Two sensors doing two different jobs
Capacitive sensing looks for a change in the dielectric behind the surface. Plasterboard with a timber stud behind it holds a different charge from plasterboard with a void behind it, and the tool reports the edge where that changes. It is what finds studs and joists, and it is why the tool has to be calibrated on a blank stretch of wall first: it is measuring a difference, so it needs to learn what nothing looks like.
Inductive sensing is a separate circuit that reacts to metal, and it does not care about density. That is what finds conduit, pipe, nails and steel framing. A multi-sensor scanner runs both and reports whichever fires, which is why it can distinguish a metal stud from a wooden one but cannot reliably tell you a copper pipe from a steel one.
Live AC detection and what it does not cover
Live cable detection senses the alternating field around a conductor carrying mains. It is genuinely valuable and it is also the function most often over-trusted. It finds live cable. It does not find a cable that is switched off, a cable in earthed metal conduit that screens the field, or a cable buried deeper than the tool can reach.
Treat a clear scan as one input rather than permission. The safe sequence is to scan, then look at where the sockets and switches are and assume the vertical and horizontal safe zones between them are occupied, then drill shallow and check. Anyone who has drilled a cable will tell you the scanner said nothing.
| Mode | Finds | Misses |
|---|---|---|
| Capacitive (stud) | Timber studs and joists, density changes | Metal, anything behind foil or wet plaster |
| Inductive (metal) | Conduit, pipe, nails, steel framing | Timber, plastic pipe |
| Live AC | Energised mains cable near the surface | Cable that is switched off, screened or deep |
Depth ratings and where they stop being true
Quoted depth is a best case on a co-operative wall: dry plasterboard, a clear target, a calibrated tool. On lath and plaster the laths themselves read as a repeating pattern of studs. Foil-backed insulation blocks capacitive sensing almost entirely. Fresh plaster and damp walls hold water, and water changes the dielectric enough to give false positives across an entire wall.
The honest workflow is to scan twice, once horizontally and once vertically, and to mark both edges of anything found rather than its centre. Two edges a stud-width apart is a stud. A single narrow reading that does not repeat at regular spacing is more likely a pipe, a cable or a fixing, and it deserves care rather than a drill.
What is worth paying for
A display that shows signal strength beats one that only beeps, because the useful information is where the reading peaks rather than whether it crossed a threshold. Auto-calibration on power-up removes the most common user error. A tool that reports what kind of target it thinks it has found is worth having, provided you treat it as a hint.
For trade use the deciding features are a case, a battery that is not a coin cell, and the ability to hold calibration while moved across a wall rather than needing a reset every 300 mm. Beyond that, extra modes add less than a careful second scan at right angles does.
