May 4, 1964 Electronics
[Table of Contents]
Wax nostalgic about and learn from the history of early electronics.
See articles from Electronics,
published 1930 - 1988. All copyrights hereby acknowledged.
The many idiosyncrasies of atmospheric phenomena that affect long distance communications are certainly more well known and understood today than they were in the early days of radio. Ionization, temperature and pressure gradients, suspended particulate contamination, and other factors have been extensively studied, measured, and modeled. Daily and seasonal patterns are somewhat predictable and exploitable for purposes of general use, but short term variability that affects long distance radar measures of distance, altitude, and speed requires near instantaneous, pulse by pulse analysis of atmospheric conditions. Research and development of methods for accommodating short term variations that skew measurements are an ongoing science. An extreme example of atmospheric variation compensation is the method used by ground-based telescopes that shine lasers into the ionosphere to create "artificial stars" whose scintillation properties can be used in both software and adaptive optics to cancel out apparent changes in position and intensity.
Refractivity of atmosphere may be approximated easily by measuring dielectric constant of the air with microwave refractometer.
Clarence H. Stewart received his bachelor's degree in engineering from the University of Kentucky in 1949, his master's from Northwestern University in 1960, and is currently pursuing studies at the University of Colorado leading to a PhD degree. From 1943 until 1951 he was employed by several radio stations. In 1951, he joined Jansky & Bailey as a project engineer. Primary duties included high frequency antenna design and propagation analysis. In 1955 he joined Bell & Gossett Co., becoming the chief engineer of the Electronics division. In 1960 he was transferred to Colorado Research Division of ITT-Bell & Gossett, Inc. (then Colorado Research Corp.) as manager of engineering and later became technical director. He holds patents for communications and switching systems and has published several articles on communications systems.
Gary J. Vincent received his bachelor of science in engineering degree from the University of Denver in 1959, and has done graduate work at the University of Utah (1960) and the University of Colorado (1963-64). From 1957 to 1959, he was employed by Denver Research Institute of the University of Denver as a staff research assistant. Primary activity was in the field of digital computer circuits. From 1959 to 1961 he was employed by Sperry Utah division of Sperry Rand, in Salt Lake City, Utah, as a project engineer on ground support equipment for the Sergeant missile system. Specific areas included digital computer logic and circuit design, digital-to-analog conversion equipment, servomechanisms, and automatic test equipment. In 1961, he joined Colorado Research division, ITT-Bell & Gossett, Inc. (then Colorado Research Corp.), as project engineer.
Refraction geometry illustrates difference in apparent and actual target locations.
By Clarence H. Stewart and Gary J. Vincent
Colorado Research Division, ITT - Bell & Gossett, Inc., Broomfield, Colo.
Radar equipment is limited in accuracy by the varying medium through which it is propagated. The Colorado Research microwave refractometer is a device capable of making very precise measurements of the propagation medium, and is so stable that readings are absolute rather than relative. By measuring the refractivity of the atmosphere, it is possible to increase the accuracy of a tracking radar.
The permittivity of the atmosphere - the major factor in tracking errors - may be easily approximated by measuring the dielectric constant or refractivity of the air. Instrumentation capable of measuring refractivity on a real-time basis is essential. For many years, the only practical method involved conventional meteorological instrumentation, combined with the use of an empirically derived formula that relates the atmospheric refractivity to the meteorological parameters; temperature, pressure, and humidity. This "indirect" method is highly subject to error, because of sensor errors and time constants, and also because of the uncertainty in the formula.
D. R. Hay of the University of Western Ontario designed a refractometer for balloon-supported soundings. Hay's instrument has two capacitors which alternately control the frequency of an oscillator. One capacitor is sealed, and acts as a reference element. The second is open and samples the atmosphere. By comparing the oscillator frequencies with the reference capacitor and the sampling capacitor, the dielectric constant (and thus the refractive index) of the sampled medium is determined directly. Using a single oscillator reduces such problems as oscillator drift and temperature variations to second-order effects. However, measurements made with the capacitive technique were not absolute, because the device did not have long-term stability.
Basic technique employed in microwave refractometer is the comparison of signals passed through a sampling cavity vented to the propagation medium and a hermetically sealed reference cavity.
The Colorado Research microwave refractometer is stable enough to perform absolute, rather than relative measurements of the dielectric constant. (This instrument is the result of a joint development effort by the National Bureau of Standards Boulder Laboratory and Colorado Research Division, ITT-Bell & Gossett, Inc.) The basic technique involves two precision microwave transmission cavities and associated circuitry. One of the cavities is ventilated, and acts as the sampling transducer. The second cavity is hermetically sealed and acts as the reference element. The output of a swept-frequency klystron oscillator is applied to the two transmission cavities (see block diagram at top of p. 74). The outputs of the cavities will have a relative timing which depends upon the resonant frequencies of the two cavities.
The introduction of atmosphere will affect the resonant frequency of the sampling cavity as in the open-capacitor tuned circuit. The comparator in the unit detects the relative timing of the detected outputs and converts this information into a suitable form for display or recording. Several methods have been implemented for performing the conversion. One of the earliest techniques involved the generation of a square wave whose duty cycle was a function of the relative timing of the two outputs. Integration of this waveform resulted in a voltage whose magnitude was proportional to refractivity. This method provided an output suitable for operating chart recorders and similar instruments, but was not optimum because of inaccuracies introduced by amplifier drift, variation of klystron tuning rate with age, and related problems associated with analog circuits. The final unit (see diagram below, left) employed an X-band klystron frequency modulated by a 400-cps sine wave from a resonant-reed oscillator. The two cavity outputs are combined in the comparator, which produces two signals: one consists of the difference of the two cavity outputs; and the second, the sum. If both cavities resonate at the same frequency, the "difference" output is zero, and the sum output is an 800-cps signal.
Complete refractometer illustrates use of servo motor to match cavity resonant frequencies, and afc loop to keep klystron centered on proper frequency.
If the dielectric constant of the contents of the sampling cavity changes, the difference output will be a signal containing a 400-cps component, the phase of which is determined by the direction in which the cavity resonant frequency was changed. This is applied, via a narrow-band amplifier, to the control phase of the servo motor. The motor drives a small tuning probe in the reference cavity, causing the resonant frequency of the reference cavity to exactly match that of the sampling cavity. If the center frequency of the klystron is not midway between the resonant frequencies of the two cavities, the sum output of the comparator will also contain a 400-cps component. This output, fed back through an AFC loop, maintains the center frequency at the correct point.
The motion of the tuning probe is converted into refractivity data by a shaft angle encoder - for digital output - or a precision potentiometer - for an analog output. The range of the digital readout is 0-400 N units, with a resolution of 0.1 N unit. This range is sufficient to cover virtually any condition from a hard vacuum to moist, sea-level atmospheres.
Calibration of the system is accomplished with a reference cavity assembly fitted with a multi-turn dial. This cavity has been calibrated by the National Bureau of Standards. It indicates resonant frequency as a function of dial setting. After the system is installed, the sampling cavity is evacuated and the mechanical portions of the servo system are adjusted for a read of 0.0 N. The accurate operation of the refractometer is dependent primarily on the mechanical stability of the microwave cavities. These are carefully fabricated, using special alloys, and then plated to obtain the desired electrical properties. Additionally, each cavity is temperature compensated to provide an extremely low temperature coefficient (approximately 3 x 10-8°C). The resulting instrument is relatively insensitive to variations in environmental conditions.
Need for Computation Procedure
In addition to the instrumentation requirements, it is necessary to determine computations necessary to correct the raw data. The necessary calculations include curve-fitting, application of Snell's Law, and numerical integration, which are all readily performed by conventional scientific computers - for which programs have already been written.
Definition of measuring techniques to be used is a more difficult task; in part, because requirements for high-accuracy correction have not existed for long. Therefore, definitive measurement procedures have not yet been developed, and there is considerable disagreement among the various researchers regarding the optimum techniques. Proponents of predictive techniques maintain that one-point (usually ground level) correlation of "average" refractivity functions will provide adequate precision in the majority of cases. At the opposite extreme would be the procedure of actually measuring the refractivity gradient along the entire propagation path with an aircraft carrying refractivity instrumentation.
Sampling cavity is vented to atmosphere; resonant frequency varies in proportion to the atmosphere's dielectric constant.
Generally, the optimum technique would lie somewhere between these extremes. A typical experiment would employ a refractometer at the ground station, plus one or two at intermediate points along the propagation path. Additional measurements could be provided by a second refractometer in the target and others obtained either by refractometers in aircraft or balloons. Multi-point correlation of the refractivity function can afford a much higher confidence level in the precision of the final results. The number of intermediate samples which should be taken is determined primarily by the prevailing meteorological conditions. During clear weather, particularly in dry, inland climates, relatively few measurements are required because of the spatial and temporal stability of such atmospheres. Coastal regions, or regions in which frontal activity is present, require an increased number of measurements since there is a high probability of significant refractivity gradients over short distances within much moist, unstable atmospheres.
One example of such a situation is a very high gradient condition known as a "duct." In a duct, the vertical gradient is sufficient to cause the propagation path to follow a course parallel to the surface of the earth. In tropical climates this condition may exist approximately 10% of the time, and may occur at any location at least occasionally. Ducts cannot yet be predicted with any degree of confidence, and may appear and disappear with little or no apparent warning. Another abnormal condition which may exist is a refractivity inversion, usually associated with a humidity inversion. Such a condition will cause the propagation path to bend upward. Depending upon the extent of the inversion, this condition could result in an elevation error greatly different from that predicted by any of the statistical techniques.
Posted June 6, 2019