
- A thickness is the output of a calculation, not an input: the design pressure, the load duration, the pane geometry, the glass type, the edge condition and the support condition all have to be stated before any chart or formula means anything.
- Glass has two separate acceptance criteria - resistance to breakage under the ultimate wind action, and deflection under serviceability conditions - and a pane that satisfies the first can still pump the perimeter seal, dish visibly in reflection or lose gasket contact at mid-span.
- ASTM E1300 works from a probability of breakage at a defined load duration, while EN 16612 works from a characteristic strength modified by load duration, surface and edge condition; the two philosophies differ, but both fail immediately if the schedule omits edge work and duration.
- Load sharing inside a sealed insulating glass unit is real under a three-second gust and largely disappears under sustained load, so a cold-side pane sized on a 50/50 assumption is the usual explanation for a breakage that gets reported as spontaneous.
- Laminate interlayer stiffness is time and temperature dependent: a unit sized on short-duration wind can be under-designed for a sustained snow pocket, and post-breakage retention is a different calculation from either.
- Edge damage during handling and glazing removes more strength than an extra millimetre of thickness adds, which is why rejection criteria at the point of glazing matter as much as the arithmetic.
- The frame and the glass must be checked against consistent criteria; a flexible frame paired with a thin pane converts wind load into sealant strain and water ingress long before either element reaches its limit.
1. The Pane Is the First Element in the Load Path, Not the Last
Wind pressure acts on the smallest and most flexible element in the facade, and that element is the glass. The load travels from the pane into the glazing rebate, from the rebate into the frame, from the frame into the mullions and anchors, and from there into the structure. Every one of those steps has to be checked, but in practice only one of them is normally written down as a single number. The glass schedule says eight millimetres, or six plus twelve plus six, and the reasoning that produced that figure is not recorded anywhere the next project can find it.
That is the origin of most glass sizing errors. A thickness that was correct for one opening is carried to another with a longer span, a higher design pressure, a corner zone of the same facade, or a different frame with a different cover depth, and nobody re-runs the calculation because the number looks familiar. The thickness is a conclusion. Without the inputs that produced it, it is a guess with a decimal point.
Two separate criteria sit behind the conclusion, and they are frequently confused. The first is resistance to breakage under the ultimate wind action, which is a strength question. The second is deflection under service conditions, which is not a strength question at all but a consequence question: a pane that dishes several millimetres at mid-span will visibly distort reflections in a sealed unit, will work the perimeter gasket, and on a large opening may rotate far enough to lose contact with the bead or the cover. A pane can be perfectly safe and still be the wrong pane, and a schedule that records only thickness has no way to express the difference.
2. The Design Pressure That Actually Enters the Calculation
Both the American and the European routes begin with a design wind pressure, and neither of them produces it. That number comes from the structural engineer, from the applicable loading code - EN 1991-1-4 in Europe, ASCE 7 in North America - and it is not a single figure for a building.
The pressure varies with height, with terrain and exposure category, with the size of the loaded area, and above all with position on the elevation. The corner and edge zones of a facade see considerably higher local coefficients than the field of the wall, and a parapet or a projecting canopy changes the local flow again. A schedule that quotes one pressure for the whole project is a simplification that is safe only where the glazing is genuinely in the field zone. Panes at the corner of a tall building are not.
The second trap is the factor convention. The glass charts and the design-value methods are built around different load states, and the two cannot be mixed without double counting. The American chart procedure works from an unfactored nominal wind pressure applied for a short duration; the European partial-factor route compares a design action against a design resistance. Entering a factored load into a chart that already embeds its own probability basis inflates the answer, and entering an unfactored load into a method that expects a design value deflates it. The schedule should state which convention produced the pressure it quotes.
It is also worth separating test evidence from design input. A product classification to EN 12210, or a laboratory report to ASTM E330, describes what a specimen did at a stated test pressure in a stated configuration. It is valuable evidence about the assembly, and it is routinely cited in specifications as though it sized the glass. It does not. The design pressure is a project action; the test pressure is a specimen result; the glass calculation consumes the former.
3. What ASTM E1300 Does and What It Leaves Out
ASTM E1300 is a practice for determining the load resistance of glass, and it offers two routes that should agree: published charts for common configurations and an analytical procedure for everything the charts do not cover. Either route needs the same set of inputs, and the list is longer than most schedules provide.
- Glass type - annealed, heat-strengthened or fully tempered, since the three have different resistance and, more importantly, different breakage behaviour.
- Nominal thickness and make-up - monolithic, laminated, or a sealed unit, with the individual ply thicknesses rather than the overall unit thickness.
- Load duration - the practice tabulates resistance against short-duration, sixty-second and long-duration categories, and glass is stronger under a gust than under a load that persists.
- Pane geometry - the short and long dimensions, because a narrow pane spans one way and its resistance is governed by the short dimension regardless of how long the opening is.
- Support and edge condition - how the pane is held and how its edges were finished, which is the input most often missing from a schedule.
The charts correspond to a defined probability of breakage. That is adequate for ordinary vertical glazing in a low-consequence location, and it is not adequate as an unexamined default for overhead glazing, for a canopy over an entrance, or for a pane whose failure would fall into an occupied space. Where a lower probability is required, the correct response is the analytical procedure with the target stated, not a thicker pane chosen until the number looks comfortable.
What E1300 does not do is equally important. It does not check thermal stress from shading, blind boxes or dark frames; it does not design point-supported or structurally bonded glass; it does not address blast, seismic racking or fire; and it says nothing about post-breakage behaviour. Those are separate checks, and a project that needs them needs them in addition, not instead.
4. EN 16612 and the Design-Value Route
The European calculation method approaches the same problem from the other direction. Rather than reading a chart that embeds a probability, EN 16612 derives a design resistance: a characteristic bending strength, established by test to the EN 1288 series, reduced by factors that account for how long the load acts, for the condition of the glass surface and its edges, and for the way an assembly of plies or of two panes in a sealed unit shares the load. That design resistance is then compared with the design value of the applied action.
The practical consequence is that the method forces the designer to describe things that a chart lets them ignore. Load duration has to be decided, not assumed: a wind gust, a snow pocket and a barrier load are three different durations on the same pane. Edge work has to be specified, because a polished edge and an arrised edge do not carry the same reduction, and a chipped edge carries a much larger one. Assembly has to be declared, because the load-share factor depends on whether the pane is monolithic, laminated, or one lite of a sealed unit.
Two suppliers can therefore produce different answers for the same opening without either of them making an arithmetic mistake, simply because one assumed a ground edge and a three-second gust while the other assumed an arrised edge and a sustained load. That is not a reason to distrust the method; it is a reason to put the assumptions in the specification so that the answers become comparable. The same discipline applies to the American route, where the duration category and the edge condition enter the chart reading in exactly the same way.
5. Load Sharing Inside the Sealed Cavity
An insulating glass unit is two panes with a spring between them, and the spring is the gas in the cavity. When wind arrives as a gust, the pressure on the outer pane rises faster than gas can leave the cavity, so the cavity pressure rises with it and pushes back on the inner pane. For a few seconds the two panes genuinely share the load, and the unit behaves as an assembly rather than as a single pane.
That behaviour is duration-dependent, and this is where schedules go wrong. Under a sustained load the cavity pressure equalises, by gas permeation through the seal and by temperature equilibration, and the load returns almost entirely to the loaded pane. A snow pocket on a low-slope roof light, a permanent negative pressure on a sealed building, or a long-duration load on a large pane in a cold climate are not shared in the way a gust is shared. Applying a fifty-fifty assumption to all of them puts the whole load on a pane that was sized for half of it.
There is a second, quieter mechanism. The cavity pressure at the time of manufacture is set at the factory altitude and temperature. A unit manufactured near sea level and installed at altitude, or a dark unit on a hot facade, starts its service life with the panes already bowed one way or the other. That pre-load does not usually break glass on its own, but it reduces the margin, it changes the visual flatness that architects notice in reflected light, and it works continuously on the secondary seal.
The remedy is procedural rather than clever. Require the load-share basis to be stated in the glazing schedule, check the cold-side pane as a separate element rather than as half of a unit, and treat any figure presented without a stated basis as a chart reading rather than a design. Where a supplier cannot say which pane was checked and under what duration, the number has not been verified.
6. Laminates: Interlayer Stiffness Is a Function of Time and Temperature
Laminated glass introduces a material whose stiffness is not a constant. Plasticised polyvinyl butyral, ionoplast interlayers and ethylene vinyl acetate sit in the same assembly but behave very differently, and all of them soften as temperature rises and creep as load persists. The interlayer that transfers shear between the plies under a wind gust is the same interlayer that flows slowly under a load held for months.
Under a short-duration gust a laminate behaves close to monolithic, and the full assembly thickness is a reasonable approximation. Under a sustained load the interlayer relaxes, the plies act largely independently, and the resistance approaches the sum of the individual plies rather than the resistance of the combined section. The gap between those two states is why a laminate sized on wind alone can be under-designed for a snow drift, a glass floor, a balustrade infill, or an overhead pane carrying its own weight across a deep span.
Post-breakage performance is a third, separate calculation. Fully tempered glass breaks into small fragments that can migrate within the interlayer; heat-strengthened glass breaks into larger, interlocking pieces that tend to stay engaged. For overhead glazing and for barriers, that difference often decides the specification more than the wind resistance does, and it cannot be read from a thickness chart at all.
The practical rule is that a wind calculation is not a barrier calculation, a floor calculation or a snow calculation, even when the same pane serves all four roles. Each has its own duration, its own load distribution and, in the barrier case, its own impact requirement.
7. Frame, Edge and the Interface That Decides the Real Answer
Glass deflects considerably more than the aluminium or uPVC frame that carries it at the same span, and that difference has to be absorbed somewhere. It is absorbed at the perimeter: in gasket compression, in bead movement, in the sealant line and in the clearance between the pane edge and the rebate floor. A thin pane in a flexible frame therefore converts wind load into movement at the very joint that keeps the water out.
Cover depth is the variable that receives least attention. The pane rotates as it dishes, and the cover has to be sufficient that the deflected surface still bears evenly on its gasket at mid-span rather than lifting off one face and over-compressing the other. The same reasoning governs setting block position, edge clearance and the drainage path behind the bead, which is why a glazing detail cannot be transferred between systems on the assumption that glass is glass.
Edge condition is the input that most often decides whether the calculation is meaningful. Cutting, arris quality, shelling at the cut line, and chips introduced during handling all reduce edge strength, and in a toughened pane an edge defect introduced after heat treatment is a stress concentrator that no thickness increment will compensate for. The standards that govern thermally toughened and heat-strengthened glass describe the edge work expected of the product; the site that ignores it during glazing discards the margin the calculation relied on. A simple rejection criterion at the point of glazing, applied before the bead goes on, is worth more than an extra millimetre of nominal thickness.
Finally, the frame and the glass must be checked against criteria of the same kind. Frame deflection limits are typically expressed as a fraction of the span and describe serviceability; the glass criterion combines a breakage limit with a deflection judgement. Comparing a frame deflection that has already been allowed against a glass deflection that has been assumed to be zero is a common way of producing a facade that performs on paper and rattles in a south-westerly.
8. What Belongs in the Glass Schedule Before Any Drawing Is Issued
The schedule is the only document that survives from design to site, and it should carry the inputs rather than only the conclusion. A schedule that lists the following can be checked by a third party; one that lists only thickness cannot.
| Field | What it fixes |
|---|---|
| Design pressure and its source | The code, edition and zone the figure came from, and whether it is factored |
| Load duration category | Gust, short duration or long duration, since resistance changes with it |
| Pane size and aspect ratio | Which dimension actually spans |
| Make-up by ply | Individual thicknesses, not the overall unit dimension |
| Glass type per ply | Annealed, heat-strengthened or fully tempered, and why |
| Edge work | The finish the strength assumption depends on |
| IGU load-share basis | Which pane was checked and under what duration |
| Interlayer type | The laminate behaviour assumed, at service temperature |
| Deflection criterion | The serviceability limit the frame and pane are both held to |
| Cover and edge clearance | What the glazing detail must provide for the assumption to hold |
Alongside the schedule, the submittal should be a calculation rather than a screenshot of a chart. It should identify the party responsible for it, state the probability basis or the partial factors used, and show the cold-side pane of every sealed unit. On delivery, a thickness reading at the edge confirms the make-up actually supplied, and a spare pane of the same make-up should be retained on site, because a replacement cut from a different build-up does not break, deflect or colour-match like the original.
None of this is exotic. It is the ordinary discipline of a facade package applied to the one component that is usually specified by habit. The number in the schedule is a residue of these decisions, and when the decisions are recorded, the number can be defended on the next project instead of being copied into it.
Planning a Commercial or Villa Project?
Send your window schedules or architectural CAD drawings to the OPPGATE engineering team for a prompt, certified quotation and structural review.
Request an Engineered Quotation
