Skip to content

News

4:1: is the calibration good enough for the tolerance?

  • News
  • 1 min read

A ±1 Pa room pressure transmitter, checked against a reference that is itself ±1 Pa, has not been checked.

If the two disagree by 1 Pa, either one could be wrong.

The usual rule is a test uncertainty ratio of 4:1. The tolerance being verified should be at least four times the uncertainty of the measurement used to verify it. For the ±1 Pa transmitter that means a calibration uncertainty of ±0.25 Pa or better, so the reference itself has to be better still. ANSI/NCSL Z540.3 uses 4:1 as the fallback where the probability of a false accept cannot be estimated.

Low differential pressure is where the ratio is hardest to meet. A handheld micromanometer that is comfortable at 50 Pa can carry an uncertainty close to the tolerance at 10 Pa. The options then are a better reference, a wider tolerance justified against the process, or a guard band: accept only readings inside the limit by a margin equal to the uncertainty.

On the certificate, read the uncertainty at the points tested and work out the ratio at the lowest one, where it is worst. A display resolution of 0.1 Pa says nothing about it.

What ratio does your calibration procedure ask for?