Why kerf matters more than the cut

Why the width of the removed material decides the finished dimension.

Kerf matters more than the cut itself because every cutting process, whether a saw, a laser or a waterjet, removes a strip of material along the cut line rather than simply separating the workpiece along an infinitely thin boundary, and the width of that removed strip decides how big the two resulting pieces actually turn out to be, a fact that has nothing to do with how carefully the cut line was drawn and everything to do with how wide the tool doing the cutting happens to be.

The short version

Draw a single line down the middle of a board and it is tempting to imagine that cutting along it produces two pieces that each add back up to exactly the original board. That is only true if the cutting tool has no width of its own, and no real cutting tool manages that. A saw blade, a laser beam, a waterjet stream and even a pair of scissors all remove a strip of material equal to their own width as they pass through the workpiece, and that strip does not belong to either resulting piece, it is simply gone, turned into sawdust, vapour, or eroded slurry and swept away. Two pieces cut from one board therefore never quite add back up to the board's original size, they fall short of it by exactly the width of whatever removed the material between them.

The saw-blade comparison

A carpenter marking the exact centre of a shelf board and cutting there with a handsaw does not end up with two pieces each exactly half the board's length, because the saw blade itself has real thickness, typically a couple of millimetres, and that sliver of wood along the cut line is converted entirely into sawdust rather than staying attached to either half. Cut a dozen boards this way, each meant to be divided into equal pieces, and every single one comes up short by the same small amount, invisible on any individual board but adding up steadily across a whole batch of shelving if nobody accounted for it from the start, and glaringly obvious the moment two shelves cut on different days, by different saws with different blade thicknesses, are stood side by side. A cabinetmaker who has learned this lesson marks the cut line slightly off centre on purpose, deliberately giving up a hair's width to the blade so the two finished pieces land exactly where they were actually meant to.

Why the finished dimension depends on the tool, not just the line

Because every process removes a strip of its own characteristic width, kerf is not a fixed constant that applies the same way to every job, it is a property of the specific tool and settings being used, a fine wire saw removing far less material per cut than a thick abrasive disc, a tightly focused laser removing far less than a wider waterjet stream. This means a drawing specifying only where the cut line should sit, without also specifying or accounting for the process that will actually make that cut, leaves the finished dimension genuinely undetermined, since the same line cut by two different processes can produce two pieces of noticeably different final size, both technically cut exactly on the line the drawing called for.

The one number worth remembering

A laser beam typically removes a strip only a few tenths of a millimetre wide as it cuts, while an abrasive waterjet stream commonly removes a strip several times wider than that, and a mechanical saw blade wider again, differences that sound trivial until a part with several internal cuts, each one shrinking the surrounding material by its own process's kerf, starts accumulating an error that can be many times larger than the tolerance the finished part was actually meant to hold. Switching a job from a laser to a waterjet partway through a project, perhaps because the material changed to something reflective, can therefore change the finished dimensions of every part on the drawing even though nothing about the drawing itself was ever touched.

Where kerf compounds across a nested part

A single cut losing a fraction of a millimetre is easy to dismiss, but very few real parts involve only one cut, and kerf's effect on the finished dimension scales with how many cut edges actually border the feature being measured. A rectangular hole cut into the middle of a panel loses kerf on all four sides at once, shrinking the surrounding material and enlarging the hole itself simultaneously, while an external tab or slot loses kerf on each edge that defines it. A part with several closely spaced slots can find that the material remaining between them has shrunk by several kerf widths in total, occasionally enough on its own to weaken a feature that looked perfectly adequate on the original drawing, before anyone had subtracted out what the cutting process itself was actually going to remove.

What this changes in practice

Once kerf is understood as material genuinely lost rather than a rounding error, a competent drawing either specifies the cutting process explicitly or builds in a kerf allowance that lets the finished part come out correct regardless of which machine ends up making it, offsetting the cut path outward by half the kerf width on the piece being kept and inward by the same amount on the piece being discarded. Software driving a laser or waterjet cutter routinely applies exactly this offset automatically once the kerf for a given material and process is known, quietly correcting for the fact that the cut line drawn on screen and the cut line the machine actually follows are never quite the same line at all, and a designer who understands why that correction exists is far less likely to be caught out by a batch of parts that are all, mysteriously, the same small amount too small.