Kevlar vs Dyneema: Which Stops More?

Ask which fiber is better for soft body armor — Kevlar or Dyneema — and you'll get a confident answer from almost every source. The confident answer is usually wrong, or at least incomplete, because it skips the variable that matters most. It isn't the fiber. It's the form the fiber is built into.

Here's the setup that makes this clear. Take one specific published test — a US patent that fired the same 9mm FMJ bullet at panels of the same weight, in the same lab, under the same military protocol — and compare three panels: woven Kevlar, laminated Dyneema, and woven Dyneema. The results aren't close, and they aren't in the order most people expect.

The Numbers, From One Test

Every figure below comes from a single source: US Patent 10,788,293 ("Flexible body armor"), which tested each panel against the Remington 9mm FMJ threat under the FQ/PD 07-05G military purchase description. Same bullet, same protocol, matched weight (~0.80 lb/ft²), three replicate panels averaged per result. That's what makes these comparable — most fiber comparisons online mix results from different tests and different threats, which tells you nothing.

V50 is the velocity at which a projectile has a 50% chance of fully perforating the panel. Higher is better: it means the armor stops faster rounds.

Read that bottom row again. It's the same base polymer as the top row — UHMWPE, the fiber sold as Dyneema. Laid up as a unidirectional laminate it's the best performer in the table. Woven into a fabric it's the worst, stopping barely a fifth of the velocity. No fiber comparison that ignores form can survive that result.

Why Does Form Matter More Than Fiber?

The answer is friction — specifically, how slippery UHMWPE is.

In a woven fabric, stopping a bullet depends on the crossing yarns gripping each other where they interlace. When the round hits, that grip is what lets the impact energy spread outward across many yarns instead of punching cleanly through a few. Aramid fibers like Kevlar have enough surface friction to hold that weave together under impact — the yarns catch and share the load.

UHMWPE fibers are the opposite. They are extremely slippery — one of the lowest-friction solid surfaces made. Woven into a fabric, the yarns pull through each other at the crossing points instead of gripping, and the panel fails early. That's the 469 ft/s result: the Dyneema fiber never gets a chance to use its strength because the weave can't hold.

A unidirectional (UD) laminate fixes this. Instead of relying on friction at weave points, the fibers are laid straight and parallel, bonded in place by a thin resin matrix, and stacked in cross-plied layers. Now the fiber's strength transfers directly, no weave slippage involved. This is why every real UHMWPE armor panel is a laminate and you will never see a woven Dyneema vest — it simply doesn't work.

So Between Kevlar and Dyneema — Done Right?

Compare the two in the forms actually used in armor, and Dyneema wins on weight. The UD Dyneema laminate stopped 9mm at 1997 ft/s versus 1486 ft/s for the woven Kevlar panel at the same weight — a 34% higher stopping velocity. Put the other way: to match the Dyneema panel's protection, a Kevlar panel would have to be meaningfully heavier. That weight saving is exactly why UHMWPE has displaced aramid in much of modern soft armor.

How BallisticEngine models this: both fibers use the same fiber-deformation model, but with different constants. Dyneema's lower density (970 vs 1440 kg/m³) means more fiber length per unit weight, which drives the higher stopping performance at matched areal density — the same direction the patent test shows. You can run either material against a range of handgun rounds in the calculator.

The Catch: Heat Changes the Answer

Weight isn't the only axis, and here the ranking flips. Dyneema's weakness is temperature.

Aramid fibers like Kevlar are thermally stable to roughly 450°C and do not melt — they char. UHMWPE begins to lose properties much earlier: it softens as it approaches its melt point near 145°C, and manufacturers rate it for continuous service only to about 70-80°C. That's not an abstract lab concern. A vest or panel left in a car trunk through a summer afternoon can reach temperatures that matter for UHMWPE and are irrelevant for aramid.

BallisticEngine's fiber model captures this directly: it applies a temperature penalty to Dyneema as ambient heat climbs toward its melt point, and a separate friction-melting penalty at very high impact velocities where the fiber surface heats. Kevlar, with its 450°C ceiling, takes essentially no thermal penalty across the same range. If your use case involves heat — a hot climate, storage in a vehicle, sustained sun — that stability is a real point in Kevlar's favor that the room-temperature V50 numbers don't show.

Which Should You Choose?

The honest decision comes down to what you're optimizing:

The Honest Caveats

A few limits on what these numbers can tell you, stated plainly:

These are specific products, not the whole category. "Kevlar KM2 Plus 500d" and "Dyneema SB117" are particular grades. A different aramid grade or a newer UHMWPE generation shifts the numbers — the comparison shows the mechanism reliably, not a fixed scoreboard for every product ever made.

The source is a patent, and patents make a case. US 10,788,293's "comparative examples" were chosen by an applicant with a position. The measured V50s are real — a patent is a legal document and fabricated test data carries consequences — but the selection of what to compare is theirs, not a neutral survey.

V50 is a probability, not a promise. It's the velocity at which half of impacts perforate. A panel with a V50 of 1997 ft/s does not "always stop" everything below that — it's a statistical threshold, and real protection standards build in margins above expected threats.

Our calculator uses this fiber model as a physics estimate, matched to the direction these tests show. Where a specific product's certified V50 exists, that certification is the authority; the model is for understanding how the variables trade off, not for replacing a lab test.