Probe Selection and Lift-Off Compensation on a Welded Aluminum Structure
You are asked to inspect a 2024-T3 aluminum aircraft skin (conductivity ~30 %IACS, relative permeability 1.0, 2.0 mm thick) for fatigue cracks emanating from fastener holes. The skin is coated with a 50 µm primer, and the fasteners are installed. The customer's draft technique uses a 500 kHz absolute pencil probe with a 0.5 mm lift-off, scanning over the fastener heads. The technician reports that the probe reads 'good' everywhere because the signal is dominated by the fastener and the coating. The customer wants a go/no-go answer on 120 fasteners in one shift.
1. Evaluate the proposed technique. What are the problems, and what probe/technique would you specify?
Read the worked answer
Scanning over installed fastener heads with an absolute pencil probe is not a valid crack-detection technique: the fastener dominates the signal, the 50 µm primer adds lift-off variation, and an absolute probe cannot separate lift-off from defect. For cracks at fastener holes in 2.0 mm 2024 aluminum, the standard approach is a bolt-hole eddy current probe (rotating or sliding) after fastener removal, or a low-frequency absolute/differential probe for surface cracks with lift-off compensation.
Specify: remove fasteners (or use a hole probe through the hole), use a differential or absolute probe at a frequency chosen so the standard depth of penetration exceeds the 2.0 mm wall (roughly 1-5 kHz for 30 %IACS aluminum, depending on the probe), and establish a lift-off compensation (mix or phase rotation) so the primer does not masquerade as a defect. Demonstrate on a reference block with EDM slots at the hole edge at 0.5 mm and 1.0 mm depth.
2. The customer wants a go/no-go in one shift. How do you handle the schedule and the validity of the result?
Read the worked answer
A one-shift go/no-go on 120 fasteners is achievable only if the technique is validated first. The Level III should: (1) build and verify the reference standard and demonstrate the technique before production scanning; (2) scan with the validated probe and record all indications; (3) classify indications as accept, re-inspect, or reject against the demonstrated detection threshold. If the schedule does not allow validation, the correct answer is to extend the schedule rather than issue an unvalidated go/no-go. Document the detection threshold and the coverage, and state that the result is valid only for the demonstrated sensitivity.