ToFD gate start and width

A ToFD A-scan is only useful if the digitising window contains everything between the lateral wave and the backwall, with margin at both ends. The lateral wave arrives at PCS/c and the backwall at 2*sqrt(S^2 + T^2)/c, and both are shifted later by the delay through the two wedges. The gate is then started a little before the lateral wave and ended a little after the backwall.

Starting the gate on the lateral wave itself is a common and expensive mistake. The wave gets clipped, the reference for near-surface depth is lost, and the software can no longer use the lateral wave to correct for probe lift-off or velocity drift during the scan. Procedures typically call for something like a microsecond of clear material either side; check the governing procedure for the figure that applies to the job.

Extending the window past the first backwall is also worth doing. Mode-converted signals arriving after the backwall carry information about flaws near the far surface, and seeing a second backwall arrival confirms that the assumed thickness and velocity are right. The cost is file size and, on some instruments, a lower maximum line rate.

Set the gate from calculated times, then verify it on the actual A-scan against a known thickness before the scan starts. If the lateral wave or backwall does not fall where predicted, fix the setup rather than moving the gate to suit.

t_LW = PCS / c + t_delay
t_BW = 2 × √(S² + T²) / c + t_delay
gate start = t_LW − pre-margin
gate end = t_BW + post-margin
gate width = gate end − gate start