What a limb actually is
A compound bow limb is a bent beam. It stores energy when deflected and releases it when allowed to return toward straight. Everything else — the cam, the cable, the axle, the pocket — is either a mechanism to control when that energy is released, or a structural interface for holding the beam in place. The limb itself is the spring. Understanding it from that starting point produces clearer answers than approaching it from the direction of archery lore.
A beam stores elastic strain energy proportional to the square of its deflection. Double the deflection, quadruple the stored energy. The limit is the maximum stress the material can sustain without failing — crack propagation, fatigue, or outright fracture. The ideal limb stores the maximum energy for a given weight and length before reaching that stress limit. Everything in limb design is a negotiation toward that target.
Three variables govern how the limb performs that job: geometry (how it is shaped and oriented), cross-section (how thick and wide it is along its length), and material (how much stress the beam can sustain per unit of strain). This article works through all three. Cams, strings, and cables are deliberately out of scope — they interact with the limb, but understanding the limb itself requires treating it as the independent problem it is.
Limb angle — what it was, what it became, and why
Early compound bows inherited their limb geometry from the recurve tradition. The limbs angled steeply from the riser — 30 to 45 degrees from horizontal was typical — because that is how bent-stave bows had always worked. The compound mechanism changed how the string and cams interacted with the limb tip, but for the first generation of designs, the limb angle stayed where tradition put it.
The progressive flattening of that angle toward horizontal — what the industry eventually labeled the parallel limb — happened because engineers worked out what the angle was costing.
The force vector argument
When the bow fires, both limbs straighten simultaneously. The limb tip accelerates through its arc. The force the limb tip exerts on the cam, and through the cam on the cables, has a direction determined by the limb geometry at that moment. For a limb angled at θ degrees from horizontal, the tip force has two components: one perpendicular to the limb (the primary bending force, doing useful work through the cam) and a secondary component that determines what the riser feels.
The critical insight is this: the vertical components of the two limb tip forces cancel. The top limb pushes up, the bottom limb pushes down, and they neutralize each other perfectly in a symmetric bow. What does not cancel is the horizontal (forward) component. For both limbs angled above horizontal, both tip forces have a component pushing forward — in the same direction. They add rather than cancel. That net forward force is felt by the shooter as hand shock and recoil.
The magnitude of this forward force scales with sin(θ): at 30° from horizontal, the forward component per limb is 50% of the tip force. At 15°, it is 26%. At 0° — parallel limbs — it is zero. The limb tips are moving purely vertically at the shot, the two forces are equal and opposite in the only plane that produces riser movement, and the riser feels nothing from the limb geometry itself. The only remaining recoil is from the arrow leaving the string, which is present regardless of limb angle.
This is not a marginal effect. At traditional 35° limb angles, the geometry-induced forward force is more than half the limb tip force — and limb tip forces are large. Moving to parallel limbs eliminates this term entirely from the recoil equation. The quieting and smoothing of the shot that archers observed as parallel limb bows replaced traditional-angle bows was not perception. It was physics working correctly for the first time.
Beyond parallel — does it keep getting better?
A natural question follows: if 0° (parallel) is better than 30°, is −5° better still? Several manufacturers explored this. Some marketed bows with limbs that angle slightly past horizontal at brace — limb tips pointing slightly upward — calling it “beyond parallel.”
The honest answer is: marginally, and only in a narrow sense. When a beyond-parallel limb fires, the tip starts above horizontal and travels through horizontal and beyond during the shot. For the portion of the power stroke while the limb is past horizontal, the tip force has a rearward component — which actually reduces recoil below zero for that interval. The net effect over the full power stroke is a slight further reduction in forward impulse compared to a true parallel limb, but the arrow has already left the string before most of that interval. The improvement is real but small enough that it falls within shot-to-shot variation for most archers.
There is also a cost. A limb angled past horizontal at brace must deflect further during draw to bring the limb tip down to the lower angles seen at full draw. That additional deflection increases peak stress at the limb base. To handle that stress without reducing limb life, either the material grade must increase, or the limb must be made thicker or wider — which adds mass and changes the dynamic behavior at the shot. The geometry gain and the stress penalty exist simultaneously. Most manufacturers who pursued beyond-parallel geometry settled back toward true parallel after working through these tradeoffs. It is not where physics runs out of ideas. It is where geometry stops being the bottleneck.
Why some manufacturers keep a steeper angle
If parallel is better, why do some bows still run a noticeable angle? The constraint is almost always axle-to-axle (ATA) length combined with limb length requirements.
A compound limb needs to be long enough to deflect smoothly without exceeding material stress limits. The stress at any point in the limb is proportional to the bending moment at that point divided by the section modulus of the cross-section. For a limb deflected to a given stored energy, a shorter limb must flex through a larger angle to store the same energy — which means higher peak curvature and higher peak stress at the base. Short limbs fail earlier than long limbs at the same energy level.
For a hunting bow designed around a 30-inch ATA, fitting an adequately long limb pair requires that the limbs angle significantly away from horizontal — if they were parallel, the limb tips would extend far beyond the riser width, making the bow physically awkward or requiring very short limbs. Angling the limbs back toward the riser keeps the limb tips contained within a reasonable width for a compact bow. The angle is not a performance choice. It is a packaging constraint the manufacturer accepted in exchange for a shorter ATA.
Target bows — which prioritize performance over packagability — trend strongly toward true parallel because ATA is not constrained. A 40-inch ATA target bow can run long, nearly horizontal limbs without any geometry compromise. The correlation between bow purpose and limb angle is not coincidental. It tracks the packaging constraint almost exactly.
A secondary constraint is cam clearance. At full draw, the limb tips have traveled significantly from their brace position. For very parallel limbs, the arc the cam sweeps through at full draw can create clearance problems with cables. Designers who have pushed limb geometry toward flat have had to solve cable routing to keep clearance adequate. It is a solvable problem, but it adds complexity to the cam and cable geometry that slightly angled limbs do not have.
Cross-section — width, thickness, and the taper problem
The limb angle governs the direction of force. The cross-section governs how well the beam stores and releases energy within the material’s stress limit. These are independent design variables.
For a rectangular beam bending in one plane, the bending stress is:
σ = M × c / I where I = w × t³ / 12, c = t / 2
Simplifying: σ = 6M / (w × t²). To reduce peak stress for a given bending moment M, you can increase width w or increase thickness t. Increasing width is more efficient — stress drops linearly with width but only as the square of thickness, while thickness also affects the bending moment distribution through its influence on limb stiffness. The practical consequence is that wider, thinner limbs store more energy at lower stress than narrower, thicker limbs of the same mass. The trend in compound bow limb design toward wider limb profiles over the past two decades is mechanically correct, not cosmetic.
The more interesting question is taper. A uniform rectangular limb along its full length distributes stress very unequally — the bending moment is maximum at the limb base (where the pocket constrains it) and approaches zero at the free tip. A uniform limb wastes material in the middle and tip sections, which are understressed while the base works near its limit. The ideal beam for energy storage is a constant-stress beam: one where the cross-section decreases from base to tip proportionally to the bending moment, so every point along the limb is working at the same fraction of its stress limit simultaneously.
For a cantilever loaded at the tip — which approximates a compound limb loaded by a cable at the cam — the constant-stress profile requires the section modulus (I/c = w × t²/6) to decrease linearly from base to tip. This can be achieved by tapering width, tapering thickness, or both. In practice, compound limbs use both: they taper in width from base to tip, and the laminate layup typically tapers in thickness as well through the grinding or pressing process. The result approaches, but does not fully achieve, the constant-stress ideal. The deviation from ideal is largest right at the limb pocket, where the geometry of the pocket interface interrupts the taper.
Solid versus split — the torsion question
Split limbs — two parallel limb halves with a gap between them — were dominant in compound design for roughly a decade beginning in the mid-2000s. The case for them was weight: removing material from the center of the limb reduces mass without proportionally reducing bending stiffness, since the center of the limb contributes less to the second moment of area than the outer material does.
The problem split limbs introduced is torsional stiffness. A wide solid limb resists twisting about its long axis effectively — the full width provides the moment arm for torsional resistance. Each half of a split limb is narrower, and narrower beams are significantly less torsionally stiff. Any asymmetry in the system — unequal cable tensions, a slightly off-center cam, a limb pocket that is not perfectly square — produces a twisting load on the limb. A solid limb resists this through its geometry. A split limb yields to it, and the tip travels slightly off the intended arc on each shot.
Torsional inconsistency in limb tip travel is exactly the kind of error that compounds across a round. It is not a random error — it is a systematic bias that shows up as a consistent left-right pattern in the group, correlated with limb twist. And because it is present on every shot, it cannot be distinguished from a tuning error without careful isolation. The archery community largely diagnosed it as a tuning problem and adjusted rather than naming the geometry. Manufacturers who returned to solid limbs observed that setups became easier to tune and more consistent, which is consistent with the torsion mechanism, even where it was not explicitly named.
The weight penalty of a solid limb over a split limb of the same width is modest for a well-tapered design. The consistency gain is not modest. The return to solid limbs in performance compound design is mechanically justified.
Materials — where we actually are
The limb laminate is a composite structure: fibers in a resin matrix, oriented and layered to produce the desired stiffness and strength. The specific combination matters because the limb sees two fundamentally different stress states simultaneously — tension on the back face (away from the archer) and compression on the belly face (toward the archer) — and common fiber materials handle these differently.
Carbon fiber is excellent in tension: tensile strength approaching 3,500 MPa in high-grade unidirectional layups. In compression, it is significantly weaker — typically 1,200–1,500 MPa for the same fiber, because the failure mechanism in compression is fiber buckling and kinking rather than fiber fracture. The belly of a compound limb, in compression at full draw, is thus the weaker side. Building an all-carbon limb pushes performance on the tension face while leaving the compression face as the limiting variable.
Glass fiber is more isotropic: comparable tension and compression properties, and better toughness against crack propagation and delamination. Gordon Glass composites — the pull-formed continuous-fiber materials used in most high-performance compound and recurve limbs — are engineered specifically for the deep-deflection, high-cycle fatigue demands of the limb application. The manufacturing process orients the fibers continuously along the limb length, which eliminates the fiber-length discontinuities that create stress concentrations in chopped-fiber or woven composites.
Most performance compound limbs today are hybrids: carbon fiber layers on both faces for stiffness and specific modulus, with glass layers in the core and as outer protection layers for toughness and compression resistance. The ratio of carbon to glass governs the stiffness-to-weight tradeoff. Stiffer, lighter limbs run more carbon. More durable, more forgiving limbs run more glass. The gradient between those extremes is where most manufacturers operate, and where they differentiate their products at the material level.
Fatigue life is the other governing material property. A competition archer shoots tens of thousands of cycles per year. A hunting bow might see fewer cycles but more abuse — temperature extremes, drops, storage under load. Published fatigue testing on compression-molded glass/carbon composite limbs shows one million stress cycles without failure as a routine specification for quality materials. The material is not close to its fatigue limit under normal use. Premature limb failures in modern bows almost always trace to manufacturing defects — voids in the laminate, fiber misalignment, pocket fit problems — not to the material reaching its intrinsic fatigue boundary.
The pocket interface — the underexamined problem
The limb pocket is where the limb base meets the riser. It is the point of highest bending moment in the entire system, and it is the point where the idealized constant-stress taper abruptly ends. The pocket constrains the limb over a short length, creating a stress concentration at the edge of the pocket where the constrained section meets the free beam. This is where compound limbs crack when they fail — not in the middle, not at the tip, but at the pocket edge, where the geometry transition is sharpest and the moment is highest.
The pocket edge radius matters enormously. A sharp pocket edge creates a stress concentration factor of two or more — meaning the local stress at the edge is twice or more what a smooth taper would predict. A radiused, well-fitted pocket edge reduces this concentration. Machining tolerances in the pocket, and how precisely the limb base fits the pocket geometry, determine how well the load transfers from limb to riser. A limb that rocks slightly in its pocket — because the fit is imprecise — creates dynamic impact loading at the edge on every shot. That impact loading cycles at peak stress, which is the worst possible location for fatigue initiation.
Limb bolt preload is the mechanism for holding the limb in its pocket. Limb bolts in a compound bow do not add draw weight directly — they compress the limb butt into the pocket, controlling how much of the limb’s stored energy at brace is transmitted to the riser rather than expressed as free deflection. The preload determines the resting limb angle, the brace height geometry, and therefore the draw weight. Equal preload on top and bottom ensures symmetric limb behavior — unequal preload is a source of tuning problems that mimics cam timing error.
Better pocket designs — ones that match the actual stress distribution at the limb base, use generous radii at the edge transition, and maintain tight fit tolerances — would improve both fatigue life and shot-to-shot consistency. This is an area where meaningful engineering improvement is still available. The laminate is well-optimized. The interface where it connects to the riser is not.
Are we there yet?
On geometry: substantially yes. Parallel limbs are physically correct, and the industry has largely converged on them for performance bows. The remaining angle variation in the market is almost entirely a packaging constraint driven by ATA length, not a performance preference. Beyond-parallel geometry offers marginal additional gain at real stress cost. The low-hanging fruit in limb angle was found and picked.
On cross-section: largely yes, with nuance. Wide, solid, tapered limbs are mechanically correct. The constant-stress taper is well-approximated by quality manufacturers. The remaining improvement available through better taper optimization — using modern finite element analysis rather than empirical iteration — is real but modest. No production compound limb has been analytically optimized for a constant-stress profile from pocket edge to tip. The tools to do it exist and have been applied to aerospace structures for decades. The archery industry has not systematically applied them.
On solid versus split: the industry is correcting course. The return to solid limbs in performance designs is mechanically justified and continuing.
On materials: this is where the most genuine improvement remains. Not in raw material properties — current glass-carbon hybrids are well-matched to the stress and fatigue demands of the application — but in manufacturing consistency. The gap between what the material is capable of and what production limbs actually deliver is determined by void content, fiber alignment, and cure uniformity. Tighter process control closes that gap. The theoretical improvement available is meaningful; the path to it runs through manufacturing engineering, not materials science.
On the pocket interface: meaningfully underoptimized. The highest-stress, most failure-prone location in the limb system receives the least analytical attention in most designs. Better pocket geometry, tighter fit tolerances, and explicit stress concentration management at the pocket edge would improve both durability and shot consistency. This is probably the single area with the most available improvement per engineering investment.
What an ideal limb looks like
From first principles, the ideal compound bow limb is:
Parallel to horizontal at brace — eliminating geometry-induced forward recoil entirely. Not beyond parallel, because the stress penalty exceeds the marginal gain. Not angled unless ATA constraints require it, in which case as close to parallel as the packaging allows.
As long as the axle-to-axle geometry permits — because longer limbs deflect the same stored energy over a greater arc, reducing peak stress per unit of stored energy and extending fatigue life. Short limbs are a packaging compromise, not a performance feature.
Wide and solid in cross-section, tapered from base to tip — approximating the constant-stress profile that puts every point in the limb at the same fraction of its stress limit simultaneously. No wasted material in the middle carrying no load while the base is at its limit.
Carbon-glass hybrid laminate on continuous fibers — with the carbon providing stiffness on both faces and glass providing compression resistance on the belly face and toughness throughout. Pull-formed or equivalent continuous-fiber process that eliminates fiber-length discontinuities. Void content minimized through process control.
Pocket interface with generous radii and tight tolerances — matching the stress state at the limb base, avoiding the sharp-edge stress concentrations that initiate most real-world limb failures.
That limb does not fully exist in production. The geometry is essentially there on performance target bows. The materials are close. The taper optimization and pocket engineering still have room. Whether those remaining improvements are worth pursuing depends entirely on how much a given manufacturer is willing to invest in processes that are invisible to the customer until a comparative test makes the consistency difference clear.
The limb itself, as a problem in mechanics, is largely solved at the geometry level and approaching solved at the material level. The remaining frontier is not discovering new physics. It is applying the physics we already have with better manufacturing precision.
Published 2026-08-17 · Axial Bowstrings