A 40-W balanced GaN HEMT class-E power ... - Wiley Online Library
2. J. Jung, Y. Kim, H. Lee, and Y. Lim, Analysis and design of stacked helix chip antenna, Inst Electron Eng Korea 43-TC (2006), 216-220. 3. J.J. Carr, Practical antenna handbook, McGraw-Hill, New York, NY. 4. R.C. Johnson, Antenna engineering handbook, McGraw-Hill, New York, NY. 5. T. Endo, Y. Sunahara, S. Satoh, and T. Katagi, Resonant frequency and radiation efficiency of meander line antennas, Electron Commun Jpn (Part 2) 83 (2000), 52-58. © 2009 Wiley Periodicals, Inc.
A 40-W BALANCED GaN HEMT CLASSE POWER AMPLIFIER WITH 71% EFFICIENCY FOR WCDMA BASE STATION Yong-Sub Lee, Mun-Woo Lee, and Yoon-Ha Jeong Department of Electronic and Electrical Engineering, Pohang University of Science and Technology, San 31, Hyoja-Dong, Nam-Gu, Pohang, Gyungbuk 790-784, Republic of Korea; Corresponding author: [email protected] Received 4 July 2008 ABSTRACT: A balanced class-E power amplifier (PA) using a pushpull GaN HEMT for high power and high efficiency is represented. For validation, a class-E PA is designed and implemented using a push-pull type GaN HEMT and tested for a single tone of 2.14 GHz. The measured results show that the balanced GaN HEMT class-E PA shows a drain efficiency and power-added efficiency (PAE) of 71% and 67.4% at an output power of 40 W with a gain of 13 dB through the significant harmonic suppression of below ⫺51 dBc. © 2009 Wiley Periodicals, Inc. Microwave Opt Technol Lett 51: 842– 845, 2009; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/mop. 24150 Key words: class-E power amplifier; efficiency; gallium nitride (GaN); harmonic Figure 10 Simulated and measured return loss bandwidths for the stacked helix monopole antennas
were 180 and 188 MHz, respectively. For the monopole antenna with a Gw of 25 mm (feeding point of “a”), the predicted and measured 10 dB return loss bandwidths were 210 and 218 MHz, respectively. 5. CONCLUSIONS
In this article, we experimented with printed meander-line and the stacked monopole antennas and discussed the variation of their impedance characteristics. The impedance characteristics of monopole antennas vary according to the feeding point and the Gw. When the position of the feeding point is shifted from the center of the ground plane to the edge, and when the Gw is decreased, the resistance of the antenna is increased. Our analysis indicates that by adjusting these variables the impedance characteristics of small monopole antennas can be improved. Also, the proposed impedance adjustment methods may be useful for the impedance matching of radiofrequency identification tag antennas and low-noise amplifier antennas. REFERENCES 1. K. Noguchi, M. Mizusawa, T. Yamaguchi, and Y. Okumura, Impedance characteristics of two-wire helical antenna in normal mode, Electron Commun Jpn (Part 1) 81 (1998), 777-783.
842
1. INTRODUCTION
Among switching-mode power amplifiers (SMPAs), which reduce thermal problems induced by the high-power PA in repeaters and base stations, the class-E PA is a promising candidate for a high-efficiency PA, which has the advantages of simple circuitry and high frequency operation [1– 6] For the high-power device for SMPAs, silicon laterally diffused metal oxide semiconductor (Si LDMOS) devices have long been employed as the technology of choice for high-power PAs because of its excellent cost and performance ratios [4, 5]. But as the limits of operability of these devices are reached, there will be a need for a semiconductor material that can fulfill the high frequency and high power requirements. Recently, gallium nitride high electron mobility transistors (GaN HEMTs) have been regarded as a promising candidate for high-power RF applications because they exhibit very high power densities, high electron saturation velocity, high operating temperature, and high cutoff frequency compared to any other technologies [6, 7]. In this letter, we report a balanced class-E PA using a push-pull GaN HEMT with high power and high efficiency for WCDMA base station. The compensation elements with a series capacitor and a shunt inductor are inserted to compensate for the internal parasitic components of the packaged transistor. For experimental validations, a class-E PA is implemented using a push-pull type GaN HEMT and tested using a single tone of 2.14 GHz. The
MICROWAVE AND OPTICAL TECHNOLOGY LETTERS / Vol. 51, No. 3, March 2009
DOI 10.1002/mop
Figure 1 Ideal circuit of the class-E PA
measured results show a high-power and high-efficiency performance for the balanced GaN HEMT class-E PA. 2. BALANCED GaN HEMT CLASS-E PA
The ideal circuit of the class-E PA is comprised of an ideal switch, parallel capacitor, simple resonant circuit, and specific load impedance, as shown in Figure 1. The fundamental operation theory of the class-E PA has been well demonstrated in previous literature [1– 6]. Theoretical drain efficiency of 100% can be achieved by avoiding the overlap between the high current and high voltage across the device. However, the ideal operation of the class-E PA is limited because the device can be no longer regarded as the ideal switch. In the packaged HEMT equivalent circuit model, the real transistor has not only intrinsic but also extrinsic parasitic components caused by the package, bond wires, and interconnection [8]. From the simplified equivalent circuit of the packaged transistor in Figure 2, the internal parasitic components seen at the drain consist of the shunt parasitic capacitance (Cd) and the package-induced series resistance (Rd) and inductance (Ld) of the commercial packaged active device [6]. The Cd of the transistor plays a role as the output parallel capacitor for the class-E operation but shows the nonlinear capacitance-voltage characteristics as well as higher capacitance than the capacitance to need the class-E operation. The internal parasitic components result in the power consumption and prevent the class-E PA from obtaining high efficiency and high power. Therefore, they should be compensated or eliminated. As shown in Figure 2, the compensation elements with a series capacitor and a shunt inductor are inserted in front of the harmonic control network. The series capacitor (Ccom) is used to compensate for the Rd and Ld. Here, the shunt inductor (Lcom) is used to compensate for the Cd. The non-ideal series LC resonant filter in Figure 1 causes the unnecessary harmonics at the output and then results in the degradation of the class-PA performance. Therefore, the output matching network (OMN) using transmission lines (TLINs) is
Figure 2 Simplified equivalent circuit of a single-ended GaN HEMT class-F PA with parasitic compensation elements
DOI 10.1002/mop
Figure 3 (a) Schematic and (b) simulated results of the OMN using TLINs
employed for the harmonic suppression and impedance transformation. The OMN should be an open circuit for all harmonic components for the harmonic termination and provide the class-E impedance, Z ⫽ R(1 ⫹ j tan 49.0524°) for the fundamental frequency to give ideal class-E operation [3]. Figure 3(a) shows the schematic of the proposed OMN using TLNs. From Agilent advance design system simulation, it is evident that the proposed OMN using TLINs can serve an open circuit to the fifth-harmonic component, as shown in Figure 3(b). Figure 4 shows the schematic of the proposed balanced GaN HEMT class-E PA. The input signal is divided by the in-phase two-way power splitter. The output signal is summed by a Wilkinson combiner. To compensate for the internal parasitic components of the packaged transistor, the compensation elements with a series capacitor and a shunt inductor is inserted in front of the Wilkinson combiner. The harmonic control network using TLINs is located after the Wilkinson combiner. The input matching networks (IMN) is implemented using the 50-⍀ TL and shunt capacitors to obtain maximum gain. 3. IMPLEMENTATION AND EXPERIMENTAL RESULTS
A balanced class-E PA was designed and implemented with a RFHIC RT440 push-pull type GaN HEMT having the P1dB of 20 W at 2.14 GHz. The push-pull device reduces the circuit complexity and board size. Figure 5 shows the photograph of the fabricated balanced GaN HEMT class-E PA (the size of 100 ⫻ 50 mm2). The RF35 (r ⫽ 3.5, H ⫽ 0.5 mm) circuit board has been used as the substrate on which the circuit is implemented and provides a width of 1.1 mm for all TLs with 50-⍀ characteristic impedance. The in-phase two-way power splitter is used Anaren 4A1305. The same components are used to obtain the identical performance for each transistor of the push-pull GaN HEMT. The IMN is implemented
Figure 4 The schematic of the balanced class-E PA using a push-pull GaN HEMT
MICROWAVE AND OPTICAL TECHNOLOGY LETTERS / Vol. 51, No. 3, March 2009
843
Figure 5 The photograph of the balanced class-E PA using a push-pull GaN HEMT. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com]
with two capacitors of 0.5 and 1.5 pF. The RF choke (RFC) uses the 50-⍀ TLIN with /4 length. The 100 ⍀ in the Wilkinson combiner is set to two 200 ⍀ with the 1210 size in parallel. The series capacitor (Ccom) and shunt inductor (Lcom) in the compensation circuit are a capacitance of 1.5 pF and an inductance of 1.1 nH, respectively. To further improve the performance, the 0.2 pF is inserted after the OMN using TLINs. Figure 6 shows the measured output power, efficiency, and gain characteristics according to input power levels for a single tone of 2.14 GHz. With a drain bias voltage (VDD) of 25 V, each gate bias (VGS) is set to the class-AB biases of ⫺1.55 V (quiescent current, IDQ, of 150 mA) and ⫺1.62 V (IDQ of 135 mA), which are higher than the pinch-off voltage (Vpinch-off) of ⫺1.84 V. From the measured results, the maximum power-added efficiency (PAE) of 67.4% with drain efficiency of 71% is obtained at an output power of 46 dBm with a gain of 13 dB. Figure 7 shows the measured harmonic power levels according to output power. The OMN using TLINs obtains the significant harmonic power suppression. The second-harmonic power level is reduced below ⫺65 dBc. Figure 8 shows the measured power spectrum density at an output power of 46 dBm. Figure 9 illustrates the measured output power and PAE characteristics according to drain bias voltage at an input power of 33 dBm. From the measured results, the balanced GaN HEMT class-E PA shows the flat response of output power and PAE characteristics according to drain bias voltage. It is especially critical that high efficiency characteristics are maintained for a wide range of
Figure 6 The measured output power, efficiency, and gain characteristics for a single tone of 2.14 GHz
844
Figure 7 The measured second- and third-harmonic power levels according to output power levels
drain bias voltage, because a flat response of the efficiency according to drain bias voltage is required in the case of the drain bias modulation applications, such as envelope elimination and restoration and envelope tracking. 4. CONCLUSIONS
In this letter, we have reported a high-power and high-efficiency balanced GaN HEMT class-E PA for WCDMA base station applications. The OMN using TLINs was used to suppress harmonic power levels and reduce losses. Also, the compensation elements with a series capacitor and a shunt inductor were inserted in front of the OMN to compensate for the internal parasitic components of the packaged transistor. To verify these methods, a balanced class-E PA was designed and implemented using a push-pull type GaN HEMT and tested for a single tone of 2.14 GHz. The drain efficiency and PAE of 71% and 67.4% with a gain of 13 dB was achieved at a Pout of 46 dBm. Additionally, the proposed amplifier showed the high power, high efficiency, and significant harmonic termination according to drain bias voltage. The results prove that
Figure 8 The measured power spectrum density at a Pout of 46 dBm. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com]
MICROWAVE AND OPTICAL TECHNOLOGY LETTERS / Vol. 51, No. 3, March 2009
DOI 10.1002/mop
FULLY EMBEDDED LUMPED LCQUADRATURE HYBRID COUPLER INTO ORGANIC PACKAGING SUBSTRATE FOR POWER SAMPLING Yi Jae Lee and Jae Yeong Park Department of Electronic Engineering, Kwangwoon University, 447-1, Wolgye-Dong, Nowon-Gu, Seoul 139-701 Korea; Corresponding author: [email protected] Received 9 July 2008
Figure 9 The measured output power and PAE characteristics according to drain bias voltage
the proposed balanced GaN HEMT class-E PA can deliver the high-power and high-efficiency performances for WCDMA base station. ACKNOWLEDGMENT
This work was partially supported by the BK21 program and the National Center for Nanomaterials Technology (NCNT) in Korea. REFERENCES 1. S.C. Cripps, Advanced techniques in RF power amplifiers design, Artech House, Norwood, MA, 2002. 2. F.H. Raab, Idealized operation of the class E tuned power amplifier, IEEE Trans Circuits Syst CAS-25 (1977), 725–735. 3. T.B. Mader and Z.B. Popovic, The transmission line high-efficiency class-E amplifier, IEEE Microwave Guided Wave Lett 5 (1995), 290 – 292. 4. A. Adahl and H. Zirath, An 1 GHz class E LDMOS power amplifier, In: Proceedings of the 33th European Microwave Conference, October 2003, Vol. 1, pp. 285–288. 5. Y.S. Lee, K.I. Jeon, and Y.H. Jeong, A 2.14 GHz class-E LDMOS power amplifier, In: Proceedings of the Asia Pacific Microwave Conference, December 2006, Vol. 2, pp. 1015–1018. 6. Y.S. Lee and Y.H. Jeong, Applications of GaN HEMTs and SiC MESFETs in high efficiency class-E power amplifier design for WCDMA applications, In: Proceedings of the IEEE MTT-S Int Microwave Symp Dig, June 2007, pp. 1099 –1102. 7. K.J. Cho, W.J. Kim, J.H. Kim, and S.P. Stapleton, 40 W gallium nitride microwave Doherty amplifier, IEEE MTT-S Int Microwave Symp Dig, June 2006, pp. 1895–1898. 8. P.M. Cabral, J.C. Pedro, and N.B. Carvalho, Nonlinear device model of microwave power GaN HEMTs for high power-amplifier design, IEEE Trans Microwave Theory Technol, November 2004, Vol. 52, No. 11, pp. 2585–2592. © 2009 Wiley Periodicals, Inc.
DOI 10.1002/mop
ABSTRACT: In this article, fully embedded lumped LC-quadrature hybrid coupler are designed, fabricated, and characterized into multilayered organic packaging substrate. This embedded device comprised high Q MIM capacitors and circular spiral stacked inductors. For realizing fully embedded high Q capacitor, barium titanate (BaTiO3) composite high dielectric film was utilized. The measured return loss and isolation were better than 25 dB. Measured insertion loss was about 0.38 – 0.42 dB at the frequencies ranged from 824 to 894 MHz. The coupler exhibited a coupling loss of 23.3–24 dB. The phase error between through and coupling ports was 1.2°–2.4°. It has a size of 2.8 mm ⫻2.95 mm ⫻0.77 mm (height), which is the smallest one in the couplers developed onto the packaging substrate. The measured performance characteristics were also well matched with the 3D EM simulated ones. © 2009 Wiley Periodicals, Inc. Microwave Opt Technol Lett 51: 845– 848, 2009; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/mop.24173 Key words: embedded quadrature hybrid coupler; high Q MIM capacitor; high Q spiral stacked inductor; barium titanate (BaTiO3) composite film; organic packaging 1. INTRODUCTION
Quadrature hybrid couplers are useful as power dividers or power combiners in various microwave circuits such as balanced amplifiers, balanced mixers, and phase shifters for achieving good input and output power matching, combining, and isolation. In CDMA handset applications, quadrature couplers are required to determine the phase error of transmitter by using QPSK modulation scheme. Thus, these couplers need to have small size, low cost, tight amplitude balance, and 90° phase difference between the coupled and direct ports. Up to now, microstrip couplers with parallel coupled transmission lines have been widely used, because they could be easily incorporated into and implemented with other circuits [1]. However, their sizes were relatively large because of the use of quarterwavelength (/4) transmission lines. Thus, size reduction became a major design consideration. In particular, at low microwave frequency bands, the size of the hybrid coupler was too large to be used in practical applications. To address these problems, two methods were proposed. The first one was to use the folded line configuration, but the resultant circuit area was still too large [2, 3]. The other was to use lumped-element components such as metalinsulator-metal (MIM) capacitors [4 – 6]. Thus, low temperature cofired ceramic (LTCC) technology has been widely investigated and applied. Although the LTCC-based couplers have good performance characteristics and high packaging density, they are limited to mass production and low cost due to the process complexity, small size, and the shrinkage occurred during the firing process. Recently, embedded passive components onto or into organic substrates have been developed due to several advantages such as low-cost fabrication and large area manufacturing. In this article, fully embedded lumped element quadrature hybrid couplers into a multilayered organic packaging substrate
MICROWAVE AND OPTICAL TECHNOLOGY LETTERS / Vol. 51, No. 3, March 2009
845
A 40-W BALANCED GaN HEMT CLASSE POWER AMPLIFIER WITH 71% EFFICIENCY FOR WCDMA BASE STATION Yong-Sub Lee, Mun-Woo Lee, and Yoon-Ha Jeong Department of Electronic and Electrical Engineering, Pohang University of Science and Technology, San 31, Hyoja-Dong, Nam-Gu, Pohang, Gyungbuk 790-784, Republic of Korea; Corresponding author: [email protected] Received 4 July 2008 ABSTRACT: A balanced class-E power amplifier (PA) using a pushpull GaN HEMT for high power and high efficiency is represented. For validation, a class-E PA is designed and implemented using a push-pull type GaN HEMT and tested for a single tone of 2.14 GHz. The measured results show that the balanced GaN HEMT class-E PA shows a drain efficiency and power-added efficiency (PAE) of 71% and 67.4% at an output power of 40 W with a gain of 13 dB through the significant harmonic suppression of below ⫺51 dBc. © 2009 Wiley Periodicals, Inc. Microwave Opt Technol Lett 51: 842– 845, 2009; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/mop. 24150 Key words: class-E power amplifier; efficiency; gallium nitride (GaN); harmonic Figure 10 Simulated and measured return loss bandwidths for the stacked helix monopole antennas
were 180 and 188 MHz, respectively. For the monopole antenna with a Gw of 25 mm (feeding point of “a”), the predicted and measured 10 dB return loss bandwidths were 210 and 218 MHz, respectively. 5. CONCLUSIONS
In this article, we experimented with printed meander-line and the stacked monopole antennas and discussed the variation of their impedance characteristics. The impedance characteristics of monopole antennas vary according to the feeding point and the Gw. When the position of the feeding point is shifted from the center of the ground plane to the edge, and when the Gw is decreased, the resistance of the antenna is increased. Our analysis indicates that by adjusting these variables the impedance characteristics of small monopole antennas can be improved. Also, the proposed impedance adjustment methods may be useful for the impedance matching of radiofrequency identification tag antennas and low-noise amplifier antennas. REFERENCES 1. K. Noguchi, M. Mizusawa, T. Yamaguchi, and Y. Okumura, Impedance characteristics of two-wire helical antenna in normal mode, Electron Commun Jpn (Part 1) 81 (1998), 777-783.
842
1. INTRODUCTION
Among switching-mode power amplifiers (SMPAs), which reduce thermal problems induced by the high-power PA in repeaters and base stations, the class-E PA is a promising candidate for a high-efficiency PA, which has the advantages of simple circuitry and high frequency operation [1– 6] For the high-power device for SMPAs, silicon laterally diffused metal oxide semiconductor (Si LDMOS) devices have long been employed as the technology of choice for high-power PAs because of its excellent cost and performance ratios [4, 5]. But as the limits of operability of these devices are reached, there will be a need for a semiconductor material that can fulfill the high frequency and high power requirements. Recently, gallium nitride high electron mobility transistors (GaN HEMTs) have been regarded as a promising candidate for high-power RF applications because they exhibit very high power densities, high electron saturation velocity, high operating temperature, and high cutoff frequency compared to any other technologies [6, 7]. In this letter, we report a balanced class-E PA using a push-pull GaN HEMT with high power and high efficiency for WCDMA base station. The compensation elements with a series capacitor and a shunt inductor are inserted to compensate for the internal parasitic components of the packaged transistor. For experimental validations, a class-E PA is implemented using a push-pull type GaN HEMT and tested using a single tone of 2.14 GHz. The
MICROWAVE AND OPTICAL TECHNOLOGY LETTERS / Vol. 51, No. 3, March 2009
DOI 10.1002/mop
Figure 1 Ideal circuit of the class-E PA
measured results show a high-power and high-efficiency performance for the balanced GaN HEMT class-E PA. 2. BALANCED GaN HEMT CLASS-E PA
The ideal circuit of the class-E PA is comprised of an ideal switch, parallel capacitor, simple resonant circuit, and specific load impedance, as shown in Figure 1. The fundamental operation theory of the class-E PA has been well demonstrated in previous literature [1– 6]. Theoretical drain efficiency of 100% can be achieved by avoiding the overlap between the high current and high voltage across the device. However, the ideal operation of the class-E PA is limited because the device can be no longer regarded as the ideal switch. In the packaged HEMT equivalent circuit model, the real transistor has not only intrinsic but also extrinsic parasitic components caused by the package, bond wires, and interconnection [8]. From the simplified equivalent circuit of the packaged transistor in Figure 2, the internal parasitic components seen at the drain consist of the shunt parasitic capacitance (Cd) and the package-induced series resistance (Rd) and inductance (Ld) of the commercial packaged active device [6]. The Cd of the transistor plays a role as the output parallel capacitor for the class-E operation but shows the nonlinear capacitance-voltage characteristics as well as higher capacitance than the capacitance to need the class-E operation. The internal parasitic components result in the power consumption and prevent the class-E PA from obtaining high efficiency and high power. Therefore, they should be compensated or eliminated. As shown in Figure 2, the compensation elements with a series capacitor and a shunt inductor are inserted in front of the harmonic control network. The series capacitor (Ccom) is used to compensate for the Rd and Ld. Here, the shunt inductor (Lcom) is used to compensate for the Cd. The non-ideal series LC resonant filter in Figure 1 causes the unnecessary harmonics at the output and then results in the degradation of the class-PA performance. Therefore, the output matching network (OMN) using transmission lines (TLINs) is
Figure 2 Simplified equivalent circuit of a single-ended GaN HEMT class-F PA with parasitic compensation elements
DOI 10.1002/mop
Figure 3 (a) Schematic and (b) simulated results of the OMN using TLINs
employed for the harmonic suppression and impedance transformation. The OMN should be an open circuit for all harmonic components for the harmonic termination and provide the class-E impedance, Z ⫽ R(1 ⫹ j tan 49.0524°) for the fundamental frequency to give ideal class-E operation [3]. Figure 3(a) shows the schematic of the proposed OMN using TLNs. From Agilent advance design system simulation, it is evident that the proposed OMN using TLINs can serve an open circuit to the fifth-harmonic component, as shown in Figure 3(b). Figure 4 shows the schematic of the proposed balanced GaN HEMT class-E PA. The input signal is divided by the in-phase two-way power splitter. The output signal is summed by a Wilkinson combiner. To compensate for the internal parasitic components of the packaged transistor, the compensation elements with a series capacitor and a shunt inductor is inserted in front of the Wilkinson combiner. The harmonic control network using TLINs is located after the Wilkinson combiner. The input matching networks (IMN) is implemented using the 50-⍀ TL and shunt capacitors to obtain maximum gain. 3. IMPLEMENTATION AND EXPERIMENTAL RESULTS
A balanced class-E PA was designed and implemented with a RFHIC RT440 push-pull type GaN HEMT having the P1dB of 20 W at 2.14 GHz. The push-pull device reduces the circuit complexity and board size. Figure 5 shows the photograph of the fabricated balanced GaN HEMT class-E PA (the size of 100 ⫻ 50 mm2). The RF35 (r ⫽ 3.5, H ⫽ 0.5 mm) circuit board has been used as the substrate on which the circuit is implemented and provides a width of 1.1 mm for all TLs with 50-⍀ characteristic impedance. The in-phase two-way power splitter is used Anaren 4A1305. The same components are used to obtain the identical performance for each transistor of the push-pull GaN HEMT. The IMN is implemented
Figure 4 The schematic of the balanced class-E PA using a push-pull GaN HEMT
MICROWAVE AND OPTICAL TECHNOLOGY LETTERS / Vol. 51, No. 3, March 2009
843
Figure 5 The photograph of the balanced class-E PA using a push-pull GaN HEMT. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com]
with two capacitors of 0.5 and 1.5 pF. The RF choke (RFC) uses the 50-⍀ TLIN with /4 length. The 100 ⍀ in the Wilkinson combiner is set to two 200 ⍀ with the 1210 size in parallel. The series capacitor (Ccom) and shunt inductor (Lcom) in the compensation circuit are a capacitance of 1.5 pF and an inductance of 1.1 nH, respectively. To further improve the performance, the 0.2 pF is inserted after the OMN using TLINs. Figure 6 shows the measured output power, efficiency, and gain characteristics according to input power levels for a single tone of 2.14 GHz. With a drain bias voltage (VDD) of 25 V, each gate bias (VGS) is set to the class-AB biases of ⫺1.55 V (quiescent current, IDQ, of 150 mA) and ⫺1.62 V (IDQ of 135 mA), which are higher than the pinch-off voltage (Vpinch-off) of ⫺1.84 V. From the measured results, the maximum power-added efficiency (PAE) of 67.4% with drain efficiency of 71% is obtained at an output power of 46 dBm with a gain of 13 dB. Figure 7 shows the measured harmonic power levels according to output power. The OMN using TLINs obtains the significant harmonic power suppression. The second-harmonic power level is reduced below ⫺65 dBc. Figure 8 shows the measured power spectrum density at an output power of 46 dBm. Figure 9 illustrates the measured output power and PAE characteristics according to drain bias voltage at an input power of 33 dBm. From the measured results, the balanced GaN HEMT class-E PA shows the flat response of output power and PAE characteristics according to drain bias voltage. It is especially critical that high efficiency characteristics are maintained for a wide range of
Figure 6 The measured output power, efficiency, and gain characteristics for a single tone of 2.14 GHz
844
Figure 7 The measured second- and third-harmonic power levels according to output power levels
drain bias voltage, because a flat response of the efficiency according to drain bias voltage is required in the case of the drain bias modulation applications, such as envelope elimination and restoration and envelope tracking. 4. CONCLUSIONS
In this letter, we have reported a high-power and high-efficiency balanced GaN HEMT class-E PA for WCDMA base station applications. The OMN using TLINs was used to suppress harmonic power levels and reduce losses. Also, the compensation elements with a series capacitor and a shunt inductor were inserted in front of the OMN to compensate for the internal parasitic components of the packaged transistor. To verify these methods, a balanced class-E PA was designed and implemented using a push-pull type GaN HEMT and tested for a single tone of 2.14 GHz. The drain efficiency and PAE of 71% and 67.4% with a gain of 13 dB was achieved at a Pout of 46 dBm. Additionally, the proposed amplifier showed the high power, high efficiency, and significant harmonic termination according to drain bias voltage. The results prove that
Figure 8 The measured power spectrum density at a Pout of 46 dBm. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com]
MICROWAVE AND OPTICAL TECHNOLOGY LETTERS / Vol. 51, No. 3, March 2009
DOI 10.1002/mop
FULLY EMBEDDED LUMPED LCQUADRATURE HYBRID COUPLER INTO ORGANIC PACKAGING SUBSTRATE FOR POWER SAMPLING Yi Jae Lee and Jae Yeong Park Department of Electronic Engineering, Kwangwoon University, 447-1, Wolgye-Dong, Nowon-Gu, Seoul 139-701 Korea; Corresponding author: [email protected] Received 9 July 2008
Figure 9 The measured output power and PAE characteristics according to drain bias voltage
the proposed balanced GaN HEMT class-E PA can deliver the high-power and high-efficiency performances for WCDMA base station. ACKNOWLEDGMENT
This work was partially supported by the BK21 program and the National Center for Nanomaterials Technology (NCNT) in Korea. REFERENCES 1. S.C. Cripps, Advanced techniques in RF power amplifiers design, Artech House, Norwood, MA, 2002. 2. F.H. Raab, Idealized operation of the class E tuned power amplifier, IEEE Trans Circuits Syst CAS-25 (1977), 725–735. 3. T.B. Mader and Z.B. Popovic, The transmission line high-efficiency class-E amplifier, IEEE Microwave Guided Wave Lett 5 (1995), 290 – 292. 4. A. Adahl and H. Zirath, An 1 GHz class E LDMOS power amplifier, In: Proceedings of the 33th European Microwave Conference, October 2003, Vol. 1, pp. 285–288. 5. Y.S. Lee, K.I. Jeon, and Y.H. Jeong, A 2.14 GHz class-E LDMOS power amplifier, In: Proceedings of the Asia Pacific Microwave Conference, December 2006, Vol. 2, pp. 1015–1018. 6. Y.S. Lee and Y.H. Jeong, Applications of GaN HEMTs and SiC MESFETs in high efficiency class-E power amplifier design for WCDMA applications, In: Proceedings of the IEEE MTT-S Int Microwave Symp Dig, June 2007, pp. 1099 –1102. 7. K.J. Cho, W.J. Kim, J.H. Kim, and S.P. Stapleton, 40 W gallium nitride microwave Doherty amplifier, IEEE MTT-S Int Microwave Symp Dig, June 2006, pp. 1895–1898. 8. P.M. Cabral, J.C. Pedro, and N.B. Carvalho, Nonlinear device model of microwave power GaN HEMTs for high power-amplifier design, IEEE Trans Microwave Theory Technol, November 2004, Vol. 52, No. 11, pp. 2585–2592. © 2009 Wiley Periodicals, Inc.
DOI 10.1002/mop
ABSTRACT: In this article, fully embedded lumped LC-quadrature hybrid coupler are designed, fabricated, and characterized into multilayered organic packaging substrate. This embedded device comprised high Q MIM capacitors and circular spiral stacked inductors. For realizing fully embedded high Q capacitor, barium titanate (BaTiO3) composite high dielectric film was utilized. The measured return loss and isolation were better than 25 dB. Measured insertion loss was about 0.38 – 0.42 dB at the frequencies ranged from 824 to 894 MHz. The coupler exhibited a coupling loss of 23.3–24 dB. The phase error between through and coupling ports was 1.2°–2.4°. It has a size of 2.8 mm ⫻2.95 mm ⫻0.77 mm (height), which is the smallest one in the couplers developed onto the packaging substrate. The measured performance characteristics were also well matched with the 3D EM simulated ones. © 2009 Wiley Periodicals, Inc. Microwave Opt Technol Lett 51: 845– 848, 2009; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/mop.24173 Key words: embedded quadrature hybrid coupler; high Q MIM capacitor; high Q spiral stacked inductor; barium titanate (BaTiO3) composite film; organic packaging 1. INTRODUCTION
Quadrature hybrid couplers are useful as power dividers or power combiners in various microwave circuits such as balanced amplifiers, balanced mixers, and phase shifters for achieving good input and output power matching, combining, and isolation. In CDMA handset applications, quadrature couplers are required to determine the phase error of transmitter by using QPSK modulation scheme. Thus, these couplers need to have small size, low cost, tight amplitude balance, and 90° phase difference between the coupled and direct ports. Up to now, microstrip couplers with parallel coupled transmission lines have been widely used, because they could be easily incorporated into and implemented with other circuits [1]. However, their sizes were relatively large because of the use of quarterwavelength (/4) transmission lines. Thus, size reduction became a major design consideration. In particular, at low microwave frequency bands, the size of the hybrid coupler was too large to be used in practical applications. To address these problems, two methods were proposed. The first one was to use the folded line configuration, but the resultant circuit area was still too large [2, 3]. The other was to use lumped-element components such as metalinsulator-metal (MIM) capacitors [4 – 6]. Thus, low temperature cofired ceramic (LTCC) technology has been widely investigated and applied. Although the LTCC-based couplers have good performance characteristics and high packaging density, they are limited to mass production and low cost due to the process complexity, small size, and the shrinkage occurred during the firing process. Recently, embedded passive components onto or into organic substrates have been developed due to several advantages such as low-cost fabrication and large area manufacturing. In this article, fully embedded lumped element quadrature hybrid couplers into a multilayered organic packaging substrate
MICROWAVE AND OPTICAL TECHNOLOGY LETTERS / Vol. 51, No. 3, March 2009
845
