The manufacturer’s specification states what is expected of the instrument when new and under the stated conditions. An in-house acceptance criterion states what error you can tolerate without your process or product decisions ceasing to be reliable. These are different questions, which is why the answers do not always coincide.
When to use the manufacturer’s tolerance
- When the instrument is sized just right for the task, so that the specification is already the requirement.
- When there is no other requirement derived from the process.
- To evaluate whether the instrument still performs as designed, for example after a repair.
When an in-house criterion is needed
- Stricter: if the product tolerance is tight, an instrument within its specification may not leave enough margin. The criterion is derived from the tolerance being controlled, leaving room for uncertainty.
- Wider: if the use is less demanding than the capability of the instrument, a larger error can be accepted without affecting the result.
- With different conditions: if the instrument operates outside the conditions for which the specification was given, the manufacturer’s tolerance loses its validity.
How to define it
- Start from the process tolerance or the customer requirement.
- Calculate how much error and measurement uncertainty you can allow while keeping decisions reliable.
- Compare with the manufacturer’s specification and adopt the one that corresponds to the use.
- Define the decision rule for uncertainty and write it into the procedure.
How to interpret the result
An instrument can be within the manufacturer’s tolerance and outside your criterion, or the reverse. In the first case it is not fit for that application even though it works well; in the second, it is worth reviewing whether the criterion was unnecessarily strict. The key is that the criterion is written down before calibrating, not chosen after the result is known.