Académique Documents
Professionnel Documents
Culture Documents
inear power amplifiers, even those at moderate power levels, are the primary element in determining the efficiency of transmitter systems. In portable communications equipment, the ability of the transmitter power stages to provide a linear signal to the antenna (while at the same time conserving power consumption) can be a particularly challenging design problem. This article presets a technique that shows how the characteristics of wide bandwidth, high linearity, and power efficiency are readily attained.
Basic premise
Bipolar transistors have two basic sources of nonlinearity: the baseemitter junction proper and the nonlinear bulk resistance of the collector region [1]. Overcoming these sources of harmonic and intermodulation (IM) distortion pose serious problems for designers of both small- and large-signal amplifiers. In the design of small-signal amplifiers, the designer has several design topologies at his disposal for linearizing amplifier stages, including dual-loop feedback [2], series-shunt feedback [3], and lossless feedback [4]. The power amplifiers, however, are a different matter. In the design of power amplifiers, the usage of feedback topologies is limited by virtue of a number of obstacles: one of which is the phase relationship beRE Q1
Common-base amplifiers
Common-base amplifiers offer some interesting properties that deserve consideration. Figure 1 illustrates a basic common-base amplifier., where the output voltage VO is determined by: Vo = Ic RL (1) where IC is the collector current and is a result of the input emitter current IE: IE hfe hfe + 1 (2) the emitter current being the result of the input voltage VS and the input resistance Ri: Ic = Ri = R E + re + rbb hfe + 1
Q1
(3)
RE
RE VS RL VS 1 N T1 RL VS M 1 T1 N RL
30
www.rfdesign.com
October 1999
-20
-15
-10
-5
10
15
20
25
12 11 10 9 8 7 6 5 4 3 2 1 0
Gain (dB)
vbe Av + 1 re = qVbe I0 T (7) that obviously shows that as the voltage gain of the amplifier is increased, the apparent emitter resistance decreases, and this process will be referred to as augmentation [6]. Both active and passive implementations of common-base amplifier augmentation are possible, but the discussion here will focus on two passive realizations because they are the most practical application of the method with regard to power amplifiers.
ve N N+1 1 ie hfe (11) that results in a substantial reduction in the input emitter resistance, as much as 95% for a transformer turns ratio of 1:3.
Gain (dB)
-20
-15
-10
-5
10
15
20
-25
-20
-15
-10
-5
10
15
20
Gain (dB)
ve " = ve (M + 1)
ve " ie "
-20
-15
-10
-5
10
15
20
25
12 11 10 9 8 7 6 5 4 3 2 1 0
Gain (dB)
(14) which is to be expected, since the primary winding of the augmentation transformer now functions as an autotransformer.
power-efficient. Clearly, an alternative method for reducing the nonlinear emitter resistance would be desirable.
Experimental results
To illustrate the effectiveness of this linearization method, three amplifiers were constructed and evaluated, using a common 2N2222 as the active device. A 2:1 wideband transformer was used in the collector to give the amplifier an initial theoretical power gain of 6.0 dB. A 47- resistor was used for RE to provide a good input impedance match. Bias conditions were set at a VCE of 10 V and IC of 10 mA, for a collector power dissipation of 100 mW. A signal fre-
Introduction to augmentation
One method for reducing this resistance is to detect the emitter signal voltage, amplify and invert it, and then apply it to the transistor base. In doing so, the base-emitter signal voltage vbe remains essentially the same, but the emitter signal voltage v e is now reduced, resulting in the apparent emitter resistance re:
32
www.rfdesign.com
October 1999
Conclusions
The application of augmentation has been shown to be an effective means of improving the linearity and power efficiency of common-base transistor amplifiers, and this technique has been applied successfully in the design of HF and VHF power amplifiers. Collector efficiencies exceeding 90% and IP3 improvements of more than 15 dB are readily achieved without the need for additional active components or the use of costly topologies such as feedforward. References 1. Trask, C., Nonlinear Distortion in Bipolar Transistors, QEX, Nov/Dec 1998, pp. 37-42. 2. Aprille, T.J., Wide-Band Matched Amplifier Design Using Dual Loop Feedback and Two Common Emitter Transistor Stages, IEEE Transactions on Circuits and Systems, vol. CAS-23, No. 7, Jul 1976, pp. 434-442. 3. Seader, L.D. and J.E. Stertett, Unit Transistor Amplifier with Matched Input and Output Impedances, US Patent 3,493,882, 3 Feb 1970. 4. Norton, D.E. and A.F. Podell, Transistor Amplifier with Impedance Matching Transformer, US Patent 3,891.934, 24 June 1975. 5. Shea, Richard F., Principles of Transistor Circuits, Wiley, 1953, pp. 136-141. 6. Patent pending. 7. Trask, C., Distortion Improvement of Lossless Feedback Amplifiers Using Augmentation, 1999 Midwest Symposium on Circuits and Systems, Las Cruces, NM, 11 Aug 1999.
quency of 10 MHz was used for gain and compression tests, while signals of 9.9 MHz and 10.1 MHz were used for intermodulation tests. Initial tests were performed on an uncompensated common-base amplifier (Figure 1) to establish a baseline for performance. The results, shown graphically in Figure 4, show that this amplifier has a gain of 4.75 dB, a 1 dB c o m p r e s s i o n p o i n t ( P 1dB) o f -2.75 dBm, and an approximate thirdorder intercept point (IP 3 ) of +12.5 dBm. Saturable power (P SAT) is just barely +3 dBm. Next, a common-base amplifier with 1:2 simple augmentation (Figure 2) was evaluated, using the same bias conditions as before. Referring now to Figure 5, we see that the performance has improved dramatically: The P1dB compression point has risen to +12.25 dBm, IP3 is +29.5 dBm, and PSAT has now risen to over +17 dBm. This latter figure indicates that the process of augmentation has rendered an amplifier with higher than 95% peak power efficiency. Power gain has risen to 5.4 dB, appreciably closer to the theoretical 6 dB design gain. Notice also that the gain remains fairly constant, having a much sharper compression characteristic than the earlier uncompensated amplifier. Finally, a common-base amplifier with 1:1:2 compound augmentation (Figure 3) was evaluated, again using the same bias conditions as with the two prior amplifiers. The additional current gain provided by the autotransformer effect of the augmentation transformer primary gives the amplifier a theoretical power gain of +12 dB. Referring now to Figure 6, we see that the performance of the compound augmented amplifier is similar to that of the previous simple augmented amplifier: The P1dB compression point is now +11.5 dBm, IP3 is +28 dBm, and PSAT is still over +17 dBm, indicating that the process of compound augmentation has also rendered an amplifier with higher
than 95% peak power efficiency. The measured power gain of 11.3 dB is again comparable to the design gain for this amplifier and also displays the fairly constant gain characteristics of the previous common-base amplifier with simple augmentation. Table 1 gives a tabulated comparison of the performance parameters mentioned here, while Figure 7 graphically compares the gain characteristics of the three example circuits, which helps to illustrate the dramatic improvement in the gain between the uncompensated common-base amplifier and the common-base amplifier with simple augmentation.
34
www.rfdesign.com
October 1999