Glass Cutting Optimization: Key Differences from Wood

Glass cutting optimization uses the same rectangular bin-packing math as wood panel cutting: you arrange every part onto standard stock sheets so the total waste and the number of sheets are as low as possible. The differences are physical, not mathematical. You score and snap glass instead of sawing it, so there is almost no kerf loss, every cut has to run edge to edge, and a bad break destroys the piece instead of leaving a usable offcut. Get three settings right (kerf near zero, guillotine-only, rotation locked for coated stock) and a woodworking optimizer will lay out glass just as well as plywood.

This guide covers what actually changes when you switch a cut-list workflow from timber to glass, with real sheet sizes, a worked shower-enclosure example, and the mistakes that cost people money and broken panes.

How glass cutting differs from wood

If you have optimized cut lists for plywood, MDF, or melamine, the concept transfers directly: fit as many parts as possible onto each stock sheet with minimal leftover. Five physical properties of glass change how you configure that process.

There is almost no kerf. This is the biggest difference. Wood is sawed, and a spinning blade removes a 3-4mm strip of material on every pass. Glass is scored: a carbide or diamond wheel scratches a shallow fissure into the surface, and you snap the sheet along that line. No material is removed. The score itself is under 0.5mm wide, so kerf loss is effectively nil. In your optimizer, set kerf to 0mm, or 1mm as a conservative safety margin. If you are used to dialing in blade width for a table saw, our saw kerf explainer shows why that 3-4mm value matters so much for wood and why glass throws it out entirely.

Every cut is a guillotine cut. A score line has to travel from one edge of the piece to the opposite edge. You cannot start a score in the middle of a sheet and stop, because glass will not fracture cleanly along a partial score. Each snap divides one rectangle into exactly two smaller rectangles, exactly like a paper guillotine. That makes guillotine mode mandatory rather than a preference. A free-nesting or true-shape layout that looks efficient on screen is simply not cuttable in glass, so any optimizer setting that allows non-guillotine placement has to be turned off.

Coated glass cannot be rotated freely. Many modern glass products carry a functional coating on one face. Low-E (low emissivity) glass has a near-invisible metallic layer that reflects infrared heat, and it must face a specific direction in the finished unit. Tinted glass carries a colored layer, mirror carries a silvered backing, and patterned or fluted glass has a directional texture. In all of these, a part cannot be rotated 90 degrees during layout without flipping the coating or breaking the pattern run. Disable rotation for coated, mirrored, or patterned stock, and only allow it for plain clear float.

Stock sheets are large. A standard float glass sheet (a "jumbo" is larger still) runs 3210 x 2440mm, well beyond a 2440 x 1220mm plywood sheet. More parts fit per sheet and the algorithm has more room to pack tightly, so glass waste percentages tend to run lower than wood for the same job. It also means one broken sheet is expensive, which raises the stakes on getting the layout right the first time.

Glass does not forgive. A bad snap on plywood leaves an undersized but usable offcut. A bad snap on glass often runs off the score and shatters the whole piece into shards. There is no re-cut. This is why cut sequence matters more than it does in wood, and why remnants below roughly 100mm are treated as scrap rather than stock: they are too small to grip and snap safely.

Standard glass sheet sizes and settings

Your stock dimensions are the starting point for any layout. The table below lists the sizes most suppliers stock, along with the optimizer settings that suit each. For how these compare with timber and board stock, see the standard sheet sizes guide.

*Laminated float can be rotated if it is clear on both faces, but any laminate built with a coated or tinted ply follows the coated rule and stays locked.

Setting up glass in your optimizer

Configuring a cut-list optimizer for glass means changing a few defaults away from their wood values:

• Kerf: 0mm for pure scoring, or 1mm as a break-safety margin. Never carry over the 3-4mm saw kerf.
• Guillotine mode: On, always. If the tool offers "guillotine only" or "panel saw" mode, enable it and confirm no part placement is nested.
• Rotation: Off for coated, mirrored, tinted, or patterned glass. On for plain clear float, where it buys the layout real flexibility.
• Edge trim: Add 5-10mm if the stock sheet arrives with chipped, ground, or arris edges, so the optimizer keeps parts off the damaged margin.
• Minimum offcut size: Set the reusable-remnant floor to about 100mm in both dimensions. Anything smaller is unsafe to handle and should count as waste, not stock.

CutPlan supports every one of these controls: set kerf per material, switch on guillotine mode, and toggle rotation for each glass type independently. You can open the optimizer and enter a glass job in a couple of minutes, no install required.

A worked example: batching two shower enclosures

Numbers make the payoff concrete. Say you are cutting two custom shower jobs from 6mm clear float (cut now, tempered later) on 3210 x 2440mm sheets.

• Job A: one fixed panel 800 x 2000mm and one door 700 x 1950mm.
• Job B: one panel 600 x 2000mm and two returns 450 x 2000mm.

Cut each job on its own and neither fills much of a jumbo sheet. Job A uses about 38 percent of the sheet, Job B roughly the same, and each leaves a big awkward remnant: that is two sheets ordered, each under 40 percent used. Now feed all five parts into one optimization and allow rotation (clear float, no coating). They drop onto a single 3210 x 2440mm sheet, stood upright in one 2000mm band, because the widths 800 + 700 + 600 + 450 + 450 add up to 3000mm and clear the sheet's 3210mm width. One long rip down the sheet plus a handful of crosscuts frees every part, and every step is a clean guillotine cut. Utilization jumps to about 76 percent and you order one sheet instead of two. On glass at 25 to 60 dollars per square meter, that saved sheet is real money, and it is the same batching logic covered in minimizing sheet-material waste: combine jobs of the same thickness and type before you optimize, not after.

Cut sequence and edge quality

A good layout is only half the job. In glass, the order you make the cuts determines whether the layout survives contact with reality.

Work from the whole sheet down: score and snap the longest full-width cuts first to reduce the big sheet into manageable strips, then break those strips into parts. This keeps you handling a large, heavy, fragile sheet as briefly as possible and means every later snap is on a piece you can control with both hands. Trying to nibble small parts off a full jumbo sheet is how sheets get dropped and cracked.

Keep the score fresh, too. Glass wants to be snapped within a minute or two of scoring, before the fissure heals microscopically and starts to wander off line. That is a practical reason to prefer a layout with fewer, longer cuts over one with many short fussy ones, even at a slight cost in waste. When you do generate reusable remnants, log them so they feed the next job rather than piling up in a corner: the same discipline in the offcut management guide applies to glass, with the extra rule that anything under about 100mm gets recycled, not stored.

Common mistakes with glass optimization

• Leaving the wood kerf in. The number-one error. 3mm of phantom kerf per cut inflates the layout and can add a sheet. Set it to 0-1mm.
• Trying to cut tempered glass. Toughened glass explodes into thousands of pebbles if scored. All sizing happens before tempering, never after. If a client changes a dimension after toughening, that panel is scrap.
• Rotating coated stock. Letting the optimizer flip a Low-E or mirror part to save 2 percent waste puts the coating on the wrong face. Lock rotation for anything coated.
• Scoring the wrong side of mirror. Always score the clear glass face, not the silvered back. Scoring the coating tears the silver and ruins the reflection at the edge.
• Ignoring edge condition. Placing a part flush to a chipped stock edge means the chip lives in your finished panel. Add edge trim.
• Optimizing after ordering. Buying a fixed count of sheets and then running the layout is backwards. Optimize first so you know the exact sheet count, then order. Otherwise you over-buy, or you discover the parts do not fit the sheets you already paid for.

Get those six right and the layout you see on screen is the layout you can actually cut. For the wider set of concepts that carry across every sheet material, see the complete guide to cut list optimization.

Frequently asked questions