Understanding Signal Loss When Your Wireless Link Loses Line of Sight
Line of sight fading is the most frustrating problem in point-to-point wireless deployment. It happens when your microwave link, WiFi bridge, or millimeter-wave connection experiences signal degradation because something blocks the direct path between the transmitter and receiver. This isn't theoretical. I spent three weeks debugging a 5.8 GHz PtMP backhaul site in late 2023 where the client kept getting 60 percent throughput drops every afternoon at exactly 4:17 PM. Turns out a newly planted row of mature trees was slowly growing into the Fresnel zone. At its core, line of sight fading occurs when the radio frequency path between two antennas is partially or fully obstructed. Unlike free-space propagation where the signal travels cleanly, real-world environments introduce reflections, diffractions, and absorptions that interfere with your signal. The result is multipath fading, where multiple copies of the same signal arrive at the receiver at slightly different times and phase states. Sometimes they add constructively. Most of the time they don't. Here is how I actually approach this in practice. First, I do a proper RF site survey using tools like Ekahau or even a simple spectrum analyzer with a directional antenna. You walk the intended path with your equipment active. I look for at least two things: the raw SNR (signal-to-noise ratio) across different times of day and any multipath signatures showing up on the frequency domain display. A clean LOS link should hold a stable SNR above 25 dB under normal conditions. If you're seeing fluctuations of more than 6 dB over a five-minute window, you have a fading problem already forming.
Then there is the Fresnel zone calculation, which most people get wrong. The first Fresnel zone radius is calculated as r equals the square root of lambda times d1 times d2 divided by d1 plus d2, all divided by 4. At 5.8 GHz with a 2-kilometer link, the first Fresnel zone is roughly 3.4 meters in radius at the midpoint. That means you need at least 60 percent clearance of that zone, ideally 80 percent, to avoid significant fading. I've seen engineers skip this entirely and wonder why their "clear line of sight" view through a binocular turns into a garbage link on the air." Let me share the exact case I ran into last year. We deployed a 12 GHz microwave bridge across an industrial park in Cincinnati. The client showed me satellite imagery confirming an unobstructed path. But the actual throughput barely hit 200 Mbps on a 500 Mbps rated link. I pulled up the link margin analysis in our planning tool and realized the elevation profile was off by about 4 meters due to outdated LiDAR data. A storage warehouse that was built six months prior was not on the GIS layers. The partial obstruction was causing knife-edge diffraction loss of approximately 18 dB. My workaround was straightforward: we raised the receiving antenna tower from 12 meters to 22 meters, which gave us the required clearance and restored full link budget margin. Now for the part nobody talks about enough. Rain fade and atmospheric attenuation are often confused with pure LOS blockage, but they behave very differently. Rain fade is frequency-dependent and hits hard above 10 GHz. A light to moderate rain event at 15 GHz can introduce 5 to 10 dB of additional attenuation. The sky is literally eating your signal. At 2 GHz, rain is basically irrelevant. If you are designing a link above 10 GHz and you are not budgeting at least 6 to 8 dB of rain fade margin for your region, you are building a system that will underperform during wet weather. This is not speculation. The ITU-R P.838 model gives you specific rainfall rate data based on your geolocation. Use it.
Another common pitfall I see repeatedly: people assume that if they can see the other tower from the ground, the RF path is clear. It is almost never that simple. Terrain elevation, vegetation, buildings, and even power lines can intrude into the Fresnel zone without being visible from a ground-level perspective. I recommend using tools like Radio Mobile or WinGreen that factor in actual terrain data and vegetation models. A quick run through either of those programs will save you a truck roll. And truck rolls in this business cost between $300 and $800 each when you include labor and equipment time. Mitigation strategies matter too. Diversity reception is the most reliable technique. I routinely specify dual receiver diversity with vertical polarization separation of about 0.3 to 0.5 wavelengths. This gives you roughly 10 to 15 dB of fading margin improvement. Adaptive coding and modulation (ACM) is another must-have feature on modern radio equipment. It dynamically adjusts the modulation scheme based on real-time channel conditions, dropping from 1024-QAM down to QPSK when the fade depth demands it. Your throughput fluctuates but your link stays up. Without ACM, your radio will either maintain high throughput and drop completely during a deep fade, or it will stay connected at miserable speeds constantly. There are scenarios where line of sight fading solutions simply cannot save you. Dense urban canyons with constant reflections from glass buildings will cause rapid multipath fluctuations that no amount of diversity or power can fully overcome. In those environments, you should consider switching to lower frequency bands like sub-6 GHz or even deploying fiber where feasible. Similarly, heavily forested paths with seasonal foliage changes require you to design for worst-case vegetation attenuation, which can add 20 to 40 dB of loss at higher frequencies during leaf-on conditions. I once overlooked this on a rural link in the Pacific Northwest and had to come back and install an additional 3-meter tower section six months later. The initial budget didn't cover it. It never does.
Get the Full Details
If you are working through a current LOS fading issue on an existing installation, start with these steps. Pull up your link report and check the current SNR and received signal level. Compare them against your design specifications. If the received signal is more than 6 dB below design, you have a fading or obstruction problem. Use a portable spectrum analyzer to identify whether the degradation is narrowband (likely multipath) or wideband (likely absorption or rainfall). Then run a Fresnel zone clearance analysis using current elevation data. If you find an obstruction, the fix is usually elevation adjustment, relocation, or adding a repeater node. Sometimes the cheapest fix is just increasing the transmit power by 3 to 5 dB if your radio has headroom. The reality is that line of sight fading is not a bug in wireless networking. It is a fundamental characteristic of radio propagation that you plan around from day one. Every professional link budget includes fade margin, diversity gain, and weather attenuation calculations. The ones who skip those steps learn about it the hard way, usually while standing on a roof in the rain trying to realign an antenna that was never going to work in the first place.
