A replacement part can match the original in every dimension and still fail within weeks if the material or its treatment is wrong. Unknown material is the single largest risk in reverse engineering, and the responsible handling is to test, to reason from duty, and to record the decision — never to quietly assume.
7 min read · Last reviewed 28 September 2026 · All resources
Dimensional errors reveal themselves at assembly. Material errors reveal themselves in service, often after the part has been installed, the machine returned to production and the original conversation forgotten.
That delay is what makes it expensive. The failure looks like a manufacturing defect, and the cost includes the downtime it causes rather than just the part.
Testing describes the part in front of you, not necessarily the specification it was made to. A part that has been in service may have been tempered by heat, work-hardened at the surface, or decarburised.
It also cannot tell you what the designer intended as a minimum. A sample that tests at a particular hardness may have been at the top or bottom of its original range, and reproducing that exact value is not the same as meeting the specification.
How the original part failed says a great deal about whether its material was right. Surface wear suggests hardness or lubrication; a fatigue crack from a corner suggests stress concentration; gross deformation suggests the material was too soft or the load higher than intended.
A part that failed by fatigue is a poor candidate for exact reproduction, because copying it faithfully copies the cause.
Whatever is chosen, it belongs on the drawing as a stated specification rather than in someone's head. That way the next person to order the part gets the same thing, and if the first replacement fails, there is something to review.
Where the duty is critical and the specification uncertain, it is reasonable to make one, run it in service and review before committing to a batch.
FAQ
The questions engineers and buyers ask before sending a drawing. If yours is not here, ask it directly — you will get an engineer's answer, not a brochure.
Still have a question?
Talk to Anita M.N, Managing Partner, or send the drawing and let the review answer it.
Not reliably. Visual and machining characteristics narrow the possibilities, but a stated specification based on duty is more defensible than a guess based on appearance.
Often, yes, and sometimes advisable where the original failed. It should be a recorded decision, because 'better' in one property can be worse in another — harder frequently means more brittle.
You do, as the person who knows the application and carries the consequence. The supplier's job is to set out the options, the assumptions and the risks clearly enough for you to decide.
Then the standard of evidence should rise accordingly, and assumption is not an adequate basis. That may mean testing, a conservative specification, or declining the work — and a supplier should say so rather than proceed quietly.
Start a project
Share what you are trying to make or fix. Our engineering team reviews every enquiry and responds with the questions that matter, a feasibility view and next steps.
