Many companies make small dual-frequency antennas, but these antennas are often poorly suited for RTK. Take the datasheet for such an antenna and look for the phrases “phase center variation”, “PCV”, or “phase center stability”, or at least the word “RTK”. If you cannot find any of them, it means that the antenna was not designed for RTK.
The phase center is the point where a signal from a satellite appears to be received. This point is different for satellites in different directions. In antennas suitable for RTK, the apparent reception point stays within a small area, regardless of the signal frequency (which differs between the center and the edges of the frequency band) or the satellite’s elevation above the horizon. The size of this variation is described by the PCV (phase center variation) parameter. When this area is small, we say that the phase center is stable. A good PCV value for RTK is less than 5 mm (0.5 cm).
When the PCV value is not stated, even in general terms (for example, with a phrase such as “high phase center stability”), it means one of three things:
1. The phase center moves over a large part of the antenna, making the antenna unsuitable for RTK. In other words, its PCV is very large.
2. The manufacturing process is not consistent, and the PCV depends on the production batch. You may find a batch, or even a single antenna within a batch, that is suitable for RTK. But the next batch may not be suitable. Sometimes the factory itself sorts and rejects antennas based on their performance. In that case, antennas suitable for RTK end up costing two to three times as much.
3. The PCV has never been calculated or measured. This happens with some Chinese copies of well-known antennas. Surprisingly, some Chinese manufacturers simply do not have the equipment needed for these measurements. There have been attempts to use selected batches of these antennas for RTK: sample antennas from a batch were tested before the batch was purchased. But when the production batch changed, the results became very poor.
In any case, if none of these important terms appears in the datasheet, it is better not to buy the antenna for RTK. RTK antennas cost more than ordinary antennas, so every seller has a reason to mention these features when the antenna has them. When it does not, it is easier to say nothing. That is why you will never see the phrase “not intended for RTK” in an antenna seller’s description. Not even in ours. :-)
A Little Physics
It is difficult to achieve a stable phase center in a small antenna. First, antennas of different sizes use ceramics with different dielectric permittivity: approximately 16 for a 25 × 25 mm antenna and approximately 4 for a 50 × 50 mm antenna (the exact values depend on the design). Different dielectric permittivity means different ceramic materials. It also seems that the manufacturing process for ceramics with high dielectric permittivity is much less predictable. This may explain the differences in PCV between production batches of small antennas.
In addition, ceramics with high dielectric permittivity have a higher quality factor (Q), so frequency has a greater effect on PCV. To receive L1 signals, an antenna needs to cover 1559.052–1605.886 MHz. For L2/E5b/B2b reception, it needs to cover 1196.910–1248.625 MHz. At first glance, the bandwidths look similar: 46.83 MHz for L1 and 51.715 MHz for L2. But the carrier frequencies differ by almost a factor of 1.5, so the relative bandwidth is larger for L2: 2.96% for L1 compared with 4.27% for L2. Also, designers often deliberately accept worse reception of GLONASS G1 signals. This allows them to design an antenna with a uniform response over a band of only 32.5 MHz, from 1559.052 to 1591.788 MHz. This is one reason why there are 35 × 35 mm L1/L2 antennas with good PCV on L1 but poor PCV on L2, while small antennas with good PCV on L2 are not available.
There is another reason. The smaller the antenna, the greater the effect of imperfections of the same physical size. An area of 1 square millimeter is 0.16% of a 25 × 25 mm antenna, but only 0.04% of a 50 × 50 mm antenna.
Another factor is the ground plane. In small antennas, this is simply a piece of metal underneath the antenna. An irregularity that reduces PCV performance occurs in a narrow strip about 2 mm wide around the edge of the antenna. For a 25 × 25 mm antenna, this strip covers 29.4% of its area. For a 50 × 50 mm antenna, it covers only 15.4%.
Conclusions
- Dual-frequency antennas smaller than 45 × 45 mm that are suitable for RTK probably do not exist at all, or they are rare and expensive.
- It is possible to find a single-frequency antenna smaller than 45 × 45 mm that is suitable for RTK. But it will be fairly expensive.
Therefore, in practice, patch antennas for L1/L2 RTK start at about 50 mm. A typical antenna of this kind measures 80–100 mm.
Now, About Single-Frequency RTK
If there is no small antenna suitable for dual-frequency RTK, why not use a single-frequency antenna and single-frequency RTK instead? Unfortunately, this solution is useful only in certain situations.
Single-frequency RTK is guaranteed to work at distances of up to 10 km and, when ionospheric conditions are good, up to 20 km. Dual-frequency RTK works over at least 50 km, and we have sometimes managed to use it over distances as long as 500 km. The distance limit for dual-frequency RTK is related to the loss of accuracy: accuracy gets worse by 1 mm for every kilometer of distance. So, at 50 km between the base station and the rover, accuracy becomes 5 cm worse (starting from an accuracy of 10 mm). There is an exception: with the rather expensive mosaic receivers, accuracy gets worse by only 0.5 mm per kilometer. This means a loss of 2.5 cm over a 50 km baseline.
The second problem with single-frequency RTK is the long time needed to obtain the first fixed RTK solution. For single-frequency receivers, RTK fix time (sometimes called “convergence time”) increases as the distance increases. For example, “less than 60 seconds” for a u-blox NEO-M8P receiver at a distance of 1 km becomes “less than 3 minutes” at 10 km, and is more likely to be 1–2 minutes with a good antenna (one with low PCV). For dual-frequency receivers, this time is almost independent of distance. For example, the u-blox ZED-F9P takes less than 10 seconds.
The third problem is accuracy. Near the base station, dual-frequency RTK has an accuracy of 10 mm (6 mm for mosaic), while the u-blox NEO-M8P—the only receiver we have that supports single-frequency RTK—has an accuracy of 25 mm.
However, if all three limitations are acceptable, single-frequency RTK can be used.
