Download

PDF Abstract Text:

SEMICONDUCTOR APPLICATION NOTE

Order this document by AN1526 / D

MOTOROLA
SEMICONDUCTOR APPLICATION NOTE
Order this document by AN1526 / D
RF Power Device Impedances: Practical Considerations
Prepared by: Alan Wood and Bob Davidson Motorola Semiconductor Products Sector
AN1526
ABSTRACT
INTRODUCTION
Many first time RF power designers, brought up on a diet of small-signal s-parameters, previously used for solving small signal text book problems, assume these same techniques are applicable to bipolar class-C and class-AB power amplifier design. They consider the best match is achieved by a simultaneous conjugate match of the input and output. However, power amplifiers provide higher power gain and better efficiency at the rated output power if the output is purposely mismatched. An added benefit of doing this is potentially unstable devices, conjugately matched, can be operated stably under these more optimum mismatched conditions. More knowledgeable designers, familiar with large-signal impedances, naively assume the published impedances are independent of operating point. They forever wonder why, although they have designed their impedance transformation networks to match the device "data book impedances, " they have to "tweak" the circuit for optimum performance. This is the basis for much of the black magic that surrounds RF power amplifier design, but the reality is the circuit designer is plagued with a paucity of good design data, and a lack of adequate tools to make the initial design "foolproof." This paper intends to enlighten these engineers to the true meaning of large-signal series equivalent device impedances. We will also show that the output impedance is, for the most part, under the control of the circuit designer, and the input device impedance can be expected to change depending upon the designers choice of output matching (or, in some cases, intentional mismatching).
DEFINITIONS
AN1526
RF Application Notes
AN1526
devices is a much stronger function of load impedance than shown for this small device. Device impedances published by vendors of RF power transistors should only be used as an approximation for a first cut circuit design. In a broadband amplifier design it is often difficult to obtain a good match over the full frequency range and in certain circumstances the input or output is deliberately mismatched to compensate for the gain increase at lower frequencies to provide a level gain response. Good design would opt for a load-line where the lower gain corresponds to a higher efficiency operating point. the output reactance of the transistor. The required peak output power and the collector bias voltage determine the operating load line. The output reactance of the device under these conditions is conjugately matched to achieve maximum power transfer, although this condition may be modified, at the expense of gain, to attain higher efficiency. The load line resistance is given approximately by:
MRF873 DEVICE IMPEDANCE COMPARISON FOR DIFFERENT MODES OF OPERATION
where VCC is the collector supply voltage, Pout is the required peak power, and VCE(sat)RF is the collector-emitter saturation voltage under the frequency of operation. The value of this parameter is particularly difficult to measure, but the normal range is 1.0 to 2.5 Volts depending on the geometry, epitaxial doping and thickness. A good approximate value for 12.5 Volt devices is 1.5 Volts and for 24 Volt transistors is 2 Volts. The load line resistance is the optimum load impedance for the internal collector node of the transistor, neglecting the junction and parasitic device capacitance. These are in parallel with the load line resistance. For transistors, operating at VHF, and above the internal collector lead inductance of the package becomes significant, and is in series with the previously defined parallel RL, Cobo network. For the CS-12 package the internal collector lead series inductance can be represented by a 0.65 nH lumped inductor. Some devices have internal collector matching, transforming the internal load line impedance to higher value for ease of broadband matching. Comparison of the impedance data taken by small-signal methods, assuming a simultaneous conjugate match, and large-signal measurements, shows dramatic shifts in input impedance (see Table 1). More subtle, but measurable differences, can be seen in the change in input impedance between class-C and class-AB data. The small-signal s-parameter data is given in Table 2 below for a collector bias current of both 50 mA and 2 A.
Table 1. Comparison of Input Impedance for Different Operating Modes
Class-C 1.10 + j3.26 1.19 + j3.24 1.24 + j3.34
Frequency (MHz) 806 S11 IC 50 mA 2A 838 50 mA 2A 870 50 mA 2A S11 0.963 0.877 0.961 0.858 0.958 0.861 172.8 170.9 172.4 172.7 172.3 175.3 S12 0.006 0.024 0.005 0.022 0.004 0.018 S12 7.72 27.9 4.12 17.5 8.35 6.31 S21 0.437 1.567 0.436 1.639 0.435 1.592 S21 - 4.95 12.7 - 11.0 - 3.30 - 17.7 - 21.7 S22 0.910 0.697 0.928 0.723 0.948 0.789 S22 - 163.5 - 174.5 - 164.5 - 169.9 - 165.4 - 166.8
RF Application Notes
AN1526
Freq. (MHz) 806 838 870 Input Z S11 0.95 + j3.15 1.00 + j3.31 1.07 + j3.37 I / P Simul. Conjugate. Match 0.48 - j3.19 0.50 - j3.41 0.57 - j3.48 Output Impedance 2.41 - j7.24 1.90 - j6.79 1.35 - j6.41 O / P Simul. Conjugate Match 1.19 + j7.15 0.94 + j6.61 0.71 + j6.27 Data Book ZOL 2.93 - j1.39 2.92 - j1.10 2.92 - j0.81 Optimum O / P Impedance From Load Pull 3.60 +j1.26 3.60 +j1.02 3.39 + j0.75
K 1.62 1.61 1.71
B1 0.336 0.270 0.197
Freq. (MHz) 806 838 870 Input Z S11 3.29 + j3.95 3.83 + j3.18 3.74 + j2.02 I / P Simul. Conjugate. Match 1.13 - j2.98 1.06 - j2.90 1.02 - j2.86 Output Impedance 8.94 - j2.31 8.09 - j4.32 5.99 - j5.72 O / P Simul. Conjugate Match 2.86 + j5.52 2.10 + j5.14 1.57 +j4.69 Data Book ZOL 2.93 - j1.39 2.92 - j1.10 2.92 - j0.81 Optimum O / P Impedance From Load Pull 3.60 +j1.26 3.60 +j1.02 3.39 + j0.75
K 1.145 1.12 1.12
B1 0.941 0.871 0.693
CURRENT METHODS OF ENSURING CONSISTENT DEVICE IMPEDANCES
Users of RF power transistors have two main concerns with regard to the long term consistency of the device, minimum gain requirements and consistent impedances. Most of the recent devices characterized for the land mobile environment utilize a broadband fixture to demonstrate performance over a range of frequencies. The recently introduced MRF650 goes a step further and specifies gain, efficiency and input return loss at three frequencies of operation in a specified test fixture. Motorola has found over the years that this is the most cost effective way of producing RF power devices with minimum variability. Of course, new testing and characterization techniques are constantly being evaluated. The engineer unfamiliar with RF power devices
RF Application Notes
AN1526
RF Application Notes
AN1526
RF Application Notes
AN1526
+ j3.0 + j1.5 + j6.0 + j9.0 + j0.6 + j0.6 70 0.6 1.5 3.0 60 6.0 50 - j30 3.5 - j0.6 4.5 - j9.0 - j9.0 5.5 20 - j15 - j0.6 30 40 - j30 - j15 9.0 15 + j1.5 + j3.0 + j6.0 + j9.0
+ j15 + j30 0.6 1.5 3.0 6.0 6.5 9.0 15
- j1.5 - j3.0
- j6.0
- j1.5 - j3.0
- j6.0
+ j3.0 + j1.5 + j6.0 + j9.0 + j0.6 20 60 70 0.0 0.6 1.5 3.0 50 - j30 - j0.6 4.5 - j9.0 5.5 30 40 - j15 - j0.6 6.0 7.5 9.0 15 0.0 0.6 1.5 + j30 3.5 6.5 + j15 + j0.6 24 + j1.5
+ j3.0 + j6.0 + j9.0 16 20 + j30 3.0 28 - j30 24 20 - j15 - j9.0 6.0 9.0 15
- j1.5 - j3.0
- j6.0
- j1.5 - j3.0
- j6.0
RF Application Notes
AN1526
+ j3.0 + j1.5 + j6.0 + j9.0 + j0.6 7.5 0.6 1.5 3.0 6.0 9.0 15 + j0.6 55 + j1.5 + j3.0 + j6.0 + j9.0 75
0.0 - j30
45 - j0.6 35 25
- j30 - j15 - j9.0
- j0.6 4.5
- j15 - j9.0
- j1.5 - j3.0
- j6.0
- j1.5 - j3.0
- j6.0
+ j3.0 + j1.5 + j6.0 + j9.0 75 6.5 7.5 3.0 6.0 9.0 15 + j15 + j30 0.0 - j30 - j0.6 5.5 4.5 - j9.0 - j15 - j0.6 0.6 18 1.5 + j0.6 + j1.5
+ j3.0 + j6.0 + j9.0 18 22 26 3.0 30 6.0 9.0 15 + j15 + j30
65 + j0.6 55 45 35 25
- j30 22 - j15 - j9.0
- j1.5 - j3.0
- j6.0
- j1.5 - j3.0
- j6.0
RF Application Notes
AN1526
+ j3.0 + j1.5 + j6.0 + j9.0 65 + j0.6 6.0 5.0 0.0 0.6 1.5 3.0 6.0 9.0 15 7.0 + j15 + j30 0.6 + j0.6 55 45 75 + j15 + j30 1.5 3.0 6.0 9.0 15 + j1.5 + j3.0 + j6.0 + j9.0
0.0 - j30
- j30 - j0.6 35 - j15 - j9.0
- j0.6
- j15 - j9.0
- j1.5 - j3.0
- j6.0
- j1.5 - j3.0
- j6.0
+ j3.0 + j1.5 + j6.0 + j9.0 + j1.5
+ j3.0 + j6.0 + j9.0 18 + j0.6
+ j0.6 55 45 35 0.0 0.6 1.5
+ j15 26 + j30 3.0 6.0 9.0 15
0.0 - j30 - j15 - j9.0 - j0.6
5.5 - j0.6
- j30 22 - j15 - j9.0
- j1.5 - j3.0
- j6.0
- j1.5 - j3.0
- j6.0
RF Application Notes
AN1526
+ j3.0 + j1.5 + j6.0 + j9.0 + j0.6
MSUB CPW MSUB CPW CAP WIRE
+ j15 + j30 0.6 1.5 3.0 6.0 9.0 15
- j30 - j0.6 - j15 - j9.0
- j1.5 - j3.0
- j6.0
Figure 18. Photograph of CS-12 Impedance Measurement Probe
Figure 19. Photograph of 1 / 2 CQ Impedance Measurement Probe
Figure 20. Photograph of MRF873 Broadband Production Test Fixture
RF Application Notes
AN1526
C1, C15 - 10 µF, 25 V Tantalum C2, C14 - 1000 pF, 350 V, Unelco C3, C12 - 43 pF, 100 Mil, ATC Chip Capacitor C5, C13 - 91 pF, Mini-Unelco C4, C11 - 0.8 - 8 pF, Johansen Gigatrim 7290 Variable C6 - 16 pF, Mini-Underwood C7, C8, C9 - 12 pF, Mini-Underwood C10 - 10 pF, Mini-Underwood NOTE: C11
0.4 down Z3 from socket edge.
Figure 21. MRF873 Boardband Production Test Fixture Schematic
Figure 22. Photograph of Break-Apart Test Fixture - Fully Assembled
Figure 23. Photograph of Break-Apart Test Fixture - Setup for Impedance Measurements
RF Application Notes
AN1526
GPIB BENCH CONTROLLER
OUTPUT POWER METER INPUT POWER METER REFLECTED POWER METER
VECTOR VOLTMETER A B
POWER SPLITTER
POWER AMPLIFIER
DUAL DIRECTIONAL COUPLER
TEST FIXTURE
DUAL DIRECTIONAL COUPLER
50 OHM LOAD TERMINATION VARIABLE POWER ATTENUATOR VARIABLE SHORTED STUB
SYNTHESIZED SIGNAL GENERATOR
BIAS SUPPLY
Figure 24. Load-Pull Test Setup
10 9 8 Gp, POWER GAIN (dB) 7 6 5 4 3 2 1 0 800 820 840 860 f, FREQUENCY, (MHz) INPUT VSWR c Gpe
2.0 INPUT VSWR
RF Application Notes
AN1526
RF Application Notes
AN1526
CONCLUSIONS
For an RF power transistor we have demonstrated that the input and output large-signal device impedances are not only frequency dependent, but are also determined by the operating conditions of the device. Because of the wide range of possible applications, it is virtually impossible for the device manufacturers to present impedance data for every eventuality. The user, therefore, is left with the choice of either measuring the device impedances under the conditions he plans to use the device, or resorting to the classical methods of tweaking the circuit impedances into approximately the optimum match. The future does hold some promise in two areas. Automated tuners will enable impedance data to be gathered faster, enabling more comprehensive data to be included on the data sheets with the eventual possibility of publishing device impedance distributions. Compact device models in conjunction with non-linear simulators hold the best hope for simulating the device under the proposed operating conditions, and then permitting the software to synthesize the optimum broadband matching networks. fixture, a vector voltmeter monitors the test fixture load reflection coefficient L. Standard three term error correction was applied to the measured reflection co-efficient and this value is then used to correct the output power, Pout. The system is calibrated over a range of frequencies and the error correction in software. The following formula can be used to correct the power reading, 16
TMSuperCompact is a trademark of Compact Software and Touchstone is a registered trademark of EESof, Inc.
APPENDIX I: Load-Pull Method and Corrections for Power Measurement in Non-50 Ohm Environment
Load-pull measurements can employ a variety of test equipment and methods. For the load-pull measurements described earlier in this paper we used readily available and inexpensive equipment. In addition to the usual equipment found on an RF power bench including a computer as the instrument controller, the only additional pieces of equipment needed are a vector voltmeter, a variety of low attenuation power attenuators, and a variable length shorted stub. Figure 24 shows a block diagram of the bench set-up. A series of load mismatch conditions was established by terminating a broadband test fixture with the attenuators and shorted stub. The shorted stub was calibrated at approximately 20° intervals to establish varying phase shifts. By varying the value of attenuation, a grid of load impedances can be presented to the device on a network of VSWR circles in the reflection coefficient plane. The system was first calibrated by using a network analyzer and a probe in the device socket to measure this series of load impedances. A vector voltmeter, with error correction, could of course have been used to measure these impedances. After system calibration, the transistor was operated with the drive level adjusted to obtain rated output power under optimum tuning for maximum gain into a matched load. With the drive level fixed at this level, the output power was remeasured over the range of calibrated load impedances. This procedure was repeated at each frequency desired. The input match was tuned for zero reflected power with the output terminated in a matched load. The input return loss, under mismatched conditions, thereby indicates changes in magnitude of the input impedance. In addition to the usual power meter to measure the output power from the test
where M is the uncorrected load reflection coefficient, e23, e21, and e22e24 are the directivity, source match and frequency response errors determined by normal vector analyzer correction techniques 32, and Pom is the measured power meter reading corrected for directional coupler coupling magnitude, attenuation and meter frequency response. Using this method, accurate measurement of power and hence efficiency can be obtained for a system in which the load impedance is perturbed from the characteristic impedance of the transmission line power meter components. Contours can be generated from this grid data by a number of commercially available software packages. An alternative system would be to use tuners in place of the attenuator / shorted stub combination. The tuners can be either manual or automated. The advantage of the latter is that with suitable software the de-embedded load impedance presented to the device is available instantaneously. Also, with suitable software, the gain and efficiency circles can be determined by contour following techniques in real time, instead of fitting contours to measurements on a grid of load mismatch points 20.
REFERENCES
RF Application Notes
AN1526
RF Application Notes
AN1526
JAPAN: Nippon Motorola Ltd. Tatsumi-SPD-JLDC, Toshikatsu Otsuki, 6F Seibu-Butsuryu-Center, 3-14-2 Tatsumi Koto-Ku, Tokyo 135, Japan. 03-3521-8315 HONG KONG: Motorola Semiconductors H.K. Ltd. 8B Tai Ping Industrial Park, 51 Ting Kok Road, Tai Po, N.T., Hong Kong. 852-26629298
AN1526 / D
AN1526 / D RF Application Notes