TFM amplitude fidelity
Amplitude fidelity is the amount by which the displayed amplitude of a reflector can change purely because of where it happens to fall on the reconstruction grid. A reflector whose true response peaks exactly on a pixel is reported at full height; the same reflector half a pixel away is reported low. ASME Section V Article 4 Mandatory Appendix XI limits that variation to 2 dB.
The loss can be estimated from the shape of the point spread function. Approximating the response near its peak by a Gaussian whose −6 dB width is w, a peak sitting half a pixel from the nearest grid point is under-reported by 20·log10(2)·(Δ/w)² decibels – the 6 dB in that expression being exactly the 6 dB that defines w. Sampling in both image axes adds the two contributions, because the worst case is a peak sitting diagonally between four pixels.
Rearranged for a 2 dB budget split over two axes, the coarsest usable pitch is about 0.41 of the −6 dB spot size. For a full aperture TFM image the spot is roughly half a wavelength, which lands close to the familiar λ/5 grid rule – the rule of thumb and the requirement are consistent, but only the requirement is enforceable.
This is an estimate that assumes a smooth, symmetric point spread function. The demonstration required by the code is a measurement: move a reflector through the zone in small steps and record the largest peak-to-peak variation in reported amplitude.
Compression or shear
Worked example
| Probe frequency | 5 MHz |
| Wave mode | compression |
| Material velocity | 5900 m/s |
| Measured −6 dB spot size | 1 mm |
| Pixel pitch, image width axis | 0.2 mm |
| Pixel pitch, image depth axis | 0.2 mm |
| Amplitude fidelity limit | asme_v_app_xi |
| Procedure limit if not the code value | 2 dB |
| Wavelength | 1.18 mm |
| Estimated −6 dB spot size (λ/2) | 0.59 mm |
| Amplitude fidelity limit applied | 2 dB |
| Worst case loss from width axis sampling | 0.241 dB |
| Estimated worst case amplitude fidelity | 0.48 dB |
| Coarsest pitch meeting the limit (both axes equal) | 0.408 mm |
With a 1.0 mm −6 dB spot and a 0.2 mm pitch, each axis costs 20·log10(2) × (0.2/1.0)² = 6.0206 × 0.04 = 0.241 dB, so the diagonal worst case is 0.48 dB - comfortably inside the 2 dB limit. The coarsest pitch that would still meet 2 dB on both axes is 1.0 × √(2/12.041) = 0.408 mm.