Mm-waves have become a well-known and desired frequency band, thanks to 5G FR2 and RADAR applications. Antenna designers have learned how tricky antenna systems can be on those frequencies and how to use simulation capability to predict that. However, the challenge of measuring a simulated system is still there, as laboratories’ update is slow and expensive. When working with mm-waves, fine-tuning of antenna parts cannot be done independently – as, for example, in cellular devices – because a complex system requires a more inclusive approach. Further, over-the-air (OTA) measurements are crucial for performance analysis and antenna tuning.
When thinking about the major differences between an anechoic chamber for cellular or IoT measurements and mm-wave one, there are basically two most important ones: the size of the chamber and internal losses. Here is the first trick – the size of the chamber must be big enough for measuring far-field, but small enough to not bring in too much cable and path losses. Figure 1 features a 3D model of the mm-chamber that our team has built at Radientum.
Figure 1. 3D model of the mm-wave anechoic chamber at Radientum with 105cm far-field distance.
The far-field region is calculated as (2D^2)/λ, where λ is wavelength and D is the maximum antenna size. This equation obviously shows, that the bigger antenna, the bigger chamber is needed. For example, a 50mm antenna array with 30GHz central frequency requires at least 50cm distance for far-field. The far-field criteria are especially crucial if the antenna’s radiation pattern has specific nulls, and they are important for its performance. Figure 2 shows the physical difference between the near and far fields of an antenna.
Figure 2. Near and far-fields of an antenna.
As clear from figure 2, the reactive near field can be used very limited on mm-wave measurements, as the probe antenna becomes a part of a measured antenna due to the small distance between them. Radiative near field has much more predictable behavior and can be used in measurements, but requires compensation post-processing in both receive and transmit paths. Finally, far-field is perfect for phase and amplitude measurements, but it introduces great path losses due to the bigger distance between the probe antenna and a measured antenna.
Unfortunately, cable losses are no more negligible at mm-waves, thus cabling from all the instruments to the antenna has to be thoroughly calculated. If the chamber is too big, cables are too long and lossy. Measurement tools and devices might be placed inside the chamber, but that will cause unwanted reflections and distortions of radiation patterns. When we were building our mm-wave chamber at Radientum, we opted on placing all the instruments outside to avoid possible measurement errors.
Figure 3. Inside view of the mm-wave chamber at Radientum.
When thinking about a regular anechoic chamber, a metal room with absorbers on the walls appears in mind. Here comes another significant difference between low and high frequencies – metal box does not work so well anymore. At Radientum, we have built an mm-wave chamber almost entirely of wood, keeping in mind that it is beneficial for high gain antenna arrays measurements. At those high frequencies, interferences from the outside of a chamber are negligible in comparison with what is happening inside the box. Further, the walls of our chamber are padded with an mm-wave absorber, which provides around 55dB attenuation at 30GHz.
Over-the-air measurements are crucial for mm-wave design verification. When directly compared, direct far-field measurements (DFF) have much more benefits than indirect far-field (IFF), taking into account the x10 cost of IFF setups.
We have been using our mm-wave chamber at Radientum in both, internal and customer cases, for verification of the simulation results and achieving confidence for more complete test house measurements.