Transmitter Architecture for CA - IEEE Xplore

3 downloads 554 Views 1MB Size Report
Jul 11, 2013 - Wenhua Chen is also with the Microwave and ... eration (4G) of wireless technology. The ... mized for the third generation of wireless standards.
Transmitter Architecture for CA Seyed Aidin Bassam, Wenhua Chen, Mohamed Helaoui, and Fadhel M. Ghannouchi

T

he long-term evolution (LTE)-advanced standard requires the possibility of having signal transmission with up to 100 MHz bandwidth in the fourth generation (4G) of wireless technology. The component carriers (CCs) of the signal in both uplink (UL) and downlink (DL) are limited to 20 MHz, introducing carrier aggregation (CA) of multiple CCs (up to maximum five CCs) has broadened the overall signal bandwidth to up to 100 MHz [1]–[4]. Having wideband OFDM M-QAM modulated signals with a high peak-to-average power ratio (PAPR) could bring numerous challenges to the design aspect of transmitter linearity, output power gain, and overall efficiency. The beauty and efficiency of the CA mechanism comes with an extra focus on the transmitter architecture design and achievable efficiency. When it comes to realization of CA techniques, the transmitter architecture is of high importance [4]. It is important to utilize an energy-efficient system where energy resources are limited and the most efficient performance per energy unit is beneficial to reduce the overall cost of the system [5]. The fact

© digital vision

Seyed Aidin Bassam ([email protected]), Wenhua Chen ([email protected]), Mohamed Helaoui ([email protected]), and Fadhel M. Ghannouchi ([email protected]) are with the iRadio Laboratory, Department of Electrical and Computer Engineering, Schulich School of Engineering, The University of Calgary, AB, T2N 1N4 Canada. Wenhua Chen is also with the Microwave and Antenna Institute, Department of Electronic Engineering, Tsinghua University, Beijing, 100084 China. Digital Object Identifier 10.1109/MMM.2013.2259399 Date of publication: 11 July 2013

78



1527-3342/13/$31.00©2013IEEE

July/August 2013

The long-term evolution–advanced standard requires the possibility of having signal transmission with up to 100-MHz bandwidth in the fourth generation of wireless technology.

that most transmitters have nonlinear behavior is well known, as distortion and nonlinearities introduced by the transmitter’s behavior on the quality of output signal have been investigated by a number of researchers with a good summary of these efforts provided in [6]. Different linearization architectures such as feedforward [7], Cartesian feedback [8], analog and digital predistortion (DPD) [9]–[11], and envelope tracking [12]–[14], to name some, have been fully explored. To some extent, these linearization schemes are optimized for the third generation of wireless standards and other legacy systems. This article reviews transmitter architectures and analyzes their applicability and energy efficiency for CA in LTE-advanced systems. The candidate transmitter architectures are also reviewed and followed with a brief discussion on the benefits and challenges of each topology. The presence of intermodulation and crossmodulation distortion in dual-band nonlinear systems is discussed. The design of the concurrent dual-band PA is briefly explained, and later the energy-efficient dual-band transmitter solution based on dual-band PA and dual-band linearization architecture is demonstrated. Finally, performance evaluation are presented and discussed in the conclusion.

architectures applicable for CA and compares them in terms of design challenges, suitability, and energy efficiency.

Transmitter Topologies There are five main transmitter topologies as shown in Figure 2. Figure 2(a) presents a conventional single-band transmitter architecture that consists of digital baseband processing unit and single-band RF front end that contains the passive and active RF components. The RF front end can be one of the conventional topologies: 1) super-heterodyne, 2) low IF, and 3) direct conversion [15], [16]. Provided adequate bandwidth, the architecture in Figure 2(a) can be used easily for intraband CA where the CCs are all transmitted within the same band of operation.

Carrier Aggregation: Intraband, Interband Modes CA is defined as the concurrent transmission of multiple carrier components from one terminal. The CA is categorized as interband and intraband CA based on the frequency band of operation. Having multiple CCs at the same frequency band results in intraband CCs, whereas concurrent transmission of CCs from different frequency bands results in interband CA. In the case of transmitting multiple CCs from the same frequency band and considering whether CCs occupying the adjacent channels or not, the intraband CA further splits into contiguous and noncontiguous modes. Figure 1 shows the three possible scenarios of the CA. Figure 1(a) and (b) show intraband CA in contiguous and noncontiguous modes with Figure 1(c) showing the interband CA.

Data

Vacant Channel Intraband CA, Contiguous Mode

Data

July/August 2013

Data Band A

Band B (a)

Intraband CA, Non-Contiguous Mode

Data

Data Band A

Candidate Transmitter Architectures for LTE-Advanced The transmitter architecture is the building block of the communication systems. It is the physical layer of the system and the actual device that send the modulated signal through the wireless channel. In this regard, it is important to investigate the impact of adding CA to the architecture design and to explore the appropriate design approaches and methodologies. This section reviews the transmitter

Occupied Channel

Band B (b) Interband CA

Data

Data Band A

Band B (c)

Figure 1. (a) and (b) The intraband CA in contiguous and noncontiguous modes and (c) the interband CA in LTE-advanced systems.



79

Baseband Processing

DAC

Up-Conversion Unit

PA RF Filter (a)

Baseband Processing

DAC

Baseband Processing

DAC

Up-Conversion Unit

PA

Up-Conversion Unit

PA

RF Filter

RF Filter (b)

Baseband Processing

DAC

Up-Conversion Unit

PA RF Multiplexer

Baseband Processing

DAC

Up-Conversion Unit

Multiband RF Filter

PA (c)

Baseband Processing

DAC

Up-Conversion Unit Power Combiner

Baseband Processing

DAC

Multiband PA

Up-Conversion Unit

Multiband RF Filter

(d) 2f1 I1 Baseband Processing Q1

Delta-Sigma Modulator 1

#

Delta-Sigma Modulator 1

#

2f1 2f2 I2 Baseband Processing Q2

1 0 -1 0 @ 4f1

+

0 1 0 -1 @ 4f1 1 0 -1 0 @ 4f2

Delta-Sigma Modulator 1

#

Delta-Sigma Modulator 1

#

2f2

Digital Combiner

DualBand PA

Dual-Band RF Filter

+

0 1 0 -1 @ 4f2 (e)

Figure 2. Candidate transmitter architectures for CA in LTE-advanced systems: (a) single-band RF transmitter, (b) multiple branch transmitter, (c) multiple branch transmitter with multiband RF filter and antenna, (d) multiband transmitter architecture, and (e) delta-sigma-based CA transmitter architecture.

80



July/August 2013

Modifying the conventional architecture for interband CA results in a multiband designing approach using alternative architectures, as presented in Figure 2(b)–(e). Figure 2(b) shows multiple branches of the conventional transmitter architectures, with each branch operating in a single frequency band. The bulky size and packaging issues might be a concern in this topology. Using multiband active and passive RF components could be an efficient approach to design a compact and efficient transmitter topology. In this architecture one multiband component working at multiple operating frequencies replaces the bank of components each working at a single operating frequency. Figure 2(c) and (d) shows two possible transmitter architectures where multiband RF components have been used. In Figure 2(c), the final stage of the transmitter, RF filter and antennas have been designed using multiband RF components. Each RF chain has its own power amplification stage. The RF signals at the output of the power amplification stages are combined by using an RF multiplexer (in the case of the double branch system, by using a diplexer) and passed through a multiband RF filter. Isolators are required at the output of the PA to avoid any load modulation effects between the different branches. The nonlinear response of the PAs produces nonlinear distortion in each of the branches that requires the use of standalone linearization technique for distortion compensation. Figure 2(d) presents a multiband architecture where the power amplification unit is also designed as one multiband power amplifier (PA). In this architecture, instead of having a single-band PA for each frequency of operation, RF signals are combined and then are amplified using a single multiband RF PA. The signal at the PA’s output is filtered using a multiband RF filter and finally passed to the antenna system. In all of these architectures, an isolator is required before the antenna system in order to suppress the effect of the antennas load variation. In full-duplex transceiver architecture, a duplexer replaces the preantenna isolator. The main differences between the transmitter architectures in Figure 2(c) and (d) are the required number of PAs and the possible position of the power combination module before or after the power amplification module. As will be explained in the following, the overall efficiency of the transmitter is affected by the position of the power combination module. Power amplification could be considered as the main building block of the transmitter RF chain; therefore, in this article the transmitter architectures in Figure 2(b) and (c) are referred to as multiplebranch transmitters (separate RF PA), and the transmitter architecture in Figure 2(d) is referred to as a multiband architecture.

July/August 2013

In the case of transmitting multiple CCs from the same frequency band and considering whether CCs occupying the adjacent channels or not, the intraband CA further splits into contiguous and noncontiguous modes. There is also another approach in using a concurrent multiband transmitter for CA, which is based on delta-sigma modulators and switching-mode PAs [17]. As shown in Figure 2(e), this topology consists of deltasigma modulators, digital up-conversion units, a combiner, and PAs. The delta-sigma modulators produce two-level (1-bit) output signals that are up-converted to RF using a digital multiplexer [17]. The nature of the signal is binary, and therefore a high speed digital multiplexer can be used instead of analog high power combiners. In addition, switching-mode PAs have been used to improve energy efficiency. The challenge in a delta-sigma-based transmitter when using CA is to prevent the noise generated by the delta-sigma modulator from interfering with the signal within each frequency band. This is mainly achievable by modification on the design of the noise transfer function (NTF) [17]. In theory, delta-sigma-based transmitters using switching model PAs could improve the overall efficiency of the system. The fact that delta-sigma modulators operate as digital modulators and upconverter units makes these topologies potential candidates for future reconfigurable and cognitive radio systems. Most published literatures on deltasigma modulators still show designs that exhibit poor power efficiency mostly because of unnecessary, yet unavoidable, amplification of out-of-band noise. In fact, because of high level of noise, the overall efficiency of delta-sigma-based transmitters is still very low [18]. The remainder of this article focuses on interband CA limited to maximum of two bands of operation [2], and will compare and analyze the double-branch and dual-band transmitter architectures in terms of design challenges, system nonlinearities, energy efficiency, and applicable linearization techniques.

Benefits and Challenges The main benefits of having the multiband transmitters in comparison with the multiple branch transmitters are the following. First, the power combination modules in both topologies [Figure 2(c) and (d)] are lossy passive modules, and there will always be loss in power combining.



81

Efficient linearization techniques need to be implemented for distortion compensation and overall system linearity improvement.

of the dual-band transmitters, and in general multiband transmitters, is better than the multiple branch transmitter architectures. Second, in interband CAs, there are scenarios where there is no active signal in one of the frequency bands. To save power consumption and improve power efficiency, a controlling mechanism is required to turn on and off the inactive RF branches. In the case of the multiple-branch architecture [Figure 2(c)], there is a need for a controlling scheme; without this, the architecture would not be energy efficient. In the multiband architecture [Figure 2(d)] though, this has been inherently considered in the design. As for the challenges, designing multiband passive and active RF components is more complex than single-band RF components. The steps required for concurrent dual-band PA design are more complex. Moreover, the nonlinear behavior of PA produces intermodulation and distortion at the amplified output signal. These nonlinearities significantly degrade the output signal quality and could generate in-band and out-of-band distortion that violates the specification requirements of signal transmission in terms of EVM, leakage within adjacent channels, spurs, and additional noise at the receiver bands. Efficient linearization techniques need to be implemented for distortion compensation and overall system linearity improvement. As an example, in Figure 4, the signal at the input and output of a nonlinear dual-band PA with thirdorder nonlinearity are shown. Figure 4(a) is the dualband two-tone excitation signal at the PA’s input and Figure 4(b) is the PA’s output signal. The PA output signal is composed of the amplified two-tone signals plus additional in-band intermodulation, out-of-band intermodulation, and cross-modulation terms [19], [20]. In dual-band nonlinear PA, in addition to in-band and out-ofband intermodulation products, there are cross-modulation terms. These distortion products have been investigated [19], [20], and it was shown that conventional DPD schemes are not able to compensate for the intermodulation and cross-modulation terms in dual-band systems.

Considering the position of the power combination module before or after the PA, the amount of power dissipated in the multiple-branch model in Figure 2(c) is always higher than the multiband transmitter in Figure 2(d). This significantly increases the power loss in the system and degrades the transmitter’s power efficiency. The typical isolator has about 0.3+0.5 dB insertion loss. A typical RF multiplexer, such as a diplexer, could have about 1.5 insertion losses, resulting in approximately a 2 dB insertion loss at the output stage of the transmitter. A simulation comparison in Agilent Advanced Design System (ADS) for the two architectures using the harmonic balance simulator and continuouswave (CW) signals was carried out to verify this analysis. Using identical dual-band PA prototypes in the topologies shown in Figure 2(c) and (d), the simulated drain efficiencies against total output power are presented. The initial phases of the two CW signals were set to zero. It is worth mentioning that different phase offsets were also simulated, which gave the same results. The efficiency performance results in Figure 3 show that the efficiency of the dual-band PA is more than double the efficiency of the double branch PAs. Since the overall efficiency of the transmitter is dominated by the efficiency of PAs, it can be concluded that the overall efficiency

45 Double-Branch PAs [Figure 2(c)]

40

Dual-Band PAs [Figure 2(d)]

Drain Efficiency (%)

35 30 25 20

Dual-Band Transmitter Architecture Demonstration

15 10 5

80

30

7.

48

30

5.

62

30

4.

46

00

3.

47

70

2.

27

70

1.

26

10

0.

64

80

0.

32

00

0.

14

00

0.

06

50

0.

02

10

00

0.

00

0.

0.

00

01

0

Total Output Power (W)

Figure 3. Comparison of the overall efficiency of the double-branch PAs [Figure 2(c)] and concurrent dual-band PAs [Figure 2(d)].

82



As described in the “Candidate Transmitter Architectures for LTE-Advanced” section, concurrent dual-band transmitter topologies are energy-efficient solutions for the realization of intraband and interband CA techniques, provided designing highly efficient concurrent dual-band PAs and developing concurrent linearization schemes are feasible.

July/August 2013

The design and efficiency improvement of concurrent dualband PAs has been considered and presented in the recent literature.

The design and efficiency improvement of concurrent dual-band PAs has been considered and presented in the recent literature. In [20]–[22], the design and implementation of a concurrent dualband Doherty PA operating at 880 and 1,960 MHz LTE frequency bands has been reported. The measured results show power-added efficiency (PAE) of 33% and 30% at 6-dB power back-off from saturation point at carrier frequencies of 880 and 1,960 MHz. The design methodology of a highly efficient concurrent dual-band Doherty PA has been further explored in [24]. The article discusses the design of the passive structures and presents different possible topologies of dual-band Doherty PA. As a validation, a concurrent dual-band PA working at frequency bands of 1,800 MHz and 2,400 MHz and PAE of 64% and 54% was designed and implemented. To broaden the bandwidth of dual-band Doherty PA, a modified dual-band parallel Doherty PA has been proposed in [25]. The design follows the use of a parallel Doherty architecture; it can provide 25–30% broadband operation around center bandwidth frequency with minimum variation of the load-network transfer characteristics. As a result, dual-band operation can be easily provided by this architecture for LTE-advanced applications. To support intraband and interband CAs in LTEadvanced, the transmitter architecture should be able to operate in both concurrent and nonconcurrent modes. In concurrent mode, and in order to satisfy the LTE standard spectrum  emission requirements, a  linearization scheme capable of compensating in-band and cross-modulation distortion is required. In nonconcur-

rent mode, the cross-modulation is eliminated due to the absence of the signal in second band, which results for only in-band distortion compensation. Generally, DPD linearization techniques are employed at the transmitter to compensate for memory effects and nonlinear behavior of the transmitter, especially the PA [23]. The linearization is also used to reduce the effect of nonlinear distortion on the overall performance. These DPD techniques are not feasible or practically not implementable in CA applications where the frequency separation between the channels of the signal is much larger than the channel bandwidths. Such signal configuration requires analog-to-digital converters (ADCs) in the feedback loop and digital-to-analog converters (DACs) in the transmitted signal path working at very high sampling rate, along with high-speed DPD signal processing unit. In most cases, the ADCs and DACs sampling rate speeds bring practical design issues and constraints, and limit the DPD techniques to signals with narrow to moderate bandwidth in the order of three to five times channel bandwidth [26]. One way to effectively decrease the minimum requirements of the ADC’s and DAC’s sampling rates is to develop frequency selective DPD architectures where only the distortion in the bands of interest is

Input Signal

~1

(a)

~2

PA Nonlinearity In-Band Intermodulation

CrossModulation

In-Band Intermodulation

Out-of-Band Intermodulation

2~1 -~2

Out-of-Band Intermodulation

~1

~2

2~2-~1

Output Signal (b)

Figure 4. Distortion due to PA third-order nonlinearity where (a) signal at the input and (b) signal at the output of the dualband nonlinear transmitter.

July/August 2013



83

One way to effectively decrease the minimum requirements of the ADCs and DACs sampling rates is to develop frequency selective DPD architectures where only the distortion in the bands of interest is compensated.

compensated. As reported in [26] and [27], these architectures could significantly lessen the requirements on the ADC’s and DAC’s sampling rate. In concurrent dual-band systems, the DPD architecture can be made even simpler since the out-ofband intermodulation is far away from the signals in designated bands of interest. They can easily be filtered out by RF filters, such as cavity filters commonly being used in duplexers. If further compensation of the out-of-band intermodulation is required, additional signal processing cells could be added, as presented in [26] and [28]. The idea of fully concentrating on distortion compensation in the bands of interests has been explored in [19], where the two-dimensional (2-D)-DPD linearization scheme is presented. The 2-D-DPD technique [19] allows the carrier components in both frequency bands to contribute to the model identification and signal predistortion process. As one of the dual-band signals is turned off (single-mode or intraband CA transmissions), the linearization process is downgraded to a one-dimensional DPD linearization technique. The 2-D-DPD linearization scheme [19] was integrated with a concurrent dual-band PA [21] for

proof of concept and functional validation of the 2-DDPD linearization scheme [20]. The results showed more than 12 dB improvements in ACPR [20]. Using a similar methodology of distortion compensation over the bands of interest, [29] presented an alternative version of the concurrent dual-band linearizer. This technique is based on a look-up table (LUT) technique, with the authors reporting more than 15 dB compensation of output band [29]. In [19] and [21], the authors explored the same linearization technique in a transmitter architecture for CA in LTE. Figure 5 depicts the complete solution that consists of baseband processing unit, digital linearization block, and dual-band active components (particularly dual-band PA). Since the signal quality at the transmitter output is the main concern, the output RF filter is not considered in the evaluation. It is assumed that RF filter provides an ideal response and any effects of the actual response of the filter should be added to the final performance evaluation. In Figure 5, the feedback loop and the DPD cell-1 and DPD cell-2 are part of 2-D-DPD linearization scheme described in details in [19]. Figure 6 shows the output power spectrum of the signal at the lower and upper bands before and after employing linearization. The 2-D-DPD linearization technique compensates for the in-band and cross-band distortion so that the out-of-band spectrum regrowth decreases by more than 15 dB. In intraband and interband CA schemes, the 2-DDPD linearization scheme is able to compensate for the out-of-band distortion of two 20 MHz LTE carrier components. The measurements and results suggest that employing a linearization scheme improves the output signal quality in the following ways:

Band A Band A CA

CC1

Dual-Band PA

DPD Cell-1

Dual-Band Antenna

Upconverter

CC2 Band B CA

CC3

Power Combiner

DPD Cell-2

Coupler

Band B Band A ADC DPD Signal Processing

Downconverter ADC Band B Feedback Receiver

Figure 5. The block diagram of the energy-efficient, dual-band transmitter for CA.

84



July/August 2013

• The linearization scheme compensates for the in-band distortion and also the nonflat frequency response of the nonlinear transmitter. It is clear from the zoom windows in Figure 6 that signals before linearization have nonflat frequency response which are compensated for after linearization. • The linearization scheme compensates for out-ofband distortion and cross-modulation products. The overall drain efficiency of the double RF chain transmitter topology in Figure 2(b) and (c) and also dual-band RF transmitter in Figure 2(d) is measured for the interband CA at the same operating conditions and total output powers. The test-modulated signals are 20 MHz LTE signals with PAPR of about 10 dB. Table 1 summarizes the measured overall efficiency of these three topologies. These measurement results show that: 1) The overall efficiency of the double branch RF transmitter [Figure 2(b)] and the dual-band RF transmitter [Figure 2(d)] is similar. This is anticipated since the two topologies can be modeled as independent systems for each operating band, assuming distortion and intermodulation are compensated. The main issue with the double branch transmitter, or in general the multiple branch RF transmitter, is the bulky size of the topology, an issue that plays a critical role in the mobile industry. On the other hand, the dual-band or multiband RF transmitters require much smaller area but at the expense of additional effort on the RF design side. 2) RF filters and antennas are bulky components. Replacing them with a multiband RF filter and

The advantages of the dual-band architecture over the double RF chain architecture are fewer RF components and the inherent power controlling mechanism to improve the energy efficiency of the system. multiband antenna as in Figure 2(c) will reduce the size of the device. Nevertheless, as summarized in Table 1, this comes at the expense of losing overall efficiency, mainly due to losses in the RF power combining circuits. The advantages of the dual-band architecture over the double RF chain architecture are fewer RF components and the inherent power controlling mechanism to improve the energy efficiency of the system, as shown in Table 1.

Table 1. Measured overall efficiency of the transmitter topologies in Figure 2(b)–(d). Total Output Power

1.21( W)

2.9 (W)

Efficiency for the double-branch 13.7 (%) RF Tx [Figure 2(b)]

24.78 (%)

Efficiency for the double-branch 8.56 (%) RF Tx [Figure 2(c)]

15.49 (%)

Efficiency for the dual-band RF Tx [Figure 2(d)]

25.99 (%)

16.12 (%)

Before Linearization After Linearization 8 0 -8 -16

15 dB

15 dB

15 dB

-32

15 dB

-24

-40 -48 840

860

880

900

920

1,940

1,960

1,980

2,000

2,020

Figure 6. Measured output spectrums before/after concurrent linearization for interband CA at lower band (880 MHz) and upper band (1978 MHz).

July/August 2013



85

Carrier aggregation successfully shows its value as an efficient way to increase the signal bandwidth within the available spectrum.

Conclusion CA has already found its place in the deployment of 4G wireless networks. It successfully shows its value as an efficient way to increase the signal bandwidth within the available spectrum. This article is focused on demonstrating and analyzing the possible transmitter architecture solutions for CA. Further, the benefits and issues of using concurrent dual-band transmitter architecture are investigated. The recent progress on deploying concurrent dualband PAs has been surveyed, where recent published literatures reported significant improvement on the concurrent dual-band PA design and achievable efficiency. The challenges present in the development of the linearization schemes for concurrent dual-band transmitters are addressed and possible solutions are studied.

References [1] S. Parkvall, A. Furuskar, and E. Dahlman, “Evolution of LTE toward IMT-advanced,” IEEE Commun. Mag., vol. 49, no. 2, pp. 84–91, Feb. 2011. [2] M. Iwamura, K. Etemad, M.-H. Fong, R. Nory, and R. Love, “Carrier aggregation framework in 3GPP LTE-advanced [WiMAX/ LTE Update],” IEEE Commun. Mag., vol. 48, no. 8, pp. 60–67, Aug. 2010. [3] G. Yuan, X. Zhang, W. Wang, and Y. Yang, “Carrier aggregation for LTE-advanced mobile communication systems,” IEEE Commun. Mag., vol. 48, no. 2, pp. 88–93, Feb. 2010. [4] T. Nakamura. LTE-advanced (3GPP release 10 and beyond).— RF aspects. presented at 3GPP 2009 Workshop for Evaluation (Dec. 17–18, 2009) [Online]. Available: http://www.3gpp.org/ ft p/workshop/2009-12-17_ITU-R_IMT-Adv_eval/docs/pdf/ REV-090006.pdf [5] P.-J. Chung. Green radio—The case for more efficient cellular base stations [Online]. Available: http://comm.ntu.edu. tw/2010UKworkshop/files/0331_Prof._Pei-Jung_Chung.pdf [6] F. H. Raab, P. Asbeck, S. Cripps, P. B. Kenington, Z. B. Popovic, N. Pothecary, J. F. Sevic, and N. O. Sokal, “Power amplifiers and transmitters for RF and microwave,” IEEE Trans. Microwave Theory Tech., vol. 50, no. 3, pp. 814–826, Mar. 2002. [7] N. Pothecary, Feedforward Linear Power Amplifiers. Norwood, MA: Artech House, 1999. [8] J. L. Dawson and T. H. Lee, Feedback Linearization of RF Power Amplifiers. Norwell, MA: Kluwer Academic, 2004. [9] Y. Jaehyok, Y. Youngoo, P. Myungkyu, K. Wonwoo, and K. Bumman, “Analog predistortion linearizer for high-power RF amplifiers,” IEEE Trans. Microwave Theory Tech., vol. 48, no. 12, pp. 2709–2713, Dec. 2000. [10] J. Kim and K. Konstantinou, “Digital predistortion of wideband signals based on power amplifier model with memory,” Electron. Lett., vol. 37, no. 23, pp. 1417–1418, Nov. 8, 2001.

86



[11] L. Ding, G. T. Zhou, D. R. Morgan, Z. Ma, J. S. Kenney, J. Kim, and C. R. Giardina, “A robust digital baseband predistorter constructed using memory polynomials,” IEEE Trans. Commun., vol. 52, no. 1, pp. 159–165, Jan. 2004. [12] F. H. Raab, B. E. Sigmon, R. G. Myers, and R. M. Jackson, “L-band transmitter using Kahn EER technique,” IEEE Trans. Microwave Theory Tech., vol. 46, pt. 2, pp. 2220–2225, Dec. 1998. [13] W. Feipeng, A. H. Yang, D. F. Kimball, L. E. Larson, and P. M. Asbeck, “Design of wide-bandwidth envelope-tracking power amplifiers for OFDM applications,” IEEE Trans. Microwave Theory Tech., vol. 53, no. 4, pp. 1244–1255, Apr. 2005. [14] B. Kim, J. Kim, D. Kim, J. Son, Y. Cho, J. Kim, and B. Park, “Push the envelope: Design concepts for envelope-tracking power amplifiers,” IEEE Microwave Mag., vol. 14, no. 3, pp. 68–81, Mar. 2013. [15] B. Razavi, “Challenges in portable RF transceiver design,” IEEE Circuits Devices Mag., vol. 12, no. 5, pp. 12–25, Sept. 1996. [16] P.-I. Mak, S.-P. U, and R. P. Martins, “Transceiver architecture selection: Review, state-of-the-art survey and case study,” IEEE Circuits Syst. Mag., vol. 7, no. 2, pp. 6–25, Second Quarter 2007. [17] T. Kitayabu and H. Ishikawa, “Generalized architecture of concurrent dual-band transmitter for spectrum aggregation system,” in Proc. IEEE 21st Int. Symp. Personal Indoor Mobile Radio Communications, Sept. 26–30, 2010, pp. 111–116. [18] M. M. Ebrahimi, M. Helaoui, and F. M. Ghannouchi, “Deltasigma-based transmitters: Advantages and disadvantages,” IEEE Microwave Mag., vol. 14, no. 1, pp. 68–78, Jan. 2013. [19] S. A. Bassam, M. Helaoui, and F. M. Ghannouchi, “2-D digital predistortion (2-D-DPD) architecture for concurrent dual-band transmitters,” IEEE Trans. Microwave Theory Tech., vol. 59, no. 10, pp. 2547–2553, Oct. 2011. [20] S. A. Bassam, W. H. Chen, M. Helaoui, F. M. Ghannouchi, and Z. H. Feng, “Linearization of concurrent dual-band power amplifier based on 2D-DPD technique,” IEEE Microwave Wireless Compon. Lett., vol. 21, no. 12, pp. 685–687, Dec. 2011. [21] W. Chen, S. A. Bassam, X. Li, Y. Liu, K. Rawat, M. Helaoui, and F. M. Ghannouchi, “Design and linearization of concurrent dualband Doherty power amplifier with frequency dependent power ranges,” IEEE Trans. Microwave Theory Tech., vol. 59, no. 10, pt. 1, pp. 2537–2546, Oct. 2011. [22] X. Li, W. Chen, Z. Zhang, Z. Feng, X. Tang, and K. Mouthaan, “A concurrent dual-band doherty power amplifier,” in Proc. AsiaPacific Microwave Conf., Yokohama, Japan, Dec. 2010. [23] F. M. Ghannouchi and O. Hammi, “Behavioral modeling and predistortion,” IEEE Microwave Mag., vol. 10, no. 7, pp. 52–64, Dec. 2009. [24] P. Saad, P. Colantonio, L. Piazzon, F. Giannini, K. Andersson, and C. Fager, “Design of a concurrent dual-band 1.8–2.4-GHz GaNHEMT Doherty power amplifier,” IEEE Trans. Microwave Theory Tech., vol. 60, no. 6, pp. 1840–1849, June 2012. [25] A. Grebennikov and J. Wong, “A dual-band parallel Doherty power amplifier for wireless applications,” IEEE Trans. Microwave Theory Tech., vol. 60, no. 10, pp. 3214–3222, Oct. 2012. [26] S. A. Bassam, M. Helaoui, and F. M. Ghannouchi, “Channel-selective multi-cell digital predistorter for multi-carrier transmitters,” IEEE Trans. Commun., vol. 60, no. 8, pp. 2344–2352, Aug. 2012. [27] P. Roblin, S. K. Myoung, D. Chaillot, Y. G. Kim, A. Fathimulla, J. Strahler, and S. Bibyk, “Frequency-selective predistortion linearization of RF power amplifiers,” IEEE Trans. Microwave Theory Tech., vol. 56, no. 1, pp. 65–76, Jan. 2008. [28] K. Jiwoo, P. Roblin, D. Chaillot, and Z. Xie, “A generalized architecture for the frequency-selective digital predistortion linearization technique,” IEEE Trans. Microwave Theory Tech., vol. 61, no. 1, pp. 596–605, Jan. 2013. [29] L. Ding, Z. Yang, and H. Gandhi, “Concurrent dual-band digital predistortion,” in 2012 IEEE MTT-S Int. Microwave Symp. Dig., June 17–22, 2012, pp. 1–3.

July/August 2013