Thanks John, W3JN, for inviting me to tell another story about the earliest space projects. This one is another quote from Reflections 3 that I hope you'll find interesting:
Sec 28.2 The TIROS WEATHER SPACECRAFT
Historical Background on the Origin of the World’s First Weather Satellite, TIROS 1
In 1957, people from the U.S. Army’s Advanced Research Projects Administration (ARPA) at Huntsville, Alabama, that included the eminent rocket scientist, Werner von Braun, approached us at the RCA Laboratories in Princeton, NJ, concerning the possibility of designing and building an orbiting spacecraft that could photograph the geographic environs of the Soviet Union. After a short study we found that we could, and we subsequently received a performance contract to build it. My assignment was to develop the antenna system for receiving commands from the ground stations and also for transmitting the pictures obtained by the cameras on-board the spacecraft back to the ground stations. After progressing with the project to the point of being close to testing the prototype, our contract was suddenly terminated. It seems that Congress had just learned of this project, which had been classified as top secret, and ordered it stopped, because, due to the timing, it was considered a politically dangerous undertaking. The timing was considered crucial diplomatically, because at that time the State Department was in the throes of quelling the fallout from Gary Francis Powers’ U2 reconnaissance plane having been shot down over the Soviet Union.
During the few weeks that followed, the Washington, D.C. rumor mill spread the news that RCA had built a spacecraft capable of photographing the Earth from an orbiting platform. People at the U.S. Weather Bureau learned of it, and came to us with the question, “Could you re-design the spacecraft to perform as an orbiting weather station?” We replied that we could, and then entered into a contract to build TIROS 1. (TIROS is the acronym for Television Infra Red Observational Satellite.) During the next few months we completed the re-design of the new spacecraft. My new assignment was to re-design the antenna system for the new spacecraft. The completed spacecraft TIROS-1, the World’s first weather satellite, was launched from Cape Canaveral on April 1, 1960, after which the first pictures of Earth and clouds ever taken by cameras in an orbiting weather satellite were transmitted back to Earth on the Author’s antennas during the first orbit. All systems were GO!
Photo 28-2.1 shows a flight model of the TIROS-1 weather satellite showing solar cells, cameras, and the Author’s multi-frequency crossed dipoles.
Overview of the Spacecraft RF System
The transmitting antenna system designed by the Author for TIROS 1, used for transmitting data obtained from the TV and IR cameras, comprised a multi-frequency array of two crossed sleeve dipoles consisting of four monopoles mounted on the bottom of the spacecraft. The antenna array was fed by four transmitters operating simultaneously on different frequencies in two frequency bands, 108 and 235 MHz; the TV transmitter on 235.0 MHz, the Infrared transmitter on 237.5 MHz, and two telemetry transmitters, one operating on 108.0 and the other on 108.03 MHz. The TV transmitter is shown in Photo 28-2.2 and the telemetry transmitter is shown in Photo 28-2.3.
The transmitters were coupled to the antenna array by a complex impedance matching network designed by the Author, comprising ring diplexers, baluns, frequency-selective filters, open- and short-circuit stubs, and phasing lines, all fabricated in 25, 50, and 70-ohm printed-circuit, sandwich-form stripline transmission lines. The network provided four outputs, one for each monopole, with progressive relative phases of 0, 90, 180, and 270° for feeding the two crossed dipoles in quadrature phase to obtain circular polarization in the 108-MHz telemetry band and the 235-MHz video band. (In later models the telemetry transmission frequencies were changed from the 108 MHz band to the then new 136 MHz band.) The two telemetry transmitters each delivered 30 milliwatts to the network, and each of the two video transmitters delivered five watts.
The well-known technique of scale modeling was used during the development of the antenna array for measuring both the radiation patterns of the antenna array and the terminal impedances. Taking many impedance measurements of various sleeve configurations at different frequencies in the experimental process leading toward the final configuration was simplified by using only one monopole mounted on a half-shell scale model over a ground plane, because it avoided the difficult obstacles involved in measuring through a complex feed system that would be required if two dipoles or four monopoles had to be fed simultaneously. Only one monopole was needed for obtaining valid impedance measurements during the development, because its image and that of the half-shell model of the spacecraft were both formed in the four-foot-square aluminum ground plane on which the model was mounted. The Author used the arrangement shown in Photo 28-2.4 to make slotted-line impedance measurements during the experimental development. The tiny sleeve monopole is just barely visible projecting upward from the top of the half-shell model. Using the techniques of scale modeling and ground-plane imaging achieves the same results as with a full-scale model. However, scale modeling was preferred here because it permitted using smaller components and a smaller work space for conducting the experiments. The model was scaled down four-to-one from the full size. Hence, the test frequencies were 430-440 MHz for the telemetry band and 930-950 MHz for the video band, four times the actual operating frequencies. The equipment used in the impedance measurements were Hewlett-Packard HP-608C and HP-612A signal generators, an HP-805A slotted line, and an HP-415B SWR indicator, seen located on the bench at the left of the ground plane. A slide rule and a Smith Chart were used for correlating the impedance data obtained from the measurements. Hand calculators had not yet been developed.
It should be noted that radiation pattern and impedance measurements obtained for the earlier spy satellite that was scrapped were also made using scale models, but with five-to-one scale. The scale models for the spy satellite appear in Photo 28-2.5, and the scale model for TIROS 1 appears in Photo 28-2.6.
Photo 28-2.7 shows the Author at the RCA Laboratories working on the crossed-dipole transmitting antenna system of TIROS-1. The spacecraft mockup in this photo is a full-scale electrical model constructed of sheet aluminum over a wood frame, used for full-scale testing all of the spacecraft antenna systems during the design phase. The model was constructed to be electrically identical to the flight models to obtain valid measurements of impedance and radiation patterns as various antenna configurations emerged during the design phase. The makeshift mounting for the model was used temporarily while the antenna test laboratory was being constructed at RCA for the TIROS project. An unused flight model of TIROS-1 with the Author’s antennas attached is on display in the spacecraft museum of the Smithsonian Institution in Washington. The writing that follows describes in detail the progress of the design of the revised antenna system.
The General Approach to the Design of the Antenna System
There were two separate, but related projects in the development of the antenna system, the physical radiators of the antenna array required to radiate on four different frequencies simultaneously, and the coupling and impedance-matching networks to couple four transmitters to the antenna array, with all transmitters operating simultaneously on different frequencies without mutual interference. We will first discuss the requirements and development of the physical radiators of the antenna array.
1. The array must radiate efficiently on four specifically assigned frequencies in two frequency bands, 108 and 236 MHz. The specific frequencies are 108, 108.03, 235, and 237.5 MHz.
2. The array must radiate on all four frequencies with a radiation pattern shape that assures a reliable signal level received at the ground stations, regardless of the attitude of the spacecraft while spinning on its axis.
3. To satisfy requirement 2, all radiation from the array must be of circular polarization.
To obtain circular polarization from the array the basic configuration of two orthogonally positioned dipoles was determined to be the correct choice, i.e. two dipoles with each half in the same plane, mounted on the bottom side of the spacecraft. However, because the third-stage rocket is attached to the spacecraft at the center of the spacecraft, to accommodate the rocket each half of the two dipoles was separated radially into four separate monopoles attached to a mounting ring surrounding the rocket-mounting location. Thus, the antenna array comprises four monopole radiators extending from the bottom section of the spacecraft at a downward angle of 45° spaced equally around the spacecraft. Each radiator comprises a sleeve section with a coaxial rod extension. The sleeve section is the primary radiator for the frequencies in the 236 MHz band and the combination of the sleeve and rod extension is the radiator for the frequencies in the 108 MHz band. Unfortunately, there is no layout diagram available for the physical characteristics of the sleeve and rod radiator. However, the four monopoles can be seen pictorially in Photos 28-2.1 and 28-2.9.
The effort in the development of the four monopoles was greatly simplified by knowing that all four in the array can be identical, thus allowing the impedance-measurements during the design phase to be focused only on a single monopole. Attempting to feed all four development monopoles simultaneously and making design changes in four instead of one would have been foolhardy and practically impossible. Further, the mutual coupling between two opposing monopoles in the array must be considered, because these two monopoles comprise a dipole whose input terminals are separated by a distance greater than in a normal dipole configuration. Consequently, additional simplification was achieved by eliminating the opposing monopole and simulating it with a single monopole appropriately positioned on a half-shell portion of the spacecraft mounted on an extensive ground plane. The mirror image of the single monopole appearing in the ground plane substituted for the opposing monopole, thus satisfying both the electrical and physical requirements for measuring the impedance of the single monopole.
We took the simplification of the design effort even further by using scale modeling during the development, thus reducing both the size of the design model and the space required to make valid impedance and radiation-pattern measurements. Accordingly, the size of the model was reduced by a ratio of four-to-one, with a corresponding increase in frequencies by a ratio of one-to-four. The simplified design and development model used for impedance measurements appears in Photo 28-2.4, just barely showing the half dipole (monopole) positioned on the four-to-one size half portion of the spacecraft mounted on a four-by-four foot square aluminum ground plane. Also shown in the photo is the Hewlett-Packard HP-805A Slotted Line, with which all impedance measurements (hundreds) were made and plotted on a Smith Chart. At that time, in 1958, the slotted line was the most advanced instrument available for measuring impedances at the frequencies involved. In addition, all calculations were performed on a slide rule, because desk-top computers and hand-held calculators had not yet been developed. The half-portion of the model is also shown in Photo 28-2.6. (Five-to-one scale models were used in the development of the antenna system for the earlier spy satellite, shown in Photo 28-2.5. The model at the top was used for measuring the radiation patterns and the half-portion shown at the bottom was used along with the aluminum ground plane for measuring impedances of the antenna, as seen in Photo 28-2.4.)
Signal generators used during the early antenna development were Hewlett-Packard HP-608C and HP-612. The HP-415B SWR indicator was used with the HP-805A slotted line for measuring the impedances of the scale model versions. The impedance measurements on the full-scale electrical test model of the spacecraft, shown in Photo 28-2.7, and all final testing of the flight model antennas, were made using the HP-608C signal generator, Polytechnic Research and Development’s Standing Wave Detector (reflectometer) PRD-219 and the HP-415B SWR indicator. The PRD-219 indicates the magnitude of the reflection coefficient as a measure of the SWR, and the angle of the reflection coefficient is indicated at the null in the SWR curve, required to obtain the complex impedance appearing at the terminals of the antenna.
Delving more deeply into the description of the radiator assembly we begin with the base mounting, which was machined to permit a tight fit for the aluminum tubing that serves as the sleeve radiator for the 236-MHz frequencies, as shown in Photo 28-2.10.
The inner portion of the base was threaded to accommodate length adjustment of the rod extension for the 108 MHz frequencies. Because the sleeve radiator is approximately /4 in length, both inside and out, in conjunction with the coaxial rod extension as the center conductor, the inner portion of the sleeve becomes a shorted /4 section of coaxial transmission line at 236 MHz. The shorted /4 inner section thus performs as a trap at 236 MHz, and isolates the portion of the rod extending beyond the sleeve at 236 MHz. Therefore, the input impedance at 236 MHz is determined largely, but not entirely, by the length and diameter of the sleeve, and the mutual coupling of the collinear relationship between the sleeve and the extended portion of the rod.
For ease in developing the coupling and matching unit it was desirable to focus the mechanical design to achieve the terminal impedance of 50 + j0 ohms for the 236-MHz radiators. Then, by designing the matching unit to exhibit characteristic impedance Z0 of 50 ohms throughout, no impedance matching was required to match the radiator to the 236-MHz transmitters. This feature simplified the design in that the only impedance matching circuitry required is for operation at 108 MHz. However, forcing the terminal impedance of the 236 MHz radiators to a resistive 50 ohms required a little fancy manipulation. First off, the bare /4-length sleeve yielded a resonant impedance of approximately 35 ohms. How can we increase it to 50 ohms? Black magic? Not really—just a little creative engineering.
It was discovered that by alternately varying the electrical length of the internal /4 shorted section of the sleeve and the length of the rod extension, we could control the mutual coupling between the rod extension and the sleeve, and thus force the input impedance of the half dipole to be 50 + j0 ohms at 236 MHz. A moveable Teflon slug was inserted into the sleeve to vary the electrical length of the internal section. A narrow slot in the sleeve allowed adjustment of the slug position to obtain the required electrical length of the shorted section. A threaded screw brazed to the end of the rod extension and screwed into the mounting base allowed varying the length of the rod extension to control the mutual impedance between the sleeve and the rod extension.
The final length of the rod extension (adjusted by turning the rod) determined by obtaining the 50 + j0 impedance at 236 MHz, simultaneously determined the input impedance at 108 MHz to be 150 – j100 ohms. This is the impedance that requires matching to 50 ohms by the circuitry in the coupling and matching network fabricated in stripline, described in the next section.
The Stripline Coupling and Matching Network
In Chapter 1 we briefly mentioned the spacecraft antenna system developed by the Author that transmitted signals on four different frequencies from a single antenna array. In this section we’ll describe and illustrate the multi-frequency antenna coupler used to couple four transmitters operating simultaneously on four different frequencies into one antenna array. Two frequency bands were used on the TIROS 1, 2, and 3 spacecraft, at 108 and 236 MHz, with two beacon frequencies at 108.0 and 108.03 MHz; the TV, or video signal frequency at 235.0 MHz, and infra red (IR) signal frequency at 237.5 MHz, one transmitter on each of the four frequencies. However, beginning with TIROS 4, these two original beacon frequencies were changed to 136.23 and 136.92 MHz on all subsequent spacecraft through TIROS 8 to avoid interference with aircraft frequencies. The corresponding sections of the later antenna couplers were redesigned to 136.57 MHz, the mean frequency between the two operating frequencies. The TV and IR frequencies remained the same as before.
The problem when feeding multiple transmitters simultaneously into one antenna array is that of isolating the transmitters from each other to prevent mutual interference. Using frequency-selective isolation filters between transmitters of the two frequency bands, and using ring hybrids to couple transmitters in the same band solved the problem. The principles used in the solution are not new, having been used for decades. For example, two AM broadcast stations operating simultaneously on different frequencies, but using a common tower antenna. It is usual in these cases to construct the filter networks with lumped capacitors and inductors.
However, our solution was radically different—no lumped components. To save on precious weight and space the isolation and impedance matching unit was constructed entirely with stripline transmission-line circuitry (not microstrip), fabricated on printed circuit board. In stripline the center conductor is a flat ribbon conductor sandwiched between two flat layers of dielectric insulating material separating it from two outer layers of thin copper foil, with the outer foil becoming the outer conductor. (Think of coaxial line compressed until it’s completely flat.) The width of the ribbon center conductor and the dielectric constant and thickness of the insulating material determine the impedance of the transmission line. In our case the thickness of each insulating layer is 1/16”, and the dielectric material is Teflon-fiberglass, with a dielectric constant of 2.42. Consequently, the completed transmission line is 1/8th inch thick.
Diagrams of the transmission-line coupling and matching circuitry for TIROS spacecrafts 1, 2, and 3, along with the operating frequencies, appear in Figs 28-2.1, 28-2.2 and 28-2.3, respectively. The inside view of the printed circuits for the 136 and 235 MHz coupling and matching transmission line layout are shown in Figs 28-2.4 and 28-2.5, respectively.
Because this project was our first experience in working with stripline, several test strips were fabricated to determine whether the characteristic impedance Z0 and electrical length could be controlled to sufficiently-close tolerances to construct a workable system using stripline printed circuitry. It would have been nice to have fabricated the entire coupler in one unit, but with our lack of experience with stripline, coupled with a tight time schedule that could not tolerate the required learning curve, the initial coupler was constructed using six separate printed circuit boards interconnected with an unbelievable number of RG-141 cables and BNC elbow connectors. The six boards consisted of two frequency filters, two hybrid rings, and four baluns. However, the coupling network was redesigned for TIROS 2 and 3. Along with the experience gained while working with the new transmission medium with the first spacecraft, we bit the bullet and designed the new coupler into a single unit. The advantage in weight saving with the new design was tremendous, because the interconnecting cables and heavy elbow BNC connectors were eliminated, plus the complex mounting hardware for the six circuit boards, and the reduction in number of boards from six to two. The two final boards were sandwiched together into one board, thus eliminating the interconnecting cables and connectors.
We now show the steps taken during the matching procedure for the 108 MHz stripline coupler of TIROS 2 and 3 that resulted in the circuit diagram appearing in Fig 28-2.2, along with the corresponding Smith Chart presentation of Fig 28-2.6.
Degrees R (°R) = reflection degrees = electrical length in degrees 2, and represents angular distance around the periphery of the Smith Chart.
Lower case y and z represent normalized values of Y and Z
G = 1 defines the unity conductance circle on Smith Chart
ZM = Input Impedance of single monopole
ZM = 4.4:1 VSWR at –18° = 150 – j100 ohms. (zM = 3 – j2)
YM = 1/ZM = 4.4:1 VSWR at +161.5° (yM = 0.24 + j0.16)
G = 1 matching point at 4.4:1 VSWR occurs at –51°
L1 = distance from YM to G = 1 circle = 212.5°R
yM at G = 1 circle = (1 – jβ) = (1 – j1.62), residual susceptance jβ = – j1.62
Residual susceptance jβ cancelled at B, E, N, and Q by short-circuited lines L2 providing susceptance jβ = + j1.62
Short circuits on L2 at 108 MHz established by open-circuit stubs at C, F, O, and R
Normalized reactance jx of L2 = reciprocal of L2 susceptance, 1/j = 1/j1.62 = – j0.617
Length of L2 = jx = 1/j = – j0.617 = 148.33° (296.65°R)
L2 = 148.33° electrical length of line fabricated in stripline matching unit.
In Photo 28-2.17 I am on the 12th story of the gantry of launch pad 17A, peering into the fairing, or nose cone, to determine if the ends of the dipoles clear the inside of the fairing.
In Photo 28-2.8 I’m preparing to take radiation patterns with an empty third-stage rocket attached, to determine whether the rocket has any effect on the radiation patterns that might disturb operations during the launch phase.
We have described the antenna array and its coupling and matching unit. However, this portion of the description is only a small part of the total story of the development of the antenna array. For those who are interested in learning more about this subject I invite you to obtain a copy of Reflections 3 for the additional material.
There are several photos, that will require additional posts to comply with the rules.
Sorry, having trouble inserting the pics.