Chapter 3: Design and Measurement of On-Package PIFA
3.2 Influence of Different Ground Size and Shielding Package
3.2.1 Ground Size
The dimensions of the radiating element, the short-citcuit strip and the feed line are fixed and the shielding package is also specified to 15 ×20 × 1.5 mm 3. We alter the ground size to 20 × 35 mm 2, 20 × 45 mm 2, and 20 × 50 mm 2.The input return loss of each case is shown in Fig 3-9. It is observed that the return loss for ground size of 20 × 35 mm 2 and 20 × 45 mm 2 are much poor. The return loss for ground size of 20 × 45 mm 2 has wider bandwidth than the original case. We also compare the radiation patterns of the PIFA with the ground size of 20 × 40 mm 2 and 20 × 45 mm 2. Their patterns are shown in Fig 3-10. The same radiation patterns occur at x-z plane.
The patterns at y-z plane are almost the same except at the theta 90 degree and -90degree. The reason is that the ground size of 20 × 45 mm 2 is longer than the order of 20 × 40 mm 2 in the y axis. Therefore, the nulls in the y axis would be more evident.
From above discussion, we realize that the structure of the original on-package PIFA do not have not to be redesigned when the ground size varies from 20 × 40 mm 2 to 20
× 45 mm 2.
Fig 3-9 Return loss of on-package PIFA with various ground size.
Fig 3-10a Radiation pattern of on-package PIFA with various ground size at x-z plane (2.45GHz)
y-z plane
Fig 3-10b Radiation pattern of on-package PIFA with various ground size at y-z plane (2.45GHz).
Fig 3-10c Radiation pattern of on-package PIFA with various ground size at x-y plane (2.45GHz).
3.2.2 Shielding Package
The original package size is 20 × 15×1.5 mm 3. In this section, we fix all parameters except for the package size. We change the package size to 15 × 15×1.5 mm 3 (case 1), 20 × 20 ×1.5 mm 3 (case 2), and 20 × 25×1.5 mm 3 (case 3). The simulated return loss is shown in Fig 3-11. It shows the bandwidth of case 1 is wider than the original one, but the other two cases are worst than that. We can see that the change of the package size has less influence than the change of ground size. The radiation patterns of case 1 and original one are both shown in Fig 3-12. The radiation pattern has no difference in any plane. The reason is the package size is less than the ground size which is fixed. We conclude that the configuration of the radiation element do not have not to be redesigned when the package changes from 15 × 15×1.5 mm 2 to 15 × 20×1.5 mm 3.
Fig 3-11 Return loss of on-package PIFA with various package size
2 2.2 2.4 2.6 2.8 3
Frequency (GHz)
S11
-30 -20 -10 0
dB
15x15
20x15
20x20
20x25
x-z plane
Fig 3-11a Radiation pattern of on-package PIFA with various package size at x-z plane (2.45GHz).
y-z plane
Fig 3-11b Radiation pattern of on-package PIFA with various package size at y-z plane (2.45GHz).
x-y plane
Fig 3-11c Radiation pattern of on-package PIFA with various package size at x-y plane (2.45GHz).
3.3 Measurement
In this section, we carry out the implementation of the on-package PIFA in air, which has been discussed in section 3.1.2, and exhibit the measured results. The on-package PIFA consists of a single folded copper plate, which is shown in Fig 3-12.
The upper and the lower palate of the original copper plate are folded as the package part and the patch part of the antenna, respectively. The final folded copper plate is shown in Fig 3-13. The sizes of the package and the radiating element are 15 × 15×
1.5 mm 3 and 14.5 × 17 mm 2, respectively. Fig 3-14 is the photograph of the realized on-package PIFA. The measured bandwidth is 150MHz from 2.37GHz to 2.52 GHz under –10dB return loss, as shown in Fig 3-14. The radiation pattern is also shown in Fig 3-16. The pattern at x-z plane is pretty omnidirectional and the maximum gain and average gain are 1.17 and -0.63dBi, respectively. The gains of all planes at 2.45GHz are listed in Table 3-2. We also investigate the frequency response of the radiation gain at x-z plane as shown in Fig 3-17. The gain has less than 2dBi variation centered at 0dBi. This on-package PIFA hold an excellently linear frequency response.
x-z plane y-z plane x-y plane
Maximum Gain 1.17dBi 3.09dBi 0.61dBi
Average Gain -0.63dBi -2.78dBi -3.17dBi
Table 3-3 Measured maximum and average gain of the on-package PIFA at 2.45 GHz.
Fig 3-12 The original copper plate.
Fig 3-13 The folded copper plate.
Fig 3-14 Photograph of implemented antenna.
Fig 3-15 Measured return loss of the on-package PIFA.
Fig 3-16a Measured radiation pattern of the on-package PIFA at x-z plane (2.45GHz).
2 2.2 2.4 2.6 2.8 3
y-z plane
Fig 3-16b Measured radiation pattern of the on-package PIFA at y-z plane (2.45GHz).
x-y plane
Fig 3-16c Measured radiation pattern of the on-package PIFA at x-y plane (2.45GHz).
( )
Frequency (GHz)
2.30 2.35 2.40 2.45 2.50 2.55 2.60
Gain (dB i)
-10 -8 -6 -4 -2 0 2
Fig 3-17 Gain vs. Frequency characteristic of the on-package PIFA at x-z plane (2.45GHz).
Chapter 4
Coupling between On-Package PIFA and RF Components
4.1 Characteristic of LTCC BPF
Basically, WLAN RF circuits consist of an antenna, a T/R switch, filters, mixers, voltage-control oscillators, a low-noise amplifier, and a power amplifier [13]-[15].
The switch is set up for TX or RX path with appropriate bias. On RX path, the band-pass filter for band selection is followed by a low-noise amplifier. The RF signal amplified by the PA radiates through an antenna when the switch is set up for the TX path. The low-pass filter can suppress the output harmonics of the PA. Among these components, the low-pass filter, the band-pass filter and the matching network of the amplifiers are all passive components. The coupling effect between the on-package PIFA and the RF passive components in the shielding package can be investigated by the 3-D full-wave EM simulator Ansoft HFSS 8.0. We select a BPF to represent the passive components which are formed with inductors and capacitors.
The band-pass filter LTB-2520-2G4H3-A2 is designed via low-temperature co-fired ceramic (LTCC) technology by MAG.. LAYERS Scientific-Technology Co., LTd. Fig 4-1 shows the top-view of the LTCC BPF with the size of 2.5×2.0×1 mm 3. More detailed dimensions are also shown in Fig 4-1.The pass band is 100MHz from 2400 to 2500 MHz. The maximum insertion loss is 2.0dB. Table 4-2 shows the detailed electrical specifications and Fig 4-5 indicates the electrical characteristic which is measured by Aglient E5071A Network Analyzer. The measured insertion loss is -1.5dB. The above characteristic is obtained from the datasheet of the BPF.
Fig 4-1 Top view and dimensions of the LTCC BPF (2520) Electrical Specifications
Pass Band 2400-2500 MHz
Insertion loss in BW 2.0 dB max
Return loss 10 dB min
Impedance 50 ohms
30dB min at 880-915MHz 30dB min at 1710-1785 MHz 28dB min at 1850-1910 MHz Attenuation
25dB min at 4800-5000 MHz
TEST
INSYRUMENT:
AGLIENT E5071A NETWORK
ANALYZER
Table 4-1 Electrical specifications of the LTCC BPF (2520)
Fig 4-2 Electrical Characteristic of the LTCC BPF (2520)
L 2.5±0.20 mm W 2.0±0.20 mm T 1.0±0.15 mm a 0.50±0.20 mm b 0.30±0.15 mm c 0.30±0.15 mm
4.2 Coupling between On-Package PIFA and LTCC BPF
In this section, we investigate the coupling effect between the on-package PIFA and LTCC BPF [16]. We place a band-pass filter LTB-2520-2G4H3-A2 in the shielding package, which is one part of the on-package PIFA. We have the BPF to be arranged at six different positions in the package to observe the phenomenon. Fig 4-3 shows the diagram of the BPF arranged at different positions. The six cases are represented as case A, case B, case C, case D, case E, and case F. In case A, the BPF is arranged 1 mm apart from the input of the antenna. The BPF is also moved backward 5 mm from position A in case B. In case C, the BPF has 13 mm right-shift from the position A. In case D, the BPF has 13 mm right-shift from the position B.
The BPF is arranged at the center of the package in case E. Finally, the BPF is rotated by -90 degree in case F. Port 1 and port 2 are the input and output of BPF, respectively, and the input of the antenna is port 3. We simulate these cases by 3-D full-wave EM simulator Ansoft HFSS 8.0. The performances in each case are also compared with that the original BPF and the on-package PIFA.
Fig 4-3 The diagram of the BPF arranged at different positions.
Fig 4-4 shows the input return loss in each case. The input return loss has merely change wherever the BPF is arranged. The insertion loss in each case is shown in Fig 4-5. Fig 4-5 shows that the insertion loss is not influenced except for the high frequency transmission zero. The insertion loss in the pass band is less than -2dB.
The reason for the variation of the transmission zero is that the designed transmission zero is seriously influenced by the parasitic inductance [17]. Fig 4-6 shows the return loss of the on-package PIFA in each case and the return loss also has merely change.
In Fig 4-7 and Fig 4-8, the isolation of S13 and S23 are at more than 30dB in WLAN band. The isolation between the antenna and BPF has a better performance when the BPF is apart from the feed point of the antenna.
We also connect the input of the on-package PIFA to the output of BPF when the BPF is arranged at the position B. Fig 4-9 shows the input return loss of the BPF. The return loss has an additional null due to the on-package PIFA. The new radiation pattern is also compared with the original on-package PIFA as shown in Fig 4-10. The shape of radiation pattern has merely change, and the gain is around 1.5dBi lower than that of original on-package PIFA due to the insertion loss of the BPF. The gains of all planes are listed in Table 4-2 and those of the original on-package PIFA are also listed in Table 4-3.
x-z plane y-z plane x-y plane
Maximum Gain 3.39dBi 3.25dBi 1.40dBi
Average Gain 1.14dBi -2.90dBi -2.62dBi
Table 4-2 The maximum and average gain of the on-package PIFA with BPF at the position B at 2.45GHz.
x-z plane y-z plane x-y plane
Maximum Gain 4.92dBi 4.91dBi 2.85dBi
Average Gain 2.53dBi -1.4dBi -1.16dBi
Table 4-3 The maximum and average gain of the original on-package PIFA at 2.45GHz.
Fig 4-4 Return loss of the BPF in all cases.
0.5 1.5 2.5 3.5 4.5 5.5
Frequency (GHz)
S11
-30 -20 -10 0
A
B
C
D
E
F
BPF
dB
Fig 4-5 Insertion loss of the BPF in all cases.
Fig 4-6 Return loss of the on-package PIFA in all cases.
0.5 1.5 2.5 3.5 4.5 5.5
Frequency (GHz)
S21
-80 -60 -40 -20 0
dB
A B C D E F BPF
0.5 1.5 2.5 3.5 4.5 5.5
Frequency (GHz)
S33
-20 -10 0
A
B
C
D
E
F
PIFA
dB
Fig 4-7 Isolation of S13 in all cases.
Fig 4-8 Isolation of S23 in all cases.
0.5 1.5 2.5 3.5 4.5 5.5
Fig 4-9 Return loss of the BPF with on-package PIFA.
Fig 4-10a Radiation pattern of the on-package PIFA with and without BPF at x-z plane (2.45GHz).
0.5 1.5 2.5 3.5 4.5 5.5
y-z plane
Fig 4-10b Radiation pattern of the on-package PIFA with and without BPF at y-z plane (2.45GHz).
x-y plane
Fig 4-10c Radiation pattern of the on-package PIFA with and without BPF at x-y plane (2.45GHz).
4.3 Measurement
In this section, we verify the characteristic of the LTCC BPF and the coupling effect between on-package PIFA and LTTC BPF in case A , B and F, which has been discussed in section 4.2. Fig 4-11 shows the photograph of the LTCC BPF. The photographs of on-package PIFA with BPF in case A and F are shown in Fig 4-12a and 4-12b, respectively. In Fig 4-13, the photograph of on-package PIFA connected to BPF at the position B is shown.
The measured return loss and insertion loss are shown in Fig 4-14 and the insertion loss is 1.54dB in WLAN band. We calibrate the connector loss and microstrip-line loss by TRL calibration. It is observed that the measured characteristic is very identical to the datasheet except for the high frequency transmission zero caused by the parasitic inductance. The PCB board we adopted is 0.8mm FR4, which is thicker than the test kit. Therefore, the parasitic inductance is larger than the original one.
In Fig 4-15, we observe that the return loss has merely change wherever the BPF is arranged at the position A or F. Fig 4-16 shows the measured insertion loss in case A and F. The difference of the insertion loss between the case F (or A) and the original one is mainly caused by the coaxial-line loss and the discontinuity. The loss of coaxial lines and the discontinuity is around 1dB, as shown in Fig 4-17. Therefore, the insertion loss in both case A and F is about 1.8dB. Fig 4-18 indicates that the return loss of on-package PIFA in case A and F is merely changed. The characteristic of BPF and on-package PIFA is not changed in both cases. The isolation between antenna and BPF has a better performance when the BPF is arranged at the position F than the position A, as shown in Fig 4-19 and 4-20. The reason is that the position F is more far away from the feed point of the antenna. Nevertheless, the isolation of S23 is
worse than that of S13 because the output of the BPF is closer to the feed point of the antenna.
The input of the on-package PIFA is connected to the output of the BPF when the BPF is arranged at the position B. The return loss has an extra null due to the on-package PIFA, as shown in Fig 4-21. The measured radiation pattern is similar to the original on-package PIFA, as shown in Fig 4-22 and the gain reduction is caused by the insertion loss of the BPF and the error in measurement process. The gains of E total at all planes are listed in Table 4-4 and those of the original on-package PIFA are also listed in Table 4-5. Fig 4-23 shows that the frequency response of the BPF occurs in the radiation gain of E phi at x-z plane.
x-z plane y-z plane x-y plane
Maximum Gain -1.34dBi 1.68dBi -1.3dBi
Average Gain -3.38dBi -5.30dBi -5.12dBi
Table 4-4 Measured maximum and average gain of the on-package PIFA with BPF at the position B at 2.45 GHz.
x-z plane y-z plane x-y plane
Maximum Gain 1.17dBi 3.09dBi 0.61dBi
Average Gain -0.63dBi -2.78dBi -3.17dBi
Table 4-5 Measured maximum and average gain of the original on-package PIFA at 2.45 GHz.
Fig 4-11 Photograph of LTCC BPF.
Fig 4-12a Photograph of on-package with BPF in case A.
Fig 4-12b Photograph of on-package PIFA in case F.
Fig 4-13 Photograph of on-package PIFA connected to BPF at the position B.
Fig 4-14 Measured return loss and insertion loss of the LTCC BPF.
Fig 4-15 Measured return loss of the BPF in case A and F.
1 2 3 4 5
Frequency (GHz)
S11
-30 -20 -10 0
A F BPF
dB
1 2 3 4 5
Frequency (GHz) S11 and S21
-60 -50 -40 -30 -20 -10 0
2.435 GHz -1.542 dB
dB
Fig 4-16 Measured insertion loss of the BPF in case A and F.
Fig 4-17 Loss due to the coaxial lines and discontinuity
1 2 3 4 5
Frequency (GHz)
S21
-60 -50 -40 -30 -20 -10 0
2.453 GHz -2.891 dB 2.425 GHz
-1.562 dB
dB
A F BPF
1 2 3 4 5
Frequency (GHz)
S21
-5 -4 -3 -2 -1 0
2.44 GHz -1.004 dB
dB
Fig 4-18 Measured return loss of on-package PIFA in case A and F.
Fig 4-19 Measured isolation S13 in case A and F.
1 2 3 4 5
Frequency (GHz)
S13
-60 -50 -40 -30 -20 -10 0
dB
F A
1 2 3 4 5
Frequency (GHz)
S33
-30 -20 -10 0
dB
A
F
PIFA
Fig 4-20 Measured isolation S23 in case A and F.
Fig 4-21 Measured return loss of BPF with on-package PIFA in case B.
1 2 3 4 5
Frequency (GHz)
S11
-30 -20 -10 0
dB
1 2 3 4 5
Frequency (GHz)
S23
-60 -50 -40 -30 -20 -10 0
dB
F
A
x-z plane
Fig 4-22a Measured radiation pattern of on-package PIFA with and without BPF at x-z plane (2.45GHz).
y-z plane
Fig 4-22b Measured radiation pattern of on-package PIFA with and without BPF.
at y-z plane (2.45GHz).
x-y plane
Fig 4-22c Measured radiation pattern of on-package PIFA with and without BPF at x-y plane (2.45GHz).
Frequency
Fig 4-23 Gain vs. Frequency characteristic of the on-package PIFA with and without BPF at x-z plane.
Chapter 5 Conclusions
In this thesis, we presented an on-package planar inverted-F antenna for RF SOP application and also investigate the coupling effect between the on-package PIFA and the RF components in the shielding package.
In chapter 3, we introduce the design methodology of the on-package PIFA with and without ceramic materials. The ceramic materials can be excluded for cost reduction or other considerations. The influence of the different ground size and shielding package is also discussed. The performance of the on-package PIFA has merely changed when the ground size changes from 20 × 40 mm 2 to 20 × 45 mm 2. The planar element does not have not to be redesigned when the package changes from 15 × 15×1.5 mm 2 to 15 × 20×1.5 mm 3.The compact on-package PIFA has achieved the impedance bandwidth of 6.55% from 2.37 to 2.53GHz and an average gain of -0.63dBi at x-z plane. In chapter 4, it was observed that the performances of the antenna and RF passive components have merely change and the best isolation between the antenna and RF passive components can be achieved when the components have been appropriately arranged in the package.
The future work is to implement a WLAN RF module with the on-package PIFA.
The WLAN RF module consists of an antenna, a T/R switch, a band-pass filter, a low- pass filter, and a power amplifier. Furthermore, a WLAN chip can also be realized with the transceiver chip and the on-package PIFA.
References
[1] Kyutae Lim; Obatoyinbo, A.; Davis, M.; Laskar, J.; Tummala, R., “Development
of planar antennas in multi-layer packages for RF system-on-a-package
applications”,
Electrical Performance of Electronic ackaging, 2001, page(s):101 – 104.[2] Steyaert, M.; Borremans, M.; Janssens, J.; de Muer, B.; Itoh, I.; Craninckx, J.;
Crols, J.; Morifuji, E.; Momose, S.; Sansen, W., “A single-chip CMOS transceiver
for DCS-1800 wireless communications”,
Solid-State Circuits Conference, 1998, page(s):48 - 49, 411.[3] Cambridge Silicon Radio: CSR’s single-chip bluetooth radio system, U.K, 2001.
[4] Y. P. Zhang, “Finite-difference time-domain analysis of integrated ceramic ball
grid array package antenna for highly integrated wireless transceivers”, Antennas
and Propagation, IEEE Transactions on, 2004, page(s):435 – 442.[5] Warren L. Stuzman; Gary A. Thiele, “Antenna theory and design”. 2nd edition, John Wiley & Sons, INC, 1998.
[6] Shyh-Jong Chung, “Antenna design for wireless communication”, class notes.
[7] Kazuhiro Hirasawa and Misao Haneishi, “Analysis, design, and measurement of
small and low-profile antennas”, Artech House, INC, 1992.
[8] Y. P. Zhang, “Integrated circuit co-fired laminated ceramic package antenna”, Solid-State and Integrated-Circuit Technology, 2001, page(s):100 - 103 vol.1.
[9] Y. P. Zhang, “Integrated circuit ceramic ball grid array package antenna”, Antennas and Propagation, IEEE Transactions on, 2004, page(s):2538 – 2544.
[10] Y. P. Zhang, “Integration of microstrip antenna on cavity-down ceramic ball grid
array package”, Electronics Letters, Oct 2002, page(s):1307 – 1308.
[12] Y. P. Zhang, “Design of on-package microstrip antennas for single-chip wireless
transceivers”, Electronics Packaging Technology Conference, 2002, page(s):40 –
44.
[13] Tentzeris, M.M.; Laskar, J.; Papapolymerou, J.; Pinel, S.; Palazzari, V.; Li, R.;
DeJean, G.; Papageorgiou, N.; Thompson, D.; Bairavasubramanian, R.; Sarkar, S.;
Lee, J.-H., “3-D-integrated RF and millimeter-wave functions and modules using
liquid crystal polymer (LCP) system-on-package technology”, Advanced
Packaging, IEEE Transactions on, 2004, page(s):332 – 340.
[14] Chang-Ho Lee; Sutono, A.; Sangwoo Han; Kyutae Lim; Pinel, S.; Tentzeris, E.M.; Laskar, J., “A compact LTCC-based Ku-band transmitter module”, Advanced Packaging, IEEE Transactions on, 2002, page(s):374 – 384.
[15] Li-Rong Zheng; Xingzhong Duo; Meigen Shen; Torrika, T.; Michielsen, W.;
Tenhunen, H.; Liu Chen; Gang Zou; Johan Liu, “Design and implementation of
system-on-package for radio and mixed-signal applications”, High Density
Microsystem Design and Packaging and Component Failure Analysis, 2004, page(s):97 – 104.[16] Sudo, T.; Sasaki, H.; Masuda, N.; Drewniak, J.L., “Electromagnetic interference
(EMI) of system-on-package (SOP)”, Advanced Packaging, IEEE Transactions
on, 2004, page(s):304 – 314.[17]Ching-Wen Tang; Yin-Ching Lin; Chi-Yang Chang, “Realization of transmission