Presets
Parameters
Live readouts
Model check
The science
Why glass is weak
Glass is brittle: it fails when a surface flaw of half-length a reaches the critical stress intensity, K = Y σ √(π a) = KIc, equivalent to Griffith's energy balance G = Y² π a σ² / E = Gc. The fracture stress therefore scales as σf = KIc / (Y √(π a)): a 1 um flaw allows over 1 GPa, a 200 um scratch only about 30 MPa. The ideal bond strength is roughly E/10, which no real window approaches.
Tempered glass keeps its surface in compression (set here as a constant offset): a flaw only opens when the applied tension exceeds that compression plus the Griffith stress. Near the tip the elastic field goes as σ ~ K / √(2π r), and an elliptical hole raises stress by 1 + 2a/b (Inglis).
What this view adds
The strength-versus-flaw-size curve on log axes with the current flaw marked, a tip stress-field inset, and a batch of 25 plates with random flaw sizes showing why glass strength scatters.
Simplifications
Mode I, linear elastic, one flaw per plate (no Weibull weakest-link over many flaws), no slow crack growth from moisture, residual stress treated as a uniform offset, handbook KIc and E rounded. The crack speed uses a fraction 0.6 of the Rayleigh speed and the displayed motion is slowed 100,000 times; branching is a cosmetic hint, not a computed pattern.