Ask ten buyers what separates 3K from 12K carbon tow and most will describe a ladder: 3K at the top for aerospace, 12K somewhere in the middle, 24K at the bottom for industrial work. There is no such ladder. K counts filaments in one bundle and nothing else. 3K means about 3,000 filaments, 12K about 12,000, 24K about 24,000. Ply thickness, drape, damage tolerance, line speed and cost all follow from that count interacting with the process you picked. None of it follows from a rank printed on the label.
The mistake runs both ways and it is expensive. A part specified at 3K because it is “high-end” can end up with four times the plies it needed. A thin cosmetic shell specified at 12K can end up too thick to drape over its own corners. Both patterns turn up in enquiries that arrive with a drawing someone copied forward years ago.
What the K in 3K and 12K Actually Counts
Tow size is a filament count per bundle, and that number sets linear density: 12K carries roughly four times the mass of 3K over the same length. Almost everything downstream comes from that one relationship.
Spread both tows to the same band width and the 12K band is about four times the areal weight of the 3K band. Weave them at the same pick count and the 12K cloth lands near four times the grammage. A 3K twill sits comfortably at 200 g/m²; a 12K plain weave naturally lands at 400 to 600 g/m². Neither is better. One gives you a thin ply, the other gives you mass per pass.
Two consequences matter. A thin ply lets you drop plies one at a time, so a laminate tapers gradually and carries load where it is needed instead of everywhere. A thick ply gets you thickness quickly, which is what a pipe wall or a pultruded profile wants and what a 1.5 mm cosmetic panel does not.
The arithmetic is easy to check on a sample. Look at a cut end under a loupe and count what is actually in the bundle, or weigh a square metre and compare it against the weave and tow you were told you had bought. Suppliers who stay vague about tow size tend to be vague about areal weight as well, and that second vagueness is the one that ruins a laminate schedule.
| Tow size | Filaments | Areal weight | Where it fits |
|---|---|---|---|
| 3K | 3,000 | 1× reference | Thin skins, visible parts, tight radii |
| 12K | 12,000 | ≈4× | Structural laminates, UD tape at volume |
| 24K | 24,000 | ≈8× | Winding, pultrusion, high-throughput conversion |
Why Aerospace Settled on Small Tow
Aerospace laminates are built from thin plies for mechanical reasons. A 3K ply fails over a smaller area when it is hit, so impact damage stays local and residual strength after impact is easier to certify. Thin plies also bend around tighter radii without the tow buckling on the inside of the bend, which is the defect that starts delamination in a thick-ply corner.
There is a second reason that gets less attention: qualification inertia. The tow sizes qualified decades ago are still the tow sizes on the drawing, and requalifying a laminate for a new tow size costs more than most programmes expect to save. When a buyer says “aerospace grade”, they are often describing a paperwork history rather than a physical requirement.
Supply is shifting at the margin. Last week a plant in northeast China that has made graphite electrodes since the 1950s put a 30 tonne per year 1K/3K aerospace-grade line into equipment commissioning, the first high-end sub-project of a relocation and upgrade programme. Thirty tonnes is too small to move a global balance. This is our assessment, not a statement of fact: what matters is not the volume but the direction, because small-tow capacity is being added outside the incumbent producers. Judge the effect over years, not quarters.
Where Large Tow Wins, and Where Small Tow Fails
Above roughly 12K the economics change character. A 24K bundle carries twice the mass of a 12K for every metre of line travel, so a winding, pultrusion or tape line converts more material per hour on the same tension hardware, with fewer splices per tonne and fewer stops. For a pressure vessel, a utility pole, a window profile or a unidirectional tape running at speed, that is the whole argument.
Small tow is a poor fit wherever throughput dominates. It needs more line hours per tonne, more packages, more splicing, and it fills a thick section slowly. Large tow is a poor fit wherever the ply has to stay thin, steer around a tight radius, or read as a smooth cosmetic surface. In a 12K twill the weave crimp is visible at a distance; in a 3K twill at the same grammage it is not.
The split in practice: thin structural skins and visible parts lean small, thick walls and high-volume conversion lean large. Hybrids are common and often correct, since a 12K structural core under a 3K face ply buys mass efficiency and surface quality at once.
Variables That Outrank Tow Size
In real orders four things decide whether the material works, and the K value is not one of them.
- Spread width and areal weight consistency: a tape quoted at a nominal width that wanders across the roll leaves gaps in automated layup and resin-rich bands in the laminate. Ask for the tolerance, not the nominal.
- Sizing chemistry: an epoxy-compatible sizing is not a polyamide sizing, and neither survives a polyarylketone melt at the temperatures those resins demand. Wrong sizing shows up as poor wet-out, voids and low interlaminar strength, and a datasheet that lists only the fibre grade will not tell you.
- Splice rate and fuzz: splices per package and broken filament inside the package drive downtime and scrap far more than a shift in filament count.
- Batch stability: a laminate qualified on one batch and produced on another is a quality escape waiting to happen. Tension, twist and finish consistency matter more than the number on the carton.
If a supplier cannot state those four in writing, the tow size is not the useful part of the conversation.
None of this is exotic. It is the list a converter checks before running a trial, and it is the difference between a material that behaves in month six the way it behaved in month one and one that quietly drifts.
How to Specify Tow Size Without Guessing
Four questions settle most of it. What is the thinnest ply the design can live with? Is any surface visible? Which process will convert the material? And what volume has to come off the line?
If the thinnest ply is under a millimetre, start small, because small tow is the only way to build thickness in controlled steps. If the constraint is wall thickness at speed, large tow wins on throughput. If the part is visible and thin, a fine weave does the work that a sanding operation would otherwise have to do. If the part is wound or pultruded, choose the tow against band width and wall build rate, not against a notion of grade.
Then write the spec around behaviour instead of a label: areal weight with tolerance, spread width with tolerance, sizing matched to your resin, splice limit per package, and the method you will use to accept a batch. A line that says “3K, aerospace grade” tells a supplier almost nothing about your process, and it invites you to pay for a word.
We supply carbon fibre tow from 3K to 24K and larger, woven into plain, twill and satin fabrics from 50 to 600 g/m² at widths up to 1,270 mm, plus thermoplastic prepregs in PP, PA6, PA12, PC, PPS and PAEK systems. If you are choosing between tow sizes for a specific part, send the ply thickness, the process and the surface requirement. We will tell you which construction fits, and which one will cost you line time for no gain.