Chapter 4 Isolation Enhancement Methods of Ground Edge Current Choke
4.2 GECC Design and Measurement
In order to block the current traveling along ground edge, the RF choke is
designed to perform an effective electrical open circuit on ground edge. Figure 4.1
illustrates the geometry of the proposed ground edge current choke (GECC), which is
implemented using a printed inductor and a capacitor. The inductor is realized by a corrugated L-shape slit at the ground plane edge on the bottom layer of the substrate, and the capacitor is by a metal strip on the top layer and the ground plane underneath.
As shown, a via is used to connect the strip and the ground plane. When a current flows upwards along the ground edge toward the structure, part of it goes along the meandered edge of the corrugated slit, thus experiencing an effective inductance; the rest climbs up to the metal strip through the via and is then capacitively coupled to the underneath ground plane. After that, both current streams merge and flow upwards together. Therefore, the proposed configuration forms an equivalent parallel LC resonator and thus exhibits an open circuit at the resonant frequency. The value of the capacitor depends on the overlapping area of the metal strip and the ground plane.
And the length of the current path along the slit edge determines the value of inductance.
Ls
Ws
Port 1
Port 2
d w
GECC
Top substrate
Bottom substrate
Ground plane
(a)
Port 2
+
-L
C Port 1
TL TL
+
-(b)
Top view Bottom view
GECC
(c)
Figure 4.2 (a)The transmission line structure for measuring the proposed RF choke. Ls = 50 mm, Ws = 20 mm, w = 5 mm, and d = 1.6 mm. (b)Equivalent circuit of measurement structure. (c)The photograph
To evaluate the proposed GECC and to study the design parameters, various sets of simulations and measurements were performed. To this end, a microstrip line structure containing the proposed GECC as shown in Figure 4.2(a) is proposed for test.
There are two FR4 substrates used for the signal line and ground plane, respectively, to form the transmission line. The signal line is printed along the edge on the upper layer of the top substrate (with thickness of 0.4 mm). And the ground plane of the microstrip line is printed on the bottom layer of the bottom substrate (0.8 mm thickness). The GECC for testing is fabricated on the ground plane and is inserted in the middle of the ground edge just under the signal line. The equivalent circuit for this test setup is illustrated in Figure 4.2(b). The current will be blocked and not go through the transmission line at the resonant frequency of the choke. By measuring the transmission coefficient (S21) of the transmission line, the property of the GECC can be observed. In order to keep impedance matching, the characteristic impedance of the choke-embedded transmission line is designed as 50 Ω. From simulation, the required width of the signal line is 5 mm and the gap between two substrates is 1.6 mm. Figure 4.2(c) shows the photograph of the proposed structure for measurement.
Two SMA connectors are used to connect the transmission line.
The operating frequency of the GECC is designed around 5 GHz. Figure 4.3 shows the frequency responses of the simulated and measured transmission coefficients (S21), for various values of the GECC dimensions a, b, c. Other structure parameters are fixed as t = 1 mm, l = 0.5 mm, and s = 0.6 mm. Three corrugation teeth as shown in Figure 4.1 are used in the L-shape slit of the GECC to provide sufficient inductance. The simulations were performed by commercial tool, Ansoft HFSS. Figure 4.3(a) shows the effect of the choke length a (= 3.5, 4.0, 4.5 mm), with choke width b and strip length c fixed at 2 mm and 3.9 mm, respectively.
The measurement results agree well with the simulation. It is observed that for each length a, the frequency response exhibits a notch, with a minimum transmission coefficient of about -14 dB. The fractional 10-dB insertion-loss (1/S21) bandwidth is about 6%. It demonstrates that the signal can be blocked by the use of the proposed GECC. The measurement notch frequencies are 5.25 GHz, 4.95GHz, and 4.75 GHz when the length is chosen as 3.5 mm, 4.0 mm, and 4.5 mm, respectively. The longer
the length a, the lower the notch frequency is. This is obvious since a larger a corresponds to a longer slit and thus a longer inductive path. The increase of the GECC inductance reduces the resonant frequency of the parallel LC resonator and thus the notch frequency. Similar phenomenon can be observed when the choke width b is increased. Figure 4.3(b) shows the results for various choke width b (=2.0, 2.5, 3.0 mm), with choke lenth a and strip length c fixed at 3.5 mm and 3.9 mm, respectively. The resonant frequency reduces from 5.25 GHz to 4.375 GHz as b increased from 2 mm to 3 mm. A total frequency shift of 875 MHz is achieved for 1 mm increase of b, which is larger than that (500 MHz) for the increase of a in the same amount.
Figure 4.3(c) shows the frequency responses of the transmission coefficient for various strip length c (= 3.4, 3.9, 4.4 mm), with a = 3.5 mm and b = 2 mm. The resonant frequency varies from 5.7 GHz to 4.8 GHz with c increased from 3.4 mm to 4.4 mm. The strip length determines the overlapping area of the metal strip and the ground, and thus affects the value of the equivalent capacitance. With the same slit size, the longer strip can has lower resonant frequency because of larger capacitance.
In practice, the capacitor of the proposed GECC can be implemented in an alternative way, i.e., by using a lumped capacitor instead of a printed one. The lumped capacitor has the advantage of larger capacitance in a small size, which can be considered when further miniaturization is required.
It is seen from the above tests that, the proposed GECC structure can effectively block the ground current flowing along the edge. Its behavior is just like that of an equivalent parallel LC circuit. Although not shown here, the effects of other structure parameters have also been checked. It is found that by tuning the parameters of the corrugated slit, the equivalent inductance can be varied. And by increasing the metal strip dimensions, the equivalent capacitance can be raised. The notch frequency of the GECC can thus be controlled and designed. Note that the proposed GECC is compact as compared to others in the open literature, and has a size as small as about 0.06 wavelength in free space.
4 4.5 5 5.5 6 Frequency (GHz)
-20 -15 -10 -5 0
Simulation Measurement Simulation Measurement
T ra n sm is si o n c o ef fi ci en t S
21(d B )
increasing
a
3.5 mm 4.5 mm 4 mm
(a)
4 4.5 5 5.5 6
Frequency (GHz) -20
-15 -10 -5 0
4 4.5 5 5.5 6
Frequency (GHz) -20
-15 -10 -5 0
T ra n sm is si o n c o ef fi ci en t S
21(d B )
increasing
b
2.5 mm 2 mm
3 mm Simulation
Measurement Simulation Measurement
(b)
4 4.5 5 5.5 6 Frequency (GHz)
-20 -15 -10 -5 0
4 4.5 5 5.5 6
Frequency (GHz) -20
-15 -10 -5 0
T ra n sm is si o n c o ef fi ci en t S
21(d B)
increasing
c
3.4 mm 3.9 mm
4.4 mm
Simulation Measurement Simulation Measurement
(c)
Figure 4.3 Simulated and measured transmission coefficients for the GECC of various sizes.
(a)Frequency responses for the chokes of different length a with b = 2 mm and c = 3.9 mm.
(b)Frequency responses for the chokes of different width b with a = 3.5 mm and c = 3.9 mm.
(c)Frequency responses for the chokes of different strip length c with a = 3.5 mm and b = 2 mm. Other structure parameters are fixed as t = 1 mm, l = 0.5 mm, and s = 0.6 mm.