by Peter Matthews, Knowles Precision Devices
What Does mmWave Mean?
Today, communication devices are constantly being pushed to transmit more data and to transmit it faster, which means these devices need to operate at higher frequencies such as “millimeterWave” (mmWave). However, the definition of mmWave is not always a clear one. A common generic definition for mmWave is the frequencies between 30 GHz to 300 GHz because the wavelengths in this range in free space can be measured, conveniently, in millimeters. We can actually see how this works if we round up the speed of light to 3×108 m/s and do a quick back-of-the-envelope calculation (Table 1).
However, it is important to note that the term “microwave” is not derived in the same way. In fact, we commonly refer to the FR2 bands of 5G new radio (NR), which is currently defined in 3GPP TS 38.104 V17.2.0 (2021-06) as the frequencies between 24.25 GHz to 52.6 GHz, as mmWave. You will notice that in this definition of mmWave, some centimeter wave frequencies have been added into the mmWave range as well. In this next section, we will dive into the reason behind this expanded “mmWave” frequency range for 5G and explore some additional applications taking advantage of mmWave frequencies.
A Brief Overview of Where mmWave Frequencies are Commonly Used Today
Operating various RF devices at mmWave frequencies is attractive for a few reasons. First, point-to-point operations in this frequency range can reduce wavelength, allowing for conveniently sized antennas to provide focused, narrow beams aimed directly at a receiving antenna. These focused beams prevent potential interference from nearby equipment using the same frequency. Additionally, with the wider bandwidths possible at these higher frequencies, military devices can achieve higher resolution for close-range targeting and SATCOM and 5G NR devices can achieve higher data-rate throughput.
Looking more closely at SATCOM applications first, Figure 1 provides an overview of some common systems relative to the IEEE frequency band nomenclature.
In Figure 1, you can see that high-throughput satellite systems such as Starlink and military advanced extremely high frequency (AEHF) start to show up at these increased frequencies in and around the mmWave range. This is because these applications are taking advantage of the “fatter pipes” that increased, and less crowded, bandwidth has to offer. Put simply, as a result of these “fatter pipes” the same amount of data can be sent faster, or we can send more data in the same amount of time.
As a result of these increased data rates at higher operating frequencies, 3GPP created two operating ranges as shown in Table 2. FR1 covers part of what 4G LTE uses and stops at ~7GHz while FR2, as we mentioned, is referred to as the mmWave band, and covers operating frequencies from 24.25 GHz to 52.6 GHz.
To fairly divide sections of these FR2 operating bands for 5G in the U.S., the Federal Communications Commission (FCC) held three spectrum auctions in recent years. Auction 101 was referred to as “28 GHz” and covered 27.500 GHz to 28.350 GHz; Auction 102 was referred to as “24 GHz” and covered 24.25 GHz to 24.45 GHz, and 24.75 GHz to 25.25 GHz; and Auction 103 was referred to as “Upper 37 GHz, 39 GHz, and 47 GHz,” and covered 37.6 GHz to 38.6 GHz, 38.6 GHz to 40 GHz, and 47.2 GHz to 48.2 GHz.
When looking at these auctions and considering the information in Figure 1 about various SATCOM device operations, it is worth noting the overlap between 5G NR FR2 and several satellite communication systems. For example, Auction 102 places 5G transmit and receive signals very close to one of the bands for a satellite application called Earth Exploration Satellite Service (EESS) that operates at 23.6 GHz to 24 GHz. Therefore, to avoid interference, high-selectivity filtering options are needed for the 5G devices operating around this frequency range (and we will explore options a little later).
The New Design Challenges of Working with Smaller Wavelengths at mmWave
In general, as frequency increases wavelength decreases. This is challenging because in phased array antennas an inter-element spacing of less than half the wavelength (<λ/2) is required to mitigate grating lobes. This means a 28 GHz band antenna for example would need approximately 5 mm of inter-element spacing. As a result, systems must be designed to be much more compact, which means some of the technologies that worked at 4G frequencies will not work at all at FR2. To better illustrate the shift in system design required for higher frequencies, Figure 2 shows the physical scale of different wavelengths compared to some familiar references, starting at 700 MHz and moving up to 39 GHz.
Since size is clearly at a premium in mmWave applications, dimensional tolerances of antenna and filter technologies are also critical. Poor tolerance encroaches on potential board space or layers that could be used for adding other devices or functionality. Another design issue in these densely packed systems is that there really is not a way to control temperature, which means frequent variations may occur and systems will run hot. Therefore, components such as filters must perform within specification over a wide range of temperatures with a temperature stability of approximately 3 ppm/°C. Let’s look at some of the other challenges specifically related to filtering.
Figuring Out the Right Filtering Technologies for mmWave
As shown in Figure 3, there is not a filtering technology available that can operate across all frequencies and all bandwidths. Instead, an application’s operating frequency will dictate the potential filtering options available. One big difference for mmWave applications versus applications operating at lower frequencies is that traditional lumped element filter designs are no longer an option. This is because at higher frequencies, it becomes difficult to manufacture the necessary discrete values without causing various other issues. As a result, distributed element approaches such as variations on a waveguide or transmission line are utilized.
As you can see in Figure 3, there is not a one-size-fits-all solution for filtering that can operate effectively across all frequencies, but there are three filtering technologies that work well at mmWave frequencies – planar transmission line, cavity, and waveguide. Depending on the power handling, size constraints, and bandwidth requirements you are working with, one of these approaches is likely to be the best fit for your 5G mmWave application.
Let’s now look more closely at the filtering options available with two different extremes – waveguide and a microstrip transmission line. Table 3 shows a traditional comparison between these two options.
Highlighted in blue in Table 3 are the main trade-offs an RF engineer needs to consider when it comes to achievable bandwidth, loss, and size. It should also be noted that this table represents a more textbook point of view of these two filtering options. There are physical implementations of waveguide, such as a substrate integrated waveguide (SIW), which can reduce size. There are also material innovations that can substantially improve loss performance of planar transmission lines. Let’s briefly look at these innovations more.
One new trend for developing waveguide filters that can address both size and fabrication cost concerns that we are digging into now at Knowles Precision Devices is the use of SIWs. An SIW is basically a rectangular waveguide in which the single conductor walls are formed by plated surfaces and vias and where transvers electric (TE) mode is used (Figure 4). With TE mode, waves propagate in the high dielectric constant materials. The key to developing an effective SIW is a thorough understanding of the materials that need to be used to replace the air dielectric as well as the ability to control the structures in a precise way that is not cost prohibitive.
The main concern with using a microstrip approach as shown in Table 3 is the loss experienced compared with a waveguide approach. However, the industry is currently innovating on various materials to help address this constraint. The evolution of high K materials, such as Knowles Precision Devices PG, CF, and CG, are making it possible for RF engineers to develop low-loss microstrip transmission line filters.
Figure 5 demonstrates how a microstrip transmission line configuration using Knowles Precision Devices’ PG ceramic reduces loss compared to the commonly used RO4350_LoPro substrate. In plot A, loss/inch is shown while plot B shows loss/wavelength. Note, the same material thickness, metal conductivity, and 50-ohm microstrip transmission line were used in both configurations.
At frequencies above 13.6 GHz, PG material overall has better loss performance. This is largely because the surface roughness of PCBs is about four times greater than that of polished ceramics. Additionally, as PCB materials are made thinner, the loss delta will become larger because surface roughness will increase. Therefore, the frequency range where polished ceramics outperform PCBs will grow. Combined with the benefits of loss/wavelength at higher frequencies, integrating passive RF components on polished ceramics is clearly the lower-loss option compared to PCB materials.
Making a Smooth Transition to 5G mmWave
As discussed, the definition of mmWave is somewhat slippery but hinges on the size of the waves being manipulated. As an engineer working on mmWave RF applications, it’s important to remember that at these higher frequencies, you will be working with much smaller wavelengths, which means using waveguide and transmission line technologies will likely be necessary. But also keep in mind that the right filter approach for your application will heavily depend on your specific needs. A specialty component supplier, such as Knowles Precision Devices, has the experience working with these technologies to help you optimize the performance of your mmWave systems.