Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

Class-J RF Power Amplifier With Wideband Harmonic Suppression

Download as pdf or txt
Download as pdf or txt
You are on page 1of 5

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/224251391

Class-J RF power amplifier with wideband harmonic suppression

Conference Paper  in  IEEE MTT-S International Microwave Symposium digest. IEEE MTT-S International Microwave Symposium · July 2011
DOI: 10.1109/MWSYM.2011.5972873 · Source: IEEE Xplore

CITATIONS READS
27 989

3 authors:

Neal Tuffy Anding Zhu


MACOM Technology Solutions University College Dublin
5 PUBLICATIONS   203 CITATIONS    148 PUBLICATIONS   2,703 CITATIONS   

SEE PROFILE SEE PROFILE

T.J. Brazil
University College Dublin
175 PUBLICATIONS   2,471 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Modelling of GaN Transistors for Power Amplifier Design View project

Trapatt diodes, characterisation, oscillators and reflection amplifiers View project

All content following this page was uploaded by Neal Tuffy on 21 May 2014.

The user has requested enhancement of the downloaded file.


Class-J RF Power Amplifier with Wideband Harmonic Suppression
Neal Tuffy, Anding Zhu, and Thomas J. Brazil
RF and Microwave Research Group
University College Dublin, Ireland

Abstract — This work outlines a design approach used for reduction of the PA bandwidth then ensues, which suggests a
achieving highly efficient power amplification over a wide practical limit for the available bandwidth given a desired
bandwidth, given narrowband passive harmonic load-pull results
at a single center frequency. Specific considerations have been level of load harmonic band suppression.
given to harmonic suppression concerns which are addressed by This work demonstrates an approach used for obtaining
appropriate choice of matching network topology. The method is moderately wide PA bandwidths with /4 narrowband
justified by amplifier fabrication using a commercially available harmonic load-pull tuner results, and discusses the topological
10W GaN HEMT transistor and achieving an efficiency of at considerations employed for obtaining adequate harmonic
least 60% over a 15% bandwidth from 2.3-2.7GHz. Output nd rd
power between 10-12W was delivered across the band, centered suppression over the entire 2 and 3 harmonic bands. Device
at 2.5GHz. Harmonic suppression of at least -41dBc and -39dBc output package and intrinsic de-embedding has been applied
was obtained across the second and third harmonic bands to obtain a representation of the impedances presented to the
respectively, largely removing the dependency on post-filtering. internal current generator. The design procedure first identifies
Index Terms — Class-J, harmonic suppression, high efficiency, the efficiency contours over the normalized fundamental and
power amplifier. nd
2 harmonic reactive plane at the intrinsic device. This
provides a design space for determining a suitable
I. INTRODUCTION fundamental frequency termination given the harmonic
terminations seen at the intrinsic drain. Independent control of
The ever increasing demand for greater data rates coupled the harmonic and fundamental matching networks is essential
with stringent spectral mask requirements and the need for for realizing a solution to the design specifications and
low power consumption provides severe challenges in the providing a systematic design approach.
field of PA design. Switch-mode amplifiers have been shown By employing the design space technique combined with
to present high efficiency, although usually at the expense of the narrowband harmonic load-pull measurements, an
linearity and/or bandwidth. Recently, the Class-J amplifier has amplifier was fabricated using a 10W Cree GaN HEMT
been proposed [1], which provides a solution for the adverse device. Results obtained demonstrate the feasibility of the
effects of switch-mode amplification, therefore potentially approach by delivering at least 10W of output power with
exhibiting high efficiency, linearity and wideband behavior greater than 60% drain efficiency across 2.3-2.7GHz. Across
simultaneously. the bandwidth, measurements revealed harmonic suppression
Class-J theory augments the traditional narrowband nd rd
in excess of -39dBc over the 2 and 3 harmonic bands at the
harmonic shorting approach associated with conventional load. Finally, measurements are shown which display the
linear amplifier modes, by allowing a reactive fundamental potential of the amplifier for incorporation within high
component and reactive terminations at the harmonics. This efficiency drain modulated transceiver architectures.
permits a wide range of solutions for a given set of harmonic
terminations which can be utilized to obtain high efficiency
across the band of interest [2]. Wright et al. [3] demonstrated II. CLASS-J AMPLIFIER THEORY
at least 60% efficiency over a 50% bandwidth with Wu et al. Class-J amplification is characterized by a complex
[4] obtaining greater than 60% efficiency over a 42% fundamental impedance combined with reactive harmonic
bandwidth. terminations at the intrinsic drain of the device. It has been
The nature of wideband amplifiers implies a difficulty in shown [1] that such a configuration can deliver power and
suppression of the harmonics associated with the in-band efficiency equivalent to the Class-B case. The consequence of
frequencies, as a near overlap of the bands can occur. These adding a reactive fundamental component and reactive
harmonic bands could be suppressed by a high order filter at harmonic terminations manifests itself as an increase in the
the output but this will be accompanied by a reduction in the magnitude of the internal drain voltage waveform. Waveform
output fundamental power and therefore a substantial decrease engineering solutions also exist, which can exploit the non-
in the effective efficiency of the amplifier. Alternatively, linearities within the device to optimize performance [5]. By
harmonic filtering can be integrated into the output matching nd rd
only considering the 2 and 3 harmonics, the normalized
network but this typically results in raising the network Q- intrinsic drain voltage waveform can be expressed as follows:
factor for appropriate harmonic rejection at the load. A

Copyright © 2011 IEEE. Reprinted from 2011 IEEE MTT-S International Microwave Symposium, Baltimore, MD, USA.
978-1-61284-757-3/11/$26.00 C2011 IEEE
This material is posted here with permission of the IEEE. Such permission of the IEEE does not in any way imply IEEE endorsement of any of
Cree’s products or services. Internal or personal use of this material is permitted. However, permission to reprint/republish this material for advertis-
ing or promotional purposes or for creating new collective works for resale or redistribution must be obtained from the IEEE by writing to
pubs-permissions@ieee.org By choosing to view this document, you agree to all provisions of the copyright laws protecting it.
Fig.1. Efficiency of the different amplifier modes against normalized Fig.3.Efficiency contours plotted over the normalized reactance plane
nd
2 harmonic reactance

III. CLASS-J AMPLIFIER DESIGN METHODOLOGY


nd
Considering only the 2 harmonic, the efficiency contours
can map out a design space in the X1, X2 reactance plane as
shown in Fig.3. The key feature of the approach taken in this
work is the isolation of the harmonic and fundamental
matching networks for reducing the complexity of the design
Fig.2. Proposed amplifier initial design consideration procedure. By presenting high impedance to harmonics at
plane C (Fig.2) the harmonic terminations are completely
determined by the device and the harmonic matching network.
��� � � � � ��� � � � ��� � � � ��� �� � ������ (1) This ability for independent harmonic and fundamental
network design allows for superior control in achieving high
Conforming to the criteria outlined by Rhodes [6], a family of efficiency, wideband amplification with full suppression of the
solutions can be exploited by updating the fundamental harmonic bands.
reactive component for a set of harmonic terminations Narrowband passive harmonic load-pull was performed on
dynamically changing across frequency. Fig.1 shows the the device to determine the optimum impedances at the center
nd
efficiency degradation with increasing 2 harmonic reactive frequency and the harmonics. The theory presented was
termination for different modes of operation. The plot utilizes utilized by initially designing a harmonic matching network
rd
optimum fundamental and 3 harmonic reactive terminations nd
which restrained the 2 harmonic reactances to within the high
for maximizing efficiency while ensuring no harmonic power efficiency contour regions. The fundamental network was then
consumption. designed by trade-off between preserving the high efficiency
In practice, difficulty arises in designing a matching trajectory, maximizing harmonic load rejection and
rd
network to avail of the improved performance with 3 maintaining at least 10W of output power.
harmonic enhancement across a wide bandwidth. Provided the
nd
normalized 2 harmonic reactance does not exceed a value of A. Harmonic Matching Network Design
approximately 1.5 (c = 0.5), maximum efficiency can still be The harmonic load network is composed of the device
maintained and thus there would be no gain in control of the intrinsic model and package parasitics terminated in a
rd
3 harmonic. It is therefore desirable to provide minimal harmonic matching network. The design criteria for the
rd
impedance across the 3 harmonic band otherwise it may harmonic matching network are outlined below:
further increase the amplitude of the drain voltage and drive 1) From (1), |�| � ������ ����� � ���� to remain in the
the device into breakdown. The effect of presenting a low high efficiency region.
rd rd
impedance 3 harmonic termination will introduce a 3 2) The fundamental impedance transformation must act to
harmonic current component and distort the ideal half- have minimal effect on the obtainable bandwidth.
sinusoidal current waveform. This was considered in final 3) Harmonic terminations at plane A must be highly reactive
efficiency calculations for strong correlation between to ensure no power generated at the harmonics.
measurement and simulated results.

978-1-61284-757-3/11/$26.00 C2011 IEEE


DC Bias DC Bias

C C
R
C

C C
RL
C C
VS Harmonic Fundamental
Matching Matching
Network Network

Fig.4. Complete amplifier schematic

4) |�| must be small enough to have a negligible effect on


efficiency with little variation over frequency, so as to be
effectively neglected at plane A.
Passive harmonic load-pull was initially implemented to
find the optimum impedances at Plane B which can then be
de-embedded to Plane A. The appropriate topology which
satisfied the above criteria is shown in Fig.4. Lumped surface
mount components were used for their superior bandwidth
performance over stubs and the application of an inductor as a
DC feed provides a preferable alternative to a narrowband
quarter-wave transmission line.
B. Fundamental Matching Network Design
Fig.5. Output matching network trajectory on the reactance plane
The function of the fundamental matching network was to
provide the appropriate fundamental impedance at plane A
over the band while presenting large harmonic band C. Input Matching Network Design
impedances. The fundamental matching network was designed For obtaining the optimum source impedance, source-pull was
by adhering to the following criteria: performed at the center frequency of 2.5GHz. The initial
1) The fundamental complex impedance presented at plane approach was to match at the centre frequency and provide
A should not allow the efficiency contour trajectory to go minimal variation in impedance around the band. This
below 60% over the band of interest. procedure was justified by performing fundamental source-
2) Fundamental power delivered to the load should be at pull over the band and observing small variations in the
least 40dBm over the desired band. optimum source impedance. Harmonic source-pull was not
nd
3) High impedance should be presented at plane C to the 2 available so the input match was designed to provide
rd
and 3 harmonics over the entire harmonic bands to allow maximum gain over the band while filtering the harmonic
independent harmonic control. bands. A capacitor was included in the network for flexibility
nd rd
4) Greater than -35dBc of 2 and 3 harmonic suppression in tuning. The complete amplifier schematic is shown in Fig 4.
should be obtained at the load.
For the design of the fundamental matching network 3D
electromagnetic simulations were utilized for accurate
prediction of the harmonic impedances. Criteria 3) and 4) can
be treated simultaneously, since meeting 3) with sufficiently
high harmonic impedance terminations should imply 4) is
satisfied. A trade-off then becomes evident as meeting criteria
3) and 4) tends to reduce the bandwidth of the fundamental
match hence compromising criterion 2). Upon satisfying
criteria 3) and 4), the trajectory can be plotted on the reactance
plane by de-embedding the impedances to plane A to test for
1) and 2). Results of the final complete output matching
network trajectory are shown in Fig.5 Fig.6. Picture of final Class-J PA

978-1-61284-757-3/11/$26.00 C2011 IEEE


Fig.7. F0, 2F0 and 3F0 Output Power and Gain across the band Fig.9. Drain efficiency and PAE obtained from 2.3-2.7GHz

IV. MEASUREMENT RESULTS providing peak efficiency of 74% with >70% efficiency over a
The final fabricated amplifier is shown in Fig.6. Continuous 9dB output power range at 2.65GHz.
Wave (CW) excitation was employed for testing from 2.3-
2.7GHz, corresponding to a 15% bandwidth. Drain efficiency, V. Conclusion
PAE, output power and harmonic power were calculated at A wideband Class-J amplifier was designed and fabricated
25MHz intervals across the band. Results are demonstrated in based on single frequency harmonic load-pull results.
Fig.7 and Fig.9. Harmonic suppression concerns were also addressed and a
Output power at the fundamental varies from 10 to 12W. systematic design methodology has been constructed for
Greater than 60% efficiency was obtained over the band with meeting the demanding design goals. Results demonstrate at
an average drain efficiency just over 62%. Over the second least 60% efficiency over a 15% bandwidth with in excess of -
harmonic band of 4.6-5.4GHz at least -44dBc of harmonic 39dBc of 3rd harmonic and -44dBc of 2nd harmonic band
rd
suppression was obtained with at least -39dBc over the 3 rejection at the load. Potential viability for integration within
harmonic band of 6.9-8.1GHz. Over the initial 10% design high efficiency EER/ET systems was demonstrated with
bandwidth at least –47dBc and -41dBc of harmonic rejection >70% efficiency over a 9dB output power range.
nd rd
was found over the 2 and 3 harmonic bands respectively.
The simulated efficiency curve in Fig.9 shows strong
agreement with the measured drain efficiency due to the ACKNOWLEDGEMENT
inclusion of a 3rd harmonic current component. For drain The authors would like to thank Cree Inc. for supplying the
modulation considerations, a single tone power sweep was device and device models. Also, the authors would like to
performed across a range of drain bias voltages. Fig.8 acknowledge the financial support of Science Foundation
illustrates the potential for dynamic drain modulation by Ireland.
REFERENCES
[1] S.C. Cripps, RF Power Amplifiers for Wireless
nd
Communications, 2 ed, Norwood, MA: Artech House, 2006.
[2] S.C. Cripps et al. “On the continuity of high efficiency modes in
linear RF power amplifiers,” IEEE Microw Wireless Compon.
Lett., vol. 19,no. 10, pp. 665-667, OCT 2009.
[3] P. Wright et al., “A methodology for realizing high efficiency
Class-J in a linear broadband PA,” IEEE Trans. Microw Theory
Tech. vol. 57,no. 12, pp. 3196-3204, Dec 2009.
[4] D.Y.-T. Wu et al., “Design of a broadband and highly efficient
45W GaN power amplifier via simplified real frequency
technique,” 2010 IEEE MTT-S Int. Microwave Symp. Dig., pp.
1090-1093, June 2010
[5] J. Moon et al., “Investigation of a class-J power amplifier with a
nonlinear Cds for optimized operation,” IEEE Trans. Microw
Theory Tech. vol. 57,no. 12, pp. 3196-3204, Dec 2009.
Fig.8. Power sweep for various drain voltages at 2.65GHz [6] J.D. Rhodes “Output universality in maximum efficiency power
amplifiers,” Int. J. Theor. Appl,, vol. 31, pp. 385-405, 2003.

978-1-61284-757-3/11/$26.00 C2011 IEEE

View publication stats

You might also like