Ref: ADC-0807201342 R0
Bruce A. Blevins
Antenna Development Corporation
Office Phone (575) 541-9319
Cell (575) 635-3528
Antennas are devices which confine accelerating charges to produce or receive far field traveling wave energy with the desired directional and polarization properties. The trick is to build a structure that is reasonably sized, light weight, robust, inexpensive, and reliable and having all of the desired electromagnetic properties. A difficulty is that these devices follow certain physical rules (described by Maxwell's equations) which are not always intuitively obvious and having the potential to be pesky.
This paper will attempt to educate the reader by presenting some of the electromagnetic concepts and by describing some properties of practical antennas. We will also describe some of the methods used to measure antenna performance and conclude with a short discussion of some of the costs associated with antenna design, manufacture, and test.
To review, let us talk about some terminology and simple physical limits which can be used as guidelines for initial antenna system design. These apply to both ground-based and satellite-based antennas.
An isotropic antenna is one that radiates equally in all directions. True isotropic antennas are hard to realize since it is very difficult, if not impossible, to have a truly spherically symmetric antenna. For example, many practical antennas require a transmission line connecting it to a power source.
The antenna’s radiation in the general direction of maximum radiation intensity. As an example an isotropic antenna has no clearly defined main beam whereas a properly designed reflector antenna has a well defined main beam.
Radiation peaks other than the main lobe. Antenna users usually want low sidelobes to prevent reception or transmission of radiation in directions off of the main beam.
The vector electric and magnetic fields of free space traveling waves are mutually orthogonal and perpendicular to the direction of travel. Polarization describes how the fields behave in time and space. For example, a "circularly polarized" wave can be thought of as having the instantaneous electric field rotating about the direction of travel. The direction of the electric field vector rotates one turn per period of the wave (either in space or time). A "linearly polarized" wave radiated in a fixed vector direction - the variations over space and time are in magnitude rather than direction.
In general an antenna radiates waves with an elliptical polarization. Measuring in a plane perpendicular to the direction of travel, the axial ratio is the ratio of the absolute value of the maximum observed electric field vector to the absolute value of the minimum observed electric field vector. Antennas with true circular polarization have an axial ratio is 1:1 (0 dB), a perfectly linearly polarized antenna would have an axial ratio of 1:0 (infinite dB).
The ratio of the main beam radiation intensity to the intensity that would be measured if the antenna radiated isotropically. The directivity is not changed by the power efficiency of the antenna. The antenna could be highly directive and radiate only a fraction of the supplied power.
The physical length of a single cycle of a single frequency wave in a given medium.
The span of frequencies over which the antenna has sufficient performance to meet specifications. There are many different types of bandwidths including VSWR, gain, and polarization. (For example, a typical bandwidth specification might be: The antenna gain with respect to isotropic must be greater than 10 dB +/- 10 MHz from a center frequency of 2092.0 MHz).
Voltage Standing Wave Ratio - the ratio of the peak detected (demodulated) voltage to the minimum detected voltage on a transmission line. The VSWR varies from 1.0 (a perfect match) to a maximum VSWR of infinity (a perfect mismatch). VSWR is a measure of the magnitude of the reflected wave caused by impedance mismatch between the load and the transmission line.
The power received in the main beam direction from the antenna under consideration divided by the power that would be received from a theoretical isotropic antenna with the same polarization supplied with the same power as the antenna under consideration. This definition accounts for antenna mismatch between the source and the antenna
The antenna gain is one of the most important antenna equations. The equation gives the gain as proportional to the "effective area" of the antenna:
Gain = 4 · pi · Ae/(wavelength*wavelength) where:
· wavelength = c/f (meters)
· c = speed of light (meters/sec)
· f = frequency of wave (Hz or cycles per second for you really old timers)
· Ae, the "effective area" is related to the cross sectional area of the antenna (like a reflector antenna or a horn antenna) and/or to the volume of the antenna (like a helix or a dipole antenna). As reference, reflector antennas typically have effective areas of about 50 % of the actual projected physical area of the reflector surface.
This relation and others are frequently not considered or understood by non-antenna design engineer when they and ask questions like "Can you deliver a high gain, omni-directional, efficient transmit antenna which operates with a frequency of 100 MHz +/- 75 MHz and which can fit on the surface of a 2 foot diameter satellite." Most antenna engineers have had variations of this question asked more than once by otherwise well informed technologists. Such a request violates several aspects of the equations and concepts we have presented:
High gain and true omni-directional patterns are mutually exclusive. The highest gain a true omni-directional antenna can have is 1 (0 dB).
High gain and a small electrical size (determined by using a wavelength as a unit of measurement) are mutually exclusive. It is possible, however, to have a highly directive electrically small antenna. Such an antenna will not be very efficient and therefore will not have high gain. (There is a continuing debate among antenna engineers as to the practicality of a super conducting electrically small high gain antenna - we will ignore such antennas in this discussion.)
This ratio is antenna gain (G) divided by the noise temperature (T) of the receiving system. G/T is an important consideration for ground-based antennas. A ground-based antenna has sidelobes which "see" the Earth. Since the Earth has a higher temperature than space (373 Kelvin for the Earth versus about 4 Kelvin for space) an antenna with large sidelobes will have a poorer G/T than one with small sidelobes. Communication system performance is usually proportional to G/T.
Since the most common application for small satellite antennas is the communication of commands to the satellite, telemetry from the satellite, and payload communications both to and from the satellite, we will briefly review some communications system considerations. We will concentrate on digital satellite communications because most modern systems are digital designs.
Important satellite communications system parameters include:
Bit error rate. A statistical measure of the number of bits in error versus the total number of bits transmitted. Some applications can tolerate relatively high bit error rates. For example, if a satellite like LANDSAT is transmitting an image of the Earth, a bit error rate of 1 in 10^5 may be acceptable. On the other hand, transmission of financial data between two banks requires a vanishingly small bit error rate.
The BER as a function of the Eb/No. The better the demodulator and bit synchronizer, the lower bit error rate for a given bit energy to noise ratio.
The ratio of total power in the modulated carrier to the noise power in a 1 Hz bandwidth. Usually measured at the input to the demodulator used to recover the digital information from the modulated signal.
The ratio of the energy per bit and the noise power in a 1 Hz bandwidth. Provides a bit error rate independent measure of the quality of the communication system.
The ratio of the actual Eb/No to the required Eb/No for the designed BER. The higher the system margin, the lower the BER and the more likely the system can withstand upsets like attenuation caused by rain or dust storms.
The above communication system parameters require detailed study prior to their application in satellite communications system design. The designer starts with some idea of the BER the system can tolerate and the required throughput (bits/sec) between the satellite and the ground. The designer is also constrained by certain regulatory requirements as to available frequency bands for space communications. Given these restrictions, a first cut estimate is then made on the size, volume, complexity, and weight available for the satellite antenna and electronics. A preliminary selection of the frequency for the link can then be made. Such a selection depends on the spacecraft assets available and the required bandwidth (depends on the data rate and modulation method). Link margins based on estimates of the satellite transmit power, satellite antenna gain, coding gain, path loss, atmospheric attenuation, ground antenna gain, and demodulator implementation loss can then be calculated. Note that the link margins are critically dependent on both the satellite-based and ground-based antenna gains as well as the G/T's of both antennas (assuming two way data flow). Details of the antenna performance are therefore important.
Small inexpensive satellites are subject to several restrictions not imposed on large expensive satellites. The key for successful small satellite antenna design is to use simple antennas with low weight and size demands. The problem is that many small satellite designs demand many of the performance capabilities common to large satellites. Some of the requirements are reviewed below:
Telemetry and Command and Communications Relay Frequencies
There are several frequency bands reserved for satellite control and monitoring, data download, and communications relay use. These include the VHF band (about 150 MHz), S-band (2.0 to 4.0 GHz), C-band (4.0 to 8.0 GHz), and Ku-band (12.0 to 18.0 GHz) as well as the 20/30 GHz bands. Please refer to official regulatory sites for definitive frequency limits.
Payload Scientific Applications Frequencies
Scientific application receive frequencies are essentially unlimited. Antennas have been designed which receive photons with frequencies from as low as 1 kHz to 1 THz (gamma rays). This paper will consider general communications antennas and will not attempt to review these sometimes exotic antenna designs.
Weight and Size
Small satellites are just that - physically small and light weight. Antennas designed for small satellites do not usually dominate the satellite profile or the weight budget.
Deployment and Pointing
To keep the total satellite package cost within reasonable bounds, deployment mechanisms must be simple and reliable. Complicated stowable-parabolic antennas like those used in NASA's Tracking and Data Relay Satellites (TDRS) are impractical for almost all small satellite designs. A restriction implied by limiting the complexity of antenna deployment mechanisms is that high gain antennas (more than 15 dB or so) are hard to realize on small satellites. Further, high gain antennas have narrow main beams. Small satellites usually cannot afford to provide the resources required to accurately point high gain antennas.
Large satellites are expensive. Extensive antenna simulations and expensive antenna range testing usually represent reasonably small fractions of the total satellite budgets. Managers can therefore relatively easily accept the premise that such simulation and test are necessary for the success of the mission. Small satellite projects do not usually have money to burn - limited simulation and test of accepted designs may be sufficient and adequate for risk reduction.
Small satellites usually require modest gain circularly polarized antennas. Historically, nearly every satellite has had custom antennas developed specifically for the satellite. Production quantities have been very low and the costs are necessarily higher than terrestrial antennas. The antennas for small satellites are generally of modest gain because the antenna has limited dimensions and cannot be actively pointed. They are usually circularly polarized since many small satellites are spin stabilized. A spin stabilized antenna is by definition rotating. If a linearly polarized antenna is used on a spinning satellite, the ground system will have to receive a so called "spinning linear" signal. This is more difficult and increases complexity or decreases system margin. We will now describe the characteristics of several types of antennas which can be used to easily generate modest gains with good linear or circular polarization performance over significant bandwidths.
A printed circuit antenna consisting of a radiating patch supported by a dielectric layer over a ground plane.
A helical wire wound with a circumference of about one wavelength and a pitch of 1/4 wavelength over a ground plane with a 1 wavelength minimum diameter.
Center-fed radiating elements each about 1/2 wavelength long. Rods are parallel to and over a ground plane of about 1/4 wavelength minimum diameter.
The canonical "whip" antenna. Similar to the standard AM car antenna. The antenna is usually a simple projecting from a ground plane. If the antenna is resonant then it is about 1/4 wavelength long, if sub-resonant then it can be much shorter than a wavelength.
Similar to a short helix in appearance, broader beam, circular polarization.
Uses a small cutout in a ground plane - can conform to the skin of the satellite.
The usual dish antenna, may be used for higher gain on small satellites or for ground system use.
Wide band - may be cavity backed, two or three dimensional.
Often used as the "feed" for reflector antennas or can be stand alone.
Small satellite ground terminals usually compensate for modest satellite antenna gains by using much higher gain ground antennas. To eliminate polarization mismatch, the ground antennas are usually circularly polarized. Ground noise is minimized by limiting the ground antenna's sidelobe levels to improve the ground system's G/T. Since the ground antenna has higher gain (and a consequently narrow beam width) the ground antenna usually tracks the satellite at it passes through the ground station's field of view. These general requirements coupled with budget forces usually drive the selection of a reflector-type ground antenna or sometimes high gain helix antennas.
Reflector antennas are the usual satellite dish antennas that are seen almost everywhere. These antennas are either front focus fed or Cassegrain fed and can be purchased from a myriad of suppliers. Commercial market forces have driven the prices for the low end reflector antennas down to very reasonable levels. Several small firms sell reasonably priced satellite tracking mounts for reflector antennas.
Ground based high gain helix antennas are essentially identical to those designed for satellite applications. Small arrays of helix elements are sometimes used to obtain even higher gain than available from a single element.
Most antenna designs are well enough known to enable accurate performance predictions -the rub comes when the customer really wants to be assured that the antenna meets specifications when on the spacecraft. Often times the antenna is placed on a platform that somehow compromises ideal antenna design or somehow restricts the space available for an ideal antenna. Then the problem becomes - "How well can analytical analysis tools predict the compromised antenna performance?" The system builders (the satellite team and the antenna team) often must resort to testing because simulation errors are generally poorly known and are therefore not trusted by the antenna user. This section of this paper will outline the various testing methods used and touch on some simulation computations available to the antenna designer/builder.
In an attempt to predict actual antenna performance in free space, the antenna engineer may use a far field antenna range. Such ranges separate the antenna under test from a monitoring antenna by a distance sufficient for the radiated fields to be nearly plane waves. The antenna under test is then rotated through various orientations and the coupling between the antenna under test and the monitoring antenna is recorded. Problems with this method include difficulties with separating the antennas by enough distance, interaction of the radiation with support structures, and interfering reflections from the ground and other nearby objects as well as costly positioners and range real estate
Near field measurements use a so-called near field scanner. The electromagnetic fields near the antenna under test are received, digitized, and extrapolated (through the Fourier transform) to the far field. The measurements must be performed to a relatively high degree of mechanical and electrical accuracy and require significantly complicated computer software analysis procedures. Small satellites have relatively modest scanner requirements - however, near field measurements are demanding and the equipment has costs comparable to those required by small far field antenna ranges.
Antenna responses may be measured in a room which has unavoidable reflectors. The problem is how to remove the effects of these unwanted scatters. One technique involves measuring the amplitude and phase response over a wide frequency span. Transformation of the data from the frequency to time domain by a Fourier transform will result in temporally separated contributions from the interfering reflectors as well as the main response at the time delay associated with the antenna under test. The unwanted time domain signals may be removed by editing the time domain data set. The resulting data set is then transformed back to the frequency domain by an inverse Fourier transform. The final result is the desired antenna response with the undesirable reflections removed. This technique has the drawbacks of requiring expensive measuring equipment as well as requiring significant computations for each antenna orientation angle.
Far field measurements can be approximated by measurements in a room lined with antireflection material. Again, the antenna under test is supported and rotated through a set of orientations to develop the antenna patterns. Problems with this approach include reflections from imperfect anechoic material and difficulty reaching the far field distance inside of the confines of the test chamber.
A refinement of anechoic chamber techniques is the compact antenna range. A large reflector is installed in the anechoic chamber and is illuminated by a feed horn. The near field region of the beam produced by the reflector closely approximates a plane wave. The antenna under test is placed in this region and is measured at various orientations. Problems with this refinement include the standard anechoic chamber reflection problems as well as the costs associated with a large precision reflector.
Both anechoic chamber techniques have low frequency limits imposed by the antireflection properties of the anechoic material.
Antenna terminal impedances must be matched to the transmission line impedance to maximize the conversion of transmission line power to radiated power. Low frequency measurements may be measured with vector voltmeters, high frequency with slotted lines or network analyzers.
Over the last 40 years or so the United States Government has funded the development of several computer codes which can compute the electromagnetic properties of antennas on satellite bodies. In general, these codes are not very user friendly and are really intended for the specialist. Commercial codes are becoming available which are more user friendly. In both cases, it is very easy to misuse the code and produce results which are misleading if not completely incorrect. We list several of the more popular codes here to bring out the flavor of such computations.
(NEC, Burke & Poggio) - Developed at Livermore National Laboratory during the 70's - now very popular. This program uses the method of moments (a finite element electromagnetic computation technique) to compute the mutual impedances of wires which have been divided into segments of less than 1/10 wavelength. A matrix equation relating the impedance matrix of the structure, the excitation voltages applied to the structure (either from transmission lines connected to the antenna or plane waves incident on the antenna), and the individual segment currents can be written. This matrix equation can be inverted to solve for the unknown currents in the segments. A solution for the currents provides the information required to calculate the far field radiation pattern. This program can be very useful for the small satellite antenna designer since many small satellites have requirements for antennas with relatively large wavelengths.
(NEC-REF, Chang, Y.C. & Ruddick, R.C., Technical Report 712242-17
(713742)) - Developed at
(NEC-BSC, Marhefka, R.J.)
- also developed by
The original wire code (NEC) has essentially entered the public domain and
is widely available both in the original government form and in several
commercial variants. The other two codes are distributed by the Department of
the Navy, Navy Regional Contracting Office,
Successful small satellite antenna design and development involves
experience, simulation, some intuition, and accurate measurements. There has
been a tendency for the non-specialist to view antennas and microwaves as a
black art. It really isn't a black art - rather, certain aspects are counter
intuitive and require knowledge of electromagnetics. The small satellite maker
usually cannot afford the costly development programs used in the past and
probably should depend on the expertise of entrepreneurs with experience.
Last Updated Sunday, 20 July, 2008