PCB-based small antenna design

The antenna design usually does not cause much attention until the end of the design cycle. The reason may be that because they are passive devices, the role played in the RF signal path does not appear to be large. It may also be because the designer wants them to always have the ability to configure the antenna design and component selection in the remaining space. It is also possible that the antenna is not the beneficiary of Moore's Law. Regardless of the specific reasons, portable wireless product designers are now faced with many new tools, new methods, and new components, adding a new compromise to the pursuit of ideal antennas. In this process, designers need to make a tough choice between “building and buying”. Unlike active devices that you have to pay for, it is entirely possible to make an antenna for free at a price of a few square centimeters on the PCB. In many cases, this is an attractive and viable option. But in many other cases, the cost of enjoying this obvious "free lunch" is too high. Now, with some newer antenna designs and devices, designers have another option.

The electrical definition of a small antenna is that the basic component has a dimension that is shorter than 1/10 of the wavelength. For a 300 MHz signal, the defined threshold is 10 cm; at 1 GHz, the value is only 3 cm.

Traditionally, small antennas offer only limited performance. If you want truly efficient antenna performance, you need to extend more metal into the air and use multiple or complex shapes to increase gain, control bandwidth, change field pattern, or reject adjacent signals. In addition, you must ensure that the antenna impedance matches the RF front end to maximize power transfer. One of the main benefits of the transition to higher frequency RF, represented by cellular handsets and Wi-Fi, is that small antennas are electrically feasible in these applications.

Although the "family spectrum" of the antenna is complicated, whether you use a PCB-based antenna or a discrete antenna, this branch containing a small antenna needs to be well compromised in the design.

The PCB antenna can be either a small piece or a small ring, and can be spiral or linear. Their BOM costs are negligible and only take up PCB space. It is worth noting that some PCB antennas are not part of the main board, but are separate devices that are usually attached to the product housing. Its performance also depends on the layout, geometry and its relative position to nearby components. In addition, the user's hand, body or head often adversely affects antenna performance. Any changes in the component or PCB layout in the product will affect the antenna performance. Therefore, this design is limited, and the modification of the final product before the finalization is not significant.

On the other hand, modifications to the antenna—whether to accommodate changes in the specification or to overcome design flaws—can be performed quickly once determined and do not affect the BOM. Modifications also affect the antenna impedance. So, you may also need to change the designed matching circuit.

In contrast, discrete antennas involve BOM costs and are typically designed by the vendor based on specific frequency bands, bandwidth, and other performance parameters. In return, this antenna occupies less PCB space than the PCB antenna, and if it is not completely unaffected, the PCB layout, adjacent devices, or the user's impact on it is much smaller. The antenna impedance is fixed by the physical design, so the matching network is also fixed and independent of the layout and device placement. These factors prevent designers from facing certain challenging constraints and eliminating the need to redesign PCB layouts and calculate BOMs.

There are many possible design approaches for small antennas that are typically based on PCBs. The most common are bright line (also known as open) structures (such as bipolar and monopole antennas), loop designs (such as loop antennas), and solid block designs.

The open antenna is actually a miniature version of the large antenna that has existed since the beginning of the radio technology. In fact, because Heinrich Hertz used a dipole antenna in his 1888 experiment, it is sometimes called a Hertz antenna. It is balanced with the ground plane, and it has faithfully fulfilled the duties of the VHF TV rabbit ear antenna before the advent of cable and satellite television.

In contrast to bipolar, monopole antennas are single-ended to ground, so a ground plane is required. In many wireless applications, monopole antennas are used as whip antennas. It is also known as the Marconi antenna, which was used in Marconi's early experiments.

Figure 1: Loop antenna (a) is easy to implement. Rectangular block antenna (b) uses well-planned PCB space (it can also be a discrete device)

Loop antennas used in mass markets such as UHF TV also have a long history. Its perimeter is approximately equal to the wavelength at which the signal can be received (see Figure 1a). In terms of electrical characteristics, a rectangular block antenna is a wide microstrip transmission line whose length is half of the operating wavelength. In Figure 1b, the wavelength is not calculated as propagation in vacuum, but is calculated from insulated PCB material. The resonant frequency band of a rectangular block is quite narrow, so its operating bandwidth is also quite narrow—about 5% of the nominal center frequency, which is good or bad, depending on the application.

All three types of antennas can be implemented with a PCB, and a multi-layer PCB can provide multiple design options, including the ground plane required for some structures. This type of antenna design is used in applications where the performance requirements are not critical, such as remote unlocking (RKE) and garage openers.

Since the design cost of the PCB antenna is negligible, when and why does it make it a design priority? Several of these dominance factors are related to front-end design and actual implementation.

First of all, the antenna design is not simple. Even with modeling programs like the Numerical Electromagne TIc Code, circuit or system engineers are new to the electromagnetic world. They are faced with a world of electromagnetic fields rather than specific voltage and current points or flows of electrons in a fixed loop.

Second, as with many engineering designs, competing and conflicting properties like center frequency, bandwidth, field mode, efficiency, and lobe and gain make it difficult to balance between them.

Third, evaluating antenna performance is not easy. It requires special test equipment, a non-reflective chamber or an open area. It also takes time, money and expertise. In addition, when assessing the impact of the user's hand on the antenna, or conversely, assessing the effect of antenna radiation on the user's hand, the correct test setup is performed, including physical copying of the human hand and head.

And these are all theoretical. In fact, there are other factors at work. The antenna of course occupies PCB space, and its performance properties are significantly affected by nearby devices as well as the user's hands, head and body. The relative dielectric constant of human tissue is 40, and the dielectric constant of the PCB component is about 25 to 85, so human tissue will excite the resonant element and affect the magnetic field.

In addition, when multiple antennas are required for multi-band operation or shape diversity design, the sympathy between several PCB-based antennas and between the antenna and nearby areas will make performance prediction very difficult and sensitive to subtle layout changes.

However, there are also rules that constrain the specific absorption rate (SAR) of the antenna field. SAR is the ratio of the ability of the mass (human tissue of this example) to absorb RF; it is usually measured by two methods: one is to measure the temperature rise due to absorption; the other is to measure the electric field of the fluid that simulates human tissue. More information is available on the Federal Communications Commission (FCC) website. The near-field and far-field performance of the antenna must be understood and analyzed, and they may be closely related.

Finally, the antenna is not isolated from the wireless device's receive front end or transmit power amplifier stage. The circuit designer must determine the impedance of the antenna and associated stages and then design a matching network to maximize power transfer over the entire target bandwidth (see Figure 2).

Figure 2: The antenna subsystem consists of a front-end receive or transmit amplifier, a matching network, and the antenna itself.

This is often a difficult design task involving professional calculations and measurements as well as specialized tools, such as the need for a Smith chart.

Fortunately, the development of materials science and antenna theory has provided design engineers with alternatives to both external and PCB-based antennas. These antennas maximize the volumetric efficiency of the antenna while overcoming or virtually eliminating the uncertainty of layout effects and matching. In contrast, block and whip wireless are two-dimensional, and their performance depends mainly on the space rather than the volume. While discrete antennas add significant cost, they often reduce size while improving or ensuring product performance.

It may sound unreasonable, and the dielectric as an insulator plays an important role in antenna design and implementation. But it is true that for more than 50 years, the dielectric has been part of the antenna design, which helps shape and manage the antenna mode electric field. Field energy accumulates at a relatively high density and is stored in the dielectric, so external objects or fields have relatively small effects and do not affect the natural resonance of the antenna.

Of course, the high relative dielectric constant is only one of the key factors for the success of dielectric based antennas. Materials also require low dielectric losses (high Q materials) and low temperature coefficients to minimize physical scale variations that can result in loss of harmonics.

For example, Sarantel's Geohelix antenna in the UK uses a unique ceramic material and shape that has the ability to reduce near-field radiation by a minimum of 90% compared to bulk and whip antennas. The near field affected by the user's hand and body is almost completely enclosed within the Sarantel antenna. The antenna is used as a bandpass filter to reject out-of-band signals while also removing the ground plane on the PCB or chassis.

The ground plane is no longer needed as a result of the balanced current of the design, since the sum of the current flowing into the antenna is zero, so its resonance is independent of the PCB or package. In contrast, a microstrip-based block or external whip antenna requires a ground plane to achieve resonance, and the current flowing into (or out of) the antenna requires a complementary current to be generated at the ground plane to generate resonance.

Similarly, another British company, Antenova, has a high dielectric antenna technology that provides a volumetric non-symmetric antenna for omnidirectional, directional and even multi-band applications with the advantage of frequency response above 10 GHz. These highly efficient devices are relatively immune to near detuning and effects. By combining these devices to create a smart antenna with excellent handling, smart antennas are increasingly being used in base stations to expand system capabilities while improving the performance of each call.

For example, Antenova has developed a dual-band hybrid IEEE 802.11a/b/g antenna for wireless LAN coverage of 2.4 to 2.5 GHz and 4.9 to 5.9 GHz dual-band with 4 & TImes; 4 & TImes; 20 mm volume (see figure 3).

Figure 3: Antenova's high dielectric hybrid antenna provides multi-band performance in a compact package

The antenna has three components: a microstrip feeder that is also matched to a 1.2mm diameter, ultra-small coaxial cable feed to the antenna; a transmitter consisting of a 1/4 wavelength block and two resonators (one per band) And a ceramic particle that is responsible for exciting the radiating element and forming a coupling between the radiating element and the feed line.

Not all of these new antennas are centered on ceramics. Barcelona's Fractus uses geometric models based on irregular fragment geometry for its antenna-in-package design. The multi-band antenna can be printed on a substrate or embedded in a chip. They offer a GSM antenna with a transmission efficiency higher than 70%, a peak gain higher than 1dBi, and a VSWR lower than 1.5:1. The antenna has a 50 ohm unbalanced impedance and is only 10 & TImes in volume; 10 x 0.9 mm.

As a comparison of block and discrete antennas from the same supplier, Centurion Wireless Technology (now part of Laird Technologies), Centurion offers a microstrip block antenna that can be attached to the product's housing. It operates in the 2.4 to 2.5 GHz band and is 43 x 43 x 1.65 mm in size. The antenna has a gain of 5.1 dBi and a VSWR of less than 2.5:1 in the operating band. The gain of the BlackChip surface-mount antenna of the same frequency band of Centurion is greater than 2dBi, and the VSWR is less than 2:1, 8×6×2.4mm.