# Class-e power amplifier with novel pre-distortion linearization technique for 4G mobile wireless communications.

I. INTRODUCTIONLong Term Evolution (LTE) protocol is a prominent solution to fulfill the continuous demand for high data rate transmission. LTE is capable in establishing a downlink peak data rate up to 326.4 Mbps and maximum data rate of 86.4 Mbps for the uplink [1]. Therefore, the demand of high output data rate results in an increased signal complexity nurturing towards the employment of multicarrier modulation standards. Owing to this signal complexity, the transmitter system, especially the power amplifier is regulated to maintain a linear operating region [2]. In fulfilling this criterion, the PA is operated at a back-off output power level from its 1 dB compression point. The operation is subjected to the degradation in the efficiency of the PA.

Several optimization methods have been reported in the effort to achieve a desired PAE for the designated PA. The most prominent is the envelope tracking method, which is reported to deliver a PAE of up to 39 %, thus complying the linearity specification for LTE signal with 10 MHz of channel bandwidth. However, in order to meet the stipulated performance criterion, a hybrid, cost ineffective dual technology has to be employed, which is a merger of CMOS and GaAs HBT [3]. An alternative approach is in realizing a RF CMOS only PA, which proves to deliver 25.8 % of PAE at a corresponding output power of 29.4 dBm [4].

In this work, a class E PA has been designed and realized in an objective to achieve a high PAE, which is measured to be 49 %. A class-E PA is categorized as a non-linear PA due to its operation at the cut-off region of the I-V curve. Hence, in order for the PA to meet the LTE linearity requirement as regulated in the 3GPP specifications [5], a novel passive linearization technique has been proposed and integrated. The linearization technique cancels out the third order intermodulation (IMD3) at high output power, thus confirming to the ACLR specifications.

This paper is organized as follows. Section II reviews the operation principle of the proposed circuit. Section III explains the theory of operation of the linearized class-E PA. In Section IV, the measured results are presented, followed by the conclusion in Section V.

II. PRINCIPLE OF OPERATION

Figure 1 illustrates the topology of the proposed PA, which integrates a Class-E PA, passive pre-distorter linearizer and an output matching network ensuring a maximum linear output power at the designated PAE. The Class-E PA encapsulates a HBT transistor and a shunt capacitor, C1. The passive pre-distorter is connected at the input of the Class-E PA, prior to the parallel RC network. The parallel RC network protects the PA from thermal runaway phenomenon [6]. The output matching network is tasked upon to transform the 50 ohm output impedance to a desired impedance point, which delivers the maximum output power. The methodology in obtaining this impedance point is explained in section III. The [G.sub.m] compensation technique [7] is adapted in the development of the biasing circuit. This technique helps to stabilize the base-emitter voltage of the biasing circuit, ensuring insensitivity towards an abrupt change of the supply voltage. The collector and base of transistor [Q.sub.b3] is shorted realizing a diode and further connected to the base of [Q.sub.b2]. Transistor, [Q.sub.b3] acts as temperature compensator alleviating significant changes in the biasing current across temperature variation.

The proposed PA is fabricated using GaAs HBT technology due to its superior electrical characteristics at high frequency operation [8]. Its inherent characteristics of low collector-emitter offset voltage and low resistance contributes to an efficient operation at low operating voltages [9]-[11].

III. THEORY OF OPERATION

A. Optimum Load Resistance

An overhead of maximum linear output power is essential in a handset design to compensate the antenna path loss. The overall maximum linear output power is determined by the load resistance of the main stage amplifier. For LTE, the desirable linear output power essential for reliable transmission by the transmitter system is 23dBm [5]. Hence, the power amplifier needs to have at least 27.5dBm of maximum linear output power overhead to compensate the path loss [12]. The optimum load resistance for a single HBT unit cell can be calculated from the following equation

[R.sub.opt] = [V.sub.dc] - [V.sub.k]/[I.sub.max], (1)

where [V.sub.dc] is the desired operating voltage, [V.sub.k] is the I-V curve knee voltage and [I.sub.max] is the maximum current as the device is biased at a class-A operating point. In order to determine the optimum load resistance for the desired maximum linear output power delivered by the PA, scaling techniques are adapted. Scaling can be realized by

[R.sub.loadopt] = [R.sub.opt]/N, (2)

where N represents the number of HBT cell. [I.sub.max] and [V.sub.k] is determined from the I-V curve of a single cell HBT transistor, as described in Fig. 2.

From (1) and (2), the optimum load resistance for the PA in this design is computed to be 6.7 [ohm]. Based on the [R.sub.loadopt] location on the Smith Chart an output matching network, as illustrated in Fig. 1 is designed and integrated to transform 50 ohm load impedance to [R.sub.loadopt]. The inherent relationship between [R.sub.loadopt] and the delivered output power is expressed in the following equation [12]

P(dBm) = 10 log [[absolute value of [(2VCC - [V.sub.sat]).sup.2]/8 x [R.sub.loadopt] x [10.sup.-3]]], (3)

where [V.sub.sat] is the saturation voltage.

B. Principle Operation of Class-E PA

In an ideal class-E PA, the transistor operates as a switch by shaping the current and voltage response not to overlap each other. This results in high efficiency, since the power dissipation has been minimized. The desirable characteristic is achieved by biasing the PA close to the cut-off region on the I-V curve. In reference to Fig. 3, the voltage and current waveform of a class-E PA when the switch is turned ON can be represented as [13]:

[v.sub.sw] = 0, (4)

[i.sub.sw]([omega]t) = [i.sub.out][sin([omega]t + [alpha]) - sin [alpha]], (5)

on the other hand, when the switch is OFF, the voltage and current is given by:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII], (6)

[i.sub.sw] = 0, (7)

where [alpha] represents the incurred phase shift.

The result of (4) to (7) is illustrated in the transient response of Fig. 4, which evidently a class-E operation.

C. Passive Pre-distortion Linearization Technique

Linearity is an essential performance parameter in power amplifier design. It defines the ability of the PA to process an input signal [14]. In this work, pre-distortion linearization technique is proposed and adapted. A pre-distorter works such that it produces anti-phase input sideband signals which tends to cancel off the unwanted sideband produced by the power amplifier. This is quantified through the amount of gain expansion and phase compression produced at the input of the PA in order to cancel out the respective gain and phase response at the desired output power.

In this work, to meet the ACLR specification at high output power, a passive pre-distorter is integrated at the input of the class-E PA to provide third order intermodulation (IMD3) cancellation. IMD3 cancellation occurs when there is a 180[degrees] phase shift [15] between the output of the pre-distorter and output of the class-E PA. This cancellation is dominant at higher output power. This phenomenon can be described from the following simplified Volterra series [16]

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII], (8)

where [a.sub.1] and [a.sub.3] represents the amplitude at fundamental frequency and IMD3 produced by the PA, respectively while [b.sub.3] is the IMD3 amplitude produced by the pre-distorter. In order to obtain an IMD3 cancellation at specific output power, the third degree terms need to have opposite signs, in the condition of [b.sub.3] > [a.sub.3]/[a.sub.1], [16].

The relationship between IMD3 and adjacent channel power ratio (ACPR) can be described as [17]-[23]

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII], (9)

where [epsilon] = mod(n/2), n represents number of tones, IMR is the multi-tone IMD to carrier ratio,

[n-1.summation over (r=1)][M.sub.1](n, r) = [n.sup.2] - [epsilon]/4 and [n-1.summation over (r=1)][N.sub.1](n, r) = 2[n.sup.3] - 3[n.sup.2]/24 + [epsilon]/8.

Figure 5 illustrates the simulated AM-AM responses of the PA prior and after linearization. The proposed novel passive pre-distorter linearizer eradicates the severe gain expansion of the class E PA and flattens it up to 28 dBm of output power.

IV. RESULTS AND DISCUSSION

The fabricated PA with a chip dimension of 1 mm x 1mm is depicted in Fig. 6. Figure 7 illustrates the simulated and measured S-parameter plot of the proposed PA. At 1.95 GHz, SI 1 and S22 are observed to be less than -10 dB. The power gain exhibited by the PA at the above mentioned frequency is 13 dB.

The measured ACLR and PAE performance at centre frequency of LTE Band 1, 1.95 GHz is depicted in Fig. 8. From Fig. 8, the third order distortion cancellation initiates at an output power of 21 dBm. Maximum cancellation is observed at 25 dBm output power. This technique helps to push the maximum linear output power to 28 dBm. The PAE measured at this power level is 49 %.

The performance of the linearized class-E PA has been summarized in Table I.

Table II summarizes the performance comparison of the proposed PA, respective to other recent reported work. It could be deduced that the proposed architecture observes an optimum PAE while satisfying the ACLR requirement in the 3GPP specification.

V. CONCLUSIONS

In this paper, a novel linearization technique has been implemented on a class-E PA. This linearization technique drives the PA to meet stringent linearity specifications for LTE compliance with 20 MHz channel bandwidth. With a PAE of 49 %, this PA serves as a good candidate in the effort increasing the battery life time of mobile phones intended for 4G wireless communications.

http://dx.doi.org/10.5755/j01.eee.20.4.3185

REFERENCES

[1] M. Rummey, LTE and the Evolution to 4G Wireless. China: Agilent Technologies, 2009, pp. 6-9.

[2] L. Larson, P. Asbeck, D. Kimball, "Next generation power amplifiers for wireless communications--Squeezing more bits out of fewer joules", IEEE RFIC Symp. Dig, 2005, pp. 417-420.

[3] D. Kang, D. Kim, J. Choi, J. Kim, Y. Cho, B. Kim, "A multimode/multiband power amplifier with a boosted supply modulator", IEEE Trans. Microwave Theory and Tech., vol. 58, pp. 2598-2608, 2010. [Online]. Available: http://dx.doi.org/10.1109/ TMTT.2010.2063851

[4] B. Francois, P. Rayneart, "A fully integrated watt-level linear 900-MHz CMOS RF power amplifier for LTE applications", IEEE Trans. Microwave Theory and Tech., vol. 60, pp. 1878-1885, 2012. [Online]. Available: http://dx.doi.org/10.1109/TMTT.2012.2189411

[5] P. N. Landin, J. Fritzin, W. V. Moer, M. Isaksson, A. Alvandpour, "Modeling and digital predistortion of class-D outphasing RF power amplifiers", IEEE Trans. Microwave Theory and Tech., vol. 60, pp. 1907-1915, 2012. [Online]. Available: http://dx.doi.org/10.1109/TMTT.2012.2187532

[6] J. Kim, P. Roblin, D. Chaillot, X. Zhijian, "A generalized architecture for the frequency selective digital predistortion linearization technique", IEEE Trans. Microwave Theory and Tech., vol. 61, pp. 596-605, 2013. [Online]. Available: http://dx.doi.org/10.1109/TMTT.2012.2229714

[7] 3GPP TS 36.101 version 9.4.0 Release 9. [Online]. Available: http://www.etsi.org/deliver/etsi_ts/136100_136199/136101/09.04.00 _60/ts_13 6101v090400p.pdf

[8] S. Maas, "Ballasting HBTs for wireless power amplifier operation", ARMMS Conf, Steventon, 2006, pp. 1-9.

[9] B. Gilbert, Analogue IC Design, the Current Mode Approach. London: IEEE Circuits and Systems Series 2, 1990, ch. 6. [10] A. A. Sweet, Designing Bipolar Transistor Radio Frequency Integrated Circuits. Norwood, NA: Artech House, 2008, pp. 43-70.

[11] T. Miura, T. Shimura, K. Mori, Y. Uneme, H. Nakano, A. Inoue, R. Hattori, N. Tanino, "High efficiency AlGaAs/GaAs power HBT's at a low supply voltage for digital cellular phones", IEEE GaAs IC Symp. Tech. Dig., pp. 91-94, 1996.

[12] Y. Tateno, H. Yamada, S. Ohara, S. Kato, H. Ohnishi, T. Fujii, J. Fukaya, "3.5V, 1W high efficiency AlGaAs/GaAs HBT's with collector launcher structure", IEEE IEDM Tech. Dig., pp. 195-198, 1994.

[13] N. Hayama, C-W. Kim, H. Takahashi, N. Goto, K. Honjo, "High efficiency, small-chip AlGaAs/GaAs power HBT's for low voltage digital cellular phones", IEEE MTT-S Int. Microwave Symp. Tech. Dig., 1997, pp. 1307-1310.

[14] K. Walsh, J. Johnson, "3G/4G multimode cellular front end challenges; Impact on power amplifier design", RFMD White Paper.

[15] A. Grebennikov, RF and Microwave Power Amplifier Design. New York: McGraw Hill, 2005, pp. 271-279.

[16] B. Zhang, Y. Z. Ziong, L. Wang, S. Hu, T. G. Lim, Y .Q. Zhuang, L. W. Li, "A D-Band power amplifier with 30 GHz bandwidth and 4.5 dBm Psat for high speed communication system", Progress In Electromagnetics Research, vol. 107, pp. 161-178, 2010. [Online]. Available: http://dx.doi.org/10.2528/PIER10060806

[17] C. I. Lee, W. C. Lin, J. M. Lin, "Low power and high linearity SiGe HBT low noise amplifier using IM3 cancellation technique", Microelectron. Eng, vol. 91, pp. 59-63, 2012. [Online]. Available: http://dx.doi.org/10.1016/j.mee.2011.10.005

[18] S. C. Cripps, Advanced Techniques in RF Power Amplifier Design, Norwood, MA: Artech House, 2002, pp. 153-195.

[19] N. B. Carvalho, J. C. Pedro, "Multi-Tone intermodulation distortion performance of 3rd order microwave circuits", IEEE MTT-S Digest, pp. 763-766, 1999.

[20] U. Karthaus, D. Sukumaran, S. Tontisirin, S. Ahles, A. Elmagrahby, L. Schmidt, H. Wagner, "Fully integrated 39 dBm 3 stage Doherty PA MMIC in a low voltage GaAs HBT technology", IEEE Microwave and Wireless Components Letters, vol. 22, pp. 94-96, 2012. [Online]. Available: http://dx.doi.org/10.1109/LMWC. 2011.2181829

[21] D. Kang, D. Kim, Y. Cho, B. Park, J. Kim, B. Kim, "Design of bandwidth enhanced Doherty power amplifiers for handset applications", IEEE Trans. Microwave Theory and Tech, vol. 59, pp. 3474-3483, 2011. [Online]. Available: http://dx.doi.org/10.1109/ TMTT.2011.2171042

[22] B. Kim, C. Kwak, J. Lee, "A dual mode power amplifier with on chip switch bias control circuits for LTE handsets", IEEE Trans. Circuit and System, vol. 58, pp. 857-861, 2011. [Online]. Available: http://dx.doi.org/10.1109/TCSII.2011.2172528

[23] J. Ding, A. Roy, N. Saxena, "Smart M2M Uplink Scheduling Algorithm over LTE", Elektronika ir Elektrotechnika (Electronics and Electrical Engineering), vol. 19, no. 10, pp. 138-144, 2009. [Online]. Available: http://dx.doi.org/10.5755/j01.eee. 19.10.5457.

U. Eswaran (1), H. Ramiah (1), J. Kanesan (1)

(1) Department of Electrical Engineering, Faculty of Engineering, University of Malaya, 50603 Kuala Lumpur, Malaysia hrkhari@um.edu.my

Manuscript received May 2, 2013; accepted December 17, 2013.

TABLE I. MEASURED PERFORMANCE SUMMARY AT 1.95 GHz. Quantity Result Technology 2 um InGaP/GaAs HBT Supply Voltage 4 V Operating Frequency 1.92 GHz-1.98 GHz LTE Channel Bandwidth 20 MHz Max Linear Output Power 28 dBm @ ACLR -30 dBc PAE 49% @ 28 dBm TABLE II. PERFORMANCE COMPARISON OF LTE PAS. Ref LTE Chan Max Linear PAE Max Pout BW (MHz) Pout (dBm) (%) (dBm) [3] 10 27.8 39 30.1 [4] 10 21.6 9 29.4 [18] 5 31.8 37 39 [19] 10 27.5 36.3 -- [20] 10 27.2 34.5 29 This 20 28 49 30.7 Work Ref Chip Area Process ([mm.sup.2]) [3] -- 2 [micro]m GaAs HBT+65 nm CMOS [4] 3 90 nm CMOS [18] 6.3 2 [micro]m GaAs HBT [19] 1.96 2 [micro]m GaAs HBT [20] 2.9 2 [micro]m GaAs HBT + 0.5 [micro]m GaAs PHEMT This 1 2 [micro]m GaAs HBT Work

Printer friendly Cite/link Email Feedback | |

Author: | Eswaran, U.; Ramiah, H.; Kanesan, J. |
---|---|

Publication: | Elektronika ir Elektrotechnika |

Geographic Code: | 1USA |

Date: | Apr 1, 2014 |

Words: | 2682 |

Previous Article: | Maximum power point tracking in solar power plants under partially shaded condition. |

Next Article: | Optimization of the fast image binarization method based on the Monte Carlo approach. |

Topics: |