Measurement Uncertainty GUM

Uncertainty budgets to the GUM (JCGM 100), with the validation record ISO/IEC 17025 asks for — between a home-made spreadsheet and a four-figure desktop licence. The mathematics is public; the pain is not. A self-built sheet invites the question every assessor asks: has it been validated (§7.11)? A desktop package costs more than a small laboratory can justify. And the ILAC P14 rounding is redone by hand every time. ▼ 30 days, every feature, free — no payment details ▼ The thinking behind it The result is never hidden: no blurred numbers, no capped component count. Nor does the app decide conformity — no "OK", no "pass". It gives the value, the reasoning and the dominant term. ▼ What it does Write the model as text — R_std * (1 + alpha * dT) + dR_res — and the app differentiates it for the sensitivity coefficients. No fixed templates, and the usual functions are supported. A syntax error points at the character instead of quietly returning zero. Type A: paste repeated observations for the mean, s, u = s/√n and n−1 degrees of freedom. Type B: U and k from a certificate, a half-width for a rectangular (a/√3), triangular (a/√6) or U-shaped (a/√2) distribution, or a standard uncertainty directly. Correlated inputs: enter r for any pair sharing a cause — the term home-made sheets most often omit. Welch–Satterthwaite effective degrees of freedom and k from the t-distribution, computed rather than read from a table. Coverage probability 68.27 % (k ≈ 1), 95 %, 95.45 %, 99 %, or any value a client specifies. Each component's share of the variance is shown. Monte Carlo (JCGM 101): propagate the distributions instead of linearising and compare both intervals; if the model is markedly non-linear, the app says so. Up to 5 000 000 trials, seed recorded. The reported line, rounded to ILAC P14: U to two significant figures, the measurand to the same decimal place, k and the coverage probability stated. ▼ The ISO/IEC 17025 §7.11 validation record Calculation software must be verified before use and the verification recorded. The app runs built-in cases from the GUM and ILAC P14 and issues a PDF: version, run time, and every case with its expected value, computed value, deviation and verdict, under your laboratory and operator names. Comparisons with published worked examples can be added; that data belongs to its publishers, so you enter it from your own copy. ▼ Outputs Budget PDF with every input value, the model, the sensitivity coefficients and the contributions, so a third party can re-derive the result. Certificate values as a separate short document. CSV with units in the column names. Excel with the formulas still live: change an estimate and the sensitivities, degrees of freedom and k recalculate. ▼ Who it is for Small calibration laboratories and technical managers; independent metrology consultants and quality auditors; laboratories working towards or holding ISO/IEC 17025; quality departments under IATF 16949 or ISO 9001. ▼ Also included Save, duplicate and rename projects, with a confirmation before deleting. Reuse components, models and units from history. Manual sensitivity coefficients. Disable a component temporarily. Free-text units. QR or transfer text moves a budget to another device. Light, dark or system appearance. Optimised for iPad. English, Japanese, German, French, Spanish. Offline; never asks for location, camera or microphone. Subscription (monthly or yearly), charged to your Apple ID and renewing automatically unless cancelled at least 24 hours before the period ends. After the trial a subscription is required; there is no reduced free mode. Saved budgets stay readable and your data can always be exported. EULA: https://www.apple.com/legal/internet-services/itunes/dev/stdeula/ Privacy: https://qconp.com/privacy A calculation aid: it does not replace engineering judgement or a conformity decision. Results must be verified by a competent person before use on a calibration certificate or test report.