Digital Instruments, Signal Processing and Controls

Learning outcomes

By the end of this chapter, the learner should be able to describe the signal path through a digital flaw detector; explain the technical effect of major controls; distinguish range, zero, delay and velocity adjustments; identify processing choices that can hide or distort evidence; and specify instrument performance checks appropriate to an application.

3.1 The signal chain

A typical pulse-echo instrument contains a timing controller, pulser, transmit-receive switching, analogue receiver, amplification and filtering stages, an analogue-to-digital converter, digital processing, gates, measurement logic, display and data storage. The timing controller coordinates excitation and sampling. The pulser energises the probe. Returning electrical signals are conditioned before digitisation, then displayed or measured.

Every stage has finite bandwidth, dynamic range and recovery time. The displayed A-scan is a processed representation of the received voltage, not the acoustic field itself.

3.2 Range, velocity, zero and display delay

Range determines the interval shown or acquired. An insufficient range can omit required reflectors; an excessive range can reduce screen resolution or data density.

Material velocity converts time to distance. Probe zero or delay compensates for travel in wear plates, delay lines, wedges, cables and switching. Display delay shifts the visible or acquired time window without necessarily changing the underlying zero reference. These controls can produce a plausible screen while giving incorrect locations if confused.

3.3 Pulser and receiver controls

Pulse voltage or energy influences excitation. Pulse width or tuning may improve transfer to a particular probe frequency. Damping changes electrical loading and can trade sensitivity against pulse duration and resolution. Receiver gain changes displayed amplitude. Analogue and digital filters can reduce unwanted frequency components but may also change pulse shape, amplitude and resolution.

Rectification may be full-wave, positive half-wave, negative half-wave or radio-frequency. Radio-frequency presentation preserves phase and polarity information but needs suitable sampling and interpretation. Reject or suppression removes low-level display content and can destroy evidence or change amplitude relationships. It should be prohibited or tightly controlled where it compromises evaluation.

3.4 Gates and measurement logic

A gate defines a time or distance region for amplitude, time-of-flight or alarm measurements. Gate start and width must contain the full expected response range. Threshold defines when a signal is measured, but different instruments may report first crossing, peak, flank or centre values. A procedure should identify the required measurement basis when it affects results.

3.5 Sampling, PRF and aliasing

The digitising rate must adequately represent the highest relevant signal frequency. Inadequate sampling can distort the waveform. Pulse repetition frequency must allow echoes from one pulse to decay before the next pulse produces ambiguous responses. Long sound paths, low attenuation materials and immersion paths can create wrap-around or ghost echoes when PRF is too high.

Encoded systems add spatial sampling. Maximum scan speed is limited by pulse rate, number of channels or focal laws, averaging and required spacing between acquired A-scans. Missing lines in a colourful image remain missing data, even if interpolation is cosmetically persuasive.

3.6 Performance characteristics

Method-level controls may include horizontal and vertical linearity, resolution, dead zone, noise, sensitivity, stability, time-base accuracy, gate accuracy, pulse repetition stability and data-storage integrity. Checks should match the intended measurement. A thickness application, high-resolution near-surface examination and long-range weld scan do not present identical risks.

Failure modes and decision control

Weak practice commonly includes using display delay to conceal a troublesome response; increasing voltage without confirming probe limitations; changing filters after calibration; using reject during evaluation; setting PRF too high for the sound path; assuming a stored setup proves the connected hardware is correct; and failing to record processing settings that materially affect the data.