Parent Category: 2022 HFE
By Yonghui Shu and Andrew Laundrie
Abstract:
The Millimeter-wave spectrum generally refers to the frequency range of 30 to 300 GHz. In recent decades it has found wide application in space, aviation, military, science, industrial, commercial and consumer systems due to its unique wavelength characteristics. It is not a new technology, although it is still considered by many to be an expensive technology because of historically high entry barriers. Advancements in semiconductor devices, packaging, and manufacturing techniques have been slow to change this attitude. To a large degree the mm-wave electronics industry continues to favor custom solutions rather than adopting more of the prototyping and manufacturing approaches that have successfully reduced hardware costs at lower frequencies. To overcome these historical barriers to new applications, the mm-wave industry is increasingly turning to standard component packages and new RF interfaces that provide alternatives to traditional coaxial and waveguide connections. Improvements in test and measurement techniques have also lowered some of the hurdles that impeded mm-wave hardware development. This article reviews significant improvements in the areas of commercial off-the-shelf (COTS) mm-wave components, novel waveguide connectors for lower-cost packaging, and contactless waveguide flanges that can streamline the measurement of waveguide components. These achievements have accelerated the development of new mm-wave systems while reducing the cost of transitioning to higher rates of production. As a result, mm-wave technology is much more accessible and more affordable for new applications.
Introduction
The mm-wave spectrum has significant advantages over its lower-frequency RF and microwave counterparts. Widely accepted to comprise 30 to 300 GHz, this frequency range offers smaller wavelengths and wider bandwidths. These features have specific advantages. Smaller wavelengths are attractive because they allow users to take advantage of additional electromagnetic spectrum. Also, for a given antenna aperture size, smaller wavelengths enable narrower beams to improve the angular resolution of radar systems used in weapons, guidance systems, aircraft, and remote-sensing satellites. Wider bandwidths support higher data rates in communications systems and improve distance resolution for target identification in radar systems. Increased bandwidth can also improve the sensitivity and resolution of passive imaging systems, as well as provide systems with greater immunity from jamming and other interference [1], [2], [3], [4]. Despite these advantages mm-wave technology is often seen as too expensive or impractical for many new applications.
Until recently the mm-wave frequency spectrum was mainly used for military, aerospace, and scientific research applications. It has been generally limited to weapon guidance, seekers, radars, military communication equipment, remote control, remote sensing, radiometry, radio telescopes, material science, and other areas of research and development. In recent years, technological advances — especially those related to simulation and design tools, semiconductor device performance, and manufacturing methods — have allowed mm-wave technologies up to 170 GHz to reach their current stages of maturity [5], [6], [7].
As mm-wave technology continues to find increasing opportunities in traditional markets, potentially explosive opportunities are seen in a wide range of new applications. There is increasing interest in topics such as improved internet connectivity, enhanced safety and security, as well as “smart” cities, homes, and appliances. As a result, mm-wave technology may experience vastly more growth in new commercial and consumer-oriented applications. Research and development is growing in many areas including the Internet of Things (IoT), high data-rate communication, passive imaging, transportation safety and management, driverless cars, security systems, commercial satellites, and test/measurement equipment. This activity confirms that the mm-wave electronics industry is healthy and growing [8], [9], [10], [11].
As the opportunities for mm-wave technologies have grown, the technical and business landscapes have often been difficult to navigate due to a lack of manufacturing standards and few stable building blocks for system integration. Reasons for not having standard building blocks in the industry include the limited bandwidths that are typically achieved using traditional design approaches. This has resulted in the continued need for customization to cover specific frequency ranges. Further, complex interconnections between building blocks can limit system performance due to high insertion loss or poor impedance matching. At higher mm-wave frequencies approaching 300 GHz, waveguide sections with low insertion loss are costly to produce. These and other factors have made the industry reluctant to develop standard building blocks. Instead, many developers of mm-wave hardware continue to favor customized design approaches. Revolutionary packaging and testing technologies are needed to speed up the development process and reduce prototyping costs so that new applications can be demonstrated and implemented as quickly as possible.
Historical Considerations
Traditionally, mm-wave hardware development required a large amount of design activity at the component level. An advanced Engineering degree with considerable training in electromagnetic theory and network analysis were commonly regarded as prerequisites for entering the field of microwave and mm-wave component design. Furthermore, the machinery required to make precision mechanical parts for mm-components was lacking while machinists with the necessary skills were limited. Experienced, hands-on technicians required many years practice to develop the “know-how” needed to compensate for imperfect designs and machining tolerances. In recent years the emergence of advanced design software has greatly reduced the need to apply electromagnetic theory and network analysis methods to the development of mm-wave components and systems. Meanwhile advanced mechanical design software such as Solidworks, as well as improved CNC machinery, have helped to reduce the need for highly experienced manufacturing personnel, improved mechanical parts accuracy, and eliminated “skill” dependency tremendously.
Now, the need to design systems at the component level is being reduced by the increased availability of COTS hardware. System developers can thus focus more of their efforts on building proof-of-concept prototypes and pre-production systems. As a result, it is increasingly possible to evaluate new applications faster and study their economic feasibility with lower up-front investment.
Another common barrier to entry for mm-wave product development is the need for advanced test equipment. Basic instrumentation includes signal generators, amplifiers, signal analyzers, and network analyzers. Traditionally such equipment was expensive to purchase and maintain, and much of it required special training and skills for proper operation. In recent years the cost of test equipment has declined significantly, but it still represents a major investment for anyone seeking to harness mm-wave technology.
For advanced instruments such as Vector Network Analyzers (VNAs), especially those operating at mm-wave frequencies, high levels of operator training have been essential to achieve reliable measurement results. Further, slow test procedures have limited the number of devices that could be tested over a given time period. Recent advancements in VNA test equipment are making it easier to perform component tests, allowing test equipment to pay for itself sooner due to increased productivity and lower operating costs.
Choosing Suppliers
For those involved in the development of mm-wave systems for new applications, some of the more difficult challenges include surveying and selecting component suppliers. This is especially true in systems where one component can impact the performance of several others. Moreover, the depth and breadth of a supplier’s understanding of these dependencies must be discovered as soon as possible to ensure that integrated systems can achieve the best possible performance. For many years Eravant, formerly SAGE Millimeter (Torrance, CA), has been investing heavily to build up its standard modular products, encompassing a wide range of broadband coaxial and waveguide components with focus on mm-wave bands from 30 to 325 GHz. These standard modules are generally designed to cover full waveguide bandwidths, or bandwidths that are as wide as possible using available technologies.
Figure 1 • Typical communication front-end block diagram
Some standard models span multiple octave bandwidths. For example, amplifier model SBB-0117033015-VFVF-E3 covers 10 MHz to 70 GHz with 30-dB nominal gain and +15 dBm output power at 1-dB compression. The broadband nature of Eravant’s standard modules offers opportunities for system integrators to find suitable building blocks that meet their system requirements without paying for non-recurring-engineering (NRE) and waiting for someone to address their particular frequency range. Eravant’s standard COTS product categories cover almost everything needed to develop mm-wave wireless systems.
Eravant has also introduced a series of novel waveguide connectors under the trademarked name, “Uni-Guide”. This product line offers an RF interface concept that is similar to coaxial connectors. The Uni-GuideTM connector series is offered to the mm-wave component and subsystem industry to reduce the need for customized device packages, and to shorten development and manufacturing cycle times [12].
In mm-wave frequency bands, rectangular waveguide is the most common transmission-line medium due to its low loss and high power capacity. Unlike coaxial cables, waveguide connections are rigid as well as polarized. As a result, custom waveguide port locations and orientations are frequently desired for system integration purposes. In addition, hermetically sealed waveguide windows are often needed. Such windows are difficult to implement and expensive to produce [13]. Alternatively, a Uni-Guide waveguide connector can preserve the hermetic seal of a component package if standard glass beads are used for RF port sealing.
The Uni-Guide waveguide connector series can also help to minimize packaging inventories, as well as simplify inventory management. Component manufacturing costs are lowered by using standard housings that were already developed for coaxial connectors. As a result, the waveguide connector is expected to be a revolutionary package technology that eliminates many custom-designed packages while reducing product development cycle times and lowering design and production costs.
Figure 2 • E-Band subsystem constructed using COTS modules
Test Challenges
Testing waveguide components and systems has always been challenging. Any misalignment, small gap, or excessive mechanical tolerance in waveguide connections can lead to inaccurate system calibration, time-consuming adjustments, and redundant efforts. There are many research papers and application notes that address these topics [15], [16], [17]. To increase testing and calibration speeds, as well as improve overall repeatability, Eravant developed the Proxi-FlangeTM contactless waveguide flange and the WaveGlideTM rail positioning system. These innovations greatly improve the productivity and reliability of mm-wave test systems [18]. Additionally, lower-cost mm-wave VNA frequency extenders and calibration kits are available to further lower the cost barriers for new and budget-limited system developers.
Millimeter-wave System Configuration
Communication and radar systems tend to include a variety of component types in their functional block diagrams. Antennas, antenna feed networks, amplifiers, up-converters, down-converters, control devices, oscillators, filters, diplexers, power distribution networks, and interconnection products are commonly used. Figure 1 shows the block diagram for a typical microwave or mm-wave communication system front end. Various common components, i.e., building blocks or modules, are included. These functional modules are largely available from the Eravant website as standard COTS products covering frequencies up to 220 GHz, with an increasing number of component selections available for frequencies approaching 325 GHz.
Eravant’s large pool of standard products, as well as its extensive design library and established manufacturing processes, support the rapid development of customized modules by altering or updating standard models at a fraction of the usual costs and lead times. This design approach speeds up system development, reduces the realization cycle time, and lowers overall program costs. Further, these products are readily configured with standard waveguide interfaces for direct connections to other components. Additional interconnection devices, such as adapters and transitions, as well as waveguide components including straights, bends and twists, may be eliminated when individual modules are offered with more configuration options.
With approximately 5,000 COTS models, Eravant stands ready to support rapid system concept demonstration, scientific apparatus development, test-lab tooling, as well as a variety of instrumentation needs. All of Eravant’s products are designed and manufactured using rigorous processes. Many are readily qualified for space and military applications after the required environmental testing is performed.
Signal Sources
Clean and stable signal sources are critical building blocks in radar and communication systems. Tuning bandwidth, tuning speed, phase noise, harmonic content, spurious levels, and frequency stability are primary considerations that often determine overall system performance. Since its beginning, Eravant has been a leader in the production of low-noise, free-running mm-wave Gunn oscillators. Their economy and excellent phase-noise performance make them ideal signal sources in many applications. Law-enforcement radar, speed sensors and security systems are a few of the main focuses of this product line. With output frequencies available from 9 to 110 GHz, mechanical tuning ranges are as wide as +/- 12 percent at various output levels. For example, a 35-GHz fixed-tuned Gunn oscillator delivers +12 dBm output power with phase noise of -95 dBc/Hz at a frequency offset of 100 kHz (Fig. 3). Frequency stability is -0.3 MHz/°C. Another example is a 90-GHz mechanically tuned oscillator that operates with +16 dBm output power and +/- 2.0 GHz tuning bandwidth. These cost-effective oscillators can be tuned within the range of 30 to 38 GHz or 90 to 100 GHz. Options include bias-voltage tuning and higher output power to meet the needs of various applications.
Voltage-tuned Gunn oscillators are widely used in range/speed radar sensors and in frequency-agile systems. They achieve industry-leading value and performance across the 9 to 96 GHz spectrum.
Figure 3A: 35 GHz low cost fixed-tuned Gunn oscillator (SOL-35312-28-G1)
Figure 3B: 90 GHz mechanically tuned Gunn oscillator (SOM-90304316-10-M1)
Free-running oscillators with greater frequency stability are often required in applications such as semiconductor reliability testing and qualification. Such signal sources are typically realized using dielectric-resonator oscillators (DROs). Sources with precision frequency stability are offered as fixed-frequency Phase-Locked Dielectric Resonator Oscillators (PLDROs) or as wide-band frequency synthesizers. Both oscillator types can be locked to either an internal crystal oscillator or a more stable high-performance external clock such as a GPS frequency reference. Basic DRO and PLDRO models cover the frequency range of 3 to 40 GHz, with nominal output power of +10 to +20 dBm.
Standard frequency synthesizer modules can be tuned from 0.1 to 10 GHz or 0.2 to 20 GHz with 10-kHz frequency resolution. Models configured for high-speed tuning can achieve settling times of 3 microseconds. These oscillators use a coaxial connector for the RF output port. Sources with higher frequency or higher output power can be achieved by adding frequency multipliers and amplifiers. DROs, PLDROs, and frequency synthesizers can be hermetically sealed and can cover wide temperature ranges to meet aerospace and millitary standards (Figure 4).
Figure 4A: 22 GHz free-running DRO model SOD-22301215-SF-S1
Figure 4B: 28 GHz phase-locked DRO, model SOP-28310113-KF-BB
Figure 4C: 0.2 to 20 GHz high-speed synthesizer, model SOT-02220313200-SF-B6
With add-on frequency multipliers, these oscillators can readily reach 220 GHz.
Frequency Multipliers
Frequency multipliers are often used to achieve higher operating frequencies. Both passive and active frequency multipliers are available. Passive multipliers typically employ a Schottky diode and generally use multiplication factors of 2, 3, or 4. They can provide frequency coverage over full waveguide bands up to 220 GHz. Passive multipliers require no DC power and have relatively straightforward implementations, but they exhibit significant conversion loss. In contrast, active multipliers combine passive multiplier circuits, filters and amplifiers to provide higher output power and higher multiplication factors. Most of Eravant’s active frequency multipliers can cover full waveguide bandwidths. Alternatively, units with narrower frequency coverage can be designed to reach other performance goals.
One of the consequences of frequency multiplications is that noise sidebands contained in the primary signal are also multiplied, and are extended to higher offset frequencies. The carrier-to-noise ratio at a given offset frequency is degraded by 20 Log(N), where N is the multiplication factor. Harmonic and sub-harmonic content may also be a consideration. To minimize these effects, innovative circuit designs and rigorous manufacturing processes are incorporated into Eravant’s high-performance frequency multipliers (Figure 5).
Figure 5A: 110 to 170 GHz passive X2 multiplier, model SFP-06212-S2
Figure 5B: 75 to 110 GHz active X6 multiplier, model SFA-753114616-10SF-E1
Frequency Converters
Frequency converters are key functional blocks in most microwave and mm-wave systems. They typically translate the spectrum of an input signal from an Intermediate Frequency (IF) to a Radio Frequency (RF), or vice-versa. Eravant’s frequency converters support IF signals from DC to 40 GHz using Local Oscillator (LO) and RF frequencies spanning 18 to 220 GHz. Frequency converters are often referred to as “mixers”. Converter types include Fundamental (balanced), Subharmonically Pumped (LO ≈ 1/2 RF), Harmonic, Quadrature, Image-Reject, and Single-Sideband (SSB). Frequency converters are also categorized as biased or non-biased. The choice usually depends on the system frequency plan and the available DC power budget.
For example, Eravant’s W-band mixer family includes a non-biased balanced fundamental mixer, model SFB-10-N1, that covers the RF/LO frequency range of 75 to 110 GHz with an IF frequency range of DC to 40 GHz (Figure 6A). The mixer requires +13 dBm LO power, which can be difficult or costly to obtain at higher mm-wave frequencies. As an alternative, an externally biased converter, SFB-10-E2, requires LO power of 0 dBm. A subharmonically pumped converter (Figure 6B), model SFS-90311415-102FSF-N3, may be used with LO signals from a lower-frequency source that is chosen for performance or cost considerations. A possible motivation for using a subharmonically pumped mixer is to double the frequency separation between the source’s Fundamental and Spurious outputs by a factor of two. This result can ease the filtering of unwanted harmonic and spurious content. A quadrature mixer, model SFQ-75311415-1010SF-N1-M, can be configured for single-side-band rejection or for various modulation schemes (Figure 6C). By adding an external IF hybrid circuit, an image-reject mixer can be configured for communication and radar applications without using the more costly approach of incorporating two balanced mixers, a hybrid coupler, and a power combiner. The quadrature mixer approach can effectively reject spurious signals when the IF is low or is very close to the carrier frequency. Harmonic mixers such as model numbers SFH-06SFSF-A3 and STH-10SF-S1 are available for expanding the frequency range of spectrum analyzers to the W band 75 to 110 GHz and D band 110 to 170 GHz, respectively, when system performance testing is required (Figures 6D and 6E).
Figure 6A: 75 to 110 GHz non-biased balanced mixer (SFB-10-N1)
Figure 6B: 90 to 110 GHz subharmonically pumped mixer (SFS-90311415-102FSF-N3)
Figure 6C: 75 to 110 GHz quadrature (I/Q) mixer (SFQ-75311415-1010SF-N1-M)
Figure 6D: 110 to 170 GHz harmonic mixer (SFH-06SFSF-A3)
Figure 6E: 75 to 110 GHz spectrum analyzer mixer (STH-10SF)
Design considerations for mm-wave up-converters are similar to those encountered with down-converters, i.e. when using the IF port as the input and the RF port as the output. For double-sided and single-sideband modulator applications, system considerations are similar to those encountered when mixers are used as up-converters. However, modulators generally use the IF port for the modulation input, while the LO and RF ports are used for the unmodulated input and modulated output signals, respectively. The modulation signal usually requires relatively high power, typically +16 dBm, to fully bias diodes or other modulation devices. The LO port is typically fed with smaller signal, such as -20 dBm or lower.
Amplifiers
Amplifiers are essential functional blocks that make up or compensate for signal loss. They are also used to increase signal levels during transmit and receive processes. Three types of amplifiers are typically encountered. They include power amplifiers, low-noise amplifiers, and general-purpose gain blocks. Eravant designs and manufactures its amplifiers to cover full-waveguide bandwidths, or bandwidths that are as wide as possible, as the first priority. This allows COTS products to accommodate as many applications as possible. When wide bandwidth is not supported by the available technology, an application-specific bandwidth may be chosen. For general-purpose gain blocks, ultra-broad bandwidth is the main focus so that a limited number of models can support a majority of developmental programs and application needs.
Currently, low noise amplifiers cover the frequency range of several GHz up to 270 GHz. In a WR-03 waveguide package, model SBL-2242741585-0303-E1 provides 15-dB gain with 8.5-dB noise figure from 220 to 270 GHz (Figure 7A). A full-band amplifier, model SBL-1141743065-0606-E1, provides 30-dB gain from 110 to 170 GHz with 6.5-dB noise figure (Figure 7B).
Figure 7A: WR-03 low noise amplifier
Figure 7B: 110 to 170 GHz low noise amplifier
Power amplifiers typically incorporate GaAs, InP or GaN semiconductor technologies. Power combining techniques in planar circuits and in waveguide configurations are utilized to achieve the best electrical and mechanical performance. The frequency range of Eravant’s power amplifier family is 2 to 230 GHz. An E-Band model spans 71 to 76 and 81 to 86 GHz for last-mile applications (Figure 7C). The highest linear output power delivered from the E-band power amplifier is 1.5 Watts. A Ka-Band model can be used in 35-GHz radar systems (Figure 7D). The Ka-band power amplifier yields 10 Watts of output power at saturation. Several power amplifiers capable of delivery up to +10 dBm linear and +13 dBm saturated output power are available in the frequency range of 110 to 230 GHz with tens of GHz bandwidth. The model SBP-2142341507-0404-E1 covers the frequency range of 210 to 230 GHz with 15 dB typical linear gain and +10 dBm saturated output power. The photograph of the power amplifier is shown in Figure 7E.
Figure 7C: E-Band power amplifier
Figure 7D: Ka Band power amplifier
Figure 7E: WR-04 power amplifier
The third amplifier family comprises broadband gain blocks. The focus of these amplifiers is to offer moderate output power and maximum bandwidth for system gain boosting. Examples include coaxial amplifiers that operate from 10 MHz up to 70 GHz. They are offered as standard models to satisfy many system applications. Figure 7E shows model SBB-0117033015-VFVF-E3. It provides 30-dB small signal gain and 15-dBm P-1 dB output power from 10 MHz to 70 GHz with a typical noise figure of 6.0 dB. A companion model with lower gain is also available as a COTS option.
Figure 7E: 10 MHz to 70 GHz Broadband Amplifier
While configuring the standard amplifier models, two gain levels are considered for system applications, namely, the intermediate gain range of 15 to 20 dB and the higher gain range 30 to 35 dB. Other gain values can be readily configured using these standard gain values. All amplifiers are designed and manufactured with integrated voltage regulators and a single DC power input to reduce system integration complexity.
Control Devices
Control devices are typically employed in systems where control of the signal’s amplitude, phase or destination is required. The three most common control devices include attenuators, phase shifters and switches. Standard units fulfill control functions in the frequency range of 0.5 to 110 GHz.
The switch family includes electrical, mechanical, and electro-mechanical types configured as single-pole and multiple-pole models. A SPDT switch, model SKD-5031146025-1F1F-R1-M, covers the frequency range of 50 to 110 GHz using 1-mm connectors for its RF ports. Waveguide-interfaced models are available to cover major mm-wave bands in the 0.5 to 43 GHz frequency range for 5G applications. Figures 8A and 8B show typical examples. Electrical switches based on PIN diodes can provide fast switching times. Mechanical and electro-mechanical models are suitable for applications that require higher port isolation and lower insertion loss. Examples include models SWJ-28-M1-H (22 to 44 GHz) and SWJ-10-TS (75 to 110 GHz). They are shown in Figures 8C and 8D.
Figure 8A: A SPDT Switch operates from 50 to 110 GHz using 1 mm coaxial connectors
Figure 8B: A SP4T switch covers the 5G FR2 bands up to 43 GHz
Figure 8C: Model SWJ-28-M1-H is a DPDT mechanical switch that covers the 5G FR2 band, 24 to 43 GHz
Figure 8D: Model SWJ-10-TS is an electro-mechanical DPDT switch that operates from 75 to 110 GHz
Similarly there are three types of attenuators, namely electrical, mechanical, and electro-mechanical. Electrical attenuators employ PIN diodes or pHEMT devices to achieve high-speed amplitude control with moderate insertion loss and good isolation performance. Other models provide low insertion loss and high attenuation. Figures 8E, 8F, and 8G show examples of V-band attenuators. They include electric, fixed, and level-setting waveguide types. Direct reading and programmable attenuators belong to the test and measurement categories (Figures 8H and 8I). The entire attenuator family covers frequencies up to 325 GHz. The majority of attenuator products serve either a full waveguide bandwidth or an application-specific bandwidth. Electrically tunable attenuators have tuning speeds on the order of 100 ns. While the electrical tunable versions focus on tuning speed, manual or electro-mechanical tunable versions can reach attenuation values up to 60 dB with low minimum insertion loss.
Figure 8E: Attenuator model SKA-5037533030-1515-A1 supports full V-Band operation
Figure 8F: A V-Band fixed waveguide attenuator
Figure 8G: A V-Band level-setting waveguide attenuator
Figure 8H: A V-Band direct-reading attenuator, model STA-60-15-D5
Figure 8I: Programmable attenuator STA-60-15-P1 spans 50 to 75 GHz
The phase shifter is another important component in many microwave and mm-wave systems. Electrically and mechanically tuned phase shifters can be found in many test labs. Examples are shown in Figures 8J and 8K. Operating frequencies for standard phase shifters extend to 170 GHz.
Figure 8J: A digitally controlled electrical phase shifter covers 6 to 18 GHz
Figure 8K. A WR-15 mechanically adjustable phase shifter operates from 50 to 75 GHz
Other control devices such as power limiters are also developed and manufactured as COTS components.
Ferrite Devices
Ferrite non-reciprocal devices are widely used in microwave and mm-wave systems to provide port isolation or to control signal flow. Isolators are two-port devices that are mainly used for port impedance matching improvement or to guide signals in a certain direction. Circulators can be used for either port isolation or for signal duplexing. Faraday rotation isolators can generally cover full waveguide bandwidths with high levels of isolation. Junction isolators and circulators offer lighter weight and more compact size. Although junction isolators and circulators are typically optimized for specific bandwidths, recent progress has advanced some junction isolators and circulators to cover full waveguide bands in WR-42, WR-34 and WR-28 with somewhat compromised performance degradation, such as slightly higher insertion loss and lower isolation at the band edges. Figures 9A, 9B and 9C show a G-Band Faraday isolator, a W-Band narrow-bandwidth junction circulator, and a Ka-Band full waveguide band junction circulator, respectively.
A number of coaxial ferrite devices are also offered as COTS products. The frequency coverage of ferrite devices is from 8.2 to 220 GHz over specific bandwidths. Recent progress has been made to miniaturize Faraday rotation isolators. Currently, compact and miniature Faraday isolators are offered as COTS products that cover operating frequencies up to 220 GHz. Their smaller size can enhance system integrations. Figure 9D shows a G-Band miniature Faraday isolator that spans 140 to 220 GHz with 4-dB typical insertion loss and better than 20-dB isolation, with a physical length just over one-half inch.
Figure 9A: A G-Band Faraday isolator covers 140 to 220 GHz
Figure 9B. A W-Band junction circulator targets 92 to 98 GHz
Figure 9C: A Ka-Band, full waveguide band circulator spans 26.5 to 40 GHz
Figure 9D: A G-Band miniature Faraday isolator covers 140 to 220 GHz