Nowadays almost every specification for 5G device states dual-polarization as a basic and mandatory feature. Dual-polarized antennas became a part of everyday work for most RF engineers, but mmWave frequencies add even more challenges to it.

Why even bother with dual-polarization?

Short answer: to separate channels from each other and allow sending data through both channels working on one frequency without interfering.

As you can see in figure 1, each polarization has its own feed, so when one feed is excited, another one is passive.

Figure 1. 26GHz cavity antenna with orthogonal (horizontal and vertical) polarizations. E-fields.

The long version of this answer includes some basics of MIMO (Multiple Input Multiple Output) technologies. As was mentioned above, each polarization is fed separately, so the device can transmit two data streams. From the receiver’s point of view, the signal is interpreted as two discrete streams. The easiest and most popular MIMO configuration is 2X2 MIMO, with transmitter and receiver working across orthogonal polarizations. This approach also allows avoiding using two separate arrays for the same working capacity, which results in saving space and reducing manufacturing costs.

One more usage of dual-polarized antennas is the ability to adapt to the incident electromagnetic field’s polarization, in order to receive a maximum signal-to-noise ratio. Or, in case of transmitting, to match transmitted signal to a certain receiving antenna with unknown polarization.

What can go wrong with dual-polarization?

Even though each polarization has its own RF feed, they still affect each other. Theoretically, there can be a perfect antenna with completely isolated orthogonal transmissions, but in a more realistic scenario, vertical polarization will be affected by horizontal one and vice versa. This effect is called cross-polarization isolation.

In figure 2 you can see S-parameters of mmWave dual-polarized cavity antennas, showed above. S21/S12 represents cross-polarization isolation.

Figure 2. S-parameters of mmWave cavity antenna.

Usually, the -17dB level is considered sufficient for simulated isolation. It will transform to approximately -20dB in real life due to losses.

What new challenges come with mmWave 5G?

It might be easy to gain good isolation when there is only one antenna, but what if you have a multi-element array? In arrays elements couple with each other and apart from cross-polarization isolation, the engineer has to deal with inter-element isolation as well.

Extremely small sizes bring additional challenges. When the patch’s width is only 3mm, VIA size becomes significant, and it is not always possible to go for micro-VIA, because in some cases thick dielectric is your only way to achieve the required performance.

Figure 3 shows the same patch, but with different VIA sizes, using for dual-polarized feeding.

Figure 3. VIA diameters 0.5mm and 0.2mm.

Cross-polarization isolation drops significantly when moving from 0.2mm to 0.5mm VIAs, due to the dramatically increased effect of one polarization to another.

What does “poor” isolation mean for dual-polarization?

It is easy to miss the isolation aspect while concentrating on the gain or efficiency of the antenna. But in antenna theory, almost everything affects each other, so if you have poor (above -10dB) isolation between orthogonal polarizations, performance will drop respectively. Energy will leak from one polarization to another and will not effectively radiate.

MmWave 5G antenna design is a complicated and challenging process, but if the model is simulated properly – taking into account every detail, including cross-polarization isolation – it will save a lot of time and resources.

Disclaimer: The views and opinions expressed in this article are those of the author. It is intended only as a sharing of antenna design knowledge for educational purposes.

Katerina GalitskayaAntenna and RF Design Engineer