How deep — and how accurate — is concrete scanning, really?

With the 1.6 GHz antenna used for concrete, practical depth in typical structural concrete is roughly 30–45 cm — deeper than the great majority of slabs, walls, and decks you'll core. That high frequency is chosen for resolution: it separates closely-spaced rebar and identifies conduit and PT clearly. For thicker elements (transfer slabs, deep footings) a lower-frequency antenna reaches farther at the cost of fine detail, and we select the antenna to the structure.

Two accuracies. Lateral position — where a target sits on the surface — is excellent, marked to within a couple of centimetres, which is all you need to shift a core clear of rebar, conduit, or a tendon. Depth is good but an estimate calibrated to the concrete, typically within about 10–15%, and confirmed by the core itself. Anyone promising millimetre depth certainty on a single non-destructive pass is overselling; we report lateral with high confidence and depth as a calibrated estimate.

It detects well: metallic rebar and conduit, post-tension tendons and ducts, slab thickness via the back-wall reflection, and significant voids or honeycombing. It struggles with: very small non-metallic targets, the precise interior of densely congested reinforcement mats where reflections overlap, and anything beyond the antenna's depth. Honest scanning maps what radar sees clearly and flags where the slab is at the edge of the method's envelope.

Yes — it's the main real-world limiter. In a tightly-spaced reinforcement mat, individual reflections begin to overlap, making it harder to resolve a clear gap or see what lies beneath the top layer. A skilled operator still maps the pattern and cover and finds a clear window where one exists; where congestion genuinely obscures a critical penetration, the responsible call is to advise radiography rather than guess. Knowing that limit is part of competent scanning.

Yes, where the back face is within the antenna's depth range and reflects cleanly — the radar picks up the slab's far surface and thickness is read from the two-way travel time, calibrated to the concrete. For typical slabs this is reliable and one of GPR's everyday outputs; for very thick or heavily congested elements the back-wall reflection can weaken, which we flag. It's the same measurement that underpins void detection.

Radar speed through concrete depends on its dielectric properties, which vary with moisture, mix, and age — so depth estimates are calibrated to the actual structure, not a generic constant. Green (freshly cured) concrete is wetter and attenuates more; mature dry concrete reads cleaner. A 1970s post-tensioned parkade and a modern slab present differently, which is why interpretation against the structure's history matters as much as the raw scan.

On the roughly 5% of jobs where the slab is beyond GPR's honest envelope: extreme reinforcement congestion that overlaps reflections, a need to identify exact embed type and condition with sub-centimetre certainty, and situations where both faces of the element are accessible for radiography. For the other ~95%, GPR wins on one-sided access, no radiation, no evacuation, real-time results, and lower cost — see the GPR vs X-ray comparison for the full decision.

No honest scanner guarantees a zero-strike absolute — radar has physical limits, and abandoned or anomalous embeds exist. What a competent scan delivers is a dramatic risk reduction with stated confidence per location: high where the slab is clean and within range, qualified where congestion or depth pushes the envelope, and an explicit recommendation to change method where the structure demands it. That candour about confidence is exactly what makes the clearances you can trust trustworthy.