Three Mistakes That Ruin Ultrasonic Flaw Detection

Probe and couplant

A high-frequency narrow-pulse probe on coarse-grained material attenuates badly and simply will not find the defect. A small-angle probe on thick sections leaves the beam unable to cover the joint, so the edges are always missed.

Choose by grain size and thickness: thin-plate welds want a dual-crystal straight probe for near-field resolution; shrinkage and porosity in castings want a low-frequency, large-crystal longitudinal probe for coverage.

Couplant is not something you smear on. Too much runs and makes the probe slide, and adds spurious reflections; too little leaks energy and drops sensitivity. The rule is enough, evenly, pressed lightly to make the layer continuous.

Parameter setup

Gain too high and noise buries the defect; suppression too high and small but real signals are filtered out entirely; sweep range set wrong and the region of interest runs off screen; delay set wrong and near-surface defects are invisible.

Do not trust a preset. Manufacturers ship defaults, but your geometry, material and bevel are not their defaults.

A workable routine: set gain from a reference block so the reference reflector reaches a defined screen height; calibrate sweep scale against a known artificial defect; then set delay and range. Compensate velocity when the ambient temperature swings.

The human factor

Without an understanding of waveform signatures, an operator follows the procedure mechanically and misjudges anything unusual. Without an understanding of how surface finish, curvature and material variation affect the result, they cannot separate a defect echo from a geometric one.

There is no shortcut: study the physics and the defect responses, then practise under supervision – starting on flat butt welds and moving up to pressure vessels and pipe elbows.

Summary

Right probe, right parameters, readable waveform. Miss any one of the three and the conclusion will not stand.