The Ka-band (20/30GHz) satellite offers excellent potential for transportable and mobile communications due to the wide bandwidth available and the high gain achievable with small antennas. There are some difficult technical challenges in realizing this potential. While relatively small antennas, such as 500-1000mm parabolic reflectors, give high gain of the order of 40dB the beam widths are quite narrow, of the order of 1 degree, and the antenna has to operate at both the 20GHz receive band and the 30GHz transmit band.
Compact, high efficiency, high linearity power amplifiers are required to minimise the size, weight, power consumption and the heat dissipated. At 30GHz, gain, power and linearity of available FET devices are much more constrained than at other satellite bands such as X and Ku so flexible, low loss combining techniques are necessary. For the receive side, the antenna noise temperature is relatively low due to the narrow beam width, and so the noise figure of the receiver must be as low as possible to avoid degrading the signal. The performance of a satellite channel is partly determined by the phase noise of the local oscillators used to upconvert and downconvert the IF signals onto the satellite frequencies. Low phase noise becomes more difficult to achieve as the frequency increases so it is essential to employ low phase noise, high stability design techniques for the 20/30GHz oscillators used in the converters.
This article describes suitable antennas, solid state power amplifiers (SSPAs), low noise amplifiers (LNAs) and local oscillators developed for transportable and mobile Ka-band satellite terminals.
The parabolic reflector antenna offers the best compromise between antenna size, gain and bandwidth of all the various competing technologies. It has a further advantage compared to some of the planar array type antennas that have attractive flat layouts. This is that the transmit and receive beams are precisely aligned on the same bore-sight at all frequencies in the transmit and receive pass bands so that pointing/tracking is considerably simplified. A 500mm reflector is electrically large enough to permit a Cassegrain type feed to be used with good efficiency. In this case, a wideband feed horn to cover the 20-30GHz range can be designed so that a single feed is possible. Furthermore, it is possible to design the feed for linear and/or circular polarization and to make it switchable between the two which offers a degree of flexibility not often possible with other types of antennas. The full circular symmetry of the antenna and the antenna beam reduces the complexity of tracking and alignment for mobile and transportable systems.
The saturated power level of single solid state 30GHz devices currently available is less than 10W. The linear power is considerably less than this at typically 2 to 3W as 30GHz devices are significantly more non linear than X and Ku devices.
To obtain moderate levels of 10-20W of linear power, as is typically required for Ka-band terminals to provide say 2Mbps data with small (500mm) antennas, requires a number of these devices to be combined. There are several ways of doing this including free space combining, radial/parallel plate networks and rectangular waveguide networks. For combining small to moderate numbers of devices, rectangular waveguide networks are very suitable at 30GHz for both electromagnetic and mechanical reasons. The networks can be accurately designed, the insertion loss is low, additional elements such as filters, phase adjusters etc can be integrated into the networks and the physical circuit can support the modules and act as a heatsink and part of the overall package.
A conventional combining technique is to use 3dB quadrature couplers arranged in a parallel network. This has advantages for simplicity of design, good phase tracking and isolation properties and reasonable bandwidth. It is inflexible in terms of the number of modules to be combined (2, 4, 8 etc.) and the layout. A serial combining technique involves some electromagnetic compromises but offers much more flexibility in terms of the number of modules that can be combined and the physical layout of the modules. Using a combination of serial and parallel combining techniques, it is possible to combine up to 100 modules at Ka-band (>200 modules at X band) before the waveguide insertion loss becomes too high, but for small terminals more typical numbers are 4 to 10 modules.
The photographs below show power devices on a conventional combiner and the combiner network.
While it is always possible to increase the linear power by adding modules and increasing the saturated power, at least up to the limit of about 100 modules, this is not an effective technique for portable/mobile terminals where size, power consumption and heat dissipation are strong constraints. Another option is to add a linearizer to the SSPA to improve the linear power for the same saturated output power.
A 20W saturated power SSPA using 4 modules for example can be operated at 10W linear power when a linearizer is added. Without the linearizer, the saturated power would have to be increased to about 40W which would mean using 8-10 modules so doubling the size, power consumption and heat dissipation. A well adjusted linearizer can reduce the intermodulation levels by around 10 dB over a wide power range and the effective power level can be increased by 2 to 4 times depending on the particular operational parameters selected. The improvement in intermodulation level (IM) over a wide power range when a linearizer is added to the SSPA is clear from the following plot. (See diagram on the next page.)
Low Noise Amplifier (LNA)
A miniature LNA with a noise figure of 1.3dB across the 20-21GHz receive band is suitable for transportable of mobile terminals. The LNA has sufficient gain of around 30dB so that the rest of the receiver does not degrade the noise performance. To lower the noise figure further by a significant amount would require the LNA to be cooled to -20C or less which is not practical for these types of terminals.
Local Oscillators (LO)
The LOs in a mobile Ka-band terminal that operates with data rates from say 9.6kb/s to 2Mb/s must have good phase noise, high stability and low microphonics if the satellite link is to be stable with vehicle motion. To obtain good phase noise at high frequencies, fixed frequency LOs are used in a block conversion method with the IF frequency range being typically 1-2GHz. A low phase noise voltage controlled oscillator (VCO) is multiplied up to the final frequency of around 29 and 19GHz for the up and down converter respectively. A special multi loop phase lock loop design is used to reduce the vibration sensitivity (microphonics), and the VCO is locked to a 5MHz or 10MHz reference to deliver the high frequency stability necessary. All of this can be done in a compact, low power circuit consistent with the demand for small size, low power consumption and heat dissipation of a mobile terminal.
The short wavelength at Ka-band means that satellite terminals can be manufactured with small antennas and yet reasonable data rates can be achieved of 2Mbps or so, under clear sky conditions. The high frequency of Ka terminals places additional design and performance limitations on the active devices available and on the passive networks but these limitations can be significantly reduced or even used to advantage to reduce circuit size and increase packing density. Linearizers have a greater benefit at Ka-band than at lower frequencies not only because the non-linearities of the Ka-band devices are greater, but also because increasing power is much more expensive at Ka-band in terms of the number of devices required, and the increased DC power and thermal load to get even small increases in output power. These modules can be combined into compact terminals that are suitable for tracking and mobile applications where moderate data rates (~9.6kbps to 2Mbps) are required.
About EM Solutions
EM Solutions has been contracted to develop a Ka-band On-The-Move Satellite Communications System from Round 12 of Australias Defence Capability & Technology Demonstrator (CTD) Program. EM Solutions will partner with BAE Systems Australia on the CTD Project. The CTD Program supports priority defence capability development by funding Australian industry to demonstrate new technology. The technology demonstrations inform Defence of the potential performance and technical risk associated with future implementation.