A Measurement Based Approach to Avoid GNSS Interference
GNSS receivers play a critical role in commercial and military systems. They can be found in a myriad of devices and platforms including phones, watches, drones, tractors, survey equipment, automobiles, airplanes, satellites, missiles, and ships. When GNSS receivers fail, the impact ranges from nuisance scenarios such as an incorrect distance for your latest training run to potentially disastrous and life-threatening scenarios.
Figure 1. GNSS receivers are employed in a wide variety of devices and platforms.
GNSS receivers rely upon very low amplitude signals from satellite constellations to provide position, navigation, and tracking. There are multiple satellite constellations such as the United States’ GPS, Russia’s GLONASS, China’s BeiDou, and the European Union’s Galileo. All of the GNSS receivers operate in L-band (1-2 GHz).
Companies and organizations that integrate GNSS receivers into their devices or platforms need to understand both the in-band (desired) and out-of-band (undesired) performance of the GNSS receiver. Nearby transmitters can cause interference that degrades the performance of the GNSS receiver or completely jams the receiver, rendering it useless. When the interference occurs at the receiver’s designed operating frequency, we call it “in-band” interference. When the interference occurs anywhere outside of the receiver’s operating frequency, we call it “out-of-band” interference. There are times when receivers experience both in-band and out-of-band interference simultaneously, which can represent particularly challenging problems to first identify and then mitigate.
At EMA, we perform RF interference (RFI) analysis, sometimes called cosite interference analysis, for customers in a wide variety of industries and for a wide variety of applications. We use the Ansys EMIT software for these analysis projects as EMIT allows us to model every component in the RF architecture and consider extremely large problems (e.g., we recently worked on a project that involved 70 million transmit/receive channel pairs) in an efficient manner. We also use the EMA Automated Radio Measurement System (ARMS) to measure the broadband performance of transmitters and receivers and use the collected data as input to the Ansys EMIT software. An article describing this approach was published by Microwave Journal and can be found here.
The concepts described in the Microwave Journal article are also applicable to the GNSS receiver interference problem. For example, we often see cases where harmonics of transmitters fall directly in the GNSS receivers band. In Figure 2, we show an example where the 4th harmonic of a UHF transmitter falls in the GPS L1 band. The UHF transmitter operates at 390 MHz and the 4th harmonic is 1560 MHz. A simple comparison between this 4th harmonic frequency and the L1 frequency (1575.42 MHz) would indicate that there should be no interference. They are offset by 15.42 MHz. However, one needs to keep in mind that the receiver response is not ideal, and the 4th harmonic has a finite bandwidth as well. So, even though the 4th harmonic and the L1 frequency are significantly offset in frequency, there is still interference in this case.
Figure 2. The 4th harmonic of a UHF transmitter produces a large interference problem for a GPS receiver operating at L1. The black trace is a model of the receiver’s susceptibility based upon specification sheet information provide by the vendor while the red trace is measured susceptibility data collected with EMA’s ARMS.
In Figure 2, the red trace is the measured susceptibility data for the GPS receiver. The black trace is a model for the susceptibility based upon information provided by the vendor for the receiver. Typically, vendors will provide the in-band sensitivity of the receiver, processing gain (when applicable), and selectivity information. The selectivity information describes the shape of the in-band response and is typically provided as amplitude/bandwidth pairs. For example, it is common for vendors to specify the 3 dB, 20 dB, and 60 dB bandwidths for the selectivity. Specifying selectivity information in this manner results in a symmetric selectivity profile for the receiver. EMA has measured many RF receivers and the selectivity profile is rarely truly symmetrical. This can be seen in the image above for the GPS receiver. Clearly, the selectivity is not symmetrical for the measured susceptibility data (red trace).
As noted earlier, interference can also occur out-of-band. The measured data collected with ARMS for the GPS receiver shown in Figure 2 has a spurious response near 1280 MHz with an amplitude of -25 dBm. In other words, an out-of-band signal at 1280 MHz with an amplitude of -25 dBm or higher will cause degradation to the GPS receiver performance. At -25 dBm, the degradation will result in reduced carrier to noise but at higher power levels, the GPS receiver will eventually lose lock.
To avoid out-of-band interference problems, bandpass filters are often incorporated into the GNSS antenna or RF front end. These filters allow the desired signals to pass through while rejecting the out-of-band signals. This approach can be partially effective, but it is not perfect. First, this approach often provides more out-of-band rejection than is necessary at most frequencies for the receiver. It is a blanket approach as opposed to surgical approach, which can result in spending more money and requiring more weight than is necessary. Second, filter performance is not constant with frequency and will eventually degrade. For example, measured data for a lowpass filter is shown in Figure 3. The specification sheet for this filter states 50 dB of insertion loss above the cutoff frequency. While this performance is achieved from 510 MHz to 1010 MHz, the insertion loss decreases above 1010 MHz to the point where there is only about 5 dB of insertion loss at 1800 MHz. If your receiver has out-of-band responses above 1800 MHz, the filter is providing very little protection for your receiver. The moral of the story is that without measured data for both your GNSS receiver and your filters, you really do not know how your receiver is going to perform in a congested RF environment. Further, without measured data, you are going to spend a lot of time and money tracking down the source of the interference problem and mitigating it.
Figure 3. Measured insertion loss for a lowpass filter.
For GNSS receivers, one also must consider RF transmitters that operate at nearby but different frequencies. For example, both Iridium (1610 – 1626.5 MHz) and INMARSAT (1626.5 – 1660.5 MHz) operate close to the L1 band for GPS receivers. Again, when one simply compares the operating frequencies of these systems, there should be no interference. Iridium’s lowest channel (1610 MHz) is 34.58 MHz away from L1 and INMARSAT’s lowest channel (1626.5 MHz) is 51.08 MHz away from L1. At first glance, it seems like there is sufficient frequency separation to not have to worry about interference. And if you look at the specification sheet for the receiver, it probably indicates that no interference should occur at these frequencies based upon the selectivity data. However, EMA has measured several GPS receivers with ARMS and for many systems there is a definite possibility for interference unless careful filtering is used prior to the receiver. In Figure 4, we show the susceptibility of another GPS receiver with the lowest channels for Iridium and INMARSAT overlaid. For the Iridium channel, the GPS receiver has susceptibility of -50 dBm. For the INMARSAT channel, the GPS receiver has susceptibility of -34 dBm. Again, the susceptibility values mean that the receiver will start to experience degradation at these power levels and with increasing power levels, the receivers will eventually lose lock.
Figure 4. Iridium and INMARSAT operate very close to the GPS L1 band and can be sources of interference.
To further illustrate the importance of measured data collected with ARMS, consider the measured receiver susceptibility shown in Figure 5. There are several features to note in the measured data. First, the selectivity curve is not symmetric. Second, this particular GPS receiver has better susceptibility at the Iridium and INMARSAT frequencies than some of the other GPS receiver considered. So, not all GPS receivers are going to be susceptible to Iridium and INMARSAT signals. But how do you know unless you measure it? Third, this receiver has multiple out-of-band responses both below and above the L1 channel. It should be noted that ARMS employs careful filtering of the interfering signal used for susceptibility measurements. Without filtering, it is possible for harmonics and spurious emissions of the interfering signal generator to create false positives. With the filtering employed in ARMS, the out-of-band responses shown below are true out-of-band responses.
Figure 5. The broadband susceptibility of a GPS receiver indicates multiple out-of-band responses of the receiver.
The plot shown in Figure 5 reinforces the importance of measured data. With specification sheets, receiver vendors do not provide the information necessary to understand the true performance of the system. Without measured data for the receiver, system integrators will typically over engineer their system (e.g., use far more filtering than is necessary) and/or completely miss real interference problems. The system integrators are operating in the dark without measured data. EMA has seen this scenario play out repeatedly. When we have a chance to measure both transmitters and receivers for our customers, they can leverage that information to make informed decisions. For example, we have measured transmitters and found harmonic and spurious emissions that are grossly different from the values stated on the specification sheet. Spurious emissions are particularly problematic because most vendors provide a blanket threshold value indicating that spurious emissions shall not exceed that amplitude. Such threshold information is basically worthless when it comes to an RF interference analysis. It generally results in many false positive interference problems given the proximity of the transmitters and receivers. What one needs is the measured transmitter data that provides the exact frequencies and amplitudes of the spurious emissions. ARMS provides such data.
The use of RF systems in our lives and the crowding of the RF spectrum are not going to slow down. One only has to look at the recent headlines about the problems between 5G transmitters and radar altimeters to understand the severity of the problem. Think about the amount of time and money that was spent studying this problem…not to mention the revenue lost by the cellular companies. This was an avoidable problem.
If you are integrating RF systems into your device or onto your platform, contact EMA to learn how we can measure the performance of your RF systems with ARMS and use that data in the Ansys EMIT software to identify interference problems early such that cost effective solutions can be implemented.