Textile Cutting on CNC Tables: The Future of Fabrics

• 1
  How a CNC Table Removes Fabric Material
• 2
  Knife, Laser, Ultrasonic and Waterjet: What Each One Does
• 3
  Matching the Head to the Fiber and the Lay
• 4
  What Tolerances Textile Cutting Can Hold
• 5
  When Textile Cutting on a CNC Table Pays Off
• 6
  Cutting Head Selection by Fabric and Lay

A cutting table is not a milling machine with a knife bolted on. The fabric sits on a sacrificial bed, usually a bristle brush or a perforated vacuum surface. A gantry carries the cutting head across the bed in X and Y, while a controller reads a DXF vector file and drives the head along that path. The material does not spin against a cutter. It is severed by one of four energy sources: a reciprocating blade, a focused laser beam, ultrasonic vibration, or a high-pressure water jet. Each source interacts with fibers differently, and that difference decides the edge quality you get.

The vacuum bed matters more than most buyers expect. Compressed fabric shifts under the blade, so the pump holds the lay flat while the knife descends. On a 100-layer lay, a single ply that lifts by 0.5 mm produces a stepped edge on every part in the stack. Operators compensate by covering the lay with a polyethylene film before cutting. The film seals the vacuum and lets the blade pass through, which keeps the top plies locked down.

Toolpath planning is where nesting software earns its place. The controller receives vector geometry, but the nesting engine decides how those shapes sit on the roll. A good nest rotates parts, mirrors left and right hands, and fills gaps with smaller pieces. On a 1,600 mm wide roll of technical fabric, a 5% improvement in yield is worth real money across a production run. That calculation runs before the machine ever moves.

Every cut leaves a kerf, the width of material the tool destroys. A drag knife removes roughly 0.2–0.5 mm. A laser burns a wider channel and leaves a heat-affected edge. A waterjet stream, typically 0.1–0.3 mm depending on orifice and pressure, cuts cold but wets the fabric. Designers who ignore kerf get parts that are consistently undersized on the inside of a curve and fine on the outside. Compensate in the vector file, not on the shop floor.

• 1
  Vacuum holds the layPrevents ply shift and stepped edges on stacked cuts.
• 2
  Kerf must be offset0.1–0.5 mm depending on the cutting head.
• 3
  Nesting runs firstYield gains are decided before the gantry moves.

A reciprocating knife is the workhorse for woven and knitted fabric, leather, foam, and composites like carbon fiber prepreg. The blade oscillates at high frequency and shears fibers rather than melting them, so the edge stays soft and the material keeps its hand. It handles 1 to 100+ layers on a single lay. The limit is geometry: sharp internal corners under 1 mm radius need a smaller blade or a different process, because the blade body cannot turn that tightly.

A laser cuts by vaporizing material along a narrow path. It needs no vacuum hold-down for thin single plies and produces a sealed edge on synthetics, which stops fraying on nylon and polyester. That sealed edge is the reason laser dominates technical textiles and airbag-adjacent work. The tradeoff is the heat-affected zone. On aramid or glass fiber, the laser leaves a charred rim that weakens the laminate interface. On thick stacks, laser cutting is slow and the edge taper grows.

Ultrasonic cutting uses a vibrating blade at roughly 20 kHz. Friction melts the fiber locally as the blade passes, so synthetic fabric is cut and sealed in the same motion. There is almost no dust and no charred rim, which matters in cleanroom settings and medical textiles. The process is slower than a drag knife on simple shapes and the tooling is consumable. For a 20-layer polyester lay, ultrasonic gives a cleaner edge than laser with far less thermal damage.

Waterjet cuts cold with a stream of water and abrasive, or pure water for soft materials. It produces no heat-affected zone at all, which makes it the choice for aramid, glass, and pre-impregnated composites where thermal damage is unacceptable. The fabric gets wet, so drying and edge wicking become production steps. Pure-water cutting is fast on foam and felt. Abrasive waterjet handles the hard technical textiles but is slower and produces a slurry that needs handling.

• 1
  Knife for volumeBest edge hand and lowest cost per layer on woven goods.
• 2
  Laser for sealed edgesSynthetics only; watch the heat-affected zone.
• 3
  Waterjet for zero heatComposites and aramid, at the cost of drying.

Cotton and viscose take a knife cleanly. The fibers are short and the blade shears them without pulling. Laser on cotton burns the edge brown and leaves a smell that follows the roll into the next process. For apparel-grade cotton, knife cutting at 40–80 layers is the practical default. Shrinkage after washing is a separate problem, but the cut edge itself is stable.

Polyester, nylon and other thermoplastics behave differently. A knife can leave a slightly frayed edge that later sheds lint. Laser or ultrasonic melts that edge shut. If the part goes into a washing cycle or a high-vibration environment, the sealed edge is worth the tooling cost. If the part is a one-off pattern for fit testing, a drag knife on a single ply is faster and cheaper.

Aramid and glass fiber are the hard cases. Both are abrasive and both suffer from heat. A reciprocating knife dulls quickly, so blade changes climb. Laser chars the edge and can leave a conductive smear on aramid. Waterjet cuts cold and clean, which is why aerospace and ballistic textile shops default to it. The tradeoff is the wet process and the abrasive disposal, which adds handling steps a knife line does not have.

Foam, felt and nonwoven sit at the easy end. They compress under load, so the vacuum bed does most of the work. A drag knife or a straight blade cuts them at high speed with little edge damage. The main risk is compression set: if the foam is held under vacuum too long before cutting, it may not recover its thickness. Cut soon after loading, and release the vacuum as soon as the gantry clears the lay.

• 1
  Cotton and viscoseKnife at 40–80 layers; laser stains the edge.
• 2
  Polyester and nylonLaser or ultrasonic to seal against fraying.
• 3
  Aramid and glassWaterjet only; heat damages the laminate interface.

Fabric is not steel, and the tolerance story follows from that. A woven textile stretches, so the achievable tolerance depends on how well the vacuum holds it and how much the weave moves under the blade. On a single ply with a sealed vacuum bed, a knife table can hold ±0.5 mm on a 300 mm part. On a 60-layer lay, expect ±1.0 mm, because the plies shift relative to each other even under vacuum.

Laser holds tighter on thin single plies, often ±0.25 mm, because there is no mechanical contact to push the fabric. The catch is the heat-affected zone, which can add 0.2–0.5 mm of degraded material on each side of the cut. If the functional edge is the cut edge, the laser kerf must be offset and the HAZ accounted for in the design. Waterjet holds ±0.2 mm on thin material but the taper grows with thickness, so a 6 mm composite stack cuts with a slight bevel.

Repeatability is usually better than absolute accuracy. Once the vacuum, the blade, and the toolpath are fixed, a table can repeat a cut within ±0.2 mm run to run. That matters for assembly: if two parts are cut on the same setup, they will mate. If one is cut on a knife table and the other on a laser, the kerf difference alone can break the fit.

For reference, our metal side holds ±0.005 mm on machined parts and Ra 0.2–0.8 μm on fine finishes. Textile cutting cannot reach those numbers, and it should not be judged against them. The right question is whether the cut edge meets the functional requirement, not whether it matches a machining spec.

• 1
  Single ply knifeAbout ±0.5 mm on a 300 mm part.
• 2
  Laser single plyAbout ±0.25 mm, plus 0.2–0.5 mm HAZ.
• 3
  Repeatability±0.2 mm run to run on a fixed setup.

The break-even point against a steel rule die is lower than most people assume. A die costs money to make and takes time to arrive. A CNC table needs only a vector file. If a design changes twice during a season, the die is scrap and the file is not. For runs under a few thousand parts per design, the CNC route usually wins on tooling alone. Above that, a die can still be cheaper per part if the geometry is simple.

Microbatch production is the real shift. A table can cut one lay of 20 pieces, then switch to a different part on the next lay without a tooling change. That lets a manufacturer hold less inventory and respond to a design change in days rather than weeks. The limit is setup time: loading a roll, sealing the vacuum, and running a test cut takes 20–40 minutes. If the run is two pieces, that setup dominates the cost.

Material utilization is where nesting software earns its keep. Rotating and mirroring parts on a roll can lift yield by 5–15% on irregular shapes. On an expensive technical textile, that difference often pays for the cutting step. The same nest also decides how many parts fit per roll, which feeds directly into the quote. A quote that ignores nesting is a guess.

Labor changes shape rather than disappearing. One operator can tend a table that replaces several hand cutters, but that operator now needs to read a nest, check the blade, and monitor the vacuum. The skill moved from the scissors to the screen. Shops that treat the table as a push-button machine usually get inconsistent edges and blame the equipment.

• 1
  Low volume favors CNCNo die to make; design changes cost nothing.
• 2
  Setup dominates short runs20–40 minutes before the first good part.
• 3
  Nesting lifts yield5–15% on irregular shapes, sometimes more.