Setting Out & Survey Control

Definition

Setting out is the transfer of the design geometry from drawings to the ground: gridlines, levels, centre-lines and offsets, marked so the works can be built in the right place. Survey control is the stable reference framework that makes setting out repeatable — primary control stations coordinated to a national or site grid, benchmarks tied to a known datum, and secondary stations from which day-to-day work is set out. Everything built on the project inherits the accuracy, or the error, of this framework.

Why It Matters

Setting-out errors are the most expensive millimetres in construction. A pile cap cast 40 mm off grid becomes a column that will not take its baseplate, a holding-down bolt pattern that needs slotting, a steel erector standing idle, and a variation argument that outlives the defect. Vertical control is worse: a datum error discovered at level twelve can mean every cast-in channel and every façade bracket on the building is wrong. The control network is cheap — a few days of a surveyor's time at the start. Repairing what flows from its failure is not.

How It Is Established Properly

1. Verify the datum first — confirm the published benchmark actually reads what the drawing says, with a level run to a second independent benchmark.
2. Establish primary control — at least three intervisible stations in protected positions, coordinated by GNSS and total-station traverse, closed and adjusted.
3. Install stable benchmarks — on structures that will not move: existing foundations, deep-set posts, not fence lines or kerbs.
4. Commission an independent check survey — a different surveyor, different instrument, same network. Discrepancies get resolved before excavation, not after.
5. Set out from control, check by dimension — every gridline intersection checked by diagonal, every level closed back to the benchmark.
6. Protect and re-verify — stations are disturbed by plant constantly; scheduled re-checks catch movement before the works inherit it.

Real-World Example

A two-block residential scheme was set out by two subcontractors, each working diligently — from different benchmarks. Block A used the datum on the architect's drawing; Block B used a temporary benchmark a previous contractor had established 35 mm low. The error surfaced when the link bridge steel arrived and the bearing levels disagreed by 35 mm across an eight-metre span. The steel had to be re-shimmed, the façade brackets re-set, and the argument about who pays ran for nine months. The root cause was one line missing from the mobilisation checklist: a single, jointly witnessed datum verification before either block was set out. One hour with a level would have bought nine months of claim correspondence back.

Practical Lessons Learned

• One datum, one grid, one custodian. Where two grids or two datums exist on a drawing set, resolve them on paper before anyone opens an instrument case.
• Benchmarks on temporary works are temporary benchmarks. Hoarding posts, crane bases and container roofs all move; none are benchmarks.
• Diagonals catch what coordinates hide. A gridline intersection can be set from perfect stations and still be wrong through a data-entry slip — the diagonal measurement is the trapdoor check.
• GNSS lies politely under trees and beside structures. Multipath errors look like good fixes; confirm with the total station.
• Recheck control after heavy plant passes nearby. Vibration and surcharge move shallow stations by millimetres that matter.

Expert Tips

• Paint primary stations with a unique ID and a "do not disturb" board, then photograph each with a dated scale. Disputes about disturbed stations end quickly with photographs.
• Run a monthly level loop through every benchmark and log the closures. A drifting benchmark announces itself in the trend long before it announces itself in the cladding.
• Keep a setting-out request log: who asked, what drawing revision, what was set, who checked. When the wrong-revision argument starts — and it will — the log ends it.
• Transfer vertical datum up the structure by steel tape or EDM from a fixed base point, not by floor-to-floor spirit levels. Compounding a metre-stick error twelve times is how roofs end up out of level.
• Before any major pour, have the surveyor issue a signed setting-out certificate for that element. It concentrates minds wonderfully.

Common Mistakes

• Accepting the drawing datum without an independent verification run to a second benchmark.
• Setting out from a single control station with no redundancy, so one disturbed stake invalidates weeks of work.
• Using different drawing revisions across trades — the groundworks set out from revision C while the frame designer details from revision E.
• Trusting uncalibrated tapes and instruments; a five-metre tape that has been run over by a dumper is not a five-metre tape.
• No dimensional checks between set points, so systematic scale errors propagate across the whole site.

Key Takeaways

• Survey control is the project's geometric foundation — invest in it before anything is built.
• Verify the datum independently, witness it jointly, and write the result into the mobilisation record.
• One custodian owns the grid, the datum and the revision status of setting-out information.
• Check by dimension — diagonals and closures — not just by instrument readout.
• Protect stations, re-verify on a schedule, and log every setting-out request.

Related Concepts

Pairs with Pre-Construction Site Survey, Dilapidation Survey, Piling Records, and Hoarding & Site Boundary Plan.