Jia-Chi Samuel Chieh, Everly Yeo, Raif Farkouh, Randall Olsen, Alex Phipps
©SHUTTERSTOCK.COM/CHRISTOPHER PENLER
The Naval Information Warfare Center Pacific (NIWC-PAC) is located in San Diego, California, and is the only Department of Defense (DoD) laboratory that is headquartered in a major fleet concentration area.
NIWC-PAC’s mission is multifaceted and spans research and development, engineering and prototyping, test and evaluation (T&E), as well as DoD acquisition and installation as it relates to command, control, communications, computers, intelligence, surveillance, and reconnaissance (ISR), as well as cyber and space. NIWC-PAC employs over 5,000 civilians in its workforce, including over 2,500 scientists and engineers. NIWC-PAC supports capabilities across the full acquisition life cycle, including installation and support for naval operations today, engineering, development, and T&E, which supports the navy in construction for tomorrow, as well as science and technology (S&T) for future naval capabilities [1].
In this article, we will give an overview of some of the RF and microwave S&T development at NIWC-PAC, including development of RF filters and channelizers, RF power-combining modules, and phased array antennas using silicon beamforming chips.
Wideband radios and receivers have become more widespread as software-defined radios and multifunction RF becomes more ubiquitous. Wide bandwidths allow for more operational agility, but often times are more susceptible to blocking from a strong interferer [2]. In many of these wide bandwidth microwave systems, the low noise amplifier (LNA) is the most sensitive component. High-power interference or jammers can push the LNAs into compression, blocking the receiver, and in some instances can damage the front end. In many instances LNAs with limiters are employed to protect the receiver; however, most solid-state limiters are not frequency-selective and therefore short the input even if a very narrowband high-power signal is present. A more preferable solution is to develop frequency-selective limiters that can protect a wideband RF front end. One recent example of this is in [3], where a switched multiplexer is used. In [4] we proposed a different topology utilizing a two-tiered approach. The first level of protection is through the use of a frequency multiplexer or channelizer to parse the wide full bandwidth into smaller channels. The second level of protection is through the use of tunable notch filters embedded into each channel filter. These tunable notch filters can excise narrowband high-power interferences precisely, which allows for frequency-selective limiting. Since high levels of rejection are required, absorptive bandstop filters (ABSF) were used to achieve high cancellation using low-order resonators [5], [6], [7]. Figure 1 shows the proposed architecture concept. This type of architecture is especially suitable for interference mitigation from co-located frequency-hopping radios. If the hopping scheme is known a priori, then the tunable notch filters could be set accordingly to deconflict when cosite interference falls within the same subchannel.
Figure 1. Channelizer with embedded tunable notch filters.
In [4] we previously demonstrated the design and development of an X-band frequency-selective multiplexer with embedded absorptive notch filters. Operation of the channelizer was to cover from 6 to 13 GHz in the X band, with five contiguous channels. One of the critical characteristics to the operation of the channelizer is the impedance of each channel. Specifically, the input impedance of each channel filter should look like an open circuit at frequencies above or below the resonant frequency, and have matched impedance at the resonant frequency. The prototype channelizer was fabricated on 20-mil fused silica quartz substrate (sr = 3.8, i˛ = 0.00006). An inductive manifold couples to each channel filter forming an upconverting ladder network. A tubular-type filter is used to achieve the specific behavior required by the channel filter. Each channel filter was designed for ∼15% fractional bandwidth across the center frequency. Figure 2 shows the measured performance of the channelizer across the five channels, and also shows the fabricated prototype in the inset. The average insertion loss is around 3 dB for the entirety of the channelizer, and adjacent channel rejection is over 9.8 dB from channel center.
Figure 2. Measured and simulated response of five-channel channelizer (fabricated prototype is shown as inset) [4].
ABSFs have the capability of achieving high rejection using low quality factor (Q) resonators [6], [7]. A transmission line-realized two-path notch filter is used. The input signal is coupled through a bandpass filter (BPF) to the output, while a portion of the input signal is coupled through a BSF to the output. The two portions are phased such that at the designed frequency the two paths are 180° out of phase, resulting in near perfect cancellation. Flip chip varactor diodes (MAVR-000102-1441) are used to load the BPF and BSF in order to make the two filters tunable; m/4 transmission lines are used as chokes for dc biasing, and radial stubs are used for bypass capacitors.
Five separate ABSF filters are designed for each of the five channels, each designed for around 4% fractional bandwidth around the center frequency and with the capability to tune across the full bandwidth of each channel. Figure 3 shows the fabricated prototype for the ABSF for channel 2. Measurements are done using groud-signal-ground (GSG) probe launches with a custom thru-reflect-line (TRL) calibration kit to de-embed the effects of the probe launch. Figure 4 shows the measured performance of the ABSF, showing more than 35 dB of attenuation from 6.6 to 7.5 GHz with a 3-dB bandwidth of around 170 MHz. The varactor diodes are swept from 0 to 10 V and are designed to be swept simultaneously with the same voltage. The filters were also designed on a 20-mil fused silica quartz substrate. Figure 5 shows the completed frequency selective limiter with embedded absorptive tunable notch filters.
Figure 3. Fabricated prototype of absorptive tunable BSF for channel 2 [4].
Figure 4. Measurement result of ABSF for channel 2 [4].
Figure 5. Prototype of X-band frequency selective limiter using absorptive tunable notch filters embedded multiplexers. Prototype is 2.7 inches by 2.7 inches [7].
Other variations and topologies of tunable absorptive BSFs were also explored and developed in [8] and [9]. In [8], a tunable open stub L-resonator ABSF was demonstrated, shown in Figures 6 and 7. The filter operates similarly to a standard open stub L-resonator BSF, however with an additional external resistor that can absorb the reflected power. Varactor diodes are used to introduce tunability. Figure 8 shows the measured and simulated insertion loss and reflection coefficient of the tunable ABSF, and as can be seen the BSF absorbs the reflected power.
Figure 6. Proposed fully tunable open stub reflectionless BSF [8].
Figure 7. Fabricated fully tunable reflectionless BSF [8].
Figure 8. Measured S21 and S11 with varactor diode biased from 0 to 10 V [8].
In [9], a new quasi-lumped element bridged-T ABSF was presented. It was similar to Bode’s original design in [11] and [12], that minimizes the number of components, allows for more design flexibility in regards to component values, and still achieves large attenuation while maintaining a small size. This design utilizes the bridged-T network, which first found utility in delay equalization [12], [13] for telephony. In this approach, two paths are introduced between the input and the output ports and the signals are imposed to cancel each other by proper adjustment of the phase and amplitude. In this way, extremely deep notches can be realized using low-order resonators. To accomplish phase cancellation, an impedance inverter is typically used. In this work, the impedance inverter is a direct result of the bridged-T topology. As can be seen in Figure 9, the proposed notch filter is comprised of a series bandpass resonator R1L1C1, a bandstop resonator R2L2C2, and a high-pass filter comprised of two series capacitors. The two resonators are designed to resonate at the notch frequency. The high-pass section then acts like a delay or a phase shift, which is adjusted for optimal cancellation. Thus, when the input signal is “bridged” between the two passive circuit paths and both the resistance and reactance of the paths is equal, near infinite attenuation can be achieved [12]. For this design, the C-band was targeted, and at these frequencies surface mount resistors and capacitors are viable; however, most inductors have low self-resonant frequencies. For this reason, high-impedance meandered transmission lines are utilized to provide the proper inductance. The fabricated prototype is shown in Figure 10, and the filter is implemented in microstrip on a 40-mil-thick Rogers 4350B substrate (f r = 3.66, d = 0.0037). In this design, C1 and C2 are 0.1 and 0.2 pF, respectively. These are realized using AVX ACCU-P capacitors. R1 and R2 are 7.5 Ω and 3.57 Ω, respectively, and are found parametrically. These are realized using Vishay Dale CRCW0402 series resistors. Finally, the capacitor in the high-pass filter, C, is set to 4.7 pF. This is realized using Vishay RFCS series capacitors. Sonnet was used for all electromagnetic simulations on the filter structure with S-parameter models for each component [14]. The measured and simulated results are shown in Figure 11, and the correlation is good. The measured rejection is over 50 dB at 4.1 GHz, and the return loss is better than 6 dB over the whole band of operation. The measured 3- and 10-dB bandwidths are approximately 1 and 0.25 GHz, respectively. The measured 10-dB percentage bandwidth is therefore 6%. The measured insertion phase is shown in Figure 12 and shows that at the cancellation frequency, the signal goes through near 180° phase change.
Figure 9. Proposed ABSF [9].
Figure 10. Fabricated fixed-frequency prototype and TRL calibration substrate [9].
Figure 11. Measured and simulated S11 and S21 [9].
Figure 12. Measured S21 phase [9].
Many modern microwave communication systems require high-power amplifiers (HPAs). To obtain the high power (wattage) necessary, many transmitters use vacuum tube technology [15]. Recently, solid-state technology has proven to be more reliable, requires lower supply voltages, and is relatively more linear when compared to traveling wave tubes (TWTs). However, most solid-state technologies are not capable of generating the high-power required systems that typically use TWT technology. For this reason, low-loss microwave power-combining techniques have emerged—including tray-based spatial combining [16], planar transmission line combining [17], and waveguide radial/septum combiners [18]—in order to “gang-up” numerous solid-state amplifiers to support the necessary aggregate high-transmission power. Radial combining techniques are especially attractive for their ease of scalability due to their symmetry. Some recent radial combiners in literature include [19], [20], and [21]. In [22] we presented the design and development of a waveguide/coaxial radial combiner. We also report measured results from both computer numerical control (CNC)-machined and direct metal laser sintering (DMLS) stereo-lithographical fabrication techniques for this radial combiner. The benefit of using the DMLS process is a turnkey test and evaluation. The dominant mode in the parallel plate radial waveguide, the height of which is ≤m/2, is a TM00 mode, which is transverse electromagnetic (TEM). Figure 13 also shows the proposed transition. The coaxial connector (Southwest Microwave) is comprised of a metal pin, which is surrounded by a dielectric (Teflon) bead to hold the pin in place. The transition is comprised of a plurality of inverted conical tapers. The first conical taper is inset in the top of the radial waveguide with a depth of 1 mm (shown in the top purple piece in Figure 13). The second conical taper is electrically connected to the top plane of the radial waveguide and has a hollow cylindrical tube; this facilitates a smooth transition between the radial waveguide and the coaxial input [21]. The bottom of the radial waveguide has an inverted conical stub, which is electrically connected with the N-type metal pin, shown in blue in Figure 13. This piece is inset in the hollow cylindrical cavity of the second conical taper. Figure 14 shows the fabricated prototypes using both CNC machining and DMLS printing methods. The CNC machined part exhibits a smoother surface, whereas the DMLS printed part exhibits apparent surface roughness due to the 3D printing’s resolution. Figure 15 shows the simulated and measured performance of the CNC and DMLS prototypes. As can be seen, the measured 10-dB return loss bandwidth is approximately 8 GHz on the CNC prototype, but the measured and simulated return loss is not well correlated. This is mainly due to the waveguide-to-coaxial adapter, which could not be calibrated out. When time gating on the network analyzer is used, the results match fairly well. The 1.5-dB insertion loss bandwidth for the DMLS-printed prototype is 5 GHz, which is a significant reduction compared to the similar CNC-machined module.
Figure 13. Cross section and top–down view of designed radial waveguide to coaxial transition [22].
Figure 14. Manufactured prototypes using (a) DMLS stereo-lithography and (b) CNC machining [22].
Figure 15. Measured and simulated S-parameters for the designed prototype using both CNC and DMLS processes [22].
Ridge gap waveguides (RGWs) are a new transmission line modality that has gained much interest because of its small size, wideband TEM propagation, and low loss. It was first introduced in [23]. Much of the RGW work has been descended fcrom previous research on soft/hard surfaces. One interesting challenge in RGW system topologies is the interface between the RGW and the monolithic microwave integrated circuit (MMIC). Typical waveguide transitions to microstrip require an E-plane probe. This adds the requirement to design an extra interface between the waveguide and the MMIC, which can be costly, add complexity to the assembly, and finally potentially extra loss. Several groups have proposed novel RGW to microstrip transitions, including [24] and [25]. In both of these proposed techniques, the interface requires electromagnetic coupling and either requires the MMIC to adopt the coupling probe on-chip or requires an additional substrate to which the MMIC needs to be bonded. In [26] a novel approach was proposed eliminating the requirement for such an additional substrate, presenting a self-packaged solution. The RGW-to-microstrip transition was built into the metallic housing, and the MMIC is bonded directly to the metal. Wire-bonding allow for a wider variety of commercially available chips to be used in such an RF module.
Our work [27] builds on [26] and proposes several improvements. In [27] a novel substrateless RGW module using a chip carrier assembly that is implemented using both CNC and electrical discharge machining. The chip carrier is designed to be “swappable,” such that if a MMIC fails, the carrier assembly can be removed and replaced easily. We also demonstrate a two-way combiner using RGW and the novel chip carrier assembly to demonstrate multichannel combining. Simulated and measured performances are compared, with operation between 11 and 14.5 GHz.
The RGW is formed with two plates, one is a perfect electric conductor and one is a perfect magnetic conductor (PMC). The PMC surface is high impedance and is realized with a bed-of-nails approach [28]. In this structure an array of metallic posts, which are m/4 in height, are connected to a metal plate. By placing a ridge between the posts, a quasi-TEM wave can propagate between the ridge and the plate above. Figure 16 shows an exploded diagram of the proposed power combiner structure, and Figure 17 shows the manufactured prototype structure. Figure 18 shows a back-to-back microstrip carrier, representative of a power amplifier microwave MMIC. Measurements were made with WR-75 waveguide to coaxial adapters. Figure 19 shows the measured and simulated response of the microwave power-combining module. As can be seen, measurement and simulation correlate well. The measured insertion loss is less than −1.5 dB from 10.88 to 14 GHz.
Figure 16. Exploded diagram of the two-way power combined substrate-less RGW microwave module [27].
Figure 17. Top–down view of fabricated prototype [27].
Figure 18. Chip bonding assembly [27].
Figure 19. Measured and simulated results [27].
An alternative method, leveraging the same RGW substrateless transition, is presented in [29], where a current combining approach is taken, and an exploded diagram of the module is shown in Figure 20. Current mode power combining has been used extensively on power amplifiers at the chip-scale level [30], [31]. To improve the output power, multiple amplifiers/output stages need to be driven in parallel. This is typically achieved by using Wilkinson combiners, where the output of the transistor is prematched to 50 Ω. In current combining, the outputs of the N-number of transistors are directly tied such that the currents sum. The result is that the output impedance is also reduced by N. For power amplifiers at the transistor level, this can be challenging as the output impedance is quite low to start with, and so impedance matching becomes more challenging, although possible. At the module level, current combining is less used simply because solid-state power amplifier (SSPA) MMICs can be large, and tying the inputs and outputs together can be as complex and large as utilizing a standard technique, such as the Wilkinson. In [29], we utilize an RGW structure to realize a current mode power-combining scheme. This structure easily lends itself to current mode combining as the slot transition is quite long compared the width of most MMIC amplifiers. Figure 21 shows the simulated configurations, including two-, three-, and four-microstrip through transmission lines and also shows the surface current magnitude at 10 GHz for the three configurations. The surface current magnitude shows a cosine taper across the slot, such that outer elements pick up less power. Figure 22 shows the fabricated prototype, with two microstrip through transmission lines, and Figure 23 shows the chip bonding assembly. Figure 24 shows the measured and simulated response of the full back-to-back structure. The −1.5-dB insertion loss bandwidth is 10–14 GHz in measurement and 9–16 GHz in simulation. The main reason is because the impedance matching is heavily degraded beyond 14 GHz in the measured results, whereas in simulation the impedance match is excellent. This is attributed partially to the alignment between the chip carrier and the RGW and also partially to the asymmetry of the dual chip assembly, as it is highly dependent on even mode, in-phase operation. Asymmetries disrupt the in-phase operation and can degrade the even mode impedance.
Figure 20. Exploded diagram of the current mode power combining substrateless RGW microwave module [29].
Figure 21. Surface current magnitude of through microstrip for two-, three-, and four-chip configuration along the coupling slot at 10 GHz [29].
Figure 22. Fabricated prototype [29].
Figure 23. Chip bonding assembly [29].
Figure 24. Measured and simulated results [29].
In the last decade, E-band (71–86 GHz) has become a viable option for high bandwidth line of sight communications. They are used by telecommunication companies for mobile backhaul [32]. They are also utilized for low-latency rapid stock trading between various physically disparate markets [33]. More recently, W/E/V-band has been proposed for satellite communications (SATCOM) both for low Earth orbit (LEO) [34] and for geosynchronous orbit [35].
Work has been done to develop power-combining SSPA, including septum-based waveguide combiners [36] and radial waveguide combiners [37]. At millimeter wave frequencies, transmission line losses can be quite large. The mitigation of losses equals higher combining efficiencies, which is critical to “macro” level power-combining amplifiers. To reduce loss, a suspended stripline-to-rectangular waveguide transition is utilized. The designed four-way SSPA module has a measured 3-dB bandwidth of 22 GHz from 67 to 89 GHz, with a measured output power of at least 0.5 Watts at 85 GHz. The four-way combiner has a simulated efficiency of 86%.
Waveguide combining power amplifiers typically utilize what’s known as a microstrip E-plane probe in order to couple power from the waveguide to an MMIC [38]. A suspended stripline is preferable because most of the propagation energy is in the air dielectric, and the support substrate generally has a negligible impact on the attenuation and phase delay of the stripline [39]. The stripline E-plane probes are designed on a 5-mil fused silica substrate (f r ∼ 3.8, d ∼ .0002). The fused silica substrate straddles a channel on the lower half of the split-block. Small metal ledges form a support structure for the substrate, and in this way the substrate can be fully suspended.
A four-way power combined amplifier module was designed and fabricated. Commercial HPA MMICs from MACOM were used (MAAP-011106) [40]. The commercial MMIC has a Psat of 25 dBm, a P1dB of 23 dBm, and a gain of between 18 and 20 dB, and operates from 71 to 86 GHz. E-plane Y-junction waveguide power dividers/combiners were utilized for the four-way module. Each of the Y-junctions had a three-section impedance transformer, to allow for ultrawideband operation. Two sets of printed circuit boards (PCBs) were used for each MMIC, one providing gate bias for each of the four-stage amplifiers with inline 10 Ω resistors for stability, and one providing drain bias for each of the four stages with 10-nF decoupling capacitors. A 10-mL Eccosorb® BSR absorber was used in the cavities to prevent oscillation. Figure 25 shows a back-to-back model in high frequency simulation software for the full passive module, including a waveguide to suspsended stripline transition, and a suspended stripline-to-MMIC bond wire transition. The model included a passive GaAs microstrip through line that was de-embedded from the loss simulation. Simulated results from the back-to-back model show an S 11 of better than 10 dB from 71 to 100 GHz, with an average insertion loss of .65 dB, which equates to an average efficiency of 86%.
Figure 25. Full high frequency simulation software simulation model.
A fabricated prototype is shown in Figure 26. S-parameter measurements were made on an Anritsu ME7808A vector network analyzer using a waveguide thru-reflect-match (TRM) calibration kit. The amplifier was measured with gate biased to −0.2 V and drain biased at 3.5 V, with the whole amplifier module drawing 3.45 A. The measured S-parameter results are shown in Figure 27, and the amplifier module has similar gain characteristic as the individual MMIC. The minimum gain of the module occurs at 80 GHz, with a gain of 18 dB. Figure 28 shows the infrared image of the HPA module under bias, and as can be seen, the temperature where the MMICs are die-attached reach upwards of 85 °C. Large signal testing was done using the test bench shown in Figure 29. An OML S10MS module was used as an E-band signal generator, which is followed by a level-setting waveguide attenuator; a Millitech preamplifier (AMP-10-02130) drives the HPA. A 20-dB directional coupler with a Keysight W8486A power sensor was used to measure the total output power with a waveguide termination to dissipate the power delivered to the load. The developed SSPA module was a 1 W class power amplifier that operated in the E-band.
Figure 26. Fabricated prototype showing full MMIC assembly.
Figure 27. Measured S-parameters of four-way HPA.
Figure 28. Optical and infrared images of HPA module under bias.
Figure 29. Large signal test bench.
Phased array antennas for line-of-sight communications are preferable as they support agile beam steering. For SATCOM applications, mechanically scanned antennas are viable but suffer from the keyhole effect [41], and the slewing from the gimbal may require a dual antenna solution. The phased array is a preferred solution as the beam can be electronically scanned almost instantaneously so a single antenna aperture can handle the satellite handover. The availability of commercial silicon beamforming chipsets has resulted in the ability for phased array antennas to be ubiquitous in future communication systems. The fully integrated chipset eliminates the need for discrete transceiver blocks and includes a polarization switch, a transmit/receive (T/R) switch, low noise amplifier, power amplifier, phase shifters, and variable attenuators.
Circular polarized (CP) active electronically scanned arrays are of great interest for SATCOMs. Phased arrays that are capable of maintaining CP over wide scan angles are of great interest, especially for new satellite constellations that are being deployed in LEO and medium Earth orbit. LEO constellations are closer to the Earth (∼500 to 2,000 km) and therefore move overhead at a much faster rate and require a fast hand-off from horizon to horizon. Several efforts have been made in the past to develop low-profile CP phased array antennas. In [42] an X-band dual-polarized CP phased array was developed for submarine SATCOM using complex multichip-modules on low-temperature cofired ceramic material. The performance of the axial-ratio (AR) bandwidth over scan angles was not presented. In [43], a wide-scan linear phased array antenna was presented with AR < 3 dB beamwidth of 121° with co- to cross-polarization separation of 16 dB. This array, however, utilizes magnetic electric dipoles as the radiators, and therefore has a narrow impedance bandwidth (2%). In [43] and [44], a truncated corner patch phased array was realized, however demonstrated narrow axial bandwidths of 2.4% and 3.2%, respectively, at broadside. In [45] a dual-linear patch antenna array was coupled with a beamformer radio-frequency integrated circuit (RFIC), which combined the two linear polarizations on-chip with a 90° offset to generate the CP. This method yielded a 6.7% AR bandwidth at broadside; however, AR properties over the scan angles and frequency were not included.
One method that has been used to enhance the AR bandwidth in passive fixed-beam arrays is to use sequential rotation (SQR) and nested SQR [46], [47], [48], [49]. In a nested SQR approach, radiator elements within a subarray utilize SQR, and then SQR is applied at the subarray level. To support dual CPs with a wide impedance bandwidth, several radiator topologies could be adopted as described in [50]. A well-known corner truncated probe fed stacked-patch approach was utilized for the phased array design. Figure 30 shows the dimensions of the stacked patch elements as well as an exploded view of the board stack-up used for the phased array.
Figure 30. Truncated corner stacked patch antenna element. Antenna parameters are: H1= .508 mm, H2 = .599 mm, H3 = .975 mm, H4 = .127 mm, L1 = 5.79 mm, L2 = 5.41 mm, L3 = .187 mm, L4 = .61 mm, L5 = 2.69 mm [51].
We demonstrate in [51] an active phased array using this nested SQR architecture to develop a 16-element planar array. Figure 31 shows a nested SQR approach-based design where both local (#1) and nested SQR (#2) at the subarray level is applied. The interelemental spacing is 0.5mo at the design frequency. The array antenna PCB stack-up is shown in Figure 31. Typically, for passive arrays, a 90° delay transmission line is added to compensate for the SQR. Since our phased array is a fully active antenna, the compensation for the SQR is performed within the RFIC phase shifter, and therefore the delay is frequency dependent. A 16-element four-by-four phased array antenna using the nested SQR concept was designed and fabricated. Figure 32 shows the photograph of the fabricated prototype, which includes multilayer PCB and altium layout considerations; Anokiwave AWMF-0117 single-channel silicon beamforming RFICs were used. Phase compensation due to the nested SQR is applied using this chipset. Beamforming algorithm was applied through SPI controller and Labview-based graphical user interface, which controls beam peak scan angles. Radiation pattern measurements were performed at San Diego State University’s (San Diego, California, USA) far-field anechoic chamber. The presented measured beam patterns are normalized since this array has gain on receive in addition to the feed network losses. Similarly, the S-parameter is the combined effect of the active and passive components, so not included here. Figure 33 shows the measured azimuth (x–z plane) beam and elevation beam (y–z plane) scan patterns for the righthand CP (RHCP) at fo = 12.5 GHz. Measurements were performed up to ±80° due to constraints on the test setup in the anechoic chamber. The measured difference between the RHCP and the lefthand CP (LHCP) over frequency and scan angle is approximately better than 20 dB. Figure 34 shows the measured AR versus frequency for the RHCP polarization and for the AR bandwidth when the beam is scanned away from the broadside, both elevation and azimuth cut planes. The measured 3-dB AR bandwidth for scan angles up to ±30° is 24% for both cut planes. For the RHCP polarizations on the azimuth cut plane, the AR is remains below 3 dB for scan angles up to ±45°. For the RHCP polarization in the elevation cut plane, the AR for +45° and −45° are degraded to around 6–7 dB, corresponding with expected simulated results, which can still be usable for some communication applications.
Figure 31. Array architectures showing nested SQR [51].
Figure 32. Photograph of the fabricated prototype phased array antenna [51].
Figure 33. (a) Measured RHCP beam scan patterns for azimuth at 12.5 GHz and (b) measured RHCP beam scan patterns for elevation at 12.5 GHz [51].
Figure 34. (a) RHCP AR azimuth scan and (b) RHCP AR elevation scan. Here fo is 12.5 GHz [51].
The navy also utilizes common data link (CDL) for ISR applications. One of the more common airborne antennas that supports CDL is the AC-27, developed by Honeywell. This antenna operates from 14.53 to 15.35 GHz, with 27 dBic of gain, is RHCP, and maintains an AR below 2 dB [52]. We believe that future CDL antennas will leverage phased arrays, and aimed to develop a phased array antenna with similar characteristics in performance and form factor to the AC-27.
Leveraging previous work [51], we have demonstrated that utilizing corner truncated stacked patch antennas, with a sequential rotation subarray, can result in wide AR bandwidths across wide scan angles. In this work, we leverage that architecture and demonstrate a 64-element T/R phased array that operates in the CDL band. Utilizing intrinsically CP antenna elements, such as the corner truncated patch, has some benefits as compared to using dual-linear polarized patches, such as in [53], [54], [55], and [56]. When using the dual-linear polarized patch element, you can achieve arbitrary polarizations, at the expense of occupying two channels in your beamformer. In receive mode, this means that the dual-linear polarized antennas consume twice the dc power to achieve RHCP, compared to using an intrinsically RHCP antenna element. In addition, the intrinsically CP antenna element allows for simultaneous dual beams from a common shared aperture, when paired with a beamforming chip, such as the Renesas F6123. On transmit, the dual-linear polarized antenna elements can offer the advantage of twice the output power, since two beamforming channels are being combined to generate the circular polarization. In this work, intrinsically RHCP antenna elements are used, with a sequentially rotated subarray. This phased array is capable of both transmit and receive from a common aperture, with the ability to support RHCP and operate from 14.4 to 15.2 GHz. The developed phased array antenna utilizes a silicon single channel beamforming chip (AWMF-01174) from Anokiwave. This chip has five bits of phase and amplitude control, a noise figure of 3 dB, an OP1dB of +12 dBm, and operates from 10.5 to 16 GHz. There are a total of 65 beamforming chips on the array, with the 65th chip providing a gain stage compensating for the feeder losses. The 65 beamforming chips have a daisy-chained serial protocol interface (SPI), and a National Instruments USB SPI controller was used to interface and control the array. No calibrations were performed on this array. Figure 35 shows the phased array antenna under test in the anechoic chamber, and Figure 36 shows the backside of the PCB with the silicon beamforming chips exposed. On receive, the array consumes approximately 13 W of dc power, and on transmit the array consume approximately 16 W of dc power. A simple metal cold plate was attached to the array to mitigate the heat, and no active cooling was used.
Figure 35. Developed prototype array in anechoic chamber.
Figure 36. Developed prototype array, backside of printed circuit board.
Antenna measurements were made at the San Diego State University Antenna Measurement Laboratory (AML), San Diego, California, USA. CP characteristics, such as AR, were produced by taking amplitude and phase measurements in the vertical linear polarization and the horizontal linear polarization, and postprocessing. For brevity, negative scan angles were taken for azimuth scan patterns, and positive scan angles were taken for elevation scan patterns. Figure 37 shows the measured normalized azimuth scan patterns at 14.4 GHz. Measurements were taken with a Taylor taper for −20 dB sidelobes. As can be seen, the array is capable of scanning to 45° without the introduction of grating lobes, as well as < 20 dB cross-polarization levels for all scan angles. Figure 38 shows the measured elevation scan patterns at 14.4 GHz. Again, the array maintains very low cross-polarization levels for all scan angles. Figure 39 shows the measured AR versus frequency versus scan angle for both the azimuth and elevation scans. For the azimuth scan, Figure 39(a) shows that the AR is below 3 dB up to 45° scan from 13 to 14.8 GHz. With a more relaxed AR criteria of 4 dB, the array is capable of scanning up to 45° in the azimuth plane from 13 to 15.2 GHz. At 14.4 GHz, the measured compressed effective isotropic radiated power is approximately 45 dBm.
Figure 37. Measured azimuth scan patterns at 14.4 GHz.
Figure 38. Measured elevation scan patterns at 14.4 GHz.
Figure 39. Measured AR versus frequency versus scan angle for (a) azimuth scan and (b) elevation scan.
This article has summarized some S&T efforts at the NIWC-PAC as it relates to RF and microwave communications, including RF filters and channelizers, RF power-combining modules, and phased array antennas using silicon beamforming chips. Wireless communications continue to be an important S&T topic to NIWC-PAC and we plan to continue developing technologies in this area.
All photos and figures are courtesy of Naval Information Warfare Center Pacific, San Diego, California. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the Naval Information Warfare Center. Jia-Chi Samuel Chieh (jiachi.s.chieh.civ@us.navy.mil) is the corresponding author.
[1] “Naval information warfare center pacific command overview, distribution A,” Naval Information Warfare Center Pacific, San Diego, CA, USA, 2022. [Online] . Available: https://www.niwcpacific.navy.mil/wp-content/uploads/2022/02/211207_NDIA_NIWC_PAC_EOY_FY21_Bonwit_Final_DISPLAY_Distro_A_WEB.pdf
[2] E. J. Naglich and A. C. Guyette, “Frequency-selective limiters utilizing contiguous-channel double multiplexer topology,” IEEE Trans. Microw. Theory Techn., vol. 64, no. 9, pp. 2871–2882, Sep. 2016, doi: 10.1109/TMTT.2016.2590542.
[3] C. Galbraith, G. M. Rebeiz, and R. Drangmeister, “A cochlea-based preselector for UWB applications,” in Proc. IEEE Radio Freq. Integr. Circuits (RFIC) Symp., Honolulu, HI, USA, 2007, pp. 219–222, doi: 10.1109/RFIC.2007.380869.
[4] J.-C. S. Chieh and J. Rowland, “X-band frequency selective limiter using absorptive notch filters embedded multiplexers,” in Proc. IEEE MTT-S Int. Microw. Symp. (IMS), 2017, pp. 1919–1922, doi: 10.1109/MWSYM.2017.8059035.
[5] R. O’Brient et al., “A log-periodic channelizer for multichroic antenna-coupled TES-bolometers,” IEEE Trans. Appl. Supercond., vol. 21, no. 3, pp. 180–183, Jun. 2011, doi: 10.1109/TASC.2010.2093090.
[6] D. R. Jachowski, “Passive enhancement of resonator Q in microwave notch filters,” in Proc. IEEE MTT-S Int. Microw. Symp. Dig. (IEEE Cat. No. 04CH37535), 2004, vol. 3, pp. 1315–1318, doi: 10.1109/MWSYM.2004.1338808.
[7] D. R. Jachowski, “Compact, frequency-agile, absorptive bandstop filters,” in Proc. IEEE MTT-S Int. Microw. Symp. Dig., 2005, pp. 513–516, doi: 10.1109/MWSYM.2005.1516645.
[8] J.-C. S. Chieh and J. Rowland, “A fully tunable C-band reflectionless bandstop filter using L-resonators,” in Proc. 46th Eur. Microw. Conf. (EuMC), 2016, pp. 131–133, doi: 10.1109/EuMC.2016.7824295.
[9] J.-C. S. Chieh and J. Rowland, “Quasi-lumped element bridged-T absorptive bandstop filter,” IEEE Microw. Wireless Compon. Lett., vol. 26, no. 4, pp. 264–266, Apr. 2016, doi: 10.1109/LMWC.2016.2537787.
[10] B. Kim, J. Lee, J. Lee, B. Jung, and W. J. Chappell, “RF CMOS integrated on-chip tunable absorptive bandstop filter using Q-tunable resonators,” IEEE Trans. Electron Device, vol. 60, no. 5, pp. 1730–1737, May 2013, doi: 10.1109/TED.2013.2253557.
[11] H. W. Bode, “Wave filter,” U.S. Patent 2 035 258, Mar. 24, 1936.
[12] H. W. Bode, “Wave filter,” U.S. Patent 2 002 216, May 21, 1935.
[13] O. J. Zobel, “Theory and design of uniform and composite electric wave-filters,” Bell Syst. Tech. J., vol. 2, no. 1, pp. 1–46, Jan. 1923, doi: 10.1002/j.1538-7305.1923.tb00001.x.
[14] “Sonnet 12.52.” Sonnet Software. Accessed: Feb. 3, 2023. [Online] . Available: https://www.sonnetsoftware.com/support/sonnet-suites/release-notes-1252.html
[15] R. K. Parker, R. H. Abrams, B. G. Danly, and B. Levush, “Vacuum electronics,” IEEE Trans. Microw. Theory Techn., vol. 50, no. 3, pp. 835–845, Mar. 2002, doi: 10.1109/22.989967.
[16] M. P. DeLisio and R. A. York, “Quasi-optical and spatial power combining,” IEEE Trans. Microw. Theory Techn., vol. 50, no. 3, pp. 929–936, Mar 2002, doi: 10.1109/22.989975.
[17] C. Y. Law and A.-V. Pham, “A high-gain 60GHz power amplifier with 20dBm output power in 90nm CMOS,” in Proc. IEEE Int. Solid-State Circuits Conf. (ISSCC), 2009, pp. 426–427, doi: 10.1109/ISSCC.2010.5433882.
[18] J. Schellenberg, E. Watkins, M. Micovic, B. Kim, and H. Kyu, “W-band, 5W solid-state power amplifier/combiner,” in Proc. IEEE MTT-S Int. Microw. Symp. Dig., Anaheim, CA, USA, Jun. 2010, pp. 240–243, doi: 10.1109/MWSYM.2010.5517616.
[19] J.-M. Denoual, A. Peden, B. Della, and J.-P. Fraysse, “16-way radial divider/combiner for solid state power amplifiers in the K band,” in Proc. Eur. Microw. Conf. (EuMC), Oct. 2008, pp. 345–348, doi: 10.1109/EUMC.2008.4751459.
[20] S. Kaijun, F. Yong, and H. Zongrui, “Broadband radial waveguide spatial combiner,” IEEE Microw. Wireless Compon. Lett., vol. 18, no. 2, pp. 73–75, Feb. 2008, doi: 10.1109/LMWC.2007.911984.
[21] D. Yazhou, D. Shi-Wei, Z. Zhongbo, and W. Ying, “60 GHz low loss, amplitude and phase balanced radial waveguide power combiner,” in Proc. Int. Conf. Electron., Commun. Control (ICECC), Sep. 2011, pp. 4070–4073, doi: 10.1109/ICECC.2011.6066614.
[22] J.-C. S. Chieh, M. Civerolo, and A. Clawson, “A ultra wideband radial combiner for X/Ku-band using CNC and DMLS processes,” IEEE Microw. Wireless Compon. Lett., vol. 25, no. 5, pp. 286–288, May 2015, doi: 10.1109/LMWC.2015.2409806.
[23] P.-S. Kildal, E. Alfonso, A. Valero-Nogueira, and E. Rajo-Iglesias, “Local metamaterial-based waveguides in gaps between parallel metal plates,” IEEE Antennas Wireless Propag. Lett., vol. 8, pp. 84–87, 2009, doi: 10.1109/LAWP.2008.2011147.
[24] U. Nandi, A. U. Zaman, A. Vosoogh, and J. Yang, “Novel millimeter wave transition from microstrip line to groove gap waveguide for MMIC packaging and antenna integration,” IEEE Microw. Wireless Compon. Lett., vol. 27, no. 8, pp. 691–693, Aug. 2017, doi: 10.1109/LMWC.2017.2723679.
[25] Y. Shi, J. Zhang, S. Zeng, and M. Zhou, “Novel W-Band millimeter-wave transition from microstrip line to groove gap waveguide for MMIC integration and antenna application,” IEEE Trans. Antennas Propag., vol. 66, no. 6, pp. 3172–3176, Jun. 2018, doi: 10.1109/TAP.2018.2819902.
[26] B. Ahmadi and A. Banai, “Substrateless amplifier module realized by ridge gap waveguide technology for millimeter-wave applications,” IEEE Trans. Microw. Theory Techn., vol. 64, no. 11, pp. 3623–3630, Nov. 2016, doi: 10.1109/TMTT.2016.2607177.
[27] J. Samuel Chieh, “A substrate-less microwave power-combining module utilizing ridge gap waveguide,” IEEE Microw. Wireless Compon. Lett., vol. 28, no. 11, pp. 972–974, Nov. 2018, doi: 10.1109/LMWC.2018.2870932.
[28] M. G. Silveirinha, C. A. Fernandes, and J. R. Costa, “Electromagnetic characterization of textured surfaces formed by metallic pins,” IEEE Trans. Antennas Propag., vol. 56, no. 2, pp. 405–415, Feb. 2008, doi: 10.1109/TAP.2007.915442.
[29] J.-C. S. Chieh, A. Phipps, and E. Yeo, “A substrate-less current mode combining power module utilizing ridge gap waveguide,” in Proc. 50th Eur. Microw. Conf. (EuMC), 2021, pp. 751–754, doi: 10.23919/EuMC48046.2021.9338074.
[30] A. Y. Chen, Y. Baeyens, Y. Chen, and J. Lin, “An 83-GHz high-gain SiGe BiCMOS power amplifier using transmission-line current-combining technique,” IEEE Trans. Microw. Theory Techn., vol. 61, no. 4, pp. 1557–1569, Apr. 2013, doi: 10.1109/TMTT.2013.2248376.
[31] M. Bohsali and A. M. Niknejad, “Current combining 60GHz CMOS power amplifiers,” in Proc. Radio Freq. Integr. Circuits Symp., Boston, MA, USA, 2009, pp. 31–34, doi: 10.1109/RFIC.2009.5135483.
[32] G. R. MacCartney and T. S. Rappaport, “73 GHz millimeter wave propagation measurements for outdoor urban mobile and backhaul communications in New York City,” in Proc. IEEE Int. Conf. Commun., Sydney, NSW, Australia, 2014, pp. 4862–4867, doi: 10.1109/ICC.2014.6884090.
[33] “E-Band drives trading markets’ need for speed with new ultra-low latency radio for HFT customers,” Bus. Wire, 2012. [Online] . Available: https://www.businesswire.com/news/home/20121102005175/en/E-Band-Drives-Trading-Markets%E2%80%99-Need-for-Speed-With-New-Ultra-Low-Latency-Radio-for-HFT-Customers
[34] C. Bonefazi, M. Ruggieri, and A. Paraboni, “The DAVID mission in the heritage of the SIRIO and ITALSAT satellites,” IEEE Trans. Aerosp. Electron. Syst., vol. 38, no. 4, pp. 1371–1376, Oct. 2002, doi: 10.1109/TAES.2002.1145759.
[35] B. Kim, A. Tran, and J. Schellenberg, “Full W-band power amplifier/combiner utilizing GaAs technology,” in Proc. IEEE MTT-S Int. Microw. Symp. (IMS), Montreal, QC, Canada, 2012, pp. 1–3, doi: 10.1109/MWSYM.2012.6259686.
[36] J. Schellenberg, A. Tran, L. Bui, A. Cuevas, and E. Watkins, “37 W, 75–100 GHz GaN power amplifier,” in Proc. IEEE MTT-S Int. Microw. Symp. (IMS), San Francisco, CA, USA, 2016, pp. 1–4, doi: 10.1109/MWSYM.2016.7540195.
[37] B. Glance and R. Trambarulo, “A waveguide to suspended stripline transition (Letters),” IEEE Trans. Microw. Theory Techn., vol. 21, no. 2, pp. 117–118, Feb. 1973, doi: 10.1109/TMTT.1973.1127938.
[38] A. R. Kerr, “Elements for E-plane split-block waveguide circuits,” ALMA Memorandum, vol. 381, pp. 1–9, Jul. 2001. [Online] . Available: http://www.mma.nrao.edu/memos/html-memos/alma381/memo381.pdf
[39] G. L. Matthaei, L. Young, and E. M. T. Jones, Microwave Filters, Impedance-Matching Networks, and Coupling Structures. New York, NY, USA: McGraw-Hill, 1964.
[40] “MAAP-011106 datasheet,” MACOM, Lowell, MA, USA, 2016. [Online] . Available: https://cdn.macom.com/datasheets/MAAP-011106.pdf
[41] R. Gilmore, “Broadband-on-the-move satellites takes the pole position,” MilsatMagazine, pp. 52–59, May 2013. [Online] . Available: http://www.milsatmagazine.com/story.php?number=1466984736#
[42] K. M. Lee, J. Edie, R. Krueger, J. Weber, T. Brott, and W. Craig, “A low profile X-band active phased array for submarine satellite communications,” in Proc. IEEE Int. Conf. Phased Array Syst. Technol. (Cat. No.00TH8510), Dana Point, CA, USA, 2000, pp. 231–234, doi: 10.1109/PAST.2000.858947.
[43] P. Liu, Y. Li, and Z. Zhang, “Circularly polarized 2 bit reconfigurable beam-steering antenna array,” IEEE Trans. Antennas Propag., vol. 68, no. 3, pp. 2416–2421, Mar. 2020, doi: 10.1109/TAP.2019.2939669.
[44] T. Lambard, O. Lafond, M. Himdi, H. Jeuland, S. Bolioli, and L. Le Coq, “Ka-band phased array antenna for high-data-rate SATCOM,” IEEE Antennas Wireless Propag. Lett., vol. 11, pp. 256–259, Mar. 2012, doi: 10.1109/LAWP.2012.2189747.
[45] K. K. Wei Low, A. Nafe, S. Zihir, T. Kanar, and G. M. Rebeiz, “A scalable circularly-polarized 256-element Ka-band phased-array SATCOM transmitter with ±60° beam scanning and 34.5 dBW EIRP,” in Proc. IEEE MTT-S Int. Microw. Symp. (IMS), Boston, MA, USA, 2019, pp. 1064–1067, doi: 10.1109/MWSYM.2019.8701112.
[46] P. S. Hall, “Application of sequential feeding to wide bandwidth, circularly polarised microstrip patch arrays,” IEE Proc. H - Microw., Antennas Propag., vol. 136, no. 5, pp. 390–398, Oct. 1989, doi: 10.1049/ip-h-2.1989.0070.
[47] T. Teshirogi, M. Tanaka, and W. Chujo, “Wideband circularly polarised array antennas with sequential rotations and phase shift of elements,” in Proc. Int. Symp. Antennas Propag. (ISAP), Tokyo, Japan, 1985, pp. 117–120.
[48] R. R. George, A. T. Castro, and S. K. Sharma, “Comparison of a four stage sequentially rotated wideband circularly polarized high gain microstrip patch array antennas at Ku-band,” in Proc. 11th Eur. Conf. Antennas Propag. (EUCAP), Paris, France, 2017, pp. 2307–2311, doi: 10.23919/EuCAP.2017.7928343.
[49] A. Chen, Y. Zhang, Z. Chen, and C. Yang, “Development of a Ka-band wideband circularly polarized 64-element microstrip antenna array with double application of the sequential rotation feeding technique,” IEEE Antennas Wireless Propag. Lett., vol. 10, pp. 1270–1273, Nov. 2011, doi: 10.1109/LAWP.2011.2175433.
[50] P. Sharma and K. Gupta, “Analysis and optimized design of single feed circularly polarized microstrip antennas,” IEEE Trans. Antennas Propag., vol. 31, no. 6, pp. 949–955, Nov. 1983, doi: 10.1109/TAP.1983.1143162.
[51] J.-C. S. Chieh et al., “Development of flat panel active phased array antennas using 5G silicon RFICs at Ku- and Ka-bands,” IEEE Access, vol. 8, pp. 192,669–192,681, Oct. 2020, doi: 10.1109/ACCESS.2020.3032841.
[52] “AC-27 CDL antenna system, data sheet,” Honeywell, Phoenix, AZ, USA, 2016. [Online] . Available: https://aerospace.honeywell.com/content/dam/aerobt/en/documents/learn/products/satellite-communications/broucheres/N61-1551-000-000-AC-27CDL-bro.pdf
[53] K. K. W. Low, T. Kanar, S. Zihir, and G. M. Rebeiz, “A 17.7–20.2-GHz 1024-element K-band SATCOM phased-array receiver with 8.1-dB/K G/T, ±70° beam scanning, and high transmit isolation,” IEEE Trans. Microw. Theory Techn., vol. 70, no. 3, pp. 1769–1778, Mar. 2022, doi: 10.1109/TMTT.2022.3142275.
[54] K. K. W. Low, S. Zihir, T. Kanar, and G. M. Rebeiz, “A 27–31-GHz 1024-element Ka-band SATCOM phased-array transmitter with 49.5-dBW peak EIRP, 1-dB AR, and ±70° beam scanning,” IEEE Trans. Microw. Theory Techn., vol. 70, no. 3, pp. 1757–1768, Mar. 2022, doi: 10.1109/TMTT.2021.3139911.
[55] G. Gültepe and G. M. Rebeiz, “A 256-element dual-beam polarization-agile SATCOM Ku-band phased-array with 5-dB/K G/T,” IEEE Trans. Microw. Theory Techn., vol. 69, no. 11, pp. 4986–4994, Nov. 2021, doi: 10.1109/TMTT.2021.3097075.
[56] G. Gültepe, T. Kanar, S. Zihir, and G. M. Rebeiz, “A 1024-element ku-band SATCOM phased-array transmitter with 45-dBW single-polarization EIRP,” IEEE Trans. Microw. Theory Techn., vol. 69, no. 9, pp. 4157–4168, Sep. 2021, doi: 10.1109/TMTT.2021.3075678.
Digital Object Identifier 10.1109/MMM.2023.3240543