Transfer correction

Sensitivity is set on a reference block, but the component is never acoustically identical to it. Surface roughness, curvature, paint, coupling, grain structure and attenuation all differ, and every one of them costs signal. Transfer correction is the gain that must be added to the reference setting so that the same reflector would give the same screen height in the component as it did in the block.

The usual measurement is a V-path through-transmission with two identical angle probes: set the signal to a fixed height on the reference block, note the gain, repeat on the component at the same probe separation, and take the difference. Because the reading depends on where you put the probes, the measurement is repeated at several positions and averaged, and a positive difference means the component needs more gain.

If the sound paths on the block and the component are not equal, part of the difference is simple beam spread rather than material loss. In the far field a plane reflector loses 20·log₁₀(S₂/S₁) from geometry alone, and that term is removed here so the number left is the real transfer loss. If the material attenuation is known, the extra loss over the longer path — α·(S₂ − S₁) — is removed in the same way; with equal paths both terms vanish. Keep the paths equal if you can; the correction is then zero and the measurement is far more reliable.

The last line is what actually goes on the instrument: primary (reference) gain, plus transfer correction, plus whatever scanning gain the procedure calls for. A transfer correction larger than about 12 dB should not simply be dialled in — it usually means the surface preparation, the couplant or the probe contact needs attention before the inspection is valid.

ΔG = G_component − G_reference
Spread term = 20 · log₁₀( S_component / S_reference )
Transfer correction = ΔG − spread term − α · (S_component − S_reference)
Total scanning gain = primary gain + transfer correction + scanning gain