Automatically count significant figures, apply rounding, and calculate error propagation (add, subtract, multiply, divide, power, sqrt) with step-by-step display. Visualize uncertainty range on a number line.
Mode
Number or expression (e.g. 3.45 * 2.1 + 0.005)
Operators: + − * / ^(power)
Parentheses () supported. Numbers parsed with sig fig rules.
Value to round
Significant figures
Rounding mode
Measurement a ± Δa
±
Operator
Measurement b ± Δb
±
Combination method
Results
—
Result
—
Sig Figs
—
Absolute Uncertainty
—
Relative Uncertainty (%)
Step-by-Step Explanation
Enter values in the left panel and press Calculate.
Number Line (Uncertainty Visualization)
Nl
Run an uncertainty calculation to display
CAE Note
Used for elastic modulus estimation in material testing (uncertainty of Δσ/Δε), load cell accuracy evaluation, statistical analysis of dimensional tolerances, and FEM result validation. Compliant with GUM (Guide to the Expression of Uncertainty in Measurement).
Theory & Key Formulas
Error Propagation Theory
For independent measurements $a \pm \Delta a$ and $b \pm \Delta b$:
General form (Gaussian error propagation): $$\Delta f = \sqrt{\left(\frac{\partial f}{\partial a}\Delta a\right)^2 + \left(\frac{\partial f}{\partial b}\Delta b\right)^2}$$
What is Measurement Uncertainty?
🙋
What exactly is the difference between significant figures and measurement uncertainty? They both seem to be about how precise a number is.
🎓
Great question! Basically, significant figures are a quick, simplified rule of thumb for precision, while measurement uncertainty is a rigorous, quantitative statement. For instance, if you measure a length as 12.3 cm, the three significant figures imply it's roughly between 12.25 and 12.35 cm. But a formal uncertainty, like $12.3 \pm 0.1$ cm, explicitly defines that range. Try selecting "Addition" in the simulator's operation menu—you'll see how the rules for combining numbers differ for each method.
🙋
Wait, really? The rules are different? So if I add 12.11 and 0.3, what happens?
🎓
Exactly! This is a classic pitfall. By significant figures, you might think 12.11 (4 sig figs) + 0.3 (1 sig fig) = 12.41. But the rule for addition/subtraction is based on decimal places, not total figures. Since 0.3 has only one decimal place, the result must be rounded to one decimal place: 12.4. The simulator will show you this step-by-step. Now, if those numbers had uncertainties, like $12.11 \pm 0.01$ and $0.3 \pm 0.1$, the combined uncertainty is calculated totally differently!
🙋
Okay, that makes sense for adding. But what about multiplying? And what's that square root formula I see in the tool's theory section?
🎓
Ah, you're looking at error propagation! When you multiply or divide, uncertainties combine relative to the size of the numbers. That square root formula, $\frac{\Delta(a \cdot b)}{|a \cdot b|}= \sqrt{\left(\frac{\Delta a}{a}\right)^2 + \left(\frac{\Delta b}{b}\right)^2}$, gives the fractional (or percent) uncertainty in the product. It's more complex than just counting significant figures. A common case is calculating area from length and width. Change the simulator's operation to "Multiplication" and input some values with uncertainties to see it in action.
Physical Model & Key Equations
For addition or subtraction of independent measurements, the absolute uncertainty in the result is found by combining the individual absolute uncertainties in quadrature (square root of sum of squares). This assumes the errors are random and independent.
$$\Delta(a \pm b) = \sqrt{(\Delta a)^2 + (\Delta b)^2}$$
Where $\Delta a$ and $\Delta b$ are the absolute uncertainties in measurements $a$ and $b$. This method typically gives a more realistic (and smaller) uncertainty than simply adding them ($\Delta a + \Delta b$), which is a "worst-case" bound.
For multiplication or division of independent measurements, it is the fractional (or relative) uncertainties that combine in quadrature. The result is the fractional uncertainty of the product or quotient.
Here, $\frac{\Delta a}{a}$ is the fractional uncertainty in $a$. This equation tells you that when numbers are multiplied, the percentage error propagates, not the absolute error.
Real-World Applications
Material Testing (Elastic Modulus): The Young's modulus is calculated as stress divided by strain ($E = \sigma / \epsilon$). Both stress (from force and area measurements) and strain (from displacement and gauge length) have uncertainties. Using the division rule from this simulator is essential to report a reliable $E \pm \Delta E$ for engineering design.
Load Cell & Sensor Calibration: Evaluating the accuracy of a force sensor involves comparing its output to a known standard. The combined uncertainty of the measurement system, calculated through addition/subtraction of error components, determines the sensor's final rated accuracy and compliance with standards.
Statistical Analysis of Dimensional Tolerances: When assembling parts, the final clearance is the sum or difference of individual part dimensions. Using quadrature addition of tolerances (as per the uncertainty model) provides a more statistically likely assembly tolerance than the extreme worst-case stack-up, enabling smarter and more cost-effective manufacturing.
FEM (Finite Element Method) Result Validation: Computer simulations predict stresses and displacements. Engineers compare these to physical test data. Properly accounting for the uncertainty in both the test measurement and the simulation inputs (like material properties) is crucial to determine if the model is "validated" within an acceptable error margin.
Common Misconceptions and Points to Note
A common initial stumbling block in these calculations is confusing the "number of decimal places in the uncertainty" with the "number of significant figures". For example, suppose you get a result of 12.345 ± 0.12. The uncertainty 0.12 is meaningful to the second decimal place, so the main value 12.345 must also be rounded to the second decimal place, i.e., to 12.34 or 12.35. It's a mistake to report it as 12.345 just because the main value appears to have "5 digits".
Next, pay attention to the handling of "zeros". For measured values like 1.0 and 0.5, they seem to have 2 and 1 significant figures respectively, and the "5" in 0.5 is indeed a significant digit. However, in a value like 0.002, the leading zeros before the "2" are not counted as significant figures. This tool automatically excludes leading zeros and counts them correctly, so you can use it with confidence.
Finally, it's crucial to fundamentally understand that the error propagation formula is "not a universal solution". That root-sum-square formula is an approximation that holds when the errors are independent, random, and relatively small. For instance, when dealing with two correlated quantities (like temperature and length linked by thermal expansion) or when the uncertainties are very large, this simple formula becomes inadequate. First, master the basics with this tool, and also be aware of its limitations.