Overview
Measurement uncertainty is a quantified estimate of the doubt that exists about the result of any measurement.
ProCalX Pro edition lets you build an uncertainty budget directly on an Asset's record (Instruments, Test Standard, and Loops) and then apply that budget — automatically or manually — during calibration result entry. This guide explains each field on the Uncertainty tab (shown below) and how the related rules control calibration outcomes.
Uncertainty Components
The Uncertainty Components grid is where you build the uncertainty budget for an Asset. Each row represents one contributing source of uncertainty (for example, the reference standard's accuracy, instrument resolution, or repeatability from repeated readings). Select the + icon to add a component row, or the trash icon to remove one.
| Field Name | Description |
| Component Name | A free-text label identifying the source of this uncertainty contribution (e.g., "Reference Standard Accuracy," "Resolution," "Repeatability"). Required, up to 255 characters. |
| Type | Classifies how the component's value was evaluated: Type A (from statistical analysis of repeated observations) or Type B (from any other source, such as a calibration certificate, manufacturer specification, or engineering judgment). |
| Value | The numeric magnitude of the uncertainty contribution, in the unit specified. Must be numeric and greater than zero. Required. |
| Distribution | The probability distribution assumed for this component (e.g., Norm@99%, Rectangular, Triangular). ProCalX offers a preconfigured lookup list, and each distribution has a matching Divisor. |
| Divisor | The value used to convert the stated value into a standard uncertainty, based on the selected Distribution (e.g., 2.58σ for Norm@99%). This is populated automatically from the Distribution selection. |
| Sensitivity | The sensitivity coefficient — a multiplier reflecting how strongly this component influences the final measurement result. Defaults to 1 when the component affects the result one-to-one. |
| Std Unc, u(xi) | The standard uncertainty for this component, auto-calculated as Value ÷ Divisor × Sensitivity. This read-only field updates automatically as you edit the row. |
| Unit | The engineering unit that the Value is expressed in (e.g., psi, mA, °C). Selected from the Asset's configured unit list. |
Once components are defined, ProCalX can automatically pull them into a calibration result for the corresponding Unit (and for any assigned Test Standards), or you can select components manually at the time of calibration.
Type A vs. Type V Evaluation Methods
The GUM (Guide to the Expression of Uncertainty in Measurement) classifies every uncertainty contribution as Type A or Type B. This distinction describes how the component was evaluated — not how large, important, or reliable it is. A Type B contribution (e.g., from a properly issued calibration certificate) can be just as rigorous as a Type A contribution.
Type A
Evaluated using statistical analysis of a series of repeated, independent observations under the same measurement conditions. The standard uncertainty is typically the experimental standard deviation of the mean of those readings. Example: repeatability determined from 10 repeated measurements of the same test point.
Type B
Evaluated by any means other than direct statistical analysis of repeated readings — such as a reference standard's calibration certificate, a manufacturer's published specification, prior measurement history, or professional judgment applied to an assumed distribution. Example: the stated accuracy specification from a Fluke reference standard's certificate.
Once each component has been converted to a standard uncertainty (see Distribution and Divisor below), Type A and Type B values are combined identically — typically by root-sum-of-squares — into a combined standard uncertainty for the measurement.
Distribution and Divisor
Every uncertainty component is assumed to follow a probability distribution that describes how likely different values are within its stated range. The Distribution you select determines the Divisor ProCalX uses to convert the raw Value into a standard uncertainty.
| Distribution | Typical Use Case | Divisor |
| Norm@99% | Certificate-stated uncertainty with a 99% confidence coverage factor (common for reference standard certificates). | 2.58σ |
| Normal, 1s | The input value already represents one standard deviation directly — e.g., a Type A standard deviation calculated from repeated readings, or a source explicitly stated as "1‑sigma" (~68% confidence). | 1σ |
| Normal, 2s | When a source states its uncertainty as a normal distribution at roughly 95% confidence (k=2) — common for calibration certificates or engineering tolerances that cite "2‑sigma." | 2σ |
| Normal, 3s | When a source states a "worst‑case" or ~99.7% confidence normal limit (k=3) — common for manufacturer specs or process‑capability limits quoted at "3‑sigma." | 3σ |
| Quadratic | A component whose likelihood is highest at the center of its range and tapers off gradually toward the limits (a parabolic/Simpson‑type distribution) — less extreme-weighted than Rectangular but not as centrally‑peaked as Triangular. Occasionally used for resolution or linearity effects that don't fit a pure rectangular or triangular assumption. | √5 |
| Rect x 2 | If the stated tolerance or spec represents the full peak‑to‑peak range rather than a ± half‑width — i.e., the same rectangular assumption as "Rectangular" below, but applied to a full-range value instead of a half-range value. | √12 |
| Rectangular | A bounded quantity where any value within the limits is equally likely (e.g., resolution, a specification limit with no further information). | √3 |
| Triangular | A bounded quantity where values near the center of the range are more likely than the extremes. | √6 |
| U-Shaped | A quantity more likely to be near the extremes of its range than the center (e.g., sinusoidal drift). | √2 |
The Divisor field is populated automatically based on the Distribution you choose, so you do not need to look up or calculate it manually. If your organization uses a distribution not listed, contact your ProCalX administrator about extending the configured lookup list.
Uncertainty Rules
Below the Uncertainty Components grid, two settings control how uncertainty is enforced and interpreted for the Asset.
Require Uncertainty For All Test Activities
This toggle, found under Uncertainty Ruled, determines whether every test activity performed on this Asset must include an uncertainty calculation before the result can be saved as complete.
- On: When enabled, ProCalX defaults the per-test-activity Uncertainty Calculation mode to "Manual Entry" during result entry, since a mode must be selected.
- Off: When disabled, technicians can optionally choose "None" for a given test activity, and uncertainty is not required.
Decision Rules for Uncertainty
This dropdown, found under Decision Rules for Uncertainty, controls whether — and how — an out-of-spec uncertainty band affects the Pass/Fail outcome of a test activity. Three options are available:
| Option Name | Behavior |
| Uncertainty for Reference Only | Uncertainty is calculated and displayed on the result and certificate, but it has no effect on Pass/Fail logic. All Pass/Fail determination is ignored for the uncertainty band — the test result stands strictly on the measured reading versus tolerance. |
| Auto-Set Indeterminate when Uncertainty Band out of Spec | If the calculated uncertainty band extends outside the test specification's tolerance limits, ProCalX automatically sets the test activity's outcome to Indeterminate, flagging it for review rather than a firm Pass or Fail. |
| Auto-Set Fail when Uncertainty Band Out of Spec | If the calculated uncertainty band extends outside the tolerance limits, ProCalX automatically fails the test activity, treating an out-of-spec uncertainty band as equivalent to an out-of-tolerance reading. |