Mounting an antenna ten feet too low looks like a cost savings on installation day. Months later when foliage fills in and links start degrading, it looks like a hardware failure. It is neither. It is a physics problem that was locked in the moment the mount was drilled.
Antenna height at the CPE location determines one thing above everything else: how much of the radio path between the subscriber and the RN clears the obstructions in between. This is not an intuitive relationship. Most installers think about height as "getting above the roofline." The physics of fixed wireless links requires something more precise than that.
The relevant concept is the Fresnel zone, an elliptical region around the straight-line path between the RN and the CPE. When objects enter this zone, they cause diffraction loss that adds to free-space path loss and degrades the link. The standard minimum is 60% Fresnel zone clearance. Below that threshold, diffraction loss increases rapidly and becomes unpredictable. The link will work some of the time and fail, seasonally, or after vegetation growth, or gradually as the hardware ages, at a rate that no troubleshooting resolves because the root cause is not a hardware fault. It is geometry.
The first Fresnel zone radius is largest at the midpoint of the path between transmitter and receiver. It tapers to zero at each endpoint. An obstruction at any point along the path can create significant fade. The closer to the midpoint, the more damaging the obstruction.
| Link Distance | Fresnel Radius at 3.5 GHz | Fresnel Radius at 6 GHz | 60% Clearance Required (3.5 GHz) | 60% Clearance Required (6 GHz) |
|---|---|---|---|---|
| 500 m | 7.3 m | 5.6 m | 4.4 m | 3.4 m |
| 1 km | 10.4 m | 7.9 m | 6.2 m | 4.7 m |
| 2 km | 14.6 m | 11.2 m | 8.8 m | 6.7 m |
| 3 km | 17.9 m | 13.7 m | 10.8 m | 8.2 m |
| 5 km | 23.2 m | 17.7 m | 13.9 m | 10.6 m |
Read that table carefully. At a 2 km link on CBRS 3.5 GHz, you need 8.8 meters of clearance above any obstruction at the midpoint of the path. A 30-foot deciduous tree at the midpoint of a 2 km path needs a CPE mounted high enough to clear 8.8 meters above the top of that tree to achieve 60% Fresnel clearance. Not above the roofline. Above the obstruction, by nearly 9 meters.
At 6 GHz the numbers are smaller in absolute terms but the propagation penalty for obstruction is disproportionately higher. A 6 GHz link through 30% Fresnel zone obstruction loses materially more signal than a 3.5 GHz link through the same obstruction because the first Fresnel zone is smaller, which means any given obstruction represents a larger proportion of the zone being blocked.
This is where the height problem becomes an operations problem. An install completed in January in a temperate climate may have a completely different Fresnel zone obstruction profile in July. Deciduous vegetation that was bare at installation reaches full canopy by summer. Leaf-on attenuation through dense canopy can exceed 15 dB at 6 GHz. A link installed in winter with marginal clearance and 3 dB of headroom will lose that headroom and more by midsummer.
The support call arrives in August. The installer returns in August. The signal reads lower than at installation. The technician adjusts CPE alignment. Nothing improves because the obstruction is between the CPE and the RN, not at the CPE. Another truck roll in September. Same result. The subscriber churns by October having never understood why their service worked fine in winter and failed in summer.
The most common execution failure is straightforward: the installer finds acceptable signal while holding the CPE at shoulder height above the final mount location, confirms the signal looks good, and then mounts the CPE at a lower location because that is where the mounting bracket is easier to attach or where the cable run is simpler.
The signal readings taken above the eventual mount point have no operational value. They do not represent the signal environment at the location where the CPE will actually live. They represent the signal environment approximately 18 to 36 inches higher, which in many installations is the difference between Tier 2 and Tier 3, or Tier 3 and an install that should have been rejected.
Mounting location selection for a CPE is not a construction decision. It is a signal engineering decision that happens to involve a drill. The order of operations matters:
A CPE mounted in an obstruction shadow generates the following cost pattern over a typical 12-month lifecycle:
The correct mount takes the same amount of time as the incorrect one. The difference is a technician who verified signal at the correct position instead of above it.
CPE installation sign-off checklist, Fresnel zone reference tables, mounting location selection criteria, and the seasonal margin framework are all documented in the NGWA Standards.