Maximum pulse repetition frequency

A flaw detector fires, listens, then fires again. If it fires before the echoes from the previous shot have died away, those late echoes arrive during the next time base and appear as ghost echoes — indications at a depth where there is nothing. The time for one round trip of the longest sound path is t = 2·z/c; the interval between pulses must be longer than that, and in practice several times longer because multiple reflections from the same path keep returning.

The factor to use depends on how long the material rings. In a low-attenuation forging or in a small component with parallel faces, multiples persist for many round trips and a factor of 5–10 may be needed; in an attenuative or irregular part 2–3 is usually enough. The safest field check is simply to halve the PRF: if an indication moves or disappears, it was a ghost.

The other half of the problem is that PRF drives coverage. Automated and encoded scanning needs enough A-scans per millimetre of travel, and averaging multiplies the number of shots required, so PRF sets the maximum scan speed. High PRF also heats the probe and pulser and can degrade signal-to-noise on long sound paths, so it is not free.

The unambiguous range figure works the other way round: at a given PRF, it is the deepest single-path reflector whose echo can arrive before the next pulse. Any reflector beyond it will be displayed at a false, shorter range.