When a flat belt, rope, or cable wraps around a curved surface (drum, capstan, or pulley), friction causes the tension to vary exponentially around the contact arc. The governing relationship is T_tight = T_slack * e^(mu * beta), where mu is the coefficient of friction and beta is the total angle of wrap in radians. This exponential dependence means that even a modest friction coefficient over several wraps produces enormous tension amplification — the principle behind capstans, band brakes, and belt drives. For V-belts, which seat in a groove of half-angle alpha, the effective friction is amplified to mu / sin(alpha), making V-belts far more effective than flat belts for power transmission.
Always identify the direction of motion or impending motion first — the tight side is the side toward which the belt tends to slip. Express the contact angle beta in radians (common error source). Work capstan problems where a person holds one end and the load hangs from the other to see the dramatic force multiplication. Compare flat belt and V-belt results for the same geometry to appreciate the groove effect.
The belt friction equation emerges from the same Coulomb friction you already know — F ≤ μN — applied to an infinitesimally small arc element of a belt. Consider a tiny segment of belt spanning angle dθ: it is pulled by tension T on one side and T + dT on the other, and the curved surface pushes back with a normal force dN. Balancing radial forces gives dN = T dθ, and the friction force at impending slip is dF = μ dN = μT dθ. Substituting into the tangential force balance dT = dF gives dT/dθ = μT — a differential equation whose solution is the capstan equation: T_tight = T_slack · e^(μβ).
The exponential is the dramatic part. With a friction coefficient of just μ = 0.3 and two full wraps (β = 4π ≈ 12.6 rad), the tension ratio is e^(0.3 × 12.6) ≈ e^3.8 ≈ 44. A sailor holding a line with 10 N of force can resist a 440 N load — not by strength, but by wrapping the line around a cleat. This is the capstan used on sailing ships and cable cars. The mathematics is purely about accumulated small friction increments compounding like interest.
To apply the formula correctly you must first identify the tight side and slack side. The tight side is always the side toward which the belt tends to slip (the side being pulled, or downstream of the driving direction). The slack side has the lower tension. If you reverse them, your ratio inverts and predicts physically impossible results where friction amplifies in the wrong direction. The wrap angle β must be in radians — this is the single most common arithmetic error, since contact arcs are often given in degrees.
The V-belt modification replaces μ with an effective friction coefficient μ_eff = μ / sin(α), where α is the groove half-angle. When a V-belt seats in its groove, the normal forces from the two groove walls both contribute friction, and because those walls are nearly vertical, the total normal force is much larger than for the same belt tension on a flat surface. A typical V-belt groove with α = 18° gives sin(18°) ≈ 0.31, so μ_eff ≈ 3.2μ — the belt is more than three times as effective at transmitting force as a flat belt of identical geometry and material. This is why V-belts dominate industrial power transmission: they resist slip without requiring enormous pre-tension in the belt.