Microwave And Rf Wireless Systems

Setting Up a Point-to-Point Microwave Link That Actually Works

I spend more time troubleshooting failed microwave RF links than anything else. Most failures aren't design problems. They are installation problems, maintenance problems, or people not understanding how much environment matters at these frequencies. Here is how to do it right the first time. When people say "microwave" they usually mean licensed wireless bridge links operating in the 1 GHz to 40 GHz range. Not your kitchen oven. These systems carry point-to-point or point-to-multipoint data connections between towers or buildings. The "rf wireless" part just means radio frequency transmission. Everything I describe applies across that spectrum, though higher frequencies need tighter line-of-sight and deal harder with rain. The core concept is simple: dish antenna, radio unit, mount it high, align it, configure it, test it. The part nobody tells you is how much work actually happens between "configure it" and "test it."

The Real Work Before You Buy Equipment

Path analysis is where most projects fail. You need a terrain profile between your two sites. Use tools like Atoll, Radio Mobile, or even something like GPS visual with elevation data if you are keeping it lean. Calculate Fresnel zone clearance. That is the ellipsoid around your direct line of sight. If less than 60 percent of the first Fresnel zone is clear, your signal will suffer regardless of how much power you throw at it. I once spent three days trying to debug what looked like a fading interference issue. Turns out a newly grown tree was nibbling into the lower Fresnel zone by maybe two meters. We trimmed it and the bit error rate dropped to basically zero. That tree was not visible from either antenna position. That is why you need the terrain profile, not just visual confirmation.

Selecting Frequency Band

Lower bands like 6 GHz and 11 GHz are more forgiving with weather and alignment but have narrower channel widths available. Higher bands like 23 GHz and above give you wider bandwidth for backhaul but rain fade becomes a real factor, especially in tropical or subtropical regions. 60 GHz is oxygen-absorbed and essentially useless beyond a couple hundred meters outdoors. Do not buy 60 GHz gear for outdoor use unless you have a very specific indoor or short-range reason. Licensed versus unlicensed depends on your local regulations. In the United States, the 5 GHz UNII bands and 24 GHz are unlicensed but crowded. Licensed 18 GHz, 23 GHz, and 38 GHz bands give you cleaner spectrum and protection from neighbors. If you are doing carrier-grade backhaul, licensed is almost always the right call.

Antenna Selection and Mounting

Dish size matters for gain and for narrowing your beamwidth. A 60 cm dish at 11 GHz gives you roughly 33 dBi gain. The same dish at 38 GHz pushes past 40 dBi. Higher gain means tighter alignment tolerance. At 38 GHz, your dish might only be about 1.5 degrees wide. Misalignment by half a degree and you are losing several decibels. Mount rigidity is non-negotiable. Cheap pole mounts flex in wind. I have seen links drop because the tower swung enough during a 40 mph gust to push the antenna past its azimuth tolerance. Use structural mounts. Braced arms. Torque everything to spec. Vibration dampeners between the radio and dish help too if your tower is on a rooftop with HVAC units nearby.

Alignment Procedure

Do not just aim and go. Use the manufacturer's alignment tool if it has one. Most decent microwaves come with an integrated signal strength meter or a separate handheld alignment device. Start at low gain, roughly aimed, then narrow in. Adjust azimuth first, then elevation. Peak the signal. Then tighten everything down while watching the signal meter. If your signal drops more than 1 dB during tightening, you have mount movement and you need to fix that before proceeding. I learned this the hard way on a 15 km link at 18 GHz. Tightened the bolts, signal dropped 3 dB. Had to loosen, re-align, and apply thread locker. Took twenty extra minutes that saved me a callback six months later.

Power Budget and Fade Margin

Calculate your link budget properly. Transmitter output power plus antenna gain minus cable loss plus receiver antenna gain minus receiver sensitivity equals your available margin. Subtract free space path loss, which is 92.4 plus 20 times the log10 of distance in kilometers plus 20 times the log10 of frequency in GHz. That is the baseline. Then subtract atmospheric losses, ground reflection losses, and whatever fade margin your design requires. For a reliable licensed backhaul link, plan for at least 25 to 30 dB of fade margin. Rain fade at 23 GHz over a 20 km path in a heavy rain zone can eat 20 dB or more. If your budget only gives you 15 dB total margin, you will have outages during monsoon season. Everyone who designs these links knows this. Nobody plans for it until it breaks.

Configuration and Testing

Set your channel width, modulation scheme, and output power according to your fade margin calculations. Higher order modulation like 1024 QAM gives you more throughput but needs better signal quality. If your SNR is marginal, drop to 512 QAM or even 256 QAM and you will be more stable with slightly less capacity. It is better to have a stable 500 Mbps link than a flapping 1 Gbps link. Run a sustained throughput test. I usually run Iperf3 for at least thirty minutes. Check for packet loss, jitter, and any SNR fluctuations. Log the data. If the signal dips below your threshold during that window, you have a problem that will get worse over time, not better.

Common Pitfalls

Fiber optic converters eating all your margin. If you are using external SFPs or media converters between the microwave radio and your router, those add latency and introduce another point of failure. Some SFPs also do not play nicely with certain microwave radios on Ethernet handshaking. Always test the full chain. Ground loops on Ethernet cables. Shielded cable run along a metal tower with different grounding potentials at each end and you will inject noise into your data. Use fiber between the two sites if possible. If you must use copper, isolate grounds properly and consider common mode chokes. Polarization mismatch. Dual polarization is standard on modern microwave links. Make sure the polarizations match at both ends. Cross-polarization interference can degrade your signal by 20 dB or more if someone mounts one end rotated ninety degrees and nobody notices.

When Microwave Is the Wrong Choice

Not every link needs wireless. If you can run fiber and the distance is under ten kilometers, fiber will almost always win on cost per megabit over time, reliability, and latency. Microwave is great for right-of-way issues, temporary links, rural spans where trenching is prohibitive, or emergency restoration. It is not a default solution. I have seen teams install microwave links where a trenchless bore would have been cheaper and more reliable because they did not want to deal with permitting. Microwave also struggles in dense urban canyon environments with multipath. Buildings reflecting signals around corners can create constructive and destructive interference that shifts with temperature and weather. That is a different problem entirely and requires a site survey with active measurements, not just a path profile. If you are building Microwave And Rf Wireless Systems for the first time, start with a lower frequency band, a longer path, and licensed spectrum if your regulator allows it. Those three choices remove the most common failure modes. Then pay attention to the mounting hardware and the fade margin calculation. Everything after that is just configuration details.