Belt Drive Calculator & Simulator

Belt & Chain Drive — Power Transmission in Mechanical Engineering

Belt and chain drives transmit rotary motion and torque between shafts that are separated by some distance. This simulator covers open belts, crossed belts, and chain drives, computing velocity ratio, wrap angle, tension ratio (capstan equation), transmitted power and belt length in true-scale geometry for any chosen centre distance C.

Velocity Ratio and Speed Relationships

The fundamental relationship in any belt drive is the velocity ratio (VR): the ratio of driver pulley diameter to driven pulley diameter equals the ratio of driven to driver speed. If the driver has diameter D₁ and speed N₁, the driven speed (allowing for slip s) is N₂ = N₁ × D₁/D₂ × (1 − s/100). A larger driver pulley produces higher driven speed (speed increaser); a smaller driver produces speed reduction with torque multiplication.

The Capstan Equation — Belt Tensions

The ratio of tight-side tension T₁ to slack-side tension T₂ is given by the capstan (Euler–Eytelwein) equation: T₁/T₂ = e^μθ, where μ is the coefficient of friction between belt and pulley, and θ is the angle of wrap in radians. A higher wrap angle (more contact arc) and higher friction increase the drive capacity. The effective tension (T₁ − T₂) determines the power transmitted: P = (T₁ − T₂) × v, where v is belt speed.

Which Span Is the Tight Side?

Students often memorise “the lower side is tight”. It is not a rule — it is a consequence of which way the driver turns. Friction from the driving pulley adds tension to the belt in the direction the belt is travelling, so the span leaving the driver carries T₁ and the span returning to the driver carries T₂. Reverse the motor and the two swap sides. Designers usually arrange the drive so the slack side is on top, because a sagging upper span drapes onto the pulleys and slightly increases the angle of wrap; but that is a design preference, not a law.

The tension does not jump from T₂ to T₁ at a point. It climbs smoothly through the arc of contact as T = T₂e^μφ, reaching T₁ after the full wrap. The simulator draws this directly: the belt is colour-mapped from slack (blue) to tight (orange) so you can watch the tension build round the driver and fall again round the driven pulley. When the two pulleys have different wrap angles, only the smaller one is working at its full friction limit — the larger has friction in reserve.

Will the Belt Slip? Reading the Slip Margin

A belt drive does not slip because it is “loose”. It slips when the load demands a tension ratio the friction cannot supply. Friction can deliver at most T₁/T₂ = e^μθ, so the most power the drive can pass is P_max = (T₁ − T_c)(1 − e^−μθ) v. Compare that with what the driven machine is actually asking for and you get a single number worth more than any rule of thumb:

Slip margin = P_max / P_demanded. Above 1 the belt grips and only creeps — the unavoidable few tenths of a percent from the belt stretching on the tight side and relaxing on the slack side. Below 1 it slips bodily, heats, glazes and fails. Enter the demanded power in the simulator and the canvas reports the margin and a GRIPS / SLIPS verdict; when it slips it also solves the geometry backwards to tell you the idler diameter that would hold the load.

That framing makes the four fixes obvious, in order of what they cost you: raise the wrap angle θ (an idler, or a shorter centre distance); switch to a V-belt, which multiplies the effective friction by 1/sin(β/2); raise the installed tension T₁, at the cost of bearing load and belt life; or raise μ with a different belt facing. Notice that only the first two are free of a penalty elsewhere.

Idler Pulleys — Buying Wrap Angle

When the centre distance is short or the ratio is large, the arc of contact on the small pulley falls away and with it the whole capacity of the drive, because T₁/T₂ = e^μθ is exponential in θ. An idler is how you buy that angle back. It carries no torque; it simply presses on the belt and changes the geometry.

Two rules decide where it goes. First, always on the slack side: on the tight side the idler fights the full tight-side tension, which gives arm chatter, bearing failure and sometimes a snapped belt. Second, to increase the arc of contact it must be a back-side (outside) idler — mounted outside the belt loop so it pushes the slack span inward, toward the line of centres. An idler inside the loop is a take-up device: it tensions a stretched belt and steadies a long span, but it does not add wrap.

The price is reverse bending. A back-side idler bends the belt the opposite way to the pulleys, which costs fatigue life, so its diameter is limited from below: industry guidance is a minimum of 1.5 × the smallest sheave for an outside idler (an inside idler need only match the smallest sheave). The usual allowance is to add about 0.1 to the service factor for an outside idler on the slack side — against 0.2 on the tight side, which is the arithmetic behind the “slack side only” rule.

Position along the span matters as much as size. Guidance is to place an outside idler close to the small sheave, because the small pulley is the one whose wrap governs. In the simulator the idler bracket is fixed at 30 % of the slack span from the smaller pulley — near the practical optimum, since sitting any closer means the wheel collides before it can grow. On the default 300/150 mm drive at C = 600 mm, a 240 mm idler lifts the governing wrap from 165.6° to 199.5°, drops the slack tension from 840 N to 704 N and raises the transmissible power by about 12 %, at the cost of a 544 N load on the idler shaft.

That bearing load is worth computing: since the idler transmits no torque, the belt tension is the same T₂ on both sides of it, and its shaft carries only the resultant of those two pulls, F = 2T₂ sin(θ_i/2), directed along the bisector of the two spans. Grow the idler and that force grows with the angle it turns the belt through.

One more consequence of the slack side: it sags. The sag of a span under its own weight is δ = wL_s²/(8T_eff) with w = mg per metre — but for a moving belt the tension that resists sag is the effective one, T₂ − T_c, because the centrifugal part is already spent holding the belt on its own curved path. At realistic tensions the answer is a fraction of a millimetre, which surprises students who expect a visible droop; the simulator draws it to true scale and prints the value rather than exaggerating it.

Centrifugal Tension and the Maximum-Power Speed

Every belt element travelling round a pulley needs a centripetal force, and the belt supplies it out of its own tension. That component is T_c = m v², where m is the belt mass per metre and v the belt speed. It is present all the way round, so it appears equally on both sides and cancels out of the effective tension — but it still eats into the allowable T₁. With it included the capstan equation applies to the effective tensions: (T₁ − T_c)/(T₂ − T_c) = e^μθ.

This produces one of the classic results of machine design. Power P = (T₁ − T_c)(1 − e^−μθ)v rises with speed at first, then falls as T_c grows with v². Differentiating gives T₁ = 3T_c, so maximum power occurs at v_opt = √(T₁/3m). Beyond that speed a faster belt transmits less power — which is why flat and V-belt drives are rarely run above roughly 25–30 m/s. Switch the centrifugal-tension toggle on in the simulator and push the speed up to watch the transmitted power peak and then fall away.

Open vs Crossed Belt Drives

An open belt drive connects both pulleys so they rotate in the same direction. The angle of wrap on the smaller pulley — whichever of the two that is — is θ = π − 2 sin⁻¹(|R₁−R₂|/C), which is less than π (180°); the larger pulley carries π + 2 sin⁻¹(|R₁−R₂|/C). Note that this is not tied to which pulley is driving: in a speed-up drive the driver is the small one and therefore the one that governs. In a crossed belt drive, the belt crosses between the pulleys, making them rotate in opposite directions. Both pulleys then share the same wrap angle θ = π + 2 sin⁻¹((R₁+R₂)/C), which is always greater than π. Chain drives use toothed sprockets for positive (non-slip) drive, ideal for synchronous applications like camshaft timing — and the capstan equation does not apply because there is no friction-based slipping.

V-Belt Beats Flat Belt — The Wedge Effect

The capstan equation T₁/T₂ = e^μθ contains a hidden factor for V-belts. Because the belt sits in a wedge-shaped groove, the normal force pressing it against the pulley is amplified by 1/sinα where α is the groove half-angle. For a standard V-belt with α = 18°, the friction coefficient effectively increases by 1/sin18° ≈ 3.2 times compared to a flat belt of the same material. Same torque capacity from a much smaller belt cross-section, or much higher capacity from the same size.

This is why V-belts replaced flat belts in cars, factory machines, and HVAC systems by the 1950s. Modern timing belts (toothed, no slip) replaced V-belts in some applications — engine camshaft drives most notably — but for general power transmission V-belts still dominate because they self-tension and tolerate misalignment in ways chains cannot.

Sprocket Teeth, Pitch Diameter and Chordal Action

A chain drive is not a belt drive with teeth painted on. A chain has a fixed pitch p (12.7 mm for ANSI 40, 25.4 mm for ANSI 80, and so on), and a sprocket carries a whole number of teeth z. The two together fix the pitch diameter: D_p = p / sin(π/z). You cannot ask for a 297 mm sprocket — you ask for 37 teeth at 25.4 mm pitch and get 299.5 mm. That quantisation is exactly why a chain gives an exact ratio: VR = z₁/z₂ is a ratio of whole numbers, immune to the wear, stretch and creep that make a belt ratio approximate.

The price is chordal (polygonal) action. The chain wraps the sprocket as a polygon of z sides, so as each link seats the effective radius swings between D_pcos(π/z)/2 and D_p/2. Even at perfectly constant input speed the output speed ripples by 1 − cos(π/z) once per tooth — 3.4 % on an 11-tooth sprocket, 1.7 % on 15 teeth, 0.4 % on 31. This is the reason designers avoid small sprockets in fast drives, and the reason chain drives are usually kept below about 15–20 m/s. Choose a chain pitch in the simulator and the tooth counts, pitch diameters and chordal ripple are all recomputed and drawn.

Chain Speed and Chain Length — Count in Pitches, Not Millimetres

Two chain quantities catch people out because the belt formulas look like they should carry over, and they do not.

Chain speed. A chain does not ride a circle — it rides the polygon its own links form. One turn of a sprocket therefore feeds exactly z pitches of chain, so v = z p N / 60 000 (p in mm, N in rpm), not πD_pN/60 000. The circumscribed circle is bigger than the polygon by π/(z sin(π/z)): only 0.2 % at 37 teeth, but 1.7 % at 10 — and it is the same polygon that produces chordal action, so using the circle for speed while teaching chordal action contradicts itself.

Chain length. A chain is assembled from whole links, and an even number of them unless you accept a cranked offset link, which is weaker. So the length is calculated in pitches and rounded up to the next even integer:

L_p = 2C/p + (z₁+z₂)/2 + ((z₂−z₁)/2π)² × (p/C)

You then trim the centre distance to suit the chain you can actually buy — which is why chain drives are normally designed with an adjustable centre or a tensioner, and why the simulator reports both the calculated and the fitted pitch count.

Timing Belt vs Chain — Why Modern Engines Switched

• Timing belt. Quieter (no metal-on-metal). Lighter. Lower friction. But limited life — typically 100,000 km then mandatory replacement. Catastrophic failure mode: belt snap causes valve-piston collision and ruins the engine head.
• Timing chain. Longer life (often engine-lifetime). More noise, more friction. Stretches over time, causing valve timing drift. Failure mode is gradual rather than catastrophic.
• Modern preference. European premium cars (BMW, Mercedes, Audi) shifted back to chains in the 2000s for the maintenance-free aspect. Mass-market cars (Toyota, VW, Ford) often use belts because the manufacturing cost is lower. Aftermarket belt-replacement work generates substantial service revenue.

Reference Data

ANSI / ASME B29.1 Roller Chain Sizes

The chain number encodes the pitch in eighths of an inch: a #40 chain is 4/8″ = 12.70 mm, a #120 is 12/8″ = 38.10 mm. Minimum ultimate tensile strength (M.U.T.S.) is the breaking load — working load is a small fraction of it after service and speed factors.

Dimensions and strengths per ASME B29.1-2011. Note the internal pattern: M.U.T.S. ≈ 12 500 × p² lbf with p in inches, and roller diameter ≈ 5/8 of the pitch. kN values are converted at 1 lbf = 4.44822 N.

V-Belt Pulley Groove Angles

A classical V-belt is moulded with a 40° included angle, but the pulley groove is cut narrower — 32°, 34°, 36° or 38° (±0.5°) per ISO 4183. The reason is that a belt bent round a pulley bulges outward and its flank angle opens up; the smaller the pulley, the tighter the bend and the more it opens, so the groove must start narrower. Grooves are therefore selected by pulley diameter, not by preference.

This is the whole reason a V-belt out-pulls a flat belt of the same material: the wedge multiplies the normal force, so a real μ of 0.30 behaves like 0.92–1.09 in T₁/T₂ = e^μ′θ. Groove-angle-by-diameter bands per ISO 4183.

Smallest Practical Pulley Diameter by V-Belt Section

Below these diameters the belt is bent too sharply: the flanks distort, the belt runs hot and fatigue life collapses. Values taken from the smallest pulley covered by each section in a classical V-belt manufacturer's tensioning tables. Wedge sections (SPZ, SPA, SPB, SPC) allow smaller pulleys for the same power because the belt is deeper relative to its width.

Where an Idler Can Go

Only an outside (back-side) idler increases the arc of contact; an inside idler is a take-up device. Slack-side placement is not a preference — on the tight side the idler must react the full tight-side tension, which gives arm chatter, bearing failure and sometimes a snapped belt.

References

• Shigley & Mischke — Mechanical Engineering Design, 10th ed., Chapter 17 (Flexible Mechanical Elements).
• Gates Corporation — Industrial Power Transmission Drive Design Manual. The standard commercial reference.
• ISO 5292:1995 — V-belt drives — Calculation of power capacity.
• ISO 1081:2013 — Belt drives — V-belts and V-ribbed belts, and corresponding grooved pulleys.
• ISO 4183:1995 — Belt drives — Classical and narrow V-belts — Grooved pulleys (system based on datum width). Source of the 32/34/36/38° groove angles.
• ASME B29.1-2011 — Precision Power Transmission Roller Chains, Attachments and Sprockets. Source of the chain size table.
• R. S. Khurmi & J. K. Gupta — A Textbook of Machine Design, Chapter 18 (Flat Belt Drives). Initial tension, centrifugal tension, creep and the maximum-power condition.

Belt & Chain Drive Formulas

Explore Related Simulators

For more power transmission topics, explore our gear train simulator, four-bar linkage simulator, flywheel dynamics simulator, bearing selection calculator, and shaft torsion simulator for complementary topics in mechanisms and machine design.