In this guide
Between the motor and the load, most student machines — conveyors, vehicles, mills, robot drivetrains — need a flexible power-transmission stage that tolerates imperfect shaft alignment and covers a longer centre distance than gears allow. The two candidates are belt drives and chain drives. They are not interchangeable: belts slip and stretch, chains need lubrication and are noisy, and the wrong choice shows up as a stalled conveyor, a thrown chain at demo, or a belt that melts in a month. This guide compares them properly and gives you a selection procedure you can run with a calculator.
The three belt types and the one chain type that matter
V-belts (A, B, Z, SPZ sections) transmit power by friction between the belt's angled flanks and the pulley groove. They are cheap, forgiving of misalignment, and slip under overload — which is both their weakness (lost motion, heat) and their safety feature (the belt slips instead of breaking the machine). Classical V-belts are the default for mill and pump drives.
Timing (synchronous) belts — GT2, HTD-3M/5M, T5 profiles — have teeth that mesh with grooves in the pulleys: no slip, exact speed ratio, efficiency near 98%. This is what drives 3D-printer axes (GT2) and what you want whenever the motion must stay synchronised. Their limits: less shock absorption than V-belts, and the teeth wear if pulleys are misaligned.
Flat/poly-V belts appear in some machine tools; for student work you will mostly meet V and timing belts.
Roller chains (ANSI 25, 35, 40 — the number encodes pitch in eighths of an inch: 40 = 1/2" pitch) are metal link chains on toothed sprockets. Positive drive like a timing belt (no slip), very high torque capacity for the size, tolerant of heat and dirt — at the cost of needing lubrication, regular tensioning, stretching over time (chains "stretch" as pins and bushings wear, lengthening the pitch), and noise.
The one-line characterisation: belts are quiet, clean, and forgiving; chains are strong, positive, and demanding.
Head-to-head comparison
| Criterion | V-belt | Timing belt | Roller chain |
|---|---|---|---|
| Speed ratio exactness | No — 1–3% slip | Yes — toothed, no slip | Yes — toothed, no slip |
| Efficiency (well set up) | 0.90–0.95 | 0.96–0.98 | 0.95–0.98 |
| Shock-load absorption | Excellent (slips) | Fair | Poor (rigid; use a tensioner) |
| Maintenance | Re-tension periodically; replace when cracked/glazed | Minimal; check tension and alignment | Lubricate regularly; monitor stretch; re-tension |
| Noise | Quiet | Quiet | Noisy, worse as it wears |
| Dirt tolerance | Good | Fair (grit wears teeth) | Good if lubricated; grit + no lube = fast wear |
| Centre distance | Long spans fine | Long spans fine | Long spans fine (support long runs) |
| Overload behaviour | Slips (protects machine) | Teeth shear or belt snaps | Chain snaps or sprocket teeth break |
| Cost | Low | Medium | Medium (chain + sprockets + lube) |
| Typical student use | Mills, pumps, fans | 3D printers, CNC, robot drivetrains, conveyors needing sync | E-bikes, go-karts, heavy conveyors |
The overload row is the one to design around: on a parcel-sorting conveyor, a jammed parcel with a V-belt drive means the belt slips and squeals — annoying but harmless. The same jam on a chain drive with no slip clutch snaps something. If your load can jam, either choose a slipping belt or add a torque limiter to a chain drive.
The selection procedure
Run these steps in order; each one eliminates an option:
Step 1 — How much power, and what kind of load? Compute or estimate the transmitted power. Under ~0.5 kW with smooth loads, any of the three works. Above ~1 kW, or with heavy shock loads (crushers, dehuskers, vehicle drivetrains), chains pull ahead on torque density. Apply a service factor: multiply the nominal power by 1.2–1.5 for conveyors and machine tools, 1.5–2.0 for shock-loaded or stop-start duty. Size the drive for service power, not nominal.
Step 2 — Must the ratio stay exact? If the application counts revolutions — a conveyor object sorter that indexes one parcel pitch per motor revolution, a CNC axis, any synchronised multi-axis machine — V-belts are out. Timing belt or chain only.
Step 3 — What is the environment? Dusty, outdoor, no maintenance culture (a farm installation): chain with proper lubrication, or a sealed timing-belt drive. Clean indoor demo: any. Oily environment: chains tolerate oil; belts (especially V-belts) degrade in oil — keep belts away from lubricant spray.
Step 4 — Noise and cleanliness constraints? A demo in a classroom or exhibition hall argues for belts. A go-kart or e-bike does not care about chain noise.
Step 5 — Can it be maintained? Chains demand a maintenance owner: regular lubrication and stretch checks. If the machine will be handed over and never touched again, a timing belt (fit and forget for thousands of hours) is the safer engineering choice.
For most student conveyors and light vehicles, the decision lands here: timing belt for clean, quiet, exact-ratio indoor machines; roller chain for high-torque, dirty, outdoor machines; V-belt where slip-as-fuse and low cost matter most.
Sizing math: belt length and chain links
Belt drive: with driver diameter d, driven diameter D, and centre distance C, the approximate belt length is:
L ≈ 2C + π(D + d)/2 + (D − d)²/(4C)
Pick C to suit the machine layout, compute L, then buy the nearest standard belt length and adjust C slightly to suit (most designs include a slotted motor mount for tensioning — design that in from the start). Speed ratio is exact for timing belts: N_driven = N_driver × (d/D) using pitch diameters; for V-belts subtract a couple of percent for slip.
Chain drive: chain length in links must be an integer (plus the connecting link). With sprocket teeth z_1, z_2, pitch p, and centre distance C:
L_links ≈ 2C/p + (z_1 + z_2)/2 + p(z_2 − z_1)²/(40C²)
Round to an even integer where possible (even link counts allow a standard connecting link; odd counts need an offset link, which is weaker). Then solve back for the exact C. Centre distances around 30–50 pitches are the sweet spot — shorter and the chain articulates too often (fast wear); much longer and you need an idler or support.
Worked example: conveyor drive
A small sorting conveyor needs 0.4 kW at the head pulley, running 140 rpm, driven by a 1400 rpm motor — ratio 10:1. Service factor 1.4 (stop-start sorting duty) → design power 0.56 kW. Ratio must stay exact (indexing) → timing belt or chain. Indoor demo, quiet preferred, low maintenance → timing belt. A two-stage reduction keeps pulleys reasonable: 1400 → 440 rpm (HTD-5M, 3.2:1), then 440 → 140 rpm (3.1:1). Each stage's belt sized from the manufacturer's power-rating charts at the design power. The parcel-sorting conveyor project and the IR height-sensor sorter both live in this design space — study how their drives are staged.
Worked example: line-follower drive
A line-follower robot needs compact, light, exact-ratio drive from motors to wheels. The honest answer here is usually direct drive or gears, not belts or chains — but if a belt/chain stage is needed (motor mounted away from the wheel), a GT2 timing belt is the pick: light, no slip (odometry stays honest), no lubrication near electronics. Chains are overkill and messy at this scale; V-belts' slip would corrupt any encoder-based distance estimate.
Installation: where belt and chain drives actually fail
The hardware is rarely the problem — the installation is. The failure list from real student builds:
- Misaligned pulleys/sprockets. Even 1–2° of angular misalignment makes timing belts climb and shred their edges and makes chains wear one side of the sprocket teeth. Align with a straightedge across the pulley faces, not by eye.
- Wrong tension. Too loose: V-belts slip and squeal, timing belts ratchet (teeth jump), chains slap and derail. Too tight: bearing loads skyrocket and belts/chains die early. The thumb rule: at mid-span, the belt/chain should deflect about 1–2% of the span length under moderate thumb pressure. Design in a slotted or eccentric tensioning mount — "we will tension it somehow later" never works.
- No guard. A chain at speed will take a finger. A snapped belt whips. Guard every drive that a person can reach — exhibitions and viva demos put curious hands near everything.
- Chain lubrication neglected. A dry chain in dust wears 5–10× faster than a lubricated one. If the machine goes somewhere dusty, fit a simple drip or felt-pad oiler, or switch the selection to a sealed timing belt.
- Mixing worn and new. A new chain on worn sprockets (or vice versa) wears at the rate of the worn part. Replace chains and sprockets as a set.
Quick checklist
- Design power computed with service factor (1.2–1.5 smooth, 1.5–2.0 shock/stop-start).
- Exact ratio required? → timing belt or chain (V-belt excluded).
- Environment checked: dust, oil, noise, maintenance reality.
- Belt length / chain link count computed; standard size selected; centre distance adjustable via slotted mount.
- Pulleys/sprockets aligned with a straightedge; tension set to ~1–2% mid-span deflection.
- Overload path defined: belt slip, torque limiter, or shear pin — something gives before the machine breaks.
- Guard fitted over the drive before any public demo.
- For chains: lubrication plan and stretch-check interval written down; chain + sprockets replaced as a set.
Combine this with the gear ratio companion guide (same series) — most real drivetrains are a gearbox plus a belt/chain final stage, and the ratio math multiplies across both. And when the motor at the front of that drivetrain still needs choosing, the servo vs stepper companion guide (same series) closes the loop from load, through transmission, back to the motor.
Chain wear: the 12-link measurement
Chains do not stretch elastically — they wear longer as pins and bushings lose material, increasing the pitch. The standard check: measure 12 full links (13 pins) under light tension. For ANSI 40 chain (1/2" pitch), 12 links should measure 6.000". If it measures 6.090" (1.5% elongation), the chain is worn; at 3% (6.180") it is overdue for replacement and is actively wearing the sprocket teeth into hooked profiles. A steel ruler and two minutes is all this takes, and it is the difference between replacing a ₹300 chain and a ₹300 chain plus two ₹500 sprockets. Record the measurement in the machine's logbook — elongation rate tells you whether your lubrication regime is working.
Timing-belt width and tension: the two sizing details
Belt power capacity scales roughly with width: an HTD-5M belt at 15 mm width handles roughly twice the power of the same belt at 9 mm. When a drive is marginal, widening the belt is usually cheaper than jumping to a larger pitch. And tension on a timing belt is a window, not a point: too loose and teeth ratchet under load (a loud bang and lost position); too tight and bearing loads and belt fatigue soar. Most manufacturers publish a tension-vs-span specification — follow it rather than the thumb rule once you have real numbers. A simple spring scale pulling mid-span to the specified deflection is the honest way to set it.
Idlers, tensioners, and long spans
- V-belts: an idler on the slack side (never the tight side) takes up stretch. Spring-loaded idlers auto-compensate for wear.
- Timing belts: use a toothed idler inside the loop or a smooth idler outside it — a smooth idler pressing on the tooth side damages teeth. Backside idlers must be larger diameter than the minimum bend radius for the belt.
- Chains: an adjustable idler sprocket (with teeth, engaging properly) on the slack side. For very long spans, add a guide or support to stop the slack side whipping.
- Minimum pulley/sprocket sizes: every belt and chain has a minimum bend radius / minimum sprocket tooth count (e.g. 12–14 teeth minimum for small roller chain). Below it, fatigue life collapses. Respect the manufacturer's minimums — they are not suggestions.
Cost comparison: what a drive stage really costs
For a typical student conveyor stage (~0.5 kW, 3:1 reduction), indicative Indian retail:
- V-belt: two pulleys (₹200–400) + belt (₹150–300) ≈ ₹400–700. Cheapest; needs re-tensioning.
- Timing belt (HTD-5M): two pulleys (₹400–800) + belt (₹250–500) ≈ ₹700–1,300. Fit-and-forget.
- Roller chain (ANSI 40): two sprockets (₹300–600) + chain (₹250–400) + lube ≈ ₹600–1,100, plus ongoing maintenance.
The purchase price is only half the story — factor in the maintenance you will actually do. A chain drive that never gets lubricated costs more than a timing belt within a year, in replacement parts. Budget honestly with the project cost guide, and sanity-check the whole machine's fault-finding approach with the testing and debugging guide before the demo, not during it.
Troubleshooting table: what the drive is telling you
| Symptom | Likely cause | Check |
|---|---|---|
| V-belt squeals on startup | Under-tensioned or glazed belt | Re-tension; if the belt sides are shiny/glazed, replace it |
| Timing belt teeth jumping (loud bang) | Too loose, or shock load exceeding tooth shear | Tension to spec; check for jams; consider wider belt |
| Belt edges fraying | Misaligned pulleys | Straightedge across pulley faces; realign |
| Chain slapping at speed | Too loose, or worn beyond 1.5% elongation | Tension; measure 12 links; replace if elongated |
| Chain derailing | Misaligned sprockets or excessive wear | Align; check sprocket tooth profile for hooking |
| Rust-coloured dust near chain | Fretting from dry running — lubrication failure | Clean and lubricate immediately; check elongation |
| Bearings running hot near the drive | Over-tensioned belt/chain | Back tension off to spec; check bearing condition |
| Black dust under a V-belt drive | Belt wearing — misalignment or overload | Align; verify design power vs service factor |
Read the symptoms early: a drive that changes its sound is announcing a problem weeks before it becomes a failure. The five minutes spent investigating a new squeal beats the five hours rebuilding after a snapped chain — the same diagnostic habit the testing and debugging guide teaches for electronics applies verbatim to mechanisms.
Measuring drive efficiency on the bench
Catalogue efficiencies are for clean, aligned, properly tensioned drives. Yours will be worse — the question is how much worse. A simple bench test: drive the input with a known motor, load the output with a friction brake or a second motor as a generator, and measure input electrical power versus output mechanical power (torque × speed, with torque from a spring scale on a lever arm of known length). If a "95% efficient" chain stage measures 85%, something is wrong — usually tension, alignment, or lubrication. Students rarely measure this, which is why "the motor should be strong enough but the conveyor barely moves" is such a common complaint. Measure, fix the installation, measure again: the efficiency number is the report card for your build quality.
Synchronising multiple shafts: one motor or many?
A sorting line with two conveyors that must run at an exact speed ratio has two architectures:
- One motor, mechanical sync (timing belts/chains linking the conveyors): the ratio is exact forever, with zero control effort. The parcel-sorting conveyor family leans this way — mechanical sync cannot drift.
- One motor per conveyor, electronic sync (encoders + PID trimming the follower to the leader): more flexible (ratios change in software), but now you own the control problem, the sensor problem, and the failure modes of both.
For student builds, mechanical sync wins on reliability per rupee in almost every case. Choose electronic sync only when the ratio genuinely needs to change during operation.
Storage and spares
- Belts: store cool, dark, unstressed — never hung on a nail (it takes a set) or folded sharply. A timing belt with a kinked tooth section will jump at that spot forever.
- Chains: oil lightly before storage; a rusted chain is a worn chain before it ever runs.
- Keep one spare belt/chain of each size used in the machine, plus a connecting link. Drives fail at the worst possible time — the night before the demo — and local shops may not stock your exact HTD-5M length.
When gears beat belts and chains
End with the honest boundary: for short centre distances, high precision, or compact high-reduction packaging, gears (this guide's companion topic) beat both belts and chains. The experienced builder's drivetrain is often a hybrid — a gear stage for the big reduction near the motor, then a belt or chain final stage to reach a distant shaft. Size each stage with the right tool: the gear ratio companion guide (same series) for the gears, this guide for the flexible stage, and the servo vs stepper companion guide (same series) for the motor at the front.
Belt and chain in the same machine: a real conveyor layout
A realistic student sorting conveyor often uses both: a timing belt from the motor to a jackshaft (quiet, clean, near the electronics), then a roller chain from the jackshaft to the head pulley (tolerant of the dusty conveyor environment and easy to re-tension as it wears). The ratio math multiplies across both stages exactly like compound gears — i_total = i_belt × i_chain — with each stage's efficiency multiplied in. This hybrid is not indecision; it is putting each technology where its strengths matter: the belt where cleanliness and quiet count, the chain where dirt and torque dominate. Sketch the full power path from motor to load on one page before ordering parts — motor → stage 1 → stage 2 → load, with ratio, torque, and efficiency at each step. That single sketch is the document that keeps a multi-stage drivetrain coherent, and it is the first thing an examiner will ask to see.
One last habit: photograph the drive at installation — pulley alignment, tensioner position, chain routing — and keep the photo with the machine's documentation. Six months later, when something looks "off," that photo is the baseline that tells you what changed. Drives drift gradually; memory does not keep up, but a photo does.