Measurement Uncertainty

A stated ±limit is not a standard uncertainty — it has to be divided by the shape of its distribution first (√3 for a rectangular limit, √6 for triangular). Forgetting that divisor is the classic budget error, and it inflates every result. Contributions then add in quadrature, so the largest one dominates.

k = 2 gives roughly 95% coverage for a normal distribution and is the usual reporting basis. The contribution column is the useful one: chasing anything other than the dominant term is wasted effort.