as 39.6288mm.
Figure 5: S11 behaviors of parametric sweep
Matching procedure
Now, we set the length L to 39.6mm and will act on the feeding pin position ”yc” with a new Parameter Sweep analysis to find the optimum matching conditions. We will try to improve the S11 behavior and also check Z port behavior (Re(Z11) and Im(Z11)). We know that we have to move the feeding from the edge toward the center to tune it. This procedure can be achieved by increasing ”yc”. It can easily be confirmed by looking at Figure [6]. We are getting closer to the theoretical asymptotic region which is the almost vertical slope in real cases. In there, we have tried to make XA close to zero.
Figure 6: Z11 behaviors of parametric sweep
By getting closer to the center and improving the matching, we obtained better resonance at 2.45Ghz.
In the meantime, we have looked reference impedance graph if the value we found for Wf is correct. As you can see from Figure 7, it is quite close to 50ohm, the characteristic impedance we would like to achieve.
Figure 7: Reference impedance
After our effort of tuning ”yc” and ”L” parameters, we have found yc=13.9mm and L=39.48mm. In the below figures, you can see S11 and Z11 behaviors. We are able to achieve -39dB at 2.45Ghz.
Figure 8: S11 behavior of our final design
Figure 9: Z11 behavior of our final design
With the help of S11 graph, we have measured the bandwidth by referring to the maximum acceptable reflection coefficient magnitude of -10dB and -6dB. From this, we have calculated the relative bandwidths below.
Far-field and surface current results
You may look at the radiation properties of our patch in the below figures
Figure 10: 3D far field result
Figure 11: Far field result in polar coordinates
In below, you can see the surface current results at important time instances.
2. Second Step: Circularly Polarized Patch
Cutting edge solution
In this section, we have defined a new parameter ”cut” and added a chamfered edge using it. Before running an analysis, we have removed PMC(Ht=0) plane since there is no more symmetry. Then, we have done a parameter sweep with a guessed value of ”cut” to find a layout providing an Axial Ratio (AR) less than 6dB in the broadside direction.
Radiation characteristics of the satisfying result
The satisfying result has been obtained when cut=4.5mm. In order to tune and match our new patch, we have changed L=39.88mm and fc=11.5mm. The Axial Ratio angular range (the theta range where AR is less than 6dB) has measured as around 130◦. You can also see all the radiation characteristics (patterns, efficiency, and surface currents) in the below figures
Please notice how the surface current changes when the phase changes, the pointing of the current vectors make a circular shape.
Improving the efficiency without redesigning the antenna
We have decided to increase the parameter ”h” (antenna layer substrate thickness) to improve the efficiency without having to redesign the antenna because we know that the fringing effect, which radiating properties depend on, increases with increasing substrate thickness. We have verified our idea through the simulations you may see in Figures 23 and 24. When we compare these results with also Figure 18(h=1.5mm), we observed that the gain is increased.
Page updated
Google Sites
Report abuse