Quality & InspectionFree calculatorPDF, Excel, Word export

Test Accuracy Ratio (TAR) and Test Uncertainty Ratio (TUR) Calculator

Before you calibrate an instrument you need to know your standard is good enough. TAR and TUR give that answer as a single ratio, and the traditional target is 4:1. This calculator computes both and shows what it means for your acceptance limit.

Your numbers

same unit

Half of the total tolerance band of the instrument being calibrated.

same unit

Published accuracy specification of the standard at this point.

same unit

Uncertainty of the whole calibration measurement, from your budget.

Your results

Test accuracy ratio (TAR)
5
Tolerance of the UUT divided by accuracy of the standard.
Test uncertainty ratio (TUR)
6.25
Tolerance of the UUT divided by expanded uncertainty.
Meets 4:1 on TUR (1 = yes, 0 = no)
1
Largest uncertainty allowed for 4:1
0.0125
Guard banded acceptance limit (tolerance minus U)
0.042
Simple guard band: accept only readings inside this limit.
Acceptance limit as share of tolerance
84%

The 4:1 ratio is a rule of thumb. ANSI/NCSL Z540.3 requires a documented decision rule, often a TUR of at least 4:1 or a probability of false accept of 2% or less. Follow the rule in your quality system and customer contract.

Save results:

TUR vs calibration uncertainty

Sweeps Expanded uncertainty of calibration process (k=2) from 0.002 to 0.03, holding your other inputs fixed.

Build decision rules into calibration workflows

Interested in having ratios and guard bands checked automatically against your instrument master? We can scope it.

Optional. The tool and every export work without this. No spam.

How to use it

Enter the plus-or-minus tolerance of the instrument you are calibrating, the plus-or-minus accuracy specification of the reference standard, and the expanded uncertainty of the full calibration process from your uncertainty budget. The tool returns TAR, TUR, whether you meet 4:1, the biggest uncertainty you can carry and a guard banded acceptance limit. Use the same unit throughout, for example volts, millimeters or degrees C, and run it for each test point if the tolerance changes over the range.

TAR versus TUR

TAR looks only at the published accuracy of the standard, so it ignores real-world error sources such as environment, resolution and repeatability. TUR uses the expanded uncertainty of the whole calibration, which is the better measure and the one ANSI/NCSL Z540.3 builds on. A calibration can show a comfortable TAR and still fail TUR once all the uncertainty contributors are counted. When the ratio is low, you either improve the standard, reduce uncertainty, or apply a guard band to cut the risk of false acceptance.

  • TAR: tolerance over standard accuracy
  • TUR: tolerance over expanded uncertainty
  • 4:1 is the traditional minimum
  • Below 4:1: guard band or better standard

Guard banding and common mistakes

A guard band shrinks the acceptance limit so that a reading near the edge is not passed when the true value might be outside tolerance. The simple form used here subtracts the expanded uncertainty from the tolerance, similar to the approach described in ILAC G8. Mistakes to avoid: using the full tolerance band instead of plus-or-minus, mixing units, quoting the standard's accuracy where the uncertainty belongs, and forgetting that the ratio must hold at every test point, not only full scale.

Frequently Asked Questions

What is the difference between TAR and TUR?

TAR divides the tolerance of the instrument under test by the accuracy of the reference standard. TUR divides it by the expanded uncertainty of the whole calibration process. TUR is more realistic because it includes environment, resolution and repeatability.

Is 4:1 still required?

It remains the common rule of thumb and appears in many procedures. ANSI/NCSL Z540.3 instead requires a documented decision rule, with a 4:1 TUR or a probability of false accept of 2% or less. Your own contract and quality system decide.

What do I do if TUR is below 4:1?

Options are a more accurate standard, lowering uncertainty by controlling the environment or averaging readings, or applying a guard band. Document the choice and the resulting false accept risk.

Should I use plus or minus values or the full band?

Use plus-or-minus values on both sides. If the instrument tolerance is quoted as a total band, halve it. The standard's accuracy and the uncertainty must also be expressed as plus-or-minus half-widths.

If your calibration procedures and decision rules live in an ERP or QMS, Netray can help embed these checks into the workflow.