here. Requires Adobe Acrobat Reader,
It is common knowledge that efficiency is a major issue with the design of any transmitter. The benefits of high efficiency include power consumption, low power derating and high reliability, among others. In order to improve efficiency several specific amplification classes, such as class E^{1} have been described in classical literature. Nevertheless, the strict conditions required by high efficiency amplification techniques are difficult to achieved at high frequencies.
Devices intrinsic capacitances, packaging parasitics, low gain and intrinsic losses are the main limiting factors of high efficiency modes. The constraints become even more severe if high output power levels are involved. This article will discuss the design of a high efficiency VHF power amplifier (collector efficiency 0_{η} = 85%, power added efficiency (PAE) = 72 %, output power (P_{OUT}) = 20 W) based on the proper control of the load and source impedance at fundamental and harmonic frequencies.
Recent relevant works^{2} have shown the influence of proper load termination to achieve good approximations to high efficiency classes. Some practical works have shown that these approaches are feasible in practice^{3}. It also has been proven that if the strict conditions of “pure” high efficiency classes cannot be fulfilled, good efficiency results can be achieved using approximated but proper design.
High efficiency approximated approach
Class E^{1} and other “classical” high efficiency modes are usually based on heavily overdriven and saturated devices working in cutoff and forward saturation regions. Inverse active and saturation regions have not been realized with high efficiency modes except for nonidealized operation^{4,5}. Nevertheless, the most important rule to accomplish high efficiency operation lies on non coincidence of current and voltage at the same time over the active devices output. Any amplification class (pure or not) fulfilling this rule leads to high efficiency, one way or another.
Recent works^{3,6,7} have shown through loadpull measurements and simulations that high efficiency operation is possible at high frequencies using heavily overdriven devices working in inverse active and saturated regions during part of their duty cycle. The amplifier system discussed in this article forces its active device to work in reverse regions to achieve high efficiency at high frequencies.
Operating principles
The main cause precluding class E operation at high frequencies is the large intrinsic capacitances of most power transistors. The reactance of intrinsic transistor capacitances is often very low. So, to obtain optimum class E operation at high frequencies lowload impedances are needed. Unfortunately, most times low loads lead to higher output power levels and currents than specified for safe transistor operation.
Maintaining high efficiency and safe operation forces the use of load impedances different from loads for optimum class E operation. High efficiency can be retained with different loads but they may cause negative collectortoemitter voltage V_{CE} < 0 during part of the signal period. Negative V_{CE} voltage does not necessarily mean breakdown of the device or collector efficiency reduction. Using a proper base circuit and high reactive source impedance at the harmonics, the transistor can be forced to operate in inverse active and saturation regions. During this period, V_{CE} is low (zero for an ideal device) and the power derated by the transistor is also low, even for a significant collector current level. Nevertheless, because of the existence of inverse collector current the overall conduction angle of the device is effectively reduced.
Therefore, the output power P_{OUT} is also reduced compared to the output power in optimum class E conditions, PE^{1}. This power reduction can be considered a benefit in some extreme situations.
These effects are illustrated Figure 1. This figure is a simulation of the collector current I_{C} and V_{CE} for the referred load conditions (note that the intrinsic capacitances of the transistor have been extracted outside the device during the simulation in order to clarify the discussed effect).
Amplifier design
The amplifier presented to illustrate this technique is based on the well known “classical” 2N6083 RF bipolar power transistor. The transistor manufacturer announces collector efficiency 0_{η} = 65% and output power P_{OUT} = 30 W at 175 MHz for this device. The amplifier achieves 0_{C} = 85% and P_{OUT} = 20 W at 90 MHz using the same device.
In order to design the output load at the fundamental frequency, a linear equivalent value of the output capacitance C_{OUT} = 105 pF was estimated from the measured openbase collector capacitance C_{OB}. This capacitance exhibits a reactance X_{COUT} = 15.36j at 90 MHz. If this capacitance is the only reactive component used in the load network it forces an output power level P_{E} = 31.2 W at V_{CC} = 12.5 VDC for optimum class E operation^{1}. The desired output power for this amplifier is P_{OUT} = 20 W. The curves published in^{3} have been used to design the amplifier load network.
The relationship P_{OUT}/P_{E} = 0.64, (from^{3}) the optimum load for this amplifier, is derived from Z_{L} = 4.3 + 4.6j. Assuming a parasitic inductance from the package and printed circuit board L_{P} = 1.5 nH (0.84j @90 MHz), the effective load for the transistor is established as Z"_{L} = 4.3 + 3.7j. On the other hand, the source impedance at the fundamental frequency is optimized for maximum power gain.
The load and source impedances at the harmonics are set strongly reactive (at least 3 times the load at the fundamental). Discrete threeelement matching networks are used to provide both the source an load impedances at the fundamental frequency while keeping strong inductive behavior at harmonic frequencies.
Amplifier testing and optimization
The amplifier was tested and optimized using the custom testfixture shown in the photo on page 46. It not only provides a fixture for the transistor but also allows the variation of the load and source impedance while the load at the harmonics is kept strongly reactive.
The testfixture also provides ports for sampling of significant amplifier voltages and currents. This samples are measured in a broadband digital oscilloscope.
Results and measurements
Figure 2 shows collector efficiency, 0_{η}, and power added efficiency, PAE, vs. input power P_{IN}. It should be noted that an optimum input power level exists (P_{IN} = 2.75 W for this amplifier) for maximum PAE. Beyond this input drive with very low 0_{η} improvement is achieved but both, power gain and PAE. degrade very fast.
The reason for this behavior is that fast degradation of gain occurs with high driving levels. Figure 3 shows this effect, output power (P_{OUT}) and power gain (G) are plotted vs. input power. The degradation of gain with only small improvement in output power P_{OUT} is evident. Figure 4 shows the collector to emitter voltage V_{CE} and emitter current I_{E} measured at the active device working under high efficiency conditions. Note that the effect of non negligible base current appears in this measurement as well.
Conclusions
At radio frequencies classical high efficiency amplification classes cannot be easily achieved. However, careful approximations can achieve good results with low efficiency degradation. A high efficiency design using forward and inverse active and saturation regions has been presented. An example has been designed constructed and the results measured. The device was loaded with high reactive source and load impedances at harmonic frequencies. The load at the fundamental frequency was chosen to function in suboptimum class E with negative collectortoemitter voltage during part of the duty cycle. This negative V_{CE} allows inverse active conduction and saturation of the transistor.
This operational mode has shown very high collector efficiency and good PAE. with remarkable output power level at high frequencies.
References:

Nathan O. Sokal, Alan D. Sokal, “Class EA New Class of high efficiency Tuned SingleEnded Switching Power Amplifiers", IEEE Journal of SolidState Circuits, VOL. SC10, NO. 3, pp. 168176, june 1975.

Frederick H. Raab, “ClassE, classC, and classF power amplifiers based upon a finite number of harmonics", IEEE Transactions on Microwave Theory and Techniques, Volume: 49, NO. 8, pp. 14621468, Aug 2001.

Francisco Javier OrtegaGonzález, Alberto AsensioLópez, José Luis Jiménez Martin, Germán TorregrosaPenalva, “High Efficiency LoadPull Harmonic Controlled Class E Power Amplifier", IEEE Microwave and Guided Wave Letters, VOL 8, No 10, pp. 348350, october 1998.

Frederick H. Raab, “Effects of Circuit variations on the Class E Tuned Power Amplifier", IEEE Journal of SolidState Circuits, VOL. SC13, NO.2, pp. 239247, april 1978.

Willian H. Cantrell, “Tuning Analysis for the HighQ ClassE Power Amplifier", IEEE Transactions on Microwave Theory and Techniques, VOL. 48, NO. 12, pp. 23972402, december 2000.

Francisco Javier OrtegaGonzález, Alberto AsensioLópez, José Luis Jiménez Martin, Effects of Matching on RF Power Amplifier Efficiency and Output Power", Microwave Journal, pp. 6072, april 1998.

Marian K. Kazimierczuk, Wojiech A. Tabisz, “Class CE High Efficiency Tuned Power Amplifier,” IEEE Transactions on Circuit and Systems, VOL. 36, NO. 3, pp. 421428, march 1989.
About the authors
Francisco Javier Ortega González received the Ingeniero de Telecomunicación degree from Universidad Politécnica de Madrid, Madrid, Spain, and the Ph.D. degree at the Dpto. de Señales, Sistemas y Radiocomunicaciones, E.T.S.I. Telecomunicación, Universidad Politécnica de Madrid. His research activities are in the area of high frequency circuit design, electromagnetism and radar systems. Email; fjortega@diac.upm.es
Alberto Asensio received the Ingeniero de Telecomunicación degree from the Technical University of Madrid, Madrid, Spain in 1984, and the Ph.D. degree at the Signals, Systems, and Radiocommunications Department of the Technical School of Telecommunication Engineering of the same University in 1990, where he has been Associate Professor since 1991. His research activities are in the area of high frequency circuit design and radar systems. Email: vera@gmr.ssr.upm.es
Alberto Martin Consuegra received the Ingenierio Técnico de Telecomunicación degree from de Universidad Politecnica de Madrid, Madrid, Spain. Currently is working toward its Ingeniero Electrónico degree at the Alcala de Henares University, Madrid. His research activities are in the area of electronic and high frequency circuit design.
Juan Angel Ruiz was born in Madrid, Spain. Currently he is working toward its Ingeniero Técnico de Telecomunicación degree at the Universidad Politecnica de Madrid, Madrid. His research activities are in the area of high frequency circuit design and wireless networks.
Germán Torregrosa Penalva received the Ingeniero de Telecomunicación degree from Universidad Politécnica de Madrid, Madrid, Spain, in 1999, and is currently working toward the Ph.D. degree at the Dpto. de Señales, Sistemas y Radiocomunicaciones, E.T.S.I. Telecomunicación, Universidad Politécnica deMadrid. His research activities are in the area of high frequency circuitdesign and radar systems.
Acknowledgements:
This work was supported by project TIC 20013839C0301 of the Spanish National Board of Scientific and Technology Research (MCYT).