Design of embedded-fed microstrip patch antenna

By using a simple and straightforward transmission line model, it is possible to accurately model and analyze microstrip inset-fed patch antennas.

Small, low-cost antennas are used in many modern communication systems. Microstrip patch antennas represent a family of small antennas that have the advantage of conformal properties and integration with the printed circuits of communication systems. By using a simple and straightforward transmission line model, it is possible to accurately model and analyze microstrip inset-fed patch antennas. In addition, by applying curve fitting formulas, the exact inset length required for a 50Ω input impedance can be determined.

The feeding mechanism plays an important role in the design of microstrip patch antennas. Microstrip antennas can be fed by coaxial probes or embedded microstrip lines. Coaxial probe feeding has advantages in active antenna applications, while microstrip line feeding is suitable for developing high-gain microstrip array antennas. For both cases, the position of the probe or the length of the embedding will determine the size of the input impedance.

The input impedance characteristics of coaxial probe-fed patch antennas have been analyzed using a variety of models, including transmission line models and cavity models, as well as full-wave analysis. It has been demonstrated both experimentally and theoretically that the input impedance characteristics of coaxial probe-fed patch antennas obey the trigonometric function: cos2[π(y0/L)], where L is equal to the patch length and y0 is equal to the feeding position along the patch length L from the edge.

On the other hand, it is also experimentally demonstrated that on low dielectric constant materials, the input impedance of the probe antenna embedded with feeding exhibits a fourth-order characteristic of the following function: cos4[π(y0/L)].

Fortunately, a simple analytical method has been developed that uses a transmission line model to obtain the input impedance of an embedment-fed microstrip patch antenna. Using this method, a curve-fitting formula can be used to determine the embedment length to achieve a 50Ω input impedance when using modern thin dielectric circuit board materials.

FIG1 is a diagram of an embedded-fed microstrip patch antenna. The parameters εr, h, L, W, w_{f and y0 represent the dielectric constant of the substrate, thickness, patch length, patch width, feed line width and feed line embedding distance, respectively. The input impedance of the embedded-fed microstrip patch antenna depends primarily on the embedding distance y0 and, to some extent, on the embedding width (the spacing between the feed line and the patch conductor). A change in the embedding length does not produce any change in the resonant frequency, but a change in the embedding width does. Therefore, in the following discussion, the spacing between the patch conductor and the feed line is kept constant and equal to the feed line width. The change in input impedance at the resonant frequency related to the embedding length will be considered as a function of various parameters in the following discussion.}

Assuming that the patch antenna can be divided into four regions, it can be modeled as a series of transmission lines with radiation slots of different lengths (Figure 2). The table lists the parameters (width and length) of the three transmission lines and the width and length of the three radiation slots.

According to the method proposed earlier, a patch antenna with parameters of εr = 2.42, h = 0.127cm, w = 4.04cm, L = 5.94cm, and y0 = 0.99cm was analyzed. Figure 3 shows the results obtained using the transmission line model method proposed here and compares it with data obtained using a commercial computer-aided engineering electromagnetic (EM) simulator. Even though the resonant frequency shifts a little, the transmission line model can track the return loss curve predicted by the EM simulator very closely. This small shift in resonant frequency is caused by not taking into account the discontinuity between the embedded feed line and the patch.

The patch parameters were studied using a transmission line model for various values ​​of εr (2 ≤εr ≤10). Figure 4 shows that the rectangular microstrip patch antenna fed by the microstrip line has a higher input impedance at the edge (y0=0), which varies from about 150 to 450Ω for different values ​​of εr. It can also be observed that the input impedance drops rapidly when the embedding position moves from the edge of the patch to the center, compared to the patch antenna fed by the coaxial probe. These parametric studies have been used to derive the curve fitting formula as shown in Equation (1) in order to determine the exact embedding length to achieve a 50Ω input impedance on commonly used thin dielectric substrates.

The accuracy of the formula has been verified using a patch antenna with εr=5.0, h=0.127cm, W=4.1325cm, L=2.8106cm, y0=0.9009cm. To confirm the validity of the formula, the patch was analyzed using an EM simulator. Figure 5 shows a comparison between the results produced by the transmission line model and the results predicted by the EM simulator. Although there is a 1% shift in the resonant frequency between the two sets of data, the return loss curves predicted by the two methods are still very close.