Capstan Equation

The intuition: friction that compounds

Wrap a rope once around a fence post, hold one end, and tug the other. You'll feel the post "eat" most of your effort — pull hard on your end and the far end barely moves. Add a second wrap and the effect roughly squares. Add a third and your one-handed grip can hold a load that would otherwise drag you off your feet. This is not magic and it is not leverage in the lever sense: it's friction, applied over and over along the contact arc, with each tiny patch of rope shielding the patch behind it.

That compounding is what makes the capstan equation exponential rather than linear. A linear law would say "twice the wrap, twice the hold." The real law says "twice the wrap, the hold squared." A sailor leaning lightly on a line snubbed three times around a steel bollard restrains a ship displacing thousands of tonnes against wind and current. A climber's belay device, a band brake on a winch, the V-belt spinning your car's alternator — all of them live or die by this one formula.

How it works: the sliding-friction balance

Picture the rope lying against the cylinder over some arc. Tension is higher on the load side (call it T_load) and lower on the side you hold (T_hold). The rope presses into the cylinder; the cylinder pushes back with a normal force; friction acts along the contact, always opposing the tendency of the rope to slide from high tension toward low.

Take an infinitesimal slice of rope subtending angle dθ. The tension difference across it is supported partly by the geometry of the curved contact (which generates a normal force dN = T·dθ) and partly by friction (dT = μ·dN). Combine them and you get a differential equation whose only surviving variable is the angle:

Force balance on slice dθ:

  dN = T · dθ                  (curved contact → inward normal force)
  dT = μ · dN = μ · T · dθ     (friction opposes sliding)

  →  dT / T = μ · dθ

Integrate from T_hold to T_load over total angle β:

  ln(T_load / T_hold) = μ · β

  →  T_load = T_hold · e^(μ·β)        ← the capstan equation

The cylinder radius never appears. A pencil-thin pin and a metre-wide bollard give the identical tension ratio for the same wrap angle and friction. What matters is only how far around the rope wraps (β, in radians) and how grippy the contact is (μ). The exponential is the signature of a quantity that grows in proportion to itself — each radian of wrap adds friction proportional to the tension already present.

Running the numbers

Convert turns to radians (one full turn = 2π ≈ 6.283 rad), pick a friction coefficient, and the holding ratio falls straight out. With μ = 0.25 (a typical synthetic-rope-on-steel value):

Holding ratio  R = T_load / T_hold = e^(μβ)

 Wraps   β (rad)   μβ      R = e^(μβ)
 ─────   ───────   ────    ──────────
 ½       3.14      0.785   2.19
 1       6.28      1.571   4.81
 2       12.57     3.14    23.1
 3       18.85     4.71    111
 4       25.13     6.28    535
 5       31.42     7.85    2,570

Hold with 100 N (≈10 kgf, easy one-hand grip):
 3 turns  → restrains   11,100 N  ≈ 1.13 tonnes
 5 turns  → restrains  257,000 N  ≈ 26.2 tonnes

Run it the other direction to see the danger. To lower a load you must let the rope creep, which means T_hold just barely exceeds T_load·e^(−μβ). Suppose a wet rope drops μ from 0.25 to 0.12. At three wraps the ratio collapses from 111 to e^(0.12·18.85) = e^2.26 ≈ 9.6. The same grip that held a tonne now holds barely 90 kg before the line runs — which is precisely why mariners and riggers respect a wet or iced line, and add wraps when conditions turn.

Worked example: snubbing a ship on a bollard

A harbour line takes 40 kN of strain from a ferry surging on its springs. A deckhand can comfortably hold about 200 N on the bitter end. How many turns around the steel bollard does it take, at μ = 0.3 (dry synthetic on rough steel)?

Required ratio:  R = 40,000 / 200 = 200

Solve e^(μβ) = 200  for β:
  μβ = ln(200) = 5.30
  β  = 5.30 / 0.3 = 17.66 rad
  turns = 17.66 / (2π) = 2.81 turns

→ Three full turns (β = 18.85 rad) gives e^(0.3·18.85)
   = e^5.66 = 287×, comfortably above the 200× needed.

Margin: at 3 turns the hold needed is 40,000 / 287 = 139 N.

This is exactly the workboat rule of thumb — "three round turns and you can hold anything" — quantified. Note the safety logic: you choose enough wraps that the required holding force sits well below what a person can actually exert, so a momentary surge doesn't rip the line through your hands. The bollard does the work; the human just sets the friction limit by deciding how many turns to take.

Where it shows up — and the real numbers

System	What wraps what	Typical μ	Typical β	Effect
Mooring bollard / bitt	Line around steel post	0.2 – 0.3	3+ turns (18.8 rad)	One crew holds a multi-tonne surge load
Powered capstan / winch	Line around a driven drum	0.2 – 0.4	3 – 5 turns	Drum spins; tailing hand controls feed
Belay device (climbing)	Rope around grooved channel	0.2 – 0.3 (+ device geometry)	~π to 2π	Belayer holds a falling climber with a light grip
Band brake	Steel band around a drum	0.3 – 0.4 (lined 0.35)	~3π/2 to 2π	Self-energizing braking torque
Flat / V-belt drive	Belt around pulleys	0.3 flat, ~0.5 effective V	π on small pulley	Sets max transmissible torque before slip
Conveyor drive pulley	Belt around drum	0.25 – 0.45 (lagged)	π to 220°	Lagging + snub pulley raise μ and β
Rope rescue / rappel	Rope around a figure-8 or post	0.2 – 0.3	Multiple wraps	Controlled descent of a body weight

The V-belt deserves a note. A V-belt sits in a wedge-shaped groove, so the rope's inward force is reacted on two angled flanks instead of one flat surface. The effective friction coefficient becomes μ_eff = μ / sin(α), where α is the groove half-angle (typically 18°). That turns a real μ of 0.3 into an effective ~0.97 — roughly triple the grip of a flat belt for the same wrap angle, which is why V-belts transmit far more torque in the same space.

Capstan friction vs other ways to handle a load

	Capstan / belt friction	Block & tackle (pulley)	Winch drum (spooled)	Cam / ratchet cleat
Working principle	Exponential friction over wrap arc	Mechanical advantage (force ratio)	Stored rope + driven gear	Wedge / toothed grip
Tension ratio achievable	e^(μβ) — unbounded with wraps	= number of supporting parts	Limited by drum capacity	Binary: grip or release
Moves the load?	No — only holds/controls; drum must drive	Yes, at reduced speed	Yes, by spooling	No — holding only
Depends on friction?	Entirely	Ideally not (friction = loss)	For self-holding, partly	Yes
Sensitive to wet/ice	Very (μ halves → ratio collapses)	Low	Low	Moderate
Independent of size?	Yes — radius cancels	No — sheave count matters	No	No
Typical home	Bollards, belts, belays, brakes	Cranes, sail rigging, hoists	Vehicle/anchor winches	Sailboat cam cleats, tie-downs

The crucial distinction: a block and tackle moves a load with mechanical advantage but no friction help, while a capstan holds or feeds a load using friction alone and offers no advantage for moving it. They're often combined — a powered capstan drum (friction grip) tails a line that runs through fairleads and blocks (mechanical advantage) — so the two principles complement rather than compete.

When the capstan equation governs your design

• You need to hold or restrain a large load with a small input force. Mooring, belaying, and snubbing are the canonical cases — let the wrap do the work.
• You're sizing a friction belt drive. The smaller pulley's wrap angle sets the slip limit; the design tension ratio must stay below e^(μβ) with margin (typically a 1.3–1.5 service factor).
• You're designing a band or drum brake. The self-energizing band brake gets its bite directly from the capstan ratio; the direction of drum rotation determines whether the brake is self-energizing or self-de-energizing.
• You want controlled feed of a tensioned line. A powered capstan lets the load tension be enormous while the tailing tension — and therefore the control effort — stays in the human-friendly range.

Reach for a different principle when you must actually lift or move the load against gravity over distance (use a block and tackle, lead screw, or geared winch), when zero rope slippage is mandatory (use a spooled drum or a captive-fit clamp), or when the contact can't be kept clean and dry enough to trust μ (friction-only holding becomes unsafe).

Common misconceptions and pitfalls

• "A fatter post holds more." No. The radius cancels in the derivation. A thin pin and a thick bollard give the same ratio for the same wrap angle. A fatter post is gentler on the rope (larger bend radius, less wear) but adds zero holding power per radian.
• "The capstan gives mechanical advantage like a pulley." It doesn't move the load with advantage. It only restrains or controls. To haul the load you still need to drive the drum or pull rope through — the friction merely lets a small tailing force manage a huge line tension.
• "More turns are always safer." Too many turns can cause the rope to ride over itself and jam ("riding turns" or a "wrap-up"), trapping the line so it can't be eased — dangerous on a powered capstan where a fouled tail can drag a hand into the drum. Riggers add turns deliberately and keep tension on the tail to prevent overrides.
• "It works the same wet, dry, or iced." The ratio is exponential in μ, so halving the friction coefficient roughly square-roots the holding ratio. A line that holds a tonne dry may hold under 100 kg when iced — add wraps and double-check the tail before trusting a wet line.
• "Belt slip means the belt is worn out." Often it means the wrap angle is too small or the pre-tension too low for the demanded torque — the capstan limit is being exceeded. Adding an idler to increase β or correctly tensioning the belt fixes it without replacement.
• "Rope stiffness and weight don't matter." For thin flexible rope on a large drum, the simple equation is excellent. For stiff rope or wire on a small sheave, bending stiffness adds a measurable correction (the rope resists conforming to the curve), and for steep, heavy lines the rope's own weight shifts the effective tensions. Most general rigging stays well inside the simple regime.