Calculator

Weibull analysis calculator (β, η, B-life)

Yield, quality and reliability. Everything is calculated in your browser; nothing you enter is sent to a server.

KEY POINTS

The Weibull shape parameter β describes the nature of failures: β < 1 means infant mortality, β ≈ 1 random failures and β > 1 wear-out, matching the three regions of the bathtub curve. The scale parameter η is the time by which 63.2% of units have failed. In semiconductors it is used for lifetime analysis of TDDB, electromigration and solder joint fatigue.

Estimate Weibull parameters from failure times

Shape parameter β2.730
Scale parameter η (time at which 63.2% have failed)1,622
Failure typeWear-out (failure rate rises over time)
Coefficient of determination R²0.9973
Reliability R(t) at time 50096.056%
Instantaneous failure rate at time 5002.20×10⁻⁴per unit time
B10 life711.3
Mean time to failure MTTF1,443

Median rank regression assuming complete data, with every unit tested to failure. If some units were removed before failing (censored data), use Johnson's adjusted ranks or maximum likelihood instead.

Weibull probability plot

-3-2-1016.577.58ln(time)ln(−ln(1−F))
Data (median ranks)Regression line

B-life summary

Cumulative failuresTime
B1300.8
B5546.5
B10711.3
B501,418
B63.21,622

Reading the shape parameter β

The Weibull distribution describes the nature of failures with a single number, the shape parameter β. β < 1 means infant mortality, where the failure rate falls over time as manufacturing defects show up; β ≈ 1 means random failures at a constant rate; β > 1 means wear-out, where the failure rate rises over time. These map directly onto the three regions of the bathtub curve.

In semiconductors, lifetime data for gate oxide breakdown (TDDB), electromigration and thermal fatigue of solder joints are plotted on Weibull paper. The β of TDDB depends on oxide thickness: thinner oxides have smaller β and a wider spread of failure times, which makes lifetime guarantees for ultrathin films difficult.

FAQ


What is B10 life?

The time by which 10% of units have failed. It is calculated as η·(−ln 0.9)^(1/β).

Can this tool handle censored data?

No. It assumes every unit was tested to failure. Data with units removed before failure need Johnson's adjusted ranks or maximum likelihood estimation.

Related equipment and companies


More calculators


Yield, quality and reliability

Last updated: October 23, 2026