Non-destructive testing finds flaws in a finished part without damaging it, and a settlement can use four methods: close visual inspection, tapping and listening, dye penetrant for any surface crack, and magnetic powder for surface and near-surface cracks in iron and steel. Each method finds only some flaws, so match the method to the material and the likely defect, and prove it first on a reference piece with a known crack. A part that passes has not been proven flawless, only free of flaws the method can see.
Part of Testing and Quality
Non-destructive testing covers the methods that detect cracks, voids and other flaws in a part while leaving it fit for service.
Match The Method To The Material And The Flaw
| Method | Finds | Works on | Misses |
|---|---|---|---|
| Visual with magnifier | Surface cracks, pits, bad welds | Anything | Tight or hidden cracks |
| Tap and listen | Cracks and voids through the body | Ceramics, bricks, wheels, bells | Small surface flaws |
| Dye penetrant | Cracks open to the surface | Any non-porous material | Subsurface flaws, porous surfaces |
| Magnetic particle | Surface and near-surface cracks | Iron and steel only | Non-magnetic metals, deep flaws |
Radiography and ultrasound find internal flaws but need equipment beyond this section’s scope. The corpus’s Inspection Methods covers visual and load testing of structures.
Start With The Eye And The Ear
Light the part from a low angle so cracks throw shadows, clean it, and use a magnifier (see Magnifying Glass). Look at stress points: corners, holes, welds, the roots of threads and the transitions in diameter, which Failure Modes identifies as where fatigue cracks begin.
Ceramics ring or thud. The ring test for bricks in Ring Test works on the same principle for any solid ceramic or cast piece: tap with a light hammer and listen for a clear note. A cracked part gives a dull sound. This is how a grinding wheel is checked before use, as Abrasives explains.
Prove The Method On A Known Flaw
Methods can fail silently. Make a reference piece of the same material with a deliberate crack, a saw cut or a quench crack from an over-hardened sample, and run your procedure on it. If your method does not show the known flaw, it will not show real ones. Repeat when you change the dye, powder or magnetising setup.
Interpretation Needs Rules
Decide beforehand what you will do with each finding. A short surface scratch may be acceptable on a housing but a crack at a stress point on a hook never is. Write the accept and reject rules down. Test every part of a critical batch, and a sample of the rest.
Worked Example: How Many To Test
If you test only a sample, the chance of catching a bad batch depends on how many parts are bad. If 1 part in 10 has a crack and you test 5 parts picked at random, the chance that all five are good is about (0.9)⁵ = 0.59, so you miss the problem 59% of the time. Testing 20 parts gives (0.9)²⁰ = 0.12, a 12% chance of missing it. For safety-critical items such as hooks, chains and load-bearing pins, test them all.
Reagents and electric current
Penetrant liquids and developers are usually flammable solvents, so keep flame away and ventilate. Magnetising with a battery or a current through the part can arc, burn or ignite flammable liquid. Details are in the dye penetrant and magnetic particle articles.
Common Mistakes
- Trusting a clean result — the method may not detect that flaw.
- Skipping the reference piece — you cannot know the method works.
- Testing after painting or plating — coatings hide cracks.
What’s Next
- Learn next: Dye Penetrant — finding any surface crack
- Combine with: Magnetic Particle — the steel-specific method