Beamforming ICs Simplify Phased Array Antenna Design
Radar systems and wireless communications are facing increasing demands on antenna architectures to improve performance. Many new applications will only be possible with antennas that consume less power in a lower profile than traditional mechanically-steered dish antennas. These requirements are in addition to the desire to reposition quickly to a new threat or user, transmit multiple data streams, operate over longer lifetimes and meet aggressive cost targets. Some applications require nulling an incoming blocking signal and having low probability of intercept. These challenges are met with active electronically scanned array (AESA) designs that are sweeping the industry. Past disadvantages of phased array antennas are being addressed with advanced semiconductor technology, to reduce the size, weight and power of these solutions. This article describes existing antenna solutions and where AESAs have advantages. It will then discuss how semiconductor advancements are helping achieve the goals of improving SWaP-C, including examples of commercial technology making this possible.
Wireless electronic systems relying on antennas to send and receive signals have been operating for over 100 years. They continue to be improved as accuracy, efficiency and more advanced metrics become increasingly important. In past years, a dish antenna was widely used to transmit and receive signals where directivity was important, and many of those systems still work well at a relatively low-cost, reflecting years of optimization. These dish antennas with a mechanical arm to rotate the direction of radiation do have some drawbacks: slow to steer, physically large, poorer long-term reliability and only one desired radiation pattern or data stream. As a result, engineers have pushed toward advanced phased array antenna technology to improve these aspects and add new functionality. AESAs are electrically steered and offer numerous benefits compared to traditional mechanically-steered antenna, such as low profile with less volume, improved long-term reliability, fast steering and multiple beams. With these benefits, they are being adopted in military, SATCOM and 5G telecommunications applications, including connected automobiles.
Phased Array Technology
A phased array antenna is a collection of antenna elements assembled together, such that the radiation pattern of each individual element constructively combines with neighboring antennas to form an effective radiation pattern, called the main lobe. The main lobe transmits radiated energy in the desired location, while the antenna is designed to destructively interfere the signals in undesired directions, forming nulls and side lobes. The antenna array design maximizes the energy radiated in the main lobe, while reducing the energy radiated in the side lobes to an acceptable level. The direction of radiation can be manipulated by changing the phase of the signal fed into each antenna element. Figure 1 shows how adjusting the phase of the signal to each antenna in a linear array can steer the effective beam in the desired direction. Each antenna in the array has an independent phase and amplitude setting, which enables forming the desired radiation pattern.
Fast steering of the beam in a phased array is easily understood, since there are no mechanical moving parts. Semiconductor-based phase adjustments can be made in nanoseconds, such that the direction of the radiation pattern can be changed to respond to new threats or users quickly. Similarly, it is possible to change from a radiated beam to an effective null to absorb an interferer, making the object appear invisible, as would a stealth aircraft. These changes to reposition the radiation patterns or create effective nulls occur almost instantaneously because the phase settings are changed electrically with ICs, rather than mechanically.
An additional benefit of a phased array over a mechanical antenna is the ability to radiate multiple beams simultaneously, perhaps to track multiple targets or manage multiple data streams of user data. This is accomplished by digital signal processing of the multiple data streams at baseband frequencies.
The typical implementation of an AESA uses patch antenna elements configured in equally spaced rows and columns with a 4 × 4 design, equaling 16 elements (see Figure 2). Antenna arrays built from multiple 4 × 4 cells can grow quite large, e.g., more than 100,000 elements in ground-based radar systems.
There are design trade-offs with the size of the array versus the power of each radiating element, which determine the directivity of the beam, effective radiated power and other parameters. Antenna performance can be predicted from some common figures of merit: antenna gain, effective isotropic radiated power (EIRP) and Gt/Tn. The relationship among these is defined by the following equations:
The antenna gain and EIRP are directly proportional to the number of elements in the array. Achieving high EIRP can lead to the large arrays seen in ground-based radar applications.
Another key aspect of phased array antenna design is the spacing between the antenna elements. Once we have determined the system goals and set the number of elements, the physical diameter of the array is largely driven by the limit that each unit cell should be less than approximately a half-wavelength, to prevent grating lobes or energy radiated in undesired directions. This puts strict requirements on the electronics in the array, i.e., to be small, low-power and low weight. The half-wavelength spacing creates particularly challenging designs at higher frequencies, where the length of each unit cell becomes smaller. This drives the ICs at higher frequencies to be increasingly integrated, with packaging solutions that become more advanced and with simplified—though increasingly challenging—thermal management techniques.
Constructing the entire antenna poses many challenges for the array design, including control lines routing, power supply management, pulsed circuitry, thermal management and environmental considerations. A major push in the industry is toward low profile arrays that consume less volume and weight. The traditional “plank” architecture uses small PCB planks with electronics, fed perpendicularly into the backside of the antenna PCB. This approach has been improved over the past 20 years to continually reduce the size of the plank and the depth of the antenna.
Next-generation designs move from this plank architecture to a flat panel approach, where there is sufficient integration in each IC for them to fit on the backside of the antenna board, significantly reducing the depth of the antenna and enabling it to fit into portable or airborne applications. Figure 3 illustrates the flat panel approach, with the gold patch antenna elements on the topside of the PCB and the analog front-end feeding the antenna on the bottom side of the PCB. This shows only a subset of the antenna. A frequency conversion stage could be on one end of the antenna, and a distribution network routing from a single RF input to the entire array.
More integrated ICs significantly reduce the challenges in the antenna design. As the antenna becomes smaller, requiring more electronics be packed into a reduced footprint, more advanced semiconductor technology is needed to keep the architecture viable.
DIGITAL vs. ANALOG BEAMFORMING
Most AESAs that have been designed in past years use analog beamforming, where the phase adjustment occurs at the RF or IF frequencies, with only one set of data converters for the entire antenna. There is increased interest in digital beamforming, which uses one set of data converters at each antenna element, with the phase adjustment done digitally in the FPGA or data converter. Digital beamforming offers many benefits, starting with the ability to easily transmit many beams or change the number of beams almost instantly. This remarkable flexibility is attractive in many applications, which is driving adoption. Continuous improvements in the data converters—lowering power dissipation and expanding to higher frequencies, e.g., RF sampling at L- and S-Band—are making this technology a reality in radar systems.
Choosing between analog and digital beamforming requires multiple considerations, with the analysis usually driven by the number of beams required, power dissipation and cost targets. With a data converter at each element, the digital beamforming approach typically has higher power dissipation but offers flexibility in creating multiple beams. The data converters in digital beamforming must have greater dynamic range, since the beamforming that rejects blockers is done after digitization.
Analog beamforming can support multiple beams, however, it requires an additional phase adjustment channel per beam. Creating a 100-beam system, for example, multiplies the number of RF phase shifters by 100 compared to a single beam system. So the cost tradeoff between data converters and phase shifter ICs will depend on the number of beams required by the application.
Similarly, power dissipation is usually lower for an analog beamforming architecture that uses passive phase shifters; as the number of beams increases, the power dissipation will increase if additional gain stages are needed to drive the distribution network. A common compromise is hybrid beamforming, with sub-arrays of analog beamforming followed by digital combining of the sub-array signals. This is an area of growing interest and will continue to evolve.
A standard pulsed radar system transmits a signal which eventually reflects off one or more objects. The radar waits for the return pulses to map the field of view of the antenna. In prior generations, the radar front-end (see Figure 4) would use discrete components, likely fabricated in GaAs. The front-end comprises a phase shifter to adjust the phase of each antenna element and steer the antenna, an attenuator to taper the beam, a power amplifier (PA) to increase the power of the transmit signal, a low noise amplifier (LNA) to boost the receive signal and a switch to toggle between transmit (Tx) and receive (Rx). In past implementations, each of these ICs could be in a 5 mm × 5 mm package; more advanced solutions would have an integrated, single channel GaAs MMIC to achieve this functionality.
The recent proliferation of phased array antennas has been enabled by advances in semiconductor technology. The advanced nodes in SiGe BiCMOS, silicon on insulator and bulk CMOS enable the digital circuitry, used to control the steering of the beam, to be integrated with the RF, which performs phase and amplitude adjustment, in a single IC. It is possible to achieve multi-channel beamforming ICs for gain and phase adjustment, from four channels for lower frequency systems to 32 channels for mmWave designs. In some lower power applications, a silicon IC can monolithically integrate all these functions. In high-power applications, GaN PAs have significantly increased the power density sufficiently to fit into the unit cell of a phased array antenna, which traditionally would have been served by traveling wave tube (TWT) PAs or lower power GaAs PAs.
In airborne applications, the trend is toward flat panel architectures adopting the power-added efficiency (PAE) benefits of GaN technology. GaN has also enabled large ground-based radars to move to AESAs from traditional dish antennas driven by TWTs. Monolithic GaN ICs are capable of delivering greater than 100 W of power with over 50 percent PAE. Combining this level of PAE with the low duty cycle of radar applications enables surface-mount solutions to be feasible, greatly reducing the size, weight and cost of the antenna array. An additional benefit beyond the pure power capability of GaN is the size reduction compared to existing GaAs MMIC solutions. A GaN PA replacing a 6 to 8 W GaAs PA at X-Band reduces the footprint by 50 percent or more—very significant when trying to fit the electronics into the unit cell of a phased array antenna.
ANALOG PHASED ARRAY ICs
Analog Devices, among other companies, has developed integrated analog beamforming ICs aimed at a range of phased array applications, including radar, SATCOM and 5G. As one example of current technology and capability, Analog Devices’ ADAR1000 X-/Ku-Band beamforming IC covers 8 to 16 GHz, which makes it well-suited for X-Band radar and Ku-Band SATCOM applications. The four-channel device operates in time-division duplex mode, with the Tx and Rx integrated into a single IC, which can be configured to operate in Tx- or Rx-only mode. The four channels are integrated in a 7 mm × 7 mm QFN surface-mount package, compatible with integration into flat panel arrays. The IC dissipates only 240 mW per channel in Tx mode, 160 mW per channel in Rx. The Tx and Rx channels are designed to mate with a front-end module, such as those offered by Analog Devices.
Figure 5 shows the ADAR1000’s gain and phase performance. It provides full 360 degree phase coverage with phase steps less than 2.8 degrees and greater than 31 dB gain control. The device contains on-chip memory to store up to 121 beam states, where one state contains all the phase and gain settings for the entire IC. The transmitter delivers approximately 15 dBm of saturated output power with 19 dB gain. Receive gain is approximately 14 dB. Phase variation with gain control is approximately 3 degrees over 20 dB of gain control range. Similarly, the gain variation with phase control is approximately 0.25 dB over the entire 360 phase coverage. This tight distribution simplifies array calibration.
The availability of beamforming ICs is accelerating the adoption of analog and hybrid phased array architectures. Complementary improvements in front-end PA/LNA/switch modules, frequency conversion, data converters and digital signal processing are transforming the AESA, improving the performance and SWaP-C of military systems and enabling their use for commercial SATCOM and wireless systems applications.