Crack in a carbon frame: how to tell paint cracks from structural cracks

A crack on a carbon frame is, most of the time, in the paint. Less often, but often enough to take seriously, it is through the laminate. The field method for sorting one from the other turns on three things: the pattern of the crack, the location and context, and the behavior of the crack under load. None of the three are conclusive on their own, and when they point in the wrong direction the right answer is to stop the field work and go to non-destructive testing.

This post lays out the framework. It is written for the owner with a frame in a stand and a flashlight, who wants to make a defensible call about whether to ride the bike and, if not, what evidence to take to a repair shop or an inspector.

Why the question is hard#

CFRP fails in ways that metallic frames do not. Steel, aluminum, and titanium frames generally telegraph failure: they bend, deform plastically, or develop visible surface cracks. Carbon does not. It accumulates hidden internal damage and then, past a threshold, fails suddenly (Mondince Cycle). The visible surface of a carbon frame is not a reliable proxy for the laminate underneath, in either direction. A frightening web of cracks across a top tube panel can be confined entirely to the paint; a clean paint surface can sit over a shattered set of plies.

The field method exists because most cracks need to be sorted somewhere, and most are not sitting on the bench of a lab with phased array ultrasonics waiting. The framework below is what triages the call, not what answers it.

Pattern: web versus line#

Cosmetic finish cracks and structural cracks look different at the surface.

Clearcoat crazing is a dense network of multidirectional hairline cracks driven by UV degradation, weathering, normal frame flex, or thermal expansion mismatch between paint and substrate. It appears as a dull, fractured glaze, often around junctions or clamp areas. Crazing rarely follows a single direction; it covers a region in a fine, multidirectional pattern that gives the panel a slightly weathered look under raking light.

Star cracks radiate as a tiny spiderweb from a central point, typically a sharp stone strike where the brittle clearcoat fractures along radial shockwave paths. Star cracks are also usually low risk on their own, because the brittle clearcoat fractures more readily than the laminate beneath.

If the cracking is confined to the paint layers, it is low risk: the frame is sound and the cracking can be sanded and refinished, or sealed to prevent moisture ingress (TWCarbon repair services; Certify Cycle). If the same cracks extend into the structural plies, the right reading is matrix microcracking under the clearcoat, and they have to be treated as structural.

Matrix microcracking under a clearcoat looks like a diffuse network of fine lines visible under magnification. The cracks themselves are individually harmless, but they act as stress concentrators and capillary pathways for moisture, accelerating ingress into the laminate (Nairn matrix microcracking review). Under continued cycling they reach ply boundaries, trigger fiber-matrix debonding, deflect along the interfaces, and seed delamination.

Fiber fractures look completely different. A clean slit, crease, or hinge with exposed or frayed fiber ends, often along a tube circumference or at a load-path discontinuity. These are deep breaks that cut through the laminate. The visual signature alone is usually enough to retire the frame to NDT.

The single distinction that separates most cosmetic findings from most structural ones: matrix microcracks present as a diffuse network of fine lines, while fiber fractures present as a clean slit or hinge with frayed fiber ends. A tube panel covered in spider-web crazing is almost always cosmetic; a single sharp line on a tube right after a crash is almost always not.

Location and context#

A flaw near a tube junction, under a heavy clamp, or appearing immediately after a crash should be treated as potentially structural. Paint crazing that develops slowly over time is far more likely to be cosmetic. A mark that appears suddenly under stress is not.

The high-stress junctions on a typical carbon frame are the bottom bracket cluster, the head tube transitions, the seatpost insertion area, and the dropouts. A crack discovered at any of these locations gets more scrutiny than a crack on the middle of a top tube panel. Clamp zones get even more: the fork steerer in the stem clamp, the seatpost in the collar, the seatstay bridge under any rack or fender hardware. Carbon fibers are brittle and cannot tolerate being crushed; the layers crack and delaminate under clamp pressure even when the outer finish looks intact (Rat City Bikes carbon fiber warning).

The context cues that bump a crack toward structural:

• Appearance immediately after a crash or known impact, even a low-velocity one.
• Location at a tube junction, under a clamp, or at a known high-stress zone.
• A sharp, localized, single line rather than a web.
• Accompanying creaks, soft spots, or changes in ride feel.
• A faint white halo (impact halo) around an associated impact site, which is itself a reliable indirect indicator of subsurface delamination.

A paint crack with none of those flags, on a panel away from any clamp or junction, that has been there for months without progression, is almost always what it looks like.

Behavior under load: the decisive test#

The load-flexion test is the field method's strongest tool. Frame builder Markus Storck's long-cited guidance is that pressing or pinching a crack reveals its nature: a crack that stays stiff and unchanged is usually in the paint, while one that flexes, opens, closes, or grows under load or over time is through the laminate (Bicycles Stack Exchange discussion).

Two pairings make the test useful:

Thumb pressure for stiffness. Press firmly along the tube next to the crack and compare the response to an adjacent undamaged zone. Sound CFRP feels exceptionally rigid. A spongy, soft, or deflecting spot indicates subsurface delamination or fiber buckling. Where the deflection sits under the crack, the crack is far more likely structural.

Gentle flexion for behavior. Apply gentle bending or twisting near the crack and watch the crack itself. A paint-only crack stays static. A through-laminate crack visibly opens, closes, or shifts under the load, sometimes with an audible crackle or grinding under flexion.

A third, complementary test is the cotton-rag snag. Wipe a soft cotton or microfiber cloth gently over the suspect blemish. If fibers have fractured through the paint, the jagged filaments snag the cloth and leave white lint behind, confirming structural fiber breakage. The snag test is a yes-only test; the absence of snag does not exclude subsurface damage.

New creaks, soft spots, or a changed ride feel since the crack appeared all commonly accompany deeper damage. Any one of those secondary signals is reason enough to stop riding and refer the frame to NDT regardless of how the field test reads.

Why the test ride is not the answer#

A short test ride checks fit, function, and noise. It cannot validate composite structural integrity. The plies are aligned along specific load paths with significant structural redundancy, so a frame can lose roughly 40 to 50 percent of its local interlaminar shear strength in a single zone and still feel stiff and responsive under moderate riding forces (Certify Cycle).

The mechanics are simple. Internal delamination or a through-laminate crack reduces the effective load-bearing cross-section of a tube, which raises the local stress concentration. Under typical test-riding loads, that local stress stays below the ultimate compressive strength of the surrounding healthy plies, so the frame shows no abnormal flex or creak. A sudden high-force compressive event (a pothole at speed, a heavy landing, a hard out-of-the-saddle effort) drives the local stress past the failure threshold of the remaining plies and the tube collapses. The test ride read the redundancy, not the damage.

The practical rule: use the test ride to verify function, treat it as one input only. Any wobble, creak, or sudden give warrants a stop and an inspection. Any recent crash or suspicious mark warrants NDT even if the ride felt fine.

Where the field test stops and NDT takes over#

The field method is triage, not verdict. The triage is reliable in two directions and unreliable in a third.

It is reliable for sorting obvious cosmetic crazing on a panel away from any clamp or junction. The pattern is diffuse, the load-flexion test is negative, the panel has no impact history, the snag test is negative, ride feel is unchanged. This is the largest single class of carbon-frame crack questions, and the field method answers it well.

It is reliable for sorting obviously structural cracks: a sharp single line right after a crash, a flexion-positive crack with frayed fiber ends, a crack under a clamp with thumb-pressure softness. The field method does not need to verdict these; it flags them for the same NDT a less obvious case would go to.

The unreliable case is the ambiguous middle. A hairline through paint at a clamp zone with no flexion change. A faint halo around an impact site with no associated crack. A panel that snags lightly on the cotton test in a region with no visible mark. These cases are where the field method runs out, and the right action is to refer the frame to phased array ultrasonic testing (PAUT) or active infrared thermography rather than to try to push the field test further.

PAUT maps wall thickness loss, voids, and delamination beneath the paint, resolving wall thickness down to roughly one one-thousandth of an inch (Evident / Ruckus Composites). Active thermography reads the differential cooling over compromised zones because air pockets and separated plies act as thermal barriers (Certify Cycle). Both are quantitative, both are safe, and both resolve what the field method cannot.

The tap test is sometimes proposed as a middle step between field flexion and NDT. It is not a verdict either. The documented failure modes (compound curves scattering sound at junctions, ply drops near the BB sounding deadened, bonded aluminum BB shells dominating the acoustic response, BVID too tight to alter resonance) all mean that the tap test is a useful secondary screen at best, and never a substitute for PAUT or thermography (Carbon Bike Repair tap test limitations; Velo coin test).

What the inspection report records#

A complete inspection report sorts findings into Safe, Serviceable, or Unsafe (VéloColour). A confirmed clearcoat-only finding is recorded under Serviceable with a note describing the pattern and location, and logged for future monitoring; cosmetic marks are always separated from structural findings throughout the report. A confirmed structural crack is recorded under Unsafe with the location, the inferred mechanism, the NDT output (PAUT echo figures or thermographic heat maps), and an actionable remediation estimate. Where the affected component is a handlebar, stem, or lightweight steerer, the report defaults to replacement rather than repair (Carbon Bike Repair UK).

What this means for the reader#

If you have just found a crack and the bike is in your stand, the order of operations is short: pattern (web or sharp single line), location and context (junction, clamp, or post-crash), load-flexion test (stays stiff, or opens). Cosmetic findings can stay cosmetic. Structural findings go to PAUT or thermography. Ambiguous findings go the same place. When the secondary signals (creak, soft spot, halo, changed ride feel) are present, the field test no longer matters; the frame goes to NDT regardless.

Presidio Composites operates pulsed thermography NDT and returns a written report that separates cosmetic findings from structural ones and provides the evidence record an owner takes to a repair shop, an insurer, or a buyer. Presidio does not perform repair work itself; the inspection produces the documented condition and the remediation estimate.