Typical defects
Last updated 2026-09-08 · defect base revision 2
This is what we meet regularly. Some of it is on us at the furnace, some is built into the part's design and metal - and then it is fairer to say so before the work.
The full version of this base is published in machine-readable form; its schema is described on the documents page.
1. Distortion and warping (not our fault)
What it looks like. After quenching the geometry has moved: a shaft bows, a plate cups, a bore goes oval.
Where it comes from. Uneven heating or cooling, a poor orientation on the rack, and most often the part's own stresses from forging and rough machining.
What we do. Long parts hang vertically and the orientation is chosen, prone grades go into polymer rather than water, and residual bow is straightened after tempering.
How to avoid it: Stress-relief annealing after rough machining and the grinding allowance from the preparation table.
2. Quench cracking (not our fault)
What it looks like. A fine crack, usually from a sharp corner, a fillet, a hole edge or a tool mark; seen at once or within the first hours.
Where it comes from. Overheating on heat-up, cooling too severe for the grade, a stress raiser built into the design, or a delay before tempering.
What we do. Tool steels are heated in two stages with an 850 °C preheat, tempered within four hours, and complex geometry goes into oil rather than polymer.
How to avoid it: Round the corners to R 0.5 mm, remove tool marks, and do not quench a finished thread.
3. Decarburisation (not our fault)
What it looks like. The surface is 5–10 points softer than specified, while below a 0.1–0.3 mm layer the hardness is already normal.
Where it comes from. A long soak in the air atmosphere of a chamber furnace: surface carbon burns off and ferrite is left.
What we do. We hold no charge longer than the calculated soak and warn about a possible layer in advance.
How to avoid it: A 0.3–0.5 mm per-side allowance and grinding after heat treatment.
4. Low hardness (we look into it)
What it looks like. Hardness is uniformly below the target across the whole part.
Where it comes from. A low temperature or short soak, too little hardenability at this section, overheated oil, or scale on the surface.
What we do. We check the route and the bath, and where hardenability is suspect we cut a witness and read the structure through the section.
How to avoid it: Choose the grade for the section: 45 above 25 mm gives hardness only in a thin skin.
5. Patchy hardness (we look into it)
What it looks like. Hardness differs between neighbouring points, with a spread of 5–8 units.
Where it comes from. Scale and oil residue blocking heat transfer, or a dense charge where parts shield one another.
What we do. We space the charge out, clean the surface before quenching, and re-measure to the sampling rule.
How to avoid it: A clean surface and gaps in the charge; parts are not packed tight.
6. Overheating and coarse grain (not our fault)
What it looks like. Hardness is in tolerance but the part is brittle and breaks on the first impact; at 500× the grain is coarse.
Where it comes from. Heating above the needed band or too long a soak - usually an attempt to rush the work with temperature.
What we do. We hold the band from the routes table and never trade a shortage of time for temperature; coarse grain is corrected by normalising.
How to avoid it: Do not raise temperature for speed, and give the soak the section needs.
When a defect is on the protocol, we write the cause and what to do, not just "reject". How we measure and by what rules is on the control page.