Two identical towers, two different soils, two completely different foundations. Here is how our civil team reads a site before designing anything.
Ask any telecom civil engineer what causes the most rework on a rollout, and the answer is rarely the steel. It is the ground. A foundation designed on assumption instead of investigation is the fastest way to turn a four-week site into a four-month dispute.
Start with the soil, not the drawing
Before a single design is issued, we carry out a geotechnical investigation: boreholes to establish strata, standard penetration tests to gauge density, and laboratory testing for classification, shear parameters and water table depth. From this we derive the safe bearing capacity — the number every subsequent decision hangs on.
A stiff clay at 200 kN/m² supports an isolated pad-and-chimney foundation comfortably. The same tower on loose silty sand with a high water table needs a raft, or bored piles taken down to a competent layer. Designing the first for the second is how towers develop differential settlement two monsoons later.
The four foundation systems we use
- Pad and chimney — the standard four-legged GBT foundation. Economical, fast, and correct when bearing capacity and uplift resistance are adequate.
- Raft — a single mat under all four legs. Used where bearing pressure must be distributed or where individual pads would nearly touch anyway.
- Bored cast-in-situ piles — for weak upper strata, high water table or reclaimed ground. Slower and costlier, but the only honest answer on a bad site.
- Rock anchors — where hard rock is close to surface, anchored foundations save enormous excavation.
Uplift is the load people forget
Towers are dominated by wind. On a four-legged structure under lateral load, two legs go into compression and two into uplift. Compression is easy — the soil takes it. Uplift is resisted by the weight of the foundation plus the frustum of soil above the pad. Get the frustum angle wrong for the soil type, or backfill with loose material that was never compacted, and the uplift capacity you designed for does not exist on site.
This is why we insist on controlled backfill in layers with compaction testing. It is unglamorous work that quietly determines whether the design is real.
The anchor bolt template: two millimetres decide the job
Everything upstream can be perfect and still be ruined at the pour. The anchor bolt template must be set to the exact leg spacing, diagonal, level and orientation given in the tower drawing. We double-check diagonals, lock the template rigidly, and re-verify immediately before and after the concrete pour — because concrete does not negotiate.
A template out by a few millimetres means the tower base plate will not seat. The recovery options — slotting base plates, chipping and re-grouting, or breaking and recasting — are all expensive and all avoidable.
Quality records are part of the deliverable
Every foundation we hand over comes with the borelog, design calculation reference, reinforcement checklist, concrete cube test results at 7 and 28 days, level and diagonal records, and dated photographs of each stage. If a site is ever questioned years later, the file answers it.
The practical takeaway
Soil investigation costs a fraction of one percent of a tower budget and prevents the most expensive failure mode in the entire scope. Skipping it to save a week is the worst trade in telecom civil work.
