DIRECTIONAL COUPLER AND ATTENUATOR FOR MEDIUM WAVE POWER NETWORK ANALYZER MEASUREMENT SYSTEM

Transcription

1 DIRECTIONAL COUPLER AND FOR MEDIUM WAVE POWER NETWORK ANALYZER MEASUREMENT SYSTEM DIRECTIONAL COUPLER INPUT OUTPUT 50 OHMS 500 WATTS CW.2 MHZ TO 12 MHZ 30 DB MIN DIRECTIVITY S/N 0X TOROIDAL CURRENT TRANSFORMER -30 DB FWD -30 DB REF S/N Ox -30 DB OUT 12 DB RF 50 OHMS S/N 0x DEVICE PERFORMANCE DATA AND APPLICATION PAPERS NOVEMBER 2004

2 TOC MEASUREMENT NOTES AND GRAPHS PAGES SWEPT FREQUENCY TECHNIQUES FOR EVALUATING AM SYSTEM BANDWIDTH PAGES EVALUATION AND IMPROVEMENT OF ANTENNA CHARACTERISTICS FOR OPTIMAL DIGITAL PERFORMANCE PAGES The Tunwall Radio directional coupler and attenuator are built for the measurement system described in the articles listed above. The directional coupler is built according to the schematic diagram in the first article. The measurement graphs were made on the first coupler and attenuator built. All subsequent devices were constructed identically, and correct performance was verified individually. The construction and sale of the devices and reprinting of the articles are with the author s permission. Steve Tunwall 205 Ironwood Circle Aurora, OH stevetunwall@tunwallradio.com

3 MEASUREMENT NOTES, DIRECTIONAL COUPLER 1. Rounding to the nearest db, the reflected port-to-forward port ratio exceeded 30 db from 200 KHz to 12 MHz when the directional coupler was terminated in a Maury Microwave Corporation ohm calibrated reference load. Over the mediumwave range for which the design was optimized, between 500 KHz and 2 MHz, the ratio exceeded 46 db. 2. The insertion VSWR was lowest - not exceeding 1.02:1 - over the range of 500 KHz to 2 MHz. This confirms that the implementation is optimized to allow operating impedance measurements within a mediumwave system with minimal insertion effect. 3. The flatness of the reflected port-to-forward port response and phase characteristics with open-circuit and short-circuit loads is sufficient, with the calibration correction of a modern vector network analyzer employed, to allow accurate impedance measurements within the common limitations of the reflection coefficient method. 4. Impedance measurements made using it in the power network analyzer system, looking into a dummy load with power of up to 200 watts sweeping from 500 KHz to 2 MHz, showed no indication of heating effects or nonlinearity. MEASUREMENT NOTES, 1. The measured attenuation would round down to 11 db instead of up to the stated 12 db within the mediumwave frequency range of 500 KHz to 2 MHz. The theoretical attenuation for the design, considering only the pure resistances involved, is 11.5 db and, hence, the stated attenuation - rounded to the nearest db - is 12 db. In the power network analyzer system for which the attenuator is intended, such small differences are of no consequence. The attenuator does not serve a calibrated function in the system. As typical variations in value of the power resistors within the circuit can be expected to result in attenuation to vary within a small range +/- the design value of 11.5 db in future units that are constructed, 12 db is an acceptable generic value. It is not referred to as a 12.0 db attenuator, however, as that would suggest an improper level of precision. 2. The flatness of the input impedance with the attenuator terminated in a 50 ohm load is very good over the mediumwave range of 500 KHz to 2 MHz. 3. It was tested with 100 watts input and a 50 ohm dummy load termination for 15 minutes and had no significant change in its input impedance or attenuation although it did get quite hot. If the attenuator is connected to an antenna receiving enough energy from other stations to deliver significant power to a 50 ohm load, it should work well if the power network analyzer system power plus the received energy does not greatly exceed 100 watts. The simple test given to the attenuator should justify a power rating of 100 watts intermittent.

4

5

6

7

8 TUNWALL RADIO TCT RESPONSE TERMINATED 50 OHMS DB FREQUENCY MHZ

9 SWEPT FREQUENCY TECHNIQUES FOR EVALUATING AM ANTENNA SYSTEM BANDWIDTH Ronald D. Rackley, P.E. du Treil, Lundin & Rackley, Inc. Sarasota, Florida Introduction To assure high quality transmission for stations operating in the AM radio band, the bandwidth characteristics of their transmitting antenna systems must be maintained to high standards. Impedance bandwidth as presented to a transmitter s final amplifier is the primary factor that is studied today, and it can be evaluated with measurements made using traditional impedance bridging techniques. The signals radiated from antennas can be subject to other bandwidth-related factors such as the delay characteristics of the matching equipment and filters, if any, and the pattern bandwidth characteristics of directional antenna systems. Evaluation of these other factors requires measurement of more than just impedance. The advent of digital transmission within the AM band promises to focus a much higher level of attention on antenna-related bandwidth matters. Not only is it necessary to pay close attention to the symmetry of the load impedance at the final amplifier of the transmitter to minimize unintended phase modulation that can cause a conventional AM signal to interfere with its own digital counterpart, the delay and response characteristics of the networks that lie between the transmitter and the antenna element(s) can have a significant impact on the integrity of the over-the-air digital signal. Additionally, the pattern bandwidth characteristics of a directional antenna system can determine how much of a station s analog coverage area will be able to receive a decodable digital signal. Although a modern device the vector network analyzer is capable of rapidly measuring every characteristic necessary to evaluate the important aspects of RF system bandwidth performance, it has not found widespread use in the AM band. Because vector network analyzers use signals at or near the milliwatt level for their measurements, they are subject to massive interference and even damage when connected directly to AM antennas because of the high levels of interference, from other signals and static discharges, that are induced on their towers. A system has been developed that utilizes a vector network analyzer along with an external amplifier for swept-frequency measurements of AM antennas and their associated RF circuitry under power with several volts of signal at the test terminals and a high degree of isolation between the antenna and the sensitive test equipment. The system can be used to make conventional impedance sweep measurements much more rapidly than is possible with traditional bridging equipment, and can also make sweep measurements of internal operating impedances, network response characteristics, network phase characteristics, network delay characteristics, and directional antenna element phase and ratio characteristics very efficiently. About Network Analyzers Present-day network analyzer technology evolved from the sweep measurement techniques that were used to plot the VSWR-vs-frequency characteristics of antennas at VHF frequencies and above 20 or more years ago. Scalar network analyzers so named because they do not view the quantities that they measure as complex numbers made such measurements more convenient when they became available by combining the sweep generator and tracking oscilloscope functions in one box. With the advent of vector network analyzer technology, it became possible to view the complex impedances of antennas instead of just their VSWR characteristics. A vector network analyzer consists in a very basic sense of a sweep generator, three synchronized receivers that detect the relative magnitudes and phases of the signals that are fed into them, and a computer that can process the outputs of two or more of the receivers as appropriate for the type of measurement that is desired. Impedance is just one thing that vector network analyzers can measure they can be used to measure many other parameters that are described by the relative magnitudes and phases of voltages or currents within a system, such as the response and phase characteristics of the

10 currents entering and leaving a network. Some vector network analyzers have the ability to perform the necessary mathematical transform to convert impedance sweep information at the input of a transmission line into a reflection-vs-distance (time domain) plot, making it possible to use one in place of a traditional time domain reflectometer ( TDR ) for finding line faults. Vector network analyzers fall into two general categories, those which simply have the sweep generator output and three receiver inputs on their front panels for connecting to the external components that are necessary for the desired measurements (with four jacks) and those which have everything necessary for specific measurements inside (with two jacks). Vector network analyzers with four jacks generally use external test sets that are designed for them. Such test sets fall into two categories S-parameter test sets that are capable of measuring the characteristics of a device under test ( DUT ) bi-directionally and transmission/reflection test sets that measure the input and pass characteristics of a DUT with the signal flow in one direction. Vector network analyzers with two jacks are very convenient for typical laboratory applications, but they lack the flexibility to be used with custom external devices for unusual applications such as high power measurements on AM directional arrays. The High-Power Measurement System The system that has been developed for AM band measurements at high power is based on a four-jack vector network analyzer. Both a Hewlett-Packard model 8753C analyzer and an Agilent Technologies model 4395A analyzer have been employed successfully in the system, the latter having the advantage of a built-in 3.5-inch magnetic disk drive to store its screen images making it unnecessary to have a printer or plotter attached during its use. In place of a transmission/reflection test set, the system uses a linear power amplifier, followed by a 12 db power attenuator, and then a broadband, highpower directional coupler that was custom designed for optimum performance within the range of 0.5 to 2.0 MHz. The power amplifier boosts the low-level RF output of the analyzer to a much higher voltage for application to the DUT. The voltage is dropped by the attenuator before entering the directional coupler both for amplifier load isolation and to attenuate interference signals coming from the DUT, which are reflected at the amplifier, by at least 24 db before they appear at the forward port of the directional coupler. The reference receiver, which phase-locks the analyzer, is therefore isolated by any interference that is present on the DUT by at least 24 db. The 12 db attenuator that follows the amplifier should be rated for higher power than the amplifier, for continuous operation, and should also be rated for bi-directional power flow if it is to be used to measure antennas that are near other AM radio stations that induce significant voltages on them. The directional coupler can be used by itself with its reflected port connected to the analyzer s receiver A for impedance measurements of a DUT. It can also be used to derive the reference signal necessary to phase-lock the analyzer with the reference receiver connected to its forward port and receivers A and B connected in other configurations. INPUT FORWARD 50 1 T 20 T FT Figure 1 Directional Coupler 1 T 20 T REFLECTED FT OUTPUT Figure 1 shows the circuit of the directional coupler that was developed for this system. Constructed in a 3.5 X 4.5 X 2 inch metal box with minimal lead length between the two transformers to minimize strays, it provides more than 40 db directivity over the frequency range of the AM band and was designed to withstand more than 100 volts across the DUT. The transformers are wound on Amadon Associates ferrite toroids. If a directional coupler for use between 3 30 MHz is desired, FT-114A-61 ferrite toroids may be substituted for those shown. The circuit requires precision, non-inductive 50 ohm resistors for accurate measurements. Caddock MP- 820 power film resistors, which have a resistance tolerance of one percent and are rated for 2.25 watts when free-mounted, are employed in the directional coupler circuit. An ENI type 240L linear amplifier is employed by the system. It has 50 db of gain and is rated for a continuous power output of 40 watts over the frequency range of 20 KHz to 10 MHz, which is achieved with the analyzer s power output set to 4 dbm. Other amplifiers should work well in the

11 system, although their power must be sufficient to overcome the level of interference that is encountered when the system is connected to an antenna. Impedance Measurements A network analyzer measures impedance by processing reflection coefficient information from a directional coupler with the signal from its internal sweep generator passing through it into the DUT. Since that process becomes more prone to error as the measured impedance departs further from the characteristic impedance of the directional coupler typically 50 ohms a different approach is used by modern impedance analyzers which are capable of very precise measurement from a fraction of an ohm to several megohms. It has been found, however, that with high quality directional couplers network analyzers can measure impedance values within the rated accuracy of the bridges that have traditionally been used for impedance measurements at AM frequencies up to a VSWR of greater than 10:1 referenced to the characteristic impedance. Most impedances that are encountered within AM antenna systems fall within that range. RF OUTPUT RECEIVER R AMPLIFIER (+50 db) (-12 db) FWD DIRECTIONAL COUPLER REF DUT measurements are made and the length of time required for each successive frequency sweep. For impedance measurements (and most of the other measurements that are made on AM antenna equipment), an IF bandwidth setting of 300 Hz has been found to be more than satisfactory. Vector network analyzers employ very sophisticated mathematical processing to achieve a high degree of accuracy including compensation for the effects of the coaxial cable used to connect the system to the DUT after they are calibrated with known open circuit, short circuit, and characteristic impedance standards. They are capable of covering a much wider frequency range than the AM band typically from a few hundred KHz to several GHz and precision calibration kits are normally recommended for use with them. Such calibration kits have been found to be unnecessary for making accurate measurements over the relatively narrow frequency range of the AM band, as long as a precision 50 ohm non-inductive resistor is used for the terminated calibration, a low-inductance short circuit is used for the short calibration, and a low stray-capacitance open circuit is used for the open calibration. These conditions are not nearly as hard to meet at 1.0 MHz as they are at 1 GHz, or even 100 MHz so it has been found to be sufficient to open and short test clips attached to the end of the coaxial cable that is used to connect the directional coupler to the DUT and then attach them to a low-inductance precision resistor, as prompted by the analyzer during the calibration process. RECEIVER A RECEIVER B VECTOR NETWORK ANALYZER Figure 2 Impedance Measurement System Figure 2 shows a block diagram of the system configuration that may be used for impedance measurements at the input of a DUT such as at a directional antenna common point or at the termination of the final amplifier of a transmitter. For an RF output of 4 dbm from the analyzer, the attenuators limit the signal level after amplification to 12 dbm at the receiver inputs a level that is well within the range of the analyzer. The IF bandwidth that is selected for the analyzer determines both the resolution with which Figure 3 Impedance Measurements at ATU Input Figure 3 is a photograph of the system being used to measure the input impedance of a nondirectional antenna s tuning unit. Figure 4 shows both the Smith Chart plot and the tabulation of the impedance sweep that were generated by the analyzer.

12 methods, it is necessary to constantly move the equipment when setting up the networks within a phasor cabinet, for instance. With the vector network analyzer system, the equipment can be placed within view of a phasor cabinet and not moved. A test cable of sufficient length to reach all of the networks can be calibrated once and then used repeatedly during the setup process. Operating Impedance Measurements Figure 4 Impedance Measurement Results The ability to measure impedances over a wide frequency range rapidly makes the system useful for other things such as evaluating transmission lines. Even if the analyzer does not have the option that allows it to function as a TDR, a sweep of measurements such as is shown on Figure 5 can be interpreted to find the length of a transmission line that is open-circuited or short-circuited on its far end by noting the frequency spacing between successive resonant points the line being one-half wavelength, or 180 elecrtical degrees in length, at that frequency. Despite the fact that the process of measuring impedances near zero (at the resonant points) is imperfect due to the nature of the reflection coefficient method used by vector network analyzers, it has been found that line lengths may be determined with sufficient accuracy for practical directional antenna work by this method. Figure 6 shows how the system may be configured to use two directional couplers to sweep the operating impedances of a directional antenna system where the transmission lines exit its phasor cabinet. Such measurements can be very meaningful for troubleshooting both impedance bandwidth and pattern bandwidth problems in arrays. RF OUTPUT RECEIVER R RECEIVER A RECEIVER B VECTOR NETWORK ANALYZER AMPLIFIER (+50 db) (-12 db) PHASOR FWD FWD DIRECTIONAL COUPLER DIRECTIONAL COUPLER REF LINE CP Figure 6 Operating Impedance System Figure 5 Transmission Line Measurements The system is also useful for the initial setup of RF matching and phase shifting networks, since the ability to connect a coaxial cable of any practical length to the directional coupler and then calibrate it for accurate impedance measurement at its far end makes it possible to measure and adjust network branches without having to place the measurement equipment nearby. With traditional bridging The directional coupler that is used to measure the operating impedance supplies its forward and reverse samples to the B and A receiver inputs of the analyzer, respectively. Short leads must be used at its input and output to minimize its insertion effects on the antenna parameters. The analyzer is configured to use the two receivers for measurement, and the input to the reference receiver which is derived from the forward port of the directional coupler at the common point is only used to phase-lock the system. This allows the system to operate with a high degree of interference immunity even when measuring the line matches of the low-power towers of an array owing to the isolation of the reference receiver input from other signals that are picked up by the antenna.

13 Measuring Network Response Characteristics Figure 7 shows the system configured to use the A and B receiver inputs to measure the response of a network using current transformers at its input and output. The measurements may be made with the network driven by the output of the directional coupler and terminated in a load, as shown, or with the network connected for normal operation with the antenna input driven by the output of the directional coupler. Figure 9 Response Measurements RF OUTPUT AMPLIFIER (+50 db) (-12 db) DIRECTIONAL COUPLER FWD RECEIVER R RECEIVER A NETWORK RECEIVER B VECTOR NETWORK ANALYZER Figure 7 Response Measurement System The details of the current transformer design are shown on Figure 8. The secondary consists of 32 turns of RG-174 miniature coaxial cable wound onto the ferrite toroid and connected as shown. The test setup to measure the response of an antenna matching/diplexer unit, with a current transformer at one station s input and another at the combined output, is shown in Figure 9. Figure 10 is a close-up photograph showing the transformer at the unit s input. Figure 10 Current Transformer at ATU Input Figure 11 shows the measured response in db and Figure 12 shows the measured phase shift. Figure 13 shows the measured delay characteristics. In addition to phase shift and delay characteristics, vector network analyzers can also produce a screen showing a plot of the departure from linear phase a related quantity that may be convenient to use when analyzing the transparency of a network or system of networks. 1 T FT T OUTPUT Figure 8 Current Transformer Figure 11 Measured Response

14 the ratio and phase of the current flowing in a tower of an array may vary with frequency, respectively. RF OUTPUT AMPLIFIER (+50 db) (-12 db) DIRECTIONAL COUPLER CP FWD RECEIVER R RECEIVER A DA TOWER CURRENT SAMPLES Figure 12 Measured Phase Shift Sweep RECEIVER B REFERENCE TOWER CURRENT SAMPLE VECTOR NETWORK ANALYZER Figure 14 DA Phase and Ratio Measurements Figure 13 Measured Delay Sweep The system may be used in its response measurement configuration to evaluate and adjust filters such as are employed to eliminate intermodulation products from transmitters and for multiplexing more than one frequency on a single antenna. Such work goes very quickly with the system. An analyzer IF Bandwidth setting of 100 Hz has been found to be more than sufficient for filter tuning work, while a setting of 300 Hz can be used for measuring the response of simpler networks. Figure 15 Measured Ratio Sweep Measuring Directional Antenna Parameters Figure 14 shows the configuration that may be used to measure the phase and ratio-vs-frequency characteristics of a directional antenna system s elements. The common point is driven by the directional coupler which derives the signal for the analyzer s reference receiver input. The antenna monitor sampling line of the reference tower is connected to the B receiver s input. The other sampling lines are switched to the A receiver s input and the analyzer is alternately set to measure the magnitude and phase relationships of the A and B inputs for each tower. Figures 15 and 16 show how Figure 16 Measured Phase Sweep It may be necessary to use attenuator values that differ from those that are shown at the analyzer s receiver inputs on Figure 14, depending on the sensitivity of the sampling devices that are employed

15 in the antenna monitor sampling system. The voltages at the terminals of the antenna monitor can be compared to the voltage that is known to be present at the common point under normal operation to determine the db isolation between the directional antenna s input and the sampling lines when they are terminated in 50 ohm loads - and appropriate attenuator values can then be chosen. Information on how the tower ratios and phases vary with frequency may be used to evaluate pattern bandwidth. The measured antenna element parameters at carrier and sideband frequencies may be used to calculate the changes in far-field magnitude and phase that occur at different azimuths, using computer software such as the Mininec Broadcast Professional package. The far-field response may then be calculated directly from the magnitude excursions. The delay characteristics may be calculated from the phase characteristics at the azimuths of interest. Further Comments The capabilities of network analyzers far exceed the specific applications that are described herein, and anyone wishing to use the system will do well to become thoroughly familiar with them by studying their instruction manuals and any other educational materials that are available. The information provided herein is for general use only and assumes that any user will be competent in the use of vector network analyzers before the described procedures are undertaken. Although the system has been successfully used under a variety of conditions at different transmitter sites, the experience with it is not exhaustive and there may be some cases where its usefulness is limited by interference or other factors that are yet unknown. While the system has proven very effective for evaluating antenna performance and setting up networks where frequency sweep information is useful, the battery-powered convenience of current model bridging equipment still recommends it for simple single-frequency impedance measurements particularly in equipment that is located outdoors. For that convenience, and the ability to measure very high and very low impedances that are outside the accurate range of the vector network analyzer system RF bridges and generator/receiver units will continue to be standard equipment for engineers who service and adjust AM antennas. Conclusion With the high power measurement system described herein, vector network analyzers may be utilized for both conventional impedance measurements and the specialized tests that are required to evaluate the transparency of AM antenna systems for digital signal transmission. Because automatic sweptfrequency techniques are used, the bandwidth characteristics of an antenna system can be evaluated with minimal station down time. As of yet, no antenna system performance requirements have been issued by the developer of the digital transmission system that has been approved for use by AM stations in the United States. Presumably, the requirements will be forthcoming and engineers will then be able to use the vector network analyzer system to obtain the information necessary to predict the performance of antenna systems under as built conditions and, where necessary, devise corrective measures. References: 1) Using a Network Analyzer to Characterize High-Power Components, Agilent Technologies Application Note ) Exploring the Architectures of Network Analyzers, Agilent Technologies Application Note ) Understanding the Fundamental Principles of Vector Network Analysis, Agilent Technologies Application Note ) Applying Error Correction to Network Analyzer Measurements, Agilent Technologies Application Note ) Advanced Impedance Measurement Capability of the RF I-V Method Compared to the Network Analysis Method, Agilent Technologies Application Note

16 EVALUATION AND IMPROVEMENT OF AM ANTENNA CHARACTERISTICS FOR OPTIMAL DIGITAL PERFORMANCE Ronald D. Rackley, P.E. du Treil, Lundin & Rackley, Inc. Sarasota, Florida INTRODUCTION AM digital transmission places high demands on antenna system bandwidth, both from the standpoint of the input impedance at the transmitter load - which can cause noisy analog reception - and the antenna s far-field radiation characteristics, which can erode the error-correction capability of the digital signal and make its reception inconsistent. Directional antenna systems present the additional complication of sideband phase and amplitude errors resulting from changes in pattern shape with frequency, which can render digital modulation un-decodable in areas with satisfactory analog reception. A conventional AM signal is modulated in amplitude only, and has an uncomplicated spectral display that is directly related to the spectral content of the modulating audio. For single-frequency (sinusoidal) audio modulation, the RF envelope waveform is produced by two sidebands separated from the RF carrier frequency by the modulating signal s frequency. For 100-percent modulation, each sideband is one-half the amplitude of the carrier. AM In-Band-On-Channel (IBOC) digital transmission technology is currently moving from the realm of the theoretical into the "real World" and, although the body of on-the-air experience with it is limited, it is rapidly expanding. Measurements can be made to evaluate the important aspects of antenna performance and, in many cases where performance is found to be lacking, relatively simple measures can be taken to improve matters. - F(M) + F(M) MODULATION BASICS As the ultimate goal in digital transmission is to have the signal pass through the antenna system and arrive at the receiver with its components in their correct relationships, it is best to start with an understanding of the nature of those components and how they might be affected by the transmission system. Digital modulation is accomplished by varying both the amplitude and the angle of the RF carrier simultaneously to transmit the digital bits that are decoded in the receiver. In the case of IBOC transmission such as is presently being introduced in the USA, this is done while still transmitting the conventional AM signal by sending the digital information at a low level so that the resulting occupied spectrum fits within the RF mask of an AM channel as defined in the FCC Rules. AM SINUSOIDAL MODULATION SPECTRUM (FREQUENCY DOMAIN) A spectrum analyzer screen showing such modulation contains the three components of the signal plotted in terms of frequency (in the frequency domain). It is very difficult to visualize the components of even the simplest form of modulation in the time domain, since the carrier and sidebands are all varying sinusoidally over time although at different rates. Normally, complex mathematical expressions are used to explain modulation in the time domain using phasor algebra. Fortunately, it is possible to show the relative motion of the carrier and sideband phasors using vector analysis by referencing the degree of rotation of the sidebands to a fixed point in the rotation of the carrier phasor.

17 The two sidebands of a conventional singlefrequency modulated AM signal appear as equallength vectors that rotate in opposite directions relative to the stationary carrier vector and are always displaced from it by equal but opposite angles. At the point in the modulated RF envelope where its amplitude is equal to that of the carrier alone, the two sidebands oppose each other to completely cancel in the resultant of the vector addition. As the two sideband vectors fall in line with each other and the carrier vector, the resultant approaches twice the carrier magnitude to form a 100-percent positive modulation peak. R - F(M) + F(M) - F(M) + F(M) AM SINUSOIDAL MODULATION (VECTOR REPRESENTATION) As the sideband vectors rotate so that the resultant is greater than the amplitude alone, the waveform progresses toward a positive modulation peak. R AM SINUSOIDAL MODULATION (VECTOR REPRESENTATION) Similarly, when enough time has passed for the two sideband vectors to rotate to the opposite direction, the resultant will approach zero to form a 100-percent negative modulation peak (not shown). - F(M) + F(M) Except when testing with tone modulation, AM signals are much more complicated. The complex waveforms of normal programming have many frequency components. In a perfect system, however, the components will always be divided into equal, symmetrical sideband components. AM SINUSOIDAL MODULATION (VECTOR REPRESENTATION)

18 + F(M) AND ETC... FREQUENCY COMPONENTS OF MODULATING WAVEFORM F (C) - F(M) AM COMPLEX MODULATION (VECTOR REPRESENTATION) A spectrum analyzer shows the range of frequencies occupied by the various sideband components for normal program modulation. FM - PM MODULATION FIRST ORDER SIDEBANDS AM COMPLEX MODULATION SPECTRUM (FREQUENCY DOMAIN) Suppose that the sidebands vectors are symmetrical about an imaginary line running perpendicular to the carrier vector instead of to the carrier vector itself in other words, the sidebands of AM transmission are rotated 90 degrees (in quadrature) relative to the carrier. The resultant will not always fall in line with the carrier, but will vary in angle from it over time. This will be fundamentally angle modulation which, in analog transmission, may be either phase modulation or frequency modulation depending on the process applied to the audio to produce it. Careful observation reveals that angle modulation accomplished with only two sidebands for the singlefrequency, sinusoidal case would also produce secondary amplitude modulation, since the resultant follows a line tangent to the circle that would be traced by the carrier vector if only its angle were changing. That is why an FM signal has many sidebands even if only a single-frequency tone is being transmitted the infinite number of harmonically-related sidebands that trail off in amplitude and vary in phase are necessary in order to have the resultant trace the circle rather than the tangent line. Real-World conditions do not permit infinite bandwidth, of course, so at lease a small amount of distortion is theoretically unavoidable when FM is transmitted through a bandwidth-limited transmission channel. DIGITAL TRANSMISSION Digital transmission utilizes a combination of amplitude and angle modulation to send the bits of information that are decoded by receivers. Such modulation is generally referred to as I-Q modulation, as each bit has a unique combination of I (in phase with the carrier) and Q (in quadrature relative to the carrier) components assigned for it. The following diagram depicts the basic concepts involved in sending a 16-bit digital constellation with I-Q modulation superimposed on a conventional AM signal. It is a conceptual sketch only, and does not represent any actual transmission system. It is not to scale.

19 100% POSITIVE PEAK TRANSMITTER LOAD OPTIMIZATION CARRIER DIGITAL INFORMATION When a transmitter with a perfect modulation system is connected to a perfect load (one which does not change impedance with frequency), perfect transmission of the digital signal will result. Modern transmitters that have been designed to carry digital modulation approach perfection, at least for all practical purposes. Real-world antennas, however, do not. Since it is not realistic to expect antennas to have input impedances that are constant with frequency, the goal of having the effects of changing impedance at least be symmetrical about the carrier frequency must be pursued. If the load impedances at sideband frequencies differ from the carrier frequency in symmetrical fashion, cross-talk between the digital and analog signals can be minimized. AM + DIGITAL IQ MODULATION CONCEPTUAL ONLY - NOT TO SCALE It can be seen that, for the bits to arrive in the correct relationships to be decoded by a receiver, the transmission system must maintain their correct locations in terms if I and Q and they must remain isolated from the AM signal. It is important to note that the symmetry must be provided to the actual active components of the final amplifier, where the process may be most closely simulated by a perfect source. The circuitry that provides impedance matching and bandpass filtering between the final amplifier and the transmitter output port must be considered if it is impossible to make impedance measurements at the final amplifier. Assuming that the output network of a transmitter is adjusted to provide the correct load impedance for the final amplifier at carrier frequency, the most important thing to know about it for sideband symmetry analysis is its phase delay. LOAD MUST BE SYMMETRICAL HERE TRANSMITTER OUTPUT PORT FINAL AMPLIFIER OUTPUT MATCHING AND FILTERING NETWORK AM IBOC MODULATION SPECTRUM (FREQUENCY DOMAIN) CONCEPTUAL - NOT TO SCALE IBOC digital transmission occupies spectrum in addition to that of a conventional AM signal. It is most convenient to view impedances as they appear on a Smith Chart when analyzing sideband symmetry. In order to view the important characteristics of the impedances without respect to what the impedance might be at carrier frequency such as at the final amplifier of a solid state transmitter which operates into an 8 Ohm load or at the common point of a directional antenna system s phasor which might be offset in reactance by several negative ohms to correct for the insertion inductance of a contactor between it and the transmitter perunit Smith Charts (referenced to 1.0 rather than a transmission line impedance like 50 Ohms) are used.

20 45 DEG TOWARD GENERATOR +J 1.0 R=1.0 0 DEG OR 180 DEG J= DEG F(L) -J 1.0 F(H) 135 DEG SMITH CHART BASICS The process of converting an impedance sweep for per-unit analysis is known as normalizing to the carrier frequency impedance. Dividing the resistance at a sideband frequency by the carrier frequency resistance gives the sideband per-unit normalized resistance if the carrier frequency resistance is assigned a value of unity. Dividing the difference in reactance between a sideband frequency and the carrier frequency by the carrier resistance gives the sideband per-unit normalized reactance if the carrier frequency reactance is assigned a value of zero. FINAL AMPLIFIER LOAD IMPEDANCE SYMMETRY F (C) - F(M) + F(M) NORMALIZING PER-UNIT VALUES TO THE CARRIER IMPEDANCE 1. DIVIDE EACH SIDEBAND RESISTANCE BY THE CARRIER RESISTANCE 2. DIVIDE THE DIFFERENCE BETWEEN EACH SIDEBAND REACTANCE AND THE CARRIER REACTANCE BY THE CARRIER RESISTANCE EXAMPLES FREQUENCY C.P. RESISTANCE C.P. REACTANCE PER-UNIT RESISTANCE PER-UNIT REACTANCE - 15 KHZ j j 0.16 CARRIER j KHZ j j KHZ j j 0.09 CARRIER j j KHZ j j 0.12 SIDEBAND ALTERATION FROM SYMMETRICAL LOAD A symmetrical impedance sweep will appear on the Smith Chart with equal normalized resistances and equal normalized reactances of opposite sign for equal frequency deviation from the carrier frequency. For this symmetrical case, the sideband resistances are equal but lower than the carrier resistance - at both sideband frequencies and the

21 sideband reactances are symmetrical. If the final amplifier functions as a voltage source, higher power will be delivered to both sidebands than would be the case with a flat load (meaning longer sideband vectors) - resulting in pre-emphasis but not harmonic distortion in the demodulated waveform.. In addition, the opposite reactances will cause both sidebands to be offset by equal but opposite angles relative to the carrier - producing envelope delay, but not harmonic distortion, in the demodulated waveform. Both effects are perfectly acceptable for analog transmission, as long as the transmitter is not driven into nonlinearity to deliver the required sideband power. For digital transmission, the effects on the I-Q constellation will be more easily compensated than would be the case with a nonsymmetrical load. Importantly for IBOC transmission, the symmetry will prevent crosstalk between the digital and analog signals. F(L) +X -X F(H) R Digital-to-analog crosstalk may be the area of most concern for stations initially converting to IBOC transmission. Although a certain amount of hiss-type noise must be tolerated under the carrier of an AM station transmitting the IBOC signal, a very prominent bacon frying noise has been noted in certain cases. Thus far, experience indicates that this noise may be significantly decreased by establishing final amplifier load symmetry. FINAL AMPLIFIER LOAD IMPEDANCE ASSYMMETRY The following case of asymmetrical sideband load impedances shows how the two sideband vectors are altered differently in length and offset differently in angle due to the sideband impedances presented to the final amplifier. It is obvious that the resultant will follow an ellipse over a cycle of the analog modulating frequency rather than the straight line that is traced by a pure AM signal. Non-symmetrical load impedance characteristics, in addition to causing crosstalk between the digital and analog signals in IBOC transmission, also complicate the error correction process for the digital I-Q constellation - leading to higher bit error rates that erode the reliability of the digital transmission channel. For the same reasons, the analog signal sidebands are altered in such a way as to cause harmonic distortion in the demodulated waveform. F (C) - F(M) SIDEBAND ALTERATION FROM ASYMMETRICAL LOAD + F(M)

22 The asymmetrical load in this case can be made to perform identically to the symmetrical load that was considered in the first case by the addition of a -45 degree phase shifting network between the transmitter and the antenna. TRANSMITTER TRANSMITTER OUTPUT PORT - 45 DEGREE PHASE SHIFTING NETWORK TO ANTENNA Some loads may not be made symmetrical for the transmitter s final amplifier by simple phase rotation. One method to evaluate the potential for symmetry is to plot the normalized per-unit antenna input impedance on transparent overlay paper over a Smith Chart so that it can be rotated about the reference point to simulate the effects of added phase shift. It is obvious by inspection that this load will require more than the addition of a phase-shifting network to have final amplifier load symmetry. IBOC DESIRED CHARACTERISTICS Although there are no universally accepted criteria for evaluating the effects on digital transmission of finite-bandwidth antenna systems, the developer of the digital transmission system that is presently in use in the United States, Ibiquity, has recommended the following desired characteristics for antenna systems that transmit with their system: I. Load Impedance Bandwidth a. +/- 5 KHz - symmetry of the load impedance presented to the final RF amplifier within the transmitter such that the VSWR calculated for one sideband impedance, when normalized to the complex conjugate of the corresponding sideband impedance on the other side of carrier frequency, does not exceed 1.035:1 +X R +5 KHz +X - 5 KHz R -X F(H) F(L) -X ANTENNA IMPEDANCE ASYMMETRY b. +/- 10 KHz - the VSWR of the load impedance presented to the final RF amplifier within the transmitter should not exceed 1.20:1 when normalized to the carrier frequency impedance. c. +/- 15 KHz - the VSWR of the load impedance presented to the final RF amplifier within the transmitter should not exceed 1.40:1 when normalized to the carrier frequency impedance.

23 +X +10 KHz 1.2:1 VSWR R 1.4:1 VSWR Fc -15 KHz NETWORK ANALYZER MEASUREMENTS It has been found to be very convenient to use a network analyzer to make the measurements necessary for evaluating antenna system performance for digital transmission. A power-amplified system for making measurements with interference immunity has been developed for this purpose, and was described in the author s paper which appears in the NAB 2003 Engineering Conference Proceedings. -X RF OUTPUT AMPLIFIER (+50 db) (-12 db) DIRECTIONAL COUPLER DUT FWD REF II. Far-Field Radiation Pattern Bandwidth a. +/- 15 KHz - Response within +/- 2 db b. +/- 15 KHz - Group delay constant within +/- 5.0 microseconds These values are preliminary guidelines and are subject to change as more experience is gained with digital transmission. Thus far, experience has shown that the recommendations for sideband VSWR at 10 and 15 KHz from the carrier frequency are very conservative. Loads with higher sideband VSWRS have provided satisfactory performance both in the lab and in the field when the sideband impedances are made nearsymmetrical about the carrier frequency [where perfect symmetry would have equal resistance and equal-but-opposite reactance excursions on either side of the carrier frequency]. RECEIVER R RECEIVER A RECEIVER B VECTOR NETWORK ANALYZER IMPEDANCE MEASUREMENT SYSTEM As explained earlier, the symmetry of the sideband impedances at +/- 5 KHz of carrier frequency has proven to be very important for minimizing digital-to-analog crosstalk. Such crosstalk causes noise that is sometimes described as bacon frying underneath the analog signal when digital transmission is underway. IMPEDANCE MEASUREMENTS AT ATU INPUT

24 AN IMPEDANCE SWEEP A MEASURED DA TOWER RATIO SWEEP Another configuration may be used to measure the phase and ratio-vs-frequency characteristics of a directional antenna system s elements. The common point is driven by the directional coupler which derives the signal for the analyzer s reference receiver input. The antenna monitor sampling line of the reference tower is connected to the B receiver s input. The other sampling lines are switched to the A receiver s input and the analyzer is alternately set to measure the magnitude and phase relationships of the A and B inputs for each tower. A MEASURED DA TOWER PHASE SWEEP RF OUTPUT RECEIVER R RECEIVER A RECEIVER B AMPLIFIER (+50 db) (-12 db) FWD DIRECTIONAL COUPLER CP DA TOWER CURRENT SAMPLES REFERENCE TOWER CURRENT SAMPLE Information on how the tower ratios and phases vary with frequency may be used to evaluate pattern bandwidth. The measured antenna element parameters at carrier and sideband frequencies may be used to calculate the changes in far-field magnitude and phase that occur at different azimuths, using computer software such as the Mininec Broadcast Professional package. The far-field response may then be calculated directly from the magnitude excursions. The delay characteristics may be calculated from the phase characteristics at the azimuths of interest. VECTOR NETWORK ANALYZER DA PHASE AND RATIO MEASUREMENTS