P09343: Microwave Device II
/public/Simulation/

Simulation Improvement Results Phase Shifter

Re-design for Improvement of the Schiffman Phase Shifter

After the initial design, it was clear that many improvements could be made to the performance of the phase shifter, to level out the phase difference over frequency and to improve the return loss. The re-design was performed from the beginning, using the method outlined in the "Design Methodology" section. The coupled lines were designed to ensure that the square root of their product equaled the system impedance of 50ohm. As a result, when the port lines and the connection between the coupled lines was added, there was improvement in the phase difference, as well as reflections. Initially, the phase difference over frequency was approximately +/- 3deg. Flatness was achieved by varying the size and shape of the connection between the coupled lines, as well as the chamfering of the port lines. Since a complete redesign of the phase shifter was performed, from the beginning, this had to be taken into account in the timeline of the project. Once the even and odd mode analyses were complete, the entire phase shifter could be constructed and adjusted. The initial putting together did not produce desirable results, so adjustments to the connecting strip and the port lines had to be performed. Unfortunately, this is an exhaustingly iterative and time consuming process. After numerous attempts in a short time, the results obtained were not idea, but proved to be an improvement from the original design.

Figure 1: HFSS re-designed model phase difference, and transmission and reflection characteristics.

Figure 1: HFSS re-designed model phase difference, and transmission and reflection characteristics.

As can be seen in Figure 1 above, the re-design did show some improvement, most of it in transmission and reflection characteristics. The first design had reflection of less than -15.69dB over the operating frequency, while the re-design has less than -29.04dB over the same range. The phase, however, did not see as a significant improvement as was hoped. From Figure 9, it is evident that the phase is still not flat across the band. At 11GHz, the phase is centered at 44.92deg, with +2.29deg at 10GHz and -1.12deg at 12GHz. It seems that this is a reversal of what was seen in the original design, where at lower frequencies the phase difference was smaller than at higher frequencies. Table 1 below shows a comparison between the two designs.

Table 1: Table comparing the results from the original design, and the re-designed Schiffman phase shifter.

Table 1: Table comparing the results from the original design, and the re-designed Schiffman phase shifter.

As mentioned, the re-design showed the most improvement for transmission and reflection. For the phase, it was difficult to obtain a flatter response. The efforts made to improve the phase difference were successful compared to the phase results when it was first put together for the re-design. However, comparing to the original design, there is no significant improvement. Within the time allotted for the re-design, improvement in phase could not be accomplished. With additional time, it is almost certain that more improvements could be made, though simulation and though building and experimentation.