by Tim Daniels, Supplier Business Manager RF Power, RFMW and Radio Amateur G7KTP
Interference is a fact of life for radio amateurs, especially at HF frequencies below 30 MHz, where signals can travel far greater distances than those at higher frequencies. We must contend with virtually every type of emissions source, from high-power commercial shortwave broadcasters to RF interference (RFI) from power lines, switch-mode power suppliers, powerline telecommunications, solar power inverters, power over Ethernet and LED lighting and many more unidentifiable sources that can be quite bizarre.
While interference is far worse at HF frequencies, RFI can still be a problem at higher frequencies, even in the 23 cm amateur band between 1240 and 1300 MHz (L-band) where radio amateurs can make long distance contacts of a thousand kilometers or more when the right tropospheric propagation conditions allow, or experiment with digital television and even bounce signals off the moon to communicate. Radio amateurs are secondary users of this band and therefore cannot claim protection from interference from its primary users. The primary users are commercial and military radars and many other applications (Figure 1).
The radars can be a significant problem because they have Effective Isotropic Radiated Power (EIRP) levels that can reach the megawatt level. Unfortunately, I live close to a primary radar, and I was getting a lot of interference, so I decided to investigate and try to remove the interference.
The Big Emitters
Primary radars are strategically located in various regions to aid air traffic control (ATC) and surveillance. These radars are crucial in monitoring and managing commercial and military aircraft movements. Terminal radar, or airport surveillance radar (ASR), is primarily used near airports and terminals. It usually operates at S- or X-band frequencies and accurately covers local air traffic movements, monitoring aircraft during takeoff, landing, and low altitude flights. It helps controllers maintain separation between aircraft and guide them safely through the busy terminal airspace.
Enroute radar is used for air traffic control in the airspace between airports and routes or corridors in and out of countries. It covers a larger area and is designed to handle high altitude flights and long distance air traffic. The L-band radar frequencies (1.2 to 1.4 GHz) can deliver very high EIRP, use a large, high-gain antenna, and are situated on elevated hilltop locations to provide a longer range than terminal systems.
The UK has a network of about 16 civil L-band primary radars for air traffic control, and one of these sites is located about 50 km from my home on top of Titterstone Clee Hill near Ludlow in Shropshire in the west of the UK and is 533 m above sea level (Figure 2). The radar is believed to be a Raytheon ASR-23SS operated by the UK’s National Air Traffic Services. It provides air traffic control with radar coverage over the west of England, Wales, Ireland, the Irish Sea, and into the Atlantic. If you have ever flown into London’s Heathrow Airport to or from the U.S., chances are this radar will have assisted your safe transit through UK airspace.
For radio amateurs like me, living within the 150-km range of Clee Hill and operating in the 23 cm band can present challenges. This radar emits powerful signals, some as close as 9 MHz away from our operating frequencies, where we are trying to contact other amateur operators over long distances, picking out weak signals almost in the noise. When the signals are fading or weak, which they often are, the situation is worse.
There might only be a few seconds for our information exchange so this interference can prevent a successful contact. For example, the receiver’s front end suffers from gain compression during the pulse, and the Automatic Gain Control (AGC) takes a few seconds to recover. This is visualized more easily in software used for weak signal communications called WSJT-X (Figure 3).
Radio amateurs are allocated 60 MHz of spectrum between 1240 and 1300 MHz and are considered secondary users. This means we cannot interfere with other users. Of course, as RF experimenters, we can at least engineer solutions to overcome some of the interference problems we encounter, so I began looking for a suitable filter to attenuate the radar signals.
RFMW has over twenty filter suppliers. When our team receives a custom filter request, they ask several questions, select the correct filter topology, and develop a target specification to discuss with the most suitable filter suppliers. I first needed to understand the problem in as much detail as possible, so I made measurements of the interfering radar signals.
The radar signals were measured on a spectrum analyzer and found to be at 1265, 1266, 1305 and 1306 MHz, with the interference occurring in about 8-s intervals as the radar was rotating. Frequency diversity in radar is known to improve detection and signal-to-noise, so it is common to see this type of radar transmission across two pairs of frequencies. Each channel has four long pulses of 100 µs and four short pulses of 1 µs interleaved in time across the four channels.
The long pulses have an FM chirp bandwidth of 1 MHz. My receiver operates at 1296.2 MHz, so it was necessary to significantly attenuate the channels at 1305 and 1306 MHz, as seen in the plot in Figure 4. The radar is more than 30 mi. away, but I was surprised at its signals’ strength in the receiver. I decided to measure the signal strength more accurately.
I set the spectrum analyzer to zero span with an 8-s sweep time, corresponding to the radar’s antenna rotation sweep time. Setting the resolution bandwidth according to the waveform was also important to achieve accurate measurements. Rohde & Schwarz has a useful application note that explains how to measure pulsed signals.
The measurement shown in Figure 5 shows my signal strength measurement. I picked the channel at 1305 MHz, the nearest signal and the most difficult to filter. It was measured at -15 dBm, far too strong for the front end of my ICOM IC-9700 receiver, which has a very wide front end. You might also notice the additional bumps in the trace on either side of the main signal. As the radar rotates 360 deg., we see the side lobes of the antenna.
With such strong signals, I ran a path profile between my house and the Clee Hill radar site. The Solwise website has a handy tool for a basic path profile. The path between the Clee Hill radar and my house was a line of sight (Figure 6), which helped to explain why the signal was so strong.
After discussing the radar interference on the UK Microwave Group discussions board, I learned it has a maximum licensed Effective Radiated Power (ERP) of +82 dBW, a massive 158 MW. No wonder the signal is so strong at my receiver.
We can calculate the free space path loss accurately when there is a line-of-sight path, and we are far enough away from the transmitter. The free space path loss is:
Lp (dB) = 20 Log(d) + 20 Log (f) + 32.5
Lp: Path loss
d: Distance in km = 51
f: Frequency in MHz = 1305
This gives a path loss of approximately 129 dB for the 51 km path.
So, working back from the received signal strength, we can use this equation:
ERP = Sr + Cl – Gr – Lp
ERP: Radiated power from radar
Sr: Signal power at the receiver = -15 dBm
Cl: Cable losses = -1.5 dB
Gr: Estimated receiver antenna gain = 10 dBi
This gives 75.5 dBW as the radiated power.
The difference from the published +82 dBW could be for several reasons, such as:
Measurement accuracy of the pulse on my spectrum analyzer
My estimation of receiver antenna gain at this frequency
Being below the main beam of the radar
Also, the radar may not be transmitting at full power.
The most significant influence is my position underneath the antenna’s main lobe.
How Out-of-Band Signals Cause Interference
Strong out-of-band signals like these radar signals can lead to interference problems in the front end of a receiver, especially now with software-defined radios (SDRs) that have broad bandwidths and cover multiple frequency bands. The receiver’s front end is wider than the wanted band and no filter is perfect, and there are compromises and trade-offs to make that we often discuss with customers.
For example, compact filters with a steep roll-off and high out-of-band attenuation often have high insertion loss, which is undesirable in front of the first amplification stage, usually a low-noise amplifier (LNA), as any loss before the LNA is added to the system noise figure.
At RFMW, we usually select a low-loss filter with just enough out-of-band attenuation for the application in a low-loss front end requiring a low noise figure. When supplying an LNA from one of our suppliers, such as Qorvo or Guerrilla RF, we look for a high-input IP3 to ensure the front end works in its linear range and won’t saturate with strong signals. After the LNA, we can accept a filter with higher insertion loss and more out-of-band attenuation.
Target Filter Specification
Having defined the filter’s target specification with one of our filter experts, we decided a 5-pole cavity filter was required. After approaching one of our suppliers, a prototype cavity filter was ordered based on the agreed target specification RFMW provided (Table 1). To verify the performance, I took a plot of the filter and found it to be well within the target specification (Figure 7). The radar signal no longer interfered after installing the filter between the antenna and receiver.
Fortunately for me, this was a relatively simple fix (a bandpass filter with sharp skirts and high out-of-band rejection), even though the interfering signal was caused by a radar emitting massive amounts of RF power. However, every case of interference is different, and some are extremely difficult to solve, even with the best filters money can buy. That said, we’ve encountered many situations over the years that seemed impossible to remedy but ultimately were solved using various filtering techniques, so it makes sense to consult with experts who have years of experience with hundreds or thousands of interference problems.
I would like to thank my colleagues at RFMW, especially Colin Field, for his help with the filter specification and Joel Levine for agreeing to supply the prototype filter for my experiment.