Your Range Test Lied: Fresnel Zones and Why Link Budget Fails in the Field

Line of Sight Is Not the Requirement

Nearly every range test I am asked to review was flown the same way: a clear day, a long straight stretch of open ground, the aircraft flown directly away from the GCS until video degraded. The number that comes out of that test gets written into the datasheet. It is almost always wrong by a factor of three or more in real terrain.

The reason is that radio propagation does not care about the geometric line between transmitter and receiver. It cares about a three-dimensional ellipsoid around that line — the first Fresnel zone. Energy arriving at the receiver has travelled slightly different paths through that volume. Obstruct part of it and those paths recombine destructively, even when the antenna can still "see" the aircraft perfectly.

The Number Nobody Computes Before Flying

The first Fresnel zone radius at midpoint is r = 17.32 × √(D / (4f)), with D in kilometres, f in GHz, r in metres. At 5.8 GHz over a 2 km link, that is roughly 5.1 m of clearance needed at the midpoint. At 900 MHz over the same distance it is about 12.8 m. The common rule is to keep 60% of that zone clear.

This is why the same airframe behaves so differently on two ranges. On a 2 km link at 5.8 GHz you need five metres of clearance above the terrain midpoint. A gently rising field, a treeline, a parked vehicle — any of these intrudes. The operator sees full bars at 800 m, then a cliff-edge dropout at 1.2 km, and reports it as "the video link is unreliable." The link is behaving exactly as physics dictates.

Field Example: The 400-Metre Berm

On a tactical quadrotor evaluation we were asked to explain intermittent video loss that the manufacturer could not reproduce. Their range test showed 9 km. The customer was losing link at 1.6 km consistently, always in the same direction.

We walked the line. Four hundred metres from the GCS there was an earth berm roughly 3 m high — nowhere near blocking line of sight to an aircraft at 120 m altitude, and completely ignored in the site survey. But at 5.8 GHz over that geometry, the berm was cutting into the lower third of the first Fresnel zone. We re-flew with the GCS antenna raised 2.5 m on a mast. Link held to 7.4 km on the same heading. No firmware change, no hardware change — 2.5 m of antenna height.

What a Useful Range Test Actually Looks Like

Specify the terrain profile, not just the distance. Record RSSI and SNR continuously rather than noting the dropout point, because the shape of the degradation curve tells you whether you are seeing Fresnel obstruction (sharp, position-dependent, repeatable) or genuine range limit (gradual, symmetric). Fly the link in at least four headings — multipath over the same ground is directional. Log antenna height at both ends; it is the single cheapest variable you control.

And state the result honestly: "12 km line-of-sight over water, 2.5 m mast" is a specification an integrator can plan around. "12 km range" is a number that will be discovered to be false at the worst possible moment.