Chapter 4 Low-Profile Ultra-Wideband Antenna with Strong Vertical Polarization
4.2 Metal Body Effect Analysis
The antenna performance is affected by the operating environment, especially when metal bodies are nearby the antenna. In actual application environment, the antenna might be placed parallel above a metal body such as metal-like table, PC case or placed near the LCD screen edges, inserted inside a laptop, and so on. In these cases, when the antenna is operating, fringing fields are expected to be present externally; thus, coupling between the antenna and the nearby associated elements (e.g., a metal-like table under the antenna) will occur. Moreover, the EM wave radiated from the antenna will induce currents on the metal bodies, which in turn will radiate back to the antenna and thus deteriorate the antenna function. With different height and position of the antenna above the metal bodies, the image current can procure constructed or destructed radiation, according to the phase relationship between antenna current and image current induced on the metal object. Thus the performance of the antenna will be quite different depending on the placement relative to the metal body. This possible coupling will then
cause degradation effects on the antenna performances. Hence, in this section, the study on the antenna behavior when near a metal body is analyzed. A conventional printed UWB antenna is also simulated for comparison.
Figure 4.11 shows the side view of the low-profile UWB antenna placed parallel under a metal plane. In the simulation model, the metal plane whose size = 25 cm × 25 cm, the distance between the antenna and the metal plane is denoted as D. Figure 4.12 shows the simulated return losses for the proposed antenna above a metal plane of various distances between the antenna and the metal plane. In the case D = 0.5 cm, represents the distance between the antenna and metal plane almost nearby together.
From Figure 4.12, it can be observed that, there still have three resonance frequencies, at the expense of little deterioration in impedance matching. When D is larger than 3 cm, the antenna performance is similar to the case without a metal plane under the antenna.
z
x y
D short
strip feedstrip
A FR4
ground
Metal plane z
x y
z
x y
D short
strip feedstrip
A FR4
ground
Metal plane
Figure 4.11 Side view of the low-profile UWB antenna placed parallel under a metal plane.
2.5 3 3.5 4 4.5 5 5.5 Frequency (GHz)
40 30 20 10 0
Return Loss (dB)
w/o metal plane D = 0.5 cm D = 3 cm D = 1 cm
2.5 3 3.5 4 4.5 5 5.5
Frequency (GHz) 40
30 20 10 0
Return Loss (dB)
w/o metal plane D = 0.5 cm D = 3 cm D = 1 cm
Figure 4.12 Simulated return losses for the proposed antenna above a metal plane of various distance D.
Other geometric parameters are the same as given in Figure 4.2.
For comparison, a conventional printed UWB antenna is also investigated. In here, a half-circle monopole is used whose radius is 12 mm. Similarly, simulate the same arrangement as shown in Figure 4.11 and compare the antenna performance of the effect on the different antenna polarization above a metal plane. As the results in Figure 4.13, it can be seen that, in the case without metal plane, there are two resonance frequencies.
But when D is less than 1 cm, the antenna resonance phenomenon is disappeared in low band region about 3.5 GHz. However, when D is larger than 3 cm, the antenna performance is similar to the case without a metal plane under the printed antenna.
Figure 4.14 shows the comparison of the radiation patterns with and without the metal plane which is 5 mm under the proposed antenna. In Figure 4.14 (a), show the simulated xy-plane radiation patterns without metal plane at 4.41 GHz. As refer to the current distributions show in Figure 4.5, because the current distributions has the same directional on both feed and shorting strip, thus provide the high vertical polarization field in xy-plane as shown in Figure 4.14. It can obviously see that, the Eθ are omnidirectional in xy-plane. When a metal plane is placed parallel under the proposed
antenna, the simulated results are shown in the Figure 4.14 (b). The influence on the vertical polarization field is negligibly changed while adding the metal plane but the horizontal polarization field is almost deterioration. The reason is that, when a metal plane is placed parallel under the proposed antenna, it induces currents on the metal plane. By EM theory, the induced currents are equivalent to a parallel but out-of-phase image current on the other side of the metal plane and then degenerate the antenna performances. In here, the horizontal polarization field is reduced by the image current but the vertical polarization field is uninfluenced as compare in the Figure 4.14.
2.5 40 30 20 10 0
Return Loss (dB)
w/o metal plane D = 0.5 cm D = 3 cm D = 1 cm
3 3.5 4 4.5 5 5.5
Frequency (GHz) 2.5
40 30 20 10 0
Return Loss (dB)
w/o metal plane D = 0.5 cm D = 3 cm D = 1 cm
3 3.5 4 4.5 5 5.5
Frequency (GHz)
Figure 4.13 Simulated return losses for the reference antenna above a metal plane of various distance D.
Moreover, the low-profile UWB antenna is replaced a conventional printed UWB antenna and the same arrangement is placed as show in Figure 4.11. In Figure 4.15 (a), show the simulated xy-plane radiation patterns without metal plane of the printed antenna at 4.08 GHz. Because it’s inherently printed structure, the dominated current is on the horizontal direction and thus stronger horizontal field but weaker vertical polarization field in xy-plane is obtained. Similarly, while the metal plane is adding
under the printed antenna, by EM theory, the induced currents are equivalent to a parallel image current on the other side of the plane. It is noticeable that, the image current which induced on the plane has out-of-phase with the current on the printed antenna. As mentioned above, the printed antenna has high horizontal polarization field, hence, when a metal plane is placed parallel under the antenna closely. The current on the antenna will be canceled by the image current and produce destructed radiation. This phenomenon is shown in Figure 4.15 (b) and demonstrated from the simulated results.
By comparing Figure 4.15 (a) and (b), it is obvious that, when a metal plane is placed parallel under the antenna, the horizontal polarization field is reduced strongly. However, because of the weaker vertical polarization field, whatever with or without the metal plane under the printed antenna, there is not significant influence in both situations.
-35 -30 -25 -20 -15 -10 -5 0 5
Figure 4.14 Simulated xy-plane radiation patterns without metal plane at (a) 4.41 GHz and with metal plane at (b) 4.45 GHz for the proposed antenna of D = 0.5 cm. (dashed line: Eθ dotted line: Eψ)
-35 -30 -25 -20 -15 -10 -5 0 5
Figure 4.15 Simulated xy-plane radiation patterns without meta plane at (a) 4.08 GHz and with metal plane at (b) 4.45 GHz for the printed UWB antenna of D = 0.5 cm. (dashed line: Eθ dotted line: Eψ )
4.3 Experiment Results
Figure 4.16 shows the measured, as well as the simulated, return losses of the proposed antenna. The dimensions of the antenna are wa × A = 25 mm × 17 mm, W × L a
= 34 mm × 75 mm, w1 ×A1= 9 mm × 3 mm, w2 ×A2= 9.5 mm × 4 mm, w3 ×A = 4.5 3 mm × 17 mm, ws = 5.85 mm, h = 5mm, d = 2.5 mm, s = 6.75 mm, and g =1 mm. The measured impedance bandwidth, determined by a 10-dB return loss, is from 3.02 to 4.97 GHz, which is suitable for UWB low-band communication application. In Figure 4.16, it can be obviously seen that, the three resonant frequencies are located at 3.21, 3.74, and 4.65 GHz, very close to the simulated ones. The measured return losses for the proposed antenna above a metal plane of various distances between the antenna and the metal plane also show in Figure 4.16. It can be observed that, in the case D = 1 cm (dotted line), there still has three resonance frequencies, at the expense of little deterioration in impedance matching. When D = 3 cm (dash-dot line), the antenna performance is similar to the case without a metal plane under the antenna (solid line).
Good agreement between the measured and simulated results is obtained.
2.5 3 3.5 4 4.5 5 5.5 Frequency (GHz)
40 30 20 10 0
Return Loss (dB)
w/o metal plane
w/o metal plane (simulated) with metal plane (D = 3 cm) with metal plane (D = 1 cm)
2.5 3 3.5 4 4.5 5 5.5
Frequency (GHz) 40
30 20 10 0
Return Loss (dB)
w/o metal plane
w/o metal plane (simulated) with metal plane (D = 3 cm) with metal plane (D = 1 cm)
Figure 4.16 Measured and simulated return losses for proposed antenna with and without a metal plane.
Figure 4.17 (a) to (c) show the measured radiation patterns in the xy-plane at the three resonant frequencies (3.21, 3.74, and 4.64 GHz), respectively. The measured peak gains (average gains) are correspondingly 2.03 dBi (-1.10 dBi), 1.82 dBi (-0.64 dBi), and 2.34 dBi (-0.24 dBi) at the three frequencies, as summarized in Table 4.1.
Figure 4.18 shows the measured peak gain from 3.1 to 5 GHz in the xy-plane. Across the impedance bandwidth, the curve is shown to be flat and the antenna gain varies from 1.56 to 3.95 dBi. Figure 4.19 shows the photograph of the accomplished antenna. The fabricated antenna with the dimension of 5 mm height, 25 mm width and 17mm length.
It is mounted on the FR-4 substrate (εr = 4.4) with a thickness of 0.8 mm. The FR-4 substrate has a dimension of 34 mm × 75 mm, which is considered to be the ground plane of a practical application.
-35 -30 -25 -20 -15 -10 -5 0 5
Figure 4.17 Measured xy-plane radiation patterns at (a) 3.21 GHz, (b) 3.74 GHz and (c) 4.64 GHz for the proposed antenna.
Table 4.1
The Measured Peak and Average Gains at Three Resonant Frequencies Frequency (GHz) Peak gain (dBi) Average gain (dBi)
3.21 2.03 -1.10 3.74 1.82 -0.64 4.64 2.34 -0.24
-3 -1 1 3 5 7
2.5 3 3.5 4 4.5 5 5.5
Frequency (GHz)
Gain (dB)
-3 -1 1 3 5 7
2.5 3 3.5 4 4.5 5 5.5
Frequency (GHz)
Gain (dB)
Figure 4.18 Measured peak gain of the proposed antenna at xy-plane.
(a) (b)
Figure 4.19 The photographs of the fabricated antenna. (a) Front view and (b) back view.
4.4 Summary
A low-profile UWB antenna with strong vertically field has been proposed and investigated. From the simulated results, it has been demonstrated that the embedded L-shaped slits in the proposed antenna structure provide additional resonance frequencies, which leads to the appearance of three continuous resonant modes and thus
yields to the ultra-wideband performance. Moreover, because of the current flow on the feed and shorting strip has the same direction and thus high vertical polarization field as compared to the conventional printed antenna at xy-plane can be obtained. In actual operating environment, even when a metal plane placed parallel nearby and under the antenna, the proposed antenna still maintained good radiation characteristics. Finally, the measured results agree well with the simulation ones, a 10-dB return loss fractional bandwidth of 48.81% from 3.02 to 4.97 GHz. Good gain flatness with a maximum variation of 1.39 dB is observed over the entire frequency band. The proposed antenna has a compact size of 25 mm × 17 mm and low-profile of 5 mm, which is suitable for DS-UWB low-band communication applications.
Chapter 5
Analysis and Application of an On-Package Planar Inverted-F Antenna
In this chapter, a miniaturized on-package planar inverted-F antenna (PIFA) is proposed for high integration module application. The on-package PIFA consists of a single folded metal plate and has several advantages, including small size, light weight, low-cost and ease of fabrication. The antenna radiation pattern is omnidirectional in the H-plane. The coupling effect between the on-package PIFA and the RF components in the shielding package was studied. The antenna performance rarely changes and the isolation between the antenna and the RF components can be maximized when the locations of the components are appropriately arranged in the package. Finally, a wireless local area network (WLAN) front-end module (FEM) including the switch, low-pass filter (LPF), band-pass filter (BPF) and power amplifier (PA), is embedded into the shielding package of the antenna. The Error Vector Magnitude (EVM) of the resultant antenna integrated FEM, together with a WLAN card containing the baseband/medium access control (MAC) circuitry, is tested. Good performance is obtained, showing the usability of the proposed antenna configuration.
5.1 Configuration and Design
In this study, the PIFA is integrated with a package using a single folded metal plate.
Since the antenna is to satisfy the specification of IEEE 802.11b/g WLAN, the required return loss bandwidth should cover the band 2.4-2.4835 GHz. The on-package PIFA is implemented on the FR4 substrate, whose dielectric constant is 4.4, loss tangent is 0.02, and thickness is 0.8 mm.
Figure 5.1 shows the 3-D structure and side views of the on-package PIFA. The length and width of the on-package PIFA determine the resonant frequency, which is approximated by the formula
( w
a a)
f c
+A
= 4
0 (5.1)
where c is the velocity of light; Aa and wa are the length and width of the radiating element, and
f is the resonant frequency. The radiating element is connected to the
0 ground by using a short-circuit strip located near the antenna’s feed strip. Equation (5.1) yields the patch size of the antenna A × aw = 14.5 mm × 15 mm and an overall
a physical length of approximately one quarter-wavelength at the desired frequency of 2.45 GHz. The size of the shielding package is Ap × wp × h and the ground size is W ×L. The feed point is denoted as point A as shown in Figure 5.1.
h
x y
z
L
W
Aa
wa
Ap wp
Shielding package FR4
A
h
x y
z
L
W
Aa
wa
Ap wp
Shielding package FR4
A
(a)
wa
Figure 5.1 Geometry of the on-package PIFA. (a) 3-D structure, (b) side view of xz-plane, (c) side view of yz-plane.
The antenna is integrated with the shielding package using a single folded metal plate. The shorting strip is located at the corner of the patch and the feed of the on-package PIFA is located at the edge of the shielding package. The width and height of the shorting strip are 1.5 and 2 mm, respectively, while those of the feed strip are 1.5 and 3.5 mm, respectively. The input impedance of the antenna can be easily matched to 50Ω by controlling the feed position relative to the shorting strip. The proposed antenna structure is simulated using HFSS, a commercial 3-D full-wave electromagnetic (EM) simulation software. Figure 5.2 shows the return losses with various gap widths g between the shorting strip and the feed strip. The dimensions of the radiating element, the shorting strip, and the feed strip are fixed. The ground size and shielding package sizes are also specified as W × L = 20 mm × 40 mm and wp × Ap = 20 mm × 15 mm, respectively. When the gap width g is smaller, input impedance matching can be
improved. However, an extremely small gap width cannot be realized practically. Thus, the restriction of the fabrication technology must be considered and then the gap between the shorting strip and the feed strip chosen as g = 1 mm.
2 2.2 2.4 2.6 2.8 3
Frequency (GHz) 40
20 0
Return loss (dB) 10 30
mm =1 g
mm =2 g
mm =3 g
mm =4 g
Figure 5.2 Return loss of on-package PIFA with various gap widths between the shorting strip and the feed strip; Aa × wa = 14.5 mm × 15 mm, Ap × wp = 15 mm × 20 mm, W × L = 20 mm × 40 mm, h = 1.5 mm, and s = 0 mm.
The effect of the position of the shorting strip on antenna performance is investigated.
All of the antenna dimensions are fixed, except for the shorting strip position s relative to the shielding package side wall. The center frequency increases as the shorting strip moves along the –x direction (i.e., becoming farther from the side wall), because the equivalent current length decrease as s increases. In Figure 5.3, the impedance bandwidth exceeds that required for the IEEE 802.11b/g WLAN band application when
s is 0 mm. Furthermore, the shorting strip is located at the corner of the shielding
package, indicating that the feed strip is adjacent to the shielding package side wall.
Accordingly, the vertical current on the feed strip would not readily excite the cavity mode in the shielding package. Another advantage is that the size is small for a given
operating frequency.
2 2.2 2.4 2.6 2.8 3
Frequency (GHz) 40
20 0
Return loss (dB)
10
30
mm =0 s
mm =1 s
mm 5 . =1 s
mm =2 s
Figure 5.3 Return loss of on-package PIFA with various shorting strip positions. Other geometric parameters are the same as given in Figure 5.2.
The RF components can be placed within the shielding package and the baseband circuits arranged outside. The ground size and shielding package change with the size of the circuitry. However, the ground size influences the antenna performance and the shielding package is one part of the on-package PIFA. Therefore, changing the ground size and the shielding package affects the characteristics of the on-package PIFA. Hence, it is interesting to see the effect of varying the shielding package and the ground length.
Figure 5.4 shows the simulated return losses for various shielding package sizes, including wp × Ap = 15 mm × 15 mm (case 1), 20 mm × 20 mm (case 2), and 20 mm × 25 mm (case 3). The result for the original package size (15 mm × 20 mm) is also shown for comparison. Basically, the antenna frequency is little changed due to the variation of the package size. The bandwidth of case 1 exceeds the original, but those in the other two cases are worse.
2 2.2 2.4 2.6 2.8 3 Frequency (GHz)
40 20 0
Return loss (dB)
10
30
mm2
15 15×
=
× p wp A
mm2
0 2 15×
=
× p wp A
mm2
0 2 0 2 ×
=
× p wp A
mm2
5 2 0 2 ×
=
× p wp A
Figure 5.4 Return loss of on-package PIFA with various shielding package sizes. Other geometric parameters are the same as given in Figure 5.2.
The other parameter of shielding package is the package height. Figure 5.5 shows the effect of the package height hon the return loss, with h = 1.5, 2, 2.5, and 3 mm. The input impedance bandwidth is rarely changes with the package height, and the input impedance matching is improved when the package height h exceeds the original value (h = 1.5 mm). It has been observed from the simulation result that, the input reactance of the on-package PIFA with h = 1.5 mm is capacitive. Increasing the package height would lengthen the feed strip, thus increasing the input inductance. Therefore, a larger package height h is associated with better input matching. Hence, the proposed on-package PIFA is appropriate for RF components of various height without redesign when the package height h is increased from 1.5 to 3 mm.
2 2.2 2.4 2.6 2.8 3 Frequency (GHz)
40 20 0
Return loss (dB)
10
30
mm .5
=1 h
mm .0
=2 h
mm .5
=2 h
mm .0
=3 h
Figure 5.5 Return loss of on-package PIFA with various shielding package heights. Other geometric parameters are the same as given in Figure 5.2.
The effect of ground length is also examined. The ground width W is fixed to 20 mm and the ground length L varied among 35 mm, 45 mm, and 50 mm. Figure 5.6 shows the input return loss in each case. It is seen that the return loss at ground length of 35 mm is much poor, due to the truncation of the induced ground current.
2 2.2 2.4 2.6 2.8 3
Frequency (GHz) 40
20 0
Return loss (dB) 10 30
mm
=35 L
mm 0
=4 L
mm 5
=4 L
mm
=50 L
Figure 5.6 Return loss of on-package PIFA with various ground lengths. Other geometric parameters are the same as given in Figure 5.2.
As mentioned earlier, the proposed antenna structure can be made using a single metal sheet. As illustrated in Figure 5.7, first, trim the shape of the on-package PIFA and bend downward three sides of the shielding package except the one nearby the feed strip.
Then bend upward the antenna patch, followed by bending the last side of the package.
To keep the spacing between the antenna patch and the shielding package so as to remain the antenna performance, a low-loss foam with relative dielectric constant near 1.0 can be inserted in between. Figure 5.8 compares the simulated and measured input return losses of the proposed antenna for package and ground sizes of Ap × wp× h = 15 mm × 20 mm × 1.5 mm and W × L = 20 mm × 40 mm, respectively. A total impedance bandwidth of 160 MHz from 2.37 to 2.53 GHz is measured. Good agreement between the measured and simulated results is obtained.
1
Step Step2
Step4 Step3
1
Step Step2
Step4 Step3
Figure 5.7 Manufacture procedure of on-package PIFA.
2 2.2 2.4 2.6 2.8 3 Frequency (GHz)
40 20 0
Return loss (dB) 10 30
measured simulated
Figure 5.8 Measured and simulated return loss of on-package PIFA. Aa × wa = 14.5 mm × 15 mm, Ap × wp = 15 mm × 20 mm, W × L = 20 mm × 40 mm, h = 1.5 mm, g = 1 mm, and s = 0 mm.
5.2 Coupling Between On-Package PIFA and RF Components
The vertical current on the feed strip of the on-package PIFA would radiate field into the shielding package through the aperture on the package’s side wall. This might induce the coupling between the antenna and the RF components embedded in the package. It is thus interesting to check the coupling level between the on-package PIFA and the components. Essentially, the RF circuitry comprises a T/R switch, filters, a PA, and a RF transceivers [86]-[89]. Here, for simplicity, an LTCC BPF is adopted to represent a typical RF component. Since the layout of the LTCC BPF formed with inductors and capacitors was known, it is thus possible to analyze the coupling effect using the EM software HFSS.
The BPF with size of 2.5 mm × 2.0 mm × 1 mm is arranged at six locations (denoted from A to F) inside the package as shown in Figure 5.9. Ports 1 and 2 are the input and output ports of the BPF, while port 3 represents the input port of the on-package PIFA. Locations E and F are at the same position, but with the BPF rotated 90 degree. Figure 5.10 (a) shows the simulation return loss (-S ) and insertion loss (-S )
of the BPF at various locations. The results for the BPF without the package are also shown for comparison. It is seen that the performance of the BPF is not changed with the location. The insertion loss in the passband keeps under 2 dB. Figure 5.10 (b) and (c) depict the simulation isolations (-S13 and –S23) between the antenna port and the two
of the BPF at various locations. The results for the BPF without the package are also shown for comparison. It is seen that the performance of the BPF is not changed with the location. The insertion loss in the passband keeps under 2 dB. Figure 5.10 (b) and (c) depict the simulation isolations (-S13 and –S23) between the antenna port and the two