TL;DR
- Electrical tilt changes the pattern; mechanical tilt changes the pattern and the footprint shape. They are not interchangeable.
- In interference-limited networks, reducing overlap is worth more than increasing coverage, and tilt is the cheapest lever available.
- Beyond about 12° of electrical tilt the pattern distorts and the trade stops being favourable; the site is then too high for its grid position.
What is the difference between mechanical and electrical tilt?
Mechanical tilt rotates the antenna physically, so the boresight points downward. The pattern keeps its shape but the coverage footprint becomes asymmetric: the main lobe is pushed toward the ground on the boresight side and the upper sidelobes tilt with it, while the horizontal pattern seen from the side of the sector changes. Electrical tilt changes the phase relationship between array elements, which moves the main beam down while keeping the antenna vertical and the pattern symmetric around the boresight.
They are not interchangeable, and the difference matters most in interference management. Electrical tilt reduces the upper sidelobe energy that spills into neighbouring cells; mechanical tilt does so less predictably and, at large angles, distorts the horizontal pattern. Where remote electrical tilt units are installed, tilt becomes a parameter that can be tuned and re-tuned at no rigging cost, which is why RET availability changes what optimisation is economically possible.
| Situation | Starting point | Reasoning |
|---|---|---|
| New macro site, 30–40 m, 500 m spacing | 6°–8° electrical | Dense grid; the sector should not reach past its neighbours |
| New macro site, 40–50 m, 1–2 km spacing | 3°–5° electrical | Coverage matters more than overlap at this density |
| Suburban, 2–5 km spacing | 2°–3° electrical | Long sectors; heavy tilt creates holes between sites |
| High site above clutter, 60 m+ | 8°–12° electrical | High sites are interference sources before they are coverage assets |
| Coastal or over-water path | Add 1°–2° beyond the table value | Low propagation loss over water extends the footprint unexpectedly |
How do you set tilt on a new site?
Start from geometry, not from a template. The initial electrical tilt should follow from site height, inter-site distance and the antenna's vertical beamwidth: a 10° beamwidth antenna on a 35 m site at 600 m spacing needs enough downtilt that the main lobe covers the near field without overshooting the next site. A usable first approximation is the angle subtended from the antenna to roughly two-thirds of the inter-site distance, then adjusted for clutter height.
Then let the prediction decide the rest. The point of the initial value is not accuracy; it is to start the search in a region where the optimiser or the engineer is comparing plausible alternatives rather than correcting an obvious error.
Why does overlap matter more than coverage?
Best-server coverage maps are reassuring and largely uninformative. What determines SINR is how many cells serve each location at comparable strength. A location served by one cell at −95 dBm has good SINR. A location served by four cells at −95, −96, −97 and −98 dBm has poor SINR and heavy handover activity, and both maps show it as covered.
Plotting the third-best server level, or the count of cells above a threshold, exposes the difference immediately. In interference-limited networks — which is most urban LTE and nearly all urban 5G — reducing overlap is worth more than increasing coverage, and tilt is the cheapest way to do it.
What are the limits in practice?
Beyond about 12° of electrical tilt the pattern starts to distort: the horizontal beamwidth widens, gain drops, and the sidelobe structure changes in ways the antenna data sheet does not fully describe. Tilting past that point trades coverage for interference control at a rate that is no longer favourable, and the honest answer is usually that the site is too high for its grid position.
Azimuth has similar practical limits. Moving a sector by 20° to fill a coverage hole can leave a gap on the other side that no neighbouring site covers, because the original azimuth was chosen against the grid. Azimuth changes should be evaluated against the full cluster, not the single sector.
How do you verify a tilt change?
Compare the before and after measurement-report distributions for the changed cell and its neighbours, over the same time-of-day window and with comparable traffic. The expected signature of a successful tilt increase is a small drop in the cell's coverage footprint, a rise in its average SINR or RSRQ, a reduction in handover attempts to its neighbours, and no increase in dropped connections at the new edge.
If coverage drops and quality does not improve, the cell was not interference-limited and the tilt increase bought nothing. Reverting is cheaper than living with the change, and knowing which case you are in is the entire value of measuring rather than assuming.