Implementing antenna diversity inside small devices

The purpose of antenna diversity is to take advantage of the different paths of a wave propagating in a reflective environment in order to improve overall system performance. With respect to antennas, different kinds of diversity can be considered.

Spatial diversity is probably the most well known type. A simple derivation of spatial diversity is already implemented on CardBus devices and notebooks for Wi-Fi applications. In this case, two antennas are integrated into the device with at least 35 mm between both since the CardBus width is about 50 mm. The process consists of switching from one antenna to the other and selecting the antenna offering the best signal. In fact, due to multipath, the power distribution function of space presents nodes and nulls in close proximity, resulting in rapid change in power versus space. By using two antennas, even if one is in a null the other will most likely receive a much higher level of signal.

A more elaborate spatial diversity solution consists of processing the signals coming from the different antennas rather than just using the best one. By doing so, it is possible to put the different paths in phase and then add them. The theoretical optimum performance of 3 dB of gain improvement is obtained for a distance between the antennas of approximately 38% of the wavelength. In fact, this distance corresponds to a correlation factor equal to zero, which means that the delays between the different paths are perfectly determined.

In practice, distances between 17% to 38% of the wavelength are considered, reducing the gain improvement down to 1.5 dB. This simplistic method is one of the different algorithms existing today. Depending on the principals, the optimum distance may vary.

Beam diversity is slightly different from spatial diversity. For spatial diversity, the antennas are the same, with the same radiation pattern and same coverage. It is just the different position in space, which allows the antennas to receive different signals. Beam diversity uses different antenna shapes or technologies to take advantage of different radiation patterns. The coverage due to the different radiation pattern shapes allows the reception of signals coming from different directions, due to multipath. Then, the different signals are processed in a similar fashion as in a spatial diversity solution. In this case, the distance between the two antennas is theoretically not as critical. The important point is the correlation factor between the different radiation patterns, which shows how the radiation patterns overlap.

Meanwhile, polarization diversity uses the antennas' orthogonal polarizations. If the antennas are the same, they should be physically turned 90 degrees from one another. In this case, the distance between the antennas is not theoretically critical. The level of cross-polarization of each antenna is much more important for good diversity gain.

No antenna offers a perfect diversity solution. It is usually a combination of these different types of diversity, which is put into practice, especially on small, handheld devices, where the antennas are in close proximity to one another and the human body. But this is not a major issue as long as the correlation factor is low enough. The most critical parameter regarding the correlation factor in any type of diversity solution, after these parameters, is the coupling between the different antennas.

Understanding EM coupling
Antenna isolation must always be considered during the design stage. An isolation of 15 dB or more is required to keep a low correlation factor. If the isolation is lower than 15 dB, more than 3.5% of the signal received by one antenna is transmitted to the neighboring one, reducing the diversity gain. Therefore, it is important to understand the electromagnetic coupling in order to keep a high diversity gain.

Isolation refers to the way an antenna interacts with its surroundings. The less the antenna is disturbed by its surroundings, the higher the isolation. In the case of multiple antennas, the level of isolation can be determined by measuring the energy directly transmitted from one antenna to another, also called coupling, using the S21 parameter reading on a network analyzer. There are basically three ways for antennas to be coupled:

• Current Coupling. Antennas generate a high current within their immediate surroundings. If another antenna is in the same area, which is often the case in small devices, the coupling can be pretty high. If the antennas are not isolated, the correlation factor will also be high, which means there is no advantage in using diversity. Depending on the antenna technology used, the size of the area with high currents is more or less confined close to the antenna. One way to ensure that current coupling remains low, is to use a technology which confines the current.

• Free Space Coupling. In this case, the radiated wave in the near field of the antenna goes directly from one antenna to the other. Reduced coupling can be achieved by shaping the near field of the antenna away from other antennas.

• Surface Waves. Antennas may generate surface waves, electromagnetic modes that are trapped between the ground plane and the air interface dielectric. By using an antenna solution with no high dielectric material, the problem is easily solved.

By reducing the coupling between antennas, it is possible to make sure the correlation factor remains low enough to maintain high system performance. It can also help manufacturers reduce the device's size while maintaining good performance. Equally important to device performance is overall system cost.

Antennas are often an afterthought for RF designers. They are considered a 50- ohm component whose only role is signal reception. But they can actually have interesting electrical characteristics that can help reduce the component count and the overall bill of materials. For example, the selectivity of an antenna can help eliminate filters or at least reduce their specifications. Also, by using a technology that can be easily matched to different impedance values, it is possible to remove matching circuit components.

The absolute antenna size is a parameter which has long been studied. Starting with Maxwell's Equations, the basis for electromagnetic theory, it is possible to derive a formula linking the relative bandwidth, determined by the application, to the antenna mode volume. But the antenna subsystem size must also be factored in. Using an isolated antenna technology reduces the keep out area, where components cannot be placed, and reduces the ground plane size, which can be critical in smaller devices.

And, time-to-market is essential in the handheld device market. Design cycles are getting shorter, while products are getting more complex. Therefore, engineers are looking for solutions that are easy to implement and flexible enough to add onto their manufacturing lines.

Antenna isolation helps significantly in the design cycle as it provides the design engineer with greater flexibility to move components around without having to completely retune the antenna. Moreover, when multiple antennas are involved, either at different frequencies or at the same frequency as in a diversity solution, the technical challenges are even more complex. This is where a highly isolated antenna can minimize integration problems and accelerate the product design cycle.

The performance of a diversity system is highly contingent on the antenna solution, especially as it relates to coupling. Other parameters such as cost, size and time-to-market should also be taken into account by device manufacturers.