Chapter 2 360 o Reflection-Type Analog Phase Shifter
2.4 Fabrication and Measurements
2.4.2 Performance against frequency
The measured frequency response is measured in Figure 2.16 from 12.2 to 12.7 GHz.
The measurement is plotted at control voltage of 3V-9V, where a full 360o phase shifter is achieved. It shows that the maximum phase shift is larger than 360o, insertion loss is 4.8 1.6 dB, and return loss is better than 14 dB over 500MHz of the voltage stage of 3V-9V.
±
(a)
(b)
(c)
Figure 2.16: Frequency response of the 360o phase shifter. (a) Relative phase shift. (b) Insertion loss. (c) Return loss.
Chapter 3
Ku band Phased array
3.1 Introduction
Phased array plays an important role in microwave applications such as radar systems and smart antenna systems of wireless mobile communication. In a phased-array antenna system, phase shifters are used to control the radiation direction. These phase shifters have to meet high requirements for continuously adjustable phase shift.
Design of these phase shifters requires consideration of many factors such as large phase shift, low insertion loss, low phase error, bandwidth, low insertion-loss variation, simple control, low DC power consumption, and low cost for commercial wireless communication.
Reflection-type phase shifters (RTPS) have the advantages of simple control, low reflection, and low insertion loss. Using active components such as varactor diodes as reflection loads was first introduced in [1]. Thus, phase shift is restricted by tunable capacitance of varactor diodes. In order to maximize phase shift range of reflection loads, many approaches were reported. But when working on an effort to increase the phase shift range of phase shifters, insertion-loss variation increases. This
insertion-loss variation can be eliminated by adding a compensating resistance parallel to the reflection load [2]. Furthermore, in adaptive antenna receivers, the insertion loss and the insertion-loss variation can be compensated by the variable-gain low-noise amplifiers (VGLNA) located in front of the phase shifters.
In this chapter, a Ku band (12.2-12.7GHz) phased array using reflection type analog phase shifter is proposed. First, a phase shifter with large phase shift is discussed. Instead of using two varactors for each reflection load of the proposed 360o phase shifter in the previous chapter, the phase shifter uses only one varactor as the reflection load. We extends the maximum phase shift of the varactor, which was originally limited by the capacitance tuning range, from 52o to 286o by resonating the varactor with via inductance and using impedance transformers between the branch line coupler and the varactor. The circuit analysis, design procedure, and tradeoffs between phase shift range and insertion-loss variation are elaborately discussed.
Finally, a 4-element phased array with a group of dipole antennas and the proposed phase shifters is fabricated to verify the proposed phase shifter. Low noise amplifiers (LNA) are used to compensate the insertion loss of phase shifters.
3.2 Theory of Reflection-Type Phase Shifter
The basic building block of the reflection-type phase shifter composed of a 3-dB 90o
hybrid coupler and two identical reflection loads has been discussed in chapter two. In this chapter, however, the phase shifter uses only one varactor for each reflection load.
The analysis, design procedure, and fabrication of the phase shifter will be discussed.
3.2.1 Analysis
In the simplest design, a single varactor used as reflection loads is shown in Figure 3.1(a). The phase of reflection coefficient is determined by the maximum capacitance variation of the varactor diode. The phase shift can be increased by adding an inductance LS in series to the varactor, forming the series resonant circuit (Figure 3.1(b)). The maximum phase shift range is achieved at the resonance frequency
0
0
1 L CS
ω = (3.1)
where C0 represents the average capacitance value of the varactor.
To simplify this model, the parasitic resistance of the varactor and series inductance is combined as RS. Also, the inductance LS includes the parasitic inductance of the varactor. In the ideal case (RS = 0), the phase shift of reflection coefficient follows the unity circle of the Smith Chart. But for real varactor diodes with finite Q, the effective series resistance RS must be included in the circuit model. In this case, the reflection coefficient follows a constant resistance circle on the Smith Chart, and this series
resistance determines the insertion loss and insertion-loss variation with the varactor bias voltage.
(a) (b) (c)
Figure 3.1: Circuit model of reflection loads, (a) Varactor. (b) Series resonant varactor.
(c) Series resonant loads with compensating resistance.
A method for compensating the insertion-loss variation is to shunt a resistance RP to the series resonant load as shown in Figure 3.1(c). This compensating resistance is used to transform the constant resistance circle locus to a constant ρ circle on the Smith Chart. Then, the insertion loss maintains nearly constant and insensitive with the change of varactor bias voltage.
3.2.2 Design Formula
The reflection coefficient Γ of the reflection load in Figure 3.1(c) is given by
0 0
where XL is equal toωLS −1ωCvar. Thus, phase shift and return loss of the reflection
Since reflected signals are combined from two identical reflection loads at the output of the 3-dB 90o hybrid coupler, the scattering parameter S21 of the phase shifter is obtained as jα2 Γ, whereαis the loss of the 3-dB 90o hybrid coupler. Phase shift
From (3.5) and (3.6), the varactor reactance determines not only the relative phase shift but also the insertion loss. The maximum phase shift Δφ is For a given varactor capacitance range, the maximum phase shift can be increased by lowering the impedance level Z0 at the input of the reflection load, while the input and output impedance level of the 3-dB 90o hybrid coupler is unchanged.
To eliminate the insertion-loss variation, |Γ| must be constant whatever the varactor
capacitance is. Detailed equations have been derived in [9] to achieve maximum phase shift while remaining insertion loss constant. The optimal compensating
resistance is derived as
2 2
While the maximum relative phase shift with constant insertion loss is
2 2
3.3 Design Procedure of Ku Band Reflection-Type Phase Shifter
The design procedure of the proposed Ku band reflection-type phase shifter is as the following steps:
Step 1) Measure the characteristics of the varactor. Then, resonate the varactor by adding a series inductance LS.
Step 2) Change port impedance of the reflection load depending on the required maximum phase shift. Phase shift may be increased by lowering the impedance level Z0. Optimum design uses highest Z0 to produce necessary phase shift range
Step 3) Add a resistance parallel to the reflection load to compensate the insertion
3.3.1 Series Resonant Varactor
The varactor diode MA-COM MA46H120 is used as the termination load of the proposed 360o reflection-type phase shifter. In the previous chapter, we have shown to extend the phase shift from 52 o to 119o with the resonant load by adding via inductance to the varactor.
3.3.2 Changing Port Impedance of Reflection Loads
For a given varactor capacitance range, the maximum phase shift can be increased by lowering the impedance level Z0 at the input of the reflection load. Since the resonant inductance LS of the proposed phase shifter is made of via holes, the value of LS is somewhat different to the optimal resonant inductance. In order to achieve maximum phase shift, a short transmission line L1 with phase delay of 6o is added in front of the varactor load as shown in Figure 3.2(a). The transmission line changes the phase of reflection coefficient of the reflection load, making the maximum and minimum phase of reflection coefficient vertically symmetry on the Smith Chart, as shown in Figure 3.2(b).
(a) (b) Figure 3.2: (a) Varactor load with a transmission line with L1=6o. (b) Reflection coefficient on the Smith Chart.
To simplify the effect of the port impedance on the maximum relative phase shift, the cases of reflection loads without compensating resistance RP is considered. Figure 3.3 shows examples of reflection coefficient with different Z0 of 25Ω and 12.5Ω on the Smith Chart. It is shown that lowering the port impedance Z0 increases not only the maximum phase shift but also insertion-loss variation, which is the tradeoff of phase shifter design. Figure 3.4 shows design charts of maximum phase shift values and insertion-loss variations of phase shifter with respect to port impedance Z0.
(a) (b) Figure 3.3: Reflection coefficient of reflection loads at 12.45 GHz.
(a) Z0=25Ω, maximum phase shift = 195o. Insertion loss variation = 1.4dB.
(b) Z0=13Ω, maximum phase shift = 266o. Insertion loss variation = 2.8.dB.
(a)
(b)
Figure 3.4: (a) Maximum phase shift and (b) Insertion loss variation against port impedance Z0
In [9], an impedance-transforming branch line coupler is applied to the phase shifter. However, if the required impedance Z0 is very low, the impedance transforming branch line coupler is too hard to be realized. In the proposed phase shifter, Z0 was designed at 9Ω, which is impossible for microstrip line fabrication. So, two cascaded quarter wave impedance transformers are implemented between a conventional 3-dB 90o hybrid coupler and the reflection load to produce an adequately low impedance level.
Figure 3.5 shows total reflection load with impedance transformers. Comparison of
relative phase shift with and without impedance transformers is depicted in Figure 3.6.
With the help of two cascaded impedance transformers, maximum phase shift of 292o, which is 173o more than the other one, is given with the same bias voltage condition.
Figure 3.5: Reflection load with two cascaded impedance transformers. Z1=118Ω, Z0=50Ω, L1=6o
Figure 3.6: Comparison of phase shift of the reflection load with and without impedance transformers.
3.3.3 Compensation for Insertion-Loss Variation
Given in Figure 3.4, the insertion-loss variation is 4.3 dB for the phase shifter when maximum phase shift of 292o is chosen. With the help of a compensation resistance RP parallel to reflection loads shown in Figure 3.7, insertion-loss variation can be eliminated. The optimized compensating resistance RP is chosen to be 200Ω. The principle of compensation for insertion-loss variation is illustrated in Figure 3.8.
Equivalent conductance at high/low control voltage is dominated by the compensating resistance. With optimized RP, the reflection coefficient of the reflection load follows almost constant ρ circle on the Smith Chart. However, constant insertion loss is unavailable because RS varies with the bias voltage. But (3.8) is still useful for initial design.
Figure 3.7: Reflection load with compensating resistance RP. Z1=118Ω, Z0=50Ω, L1=6o, R =200Ω
Figure 3.8: Reflection coefficient of the reflection load on the Smith Chart.
Figure 3.9 shows simulated reflection coefficient of reflection loads with and without RP. With RP =200Ω, insertion-loss variation is 1 dB and maximum phase shift is 287o. The use of RP barely influences maximum phase shift range.
(a)
(b)
Figure 3.9: Comparison of simulated reflection loads with and without RP at 12.45 GHz. (a) Relative phase shift and (b) Return loss of the reflection load against bias voltage of varactor.
3.4 Fabrication and Measurements of Ku Band Reflection-Type Phase Shifter
The proposed reflection-type phase shifter was fabricated on RO4003 substrate with a dielectric constant of 3.58 and thickness of 20 mils to verify the phase shifter design procedure. The phase shifters were designed with and without the compensating resistance RP, which has been described in detail at the pervious section. Then, a 3-dB 90o hybrid coupler with the characteristic impedance of 50Ω at the center frequency
220Ω chip resistor with fan stub as ground was used as the compensating resistance RP. The photograph of the proposed phase shifters is shown in Figure 3.10.
(a) (b)
Figure 3.10: Photograph of the proposed phase shifter circuits (a) without RP (b) with RP=220Ω
3.4.1 Performance against Control Voltage
Figure 3.11 shows measured performance with respect to varactor control voltage of the proposed two phase shifters without RP and with RP =220Ω. The measured maximum phase shift of phase shifter is 286 o without RP and 283 o with RP =220Ω. In Figure 3.9(a), the simulated maximum relative phase shift, which is 292 o without RP
and 287 o with RP=200Ω, agrees well with the measured results.
The measured insertion-loss variation of phase shifter is 4 dB without RP and 3.5 dB with RP=220Ω. In Figure 3.9(b), the simulated insertion loss variation is 4.5 dB without RP and 1 dB with RP=200Ω. The measured insertion loss of phase shifter without RP agrees with the simulated results. But in the case of RP=220Ω, the compensating resistance slightly influences the insertion-loss variation. This is mainly because the unwanted package capacitance of the chip resistor is parallel to the reflection load at high frequency of 12.45 GHz. The parallel capacitance and other parasitic effect can be observed from the S21 of the phase shifter on the Smith Chart shown in Figure 3.12.
(a)
(b)
Figure 3.11: Measured results of the proposed phase shifters against bias voltage at 12.45 GHz. (a) Relative phase shift. (b) Insertion loss
Figure 3.12: Measured S21 of the proposed phase shifters at 12.45 GHz.
3.4.2 Performance against Frequency
Because the frequency response is mainly determined by the 3-dB 90o hybrid coupler and the cascaded quarter-wave impedance transformers, only the case of the one without RP is concerned. Figure 3.13 shows the measured performances such as phase shifts, insertion loss, and return loss of the phase shifters without RP over the bandwidth of 12.2 GHz to 12.7 GHz. The Figure shows that insertion-loss variation is within +2dB, return loss is better than -12dB, and phase shift is more than 280o over 500MHz at the phase stages of bias voltage from 0V to 10V.
(a)
(b)
Figure 3.13: Measured frequency performance of the proposed phase shift. (a) Insertion loss. (b) Return loss. (c) Relative phase shift.
3.5 Ku Band Phased Array
3.5.1 Theory
A phased array is a group of antennas in which the relative phases of the signals feeding the antennas are varied, so the radiation pattern of the array is reinforced in the desired direction and suppressed in undesired direction. The proposed phased shifters are used in the phased array as beamformers to provide high-speed beam-steering and high directivity. As shown in Figure 3.14, the phased array is a
combination of N antennas with equidistance, and the array factor (AF) is defined as
N ( sin )
where An represents the amplitude of the signal radiated by the n-th element; α and d represent the phase shift and distance between successive antennas, respectively.
Letψ β= dsinθ + , then α Therefore, if we want to control the angle at which the maximum power is emitted or received (main beam), we only need to adjust the phase shift α of successive antennas.
Figure 3.14: Phased array composed of N antennas with equidistance.
Now space the antennas a distance d of 1.03 λ at center frequency of 12.45 GHz.
Chosen N=4, the simulated radiation patterns with different phase shift α are shown in Figure 3.15. Each element of the phased array is assumed to be omnidirectional for simplicity. It shows that for d being approximately one wavelength, there are 4 or 5 main lobes. Considering the first quadrant (left-top), the angle of the main lobe θ is 0o, 8o, 16o, 30o, 46o, and 59o, when phase shift α is 0o, -50o, -100o, -180o, -260o, and -310o, respectively. Figure 3.16 shows the relationship between α and θ at the first quadrant.
It shows that the direction of main beam can be arbitrarily controlled if a full 360o phase shifter is used.
Figure 3.15: Simulated radiation pattern of 4-element phased array.
Figure 3.16: Relationship between the angle of main beam and phase shift between successive antennas.
3.5.2 Design
The structure of the proposed phased array with 4 elements is shown in Figure 3.17.
The series-feed phased array is used for simplicity of control, and dipole antennas are used for omnidirectional radiation pattern. Then, the received signals are coupled to the signal path by a directional coupler. With the same control voltage of phase shifter, each phase shift between successive antennas is identical to the others. However, the phase shifters accumulate insertion loss along the path. In order to receive the same power from each antenna at the output port, a low noise amplifier is added in front of each phase shifter to compensate the insertion loss.
Figure 3.17: The structure of the proposed phased array. (RTPS: reflection-type phase shifter. LNA: low noise amplifier)
3.5.3 Fabrication and Measurements
The photograph of the proposed Ku band phased array is shown in Figure 3.18, in which the components of the phased array are illustrated. The phased array is designed at the center frequency of 12.45 GHz, with its size of 10×4 cm2. Dipole
antennas are chosen for omnidirectional. The directional coupler is designed with a coupling factor of -14 dB and its through port is terminated with a 50Ω resistor. The received signal is coupled to its coupled port and direct to the output port of the phased array.
Figure 3.18: Photograph of the proposed phased array. (a) Dipole antenna. (b) Directional coupler. (c) Reflection-type phase shifter. (d) Low noise amplifier. (e) Chip resistor (100Ω). (f) Control voltage of phase shifters (0-10V). (g) LNA bias voltage (1.5V).
Figure 3.19 shows the radiation pattern of the proposed phased array against different control voltage of the phase shifter. It was measured at 12.2, 12.45, and 12.7 GHz, which is the band of broadcasting satellite service (BSS). Even if the insertion loss is compensated by LNA, the radiation gain suffers from the insertion-loss
variation with control voltage of the phase shifter. The angle of main beam against control voltage is shown in Figure 3.20.
(a)
(b)
(c)
Figure 3.19: Measured radiation pattern of the proposed Ku band phased array at (a) 12.2 GHz. (b) 12.45 GHz. (c) 12.7 GHz
Figure 3.20: The angle of main beam against control voltage.
Chapter 4
Reduced-Size Impedance-Transforming Broadband 180 o Hybrid Ring
4.1 Introduction
The 180o hybrid ring coupler, which is also known as a rat-race ring, is an essential component in microwave circuits, such as balanced mixers, multipliers, push-pull amplifiers and antenna feed network, etc. A conventional a rat-race ring consists of three λ/4 line sections and one 3λ/4 line section. The 3λ/ 4 line section works as a λ/4 line section with a phase inverter formed by a λ/2 line section. Thus, the disadvantage of the conventional rat-race ring is narrow bandwidth and large size.
Many researches [10]-[14] have attributed to make the bandwidth larger. Most of the approaches are realizing an ideal or broadband phase inverter. Replaced the 3λ/ 4 line section with a short-circuited λ/ 4 coupled line section of March [10], which not only reduce size but also widen the bandwidth of the rat-race ring. However, the tight coupled line is hard to be realized. An ideal phase inverter, which is shown in Figure 4.1(a), provides perfect amplitude and phase performance. In [11], [12], the combination of coplanar waveguide (CPW) and coplanar strips (CPS) or slot lines
was used in the 180 o hybrid ring. In these designs, frequency independent phase inverter of the hybrid ring has been implemented by CPS direct cross-type phase inverter. Another technique to increase the bandwidth is by converting baluns into 180o hybrid ring by adding an in-phase power splitter presented in [13]. In addition to realizing the broadband phase inverter, increasing the order of circuit response may widen the bandwidth as well. In [14], a broadband Chebyshev-response rat-race ring with a λ/4 unit element at each port and an ideal phase inverter is proposed.
Many approaches [15]-[19] have been presented to reduce the size of hybrid rings.
These methods include the use of folded lines [15], artificial lines [16], defected ground structure [17], lump elements [18], and others [19]. However, the above approaches are hard to achieve a wide bandwidth. In [20], both wideband and size reduction are achieved by an interdigital CPS inverter.
In this chapter, a reduced-size impedance-transforming broadband 180 o hybrid ring coupler is presented. With a unit element at each port and an ideal phase inverter, which was proposed in [14], the broadband rat-race ring with a Chebyshev response of order 3 or 4 has been developed. In [21], the author modified the coupler and developed the design and optimization method of a 180 o hybrid ring for Chebyshev equiripple functions including impedance transformation between the input and output ports, as shown in Figure 4.1(b). The values of the admittance Y1, Y2, Yt1, and
Yt2 in Figure 4.1(b) are determined by the optimization method to synthesize a broadband rat-race ring with an impedance transforming ratio of 1:3. We replace each line section by the stepped-impedance CPS/interdigital CPS structure proposed in [20]
not only to realize an ideal 180 o phase inverter but also miniaturize the circuit size. A miniaturized broadband 180o hybrid ring coupler with impedance transformation of from 40Ω to 120Ω using the stepped-impedance CPS/interdigital CPS structure is realized. The circuit analysis, design procedure, simulation and measurement results are discussed in this chapter.
(a)
(b)
Figure 4.1: Circuit schematics of 180o hybrid ring with an ideal phase inverter. (a)
Conventional 180o hybrid ring. (b) Proposed broadband 180o hybrid ring with different impedance at input and output ports.
4.2 Theory
4.2.1 Conventional 180
oHybrid Ring
The circuit schematic of an 180o hybrid ring with an ideal phase inverter is shown in Figure 4.1(a). Instead of a conventional rat-race ring composed of three λg/4 and one 3λg/4 line sections, the spacing between all adjacent ports is λg/4, and an ideal phase
The circuit schematic of an 180o hybrid ring with an ideal phase inverter is shown in Figure 4.1(a). Instead of a conventional rat-race ring composed of three λg/4 and one 3λg/4 line sections, the spacing between all adjacent ports is λg/4, and an ideal phase