grokkingstuff Home Blog Projects Wiki Calculators About

Power Transmission

date2023-02-11tags:meche:

Thesis

Belts, chains, and gears are the three families of mechanical power transmission, and the selection among them is not a matter of "which is best" but of which failure mode you can tolerate. Belts slip gracefully and quietly, but they slip — power is lost and, under shock, transmitted torque is capped by friction. Chains carry high torque at modest speed without slipping. They are noisy, need lubrication, and fail violently. Gears are precise, compact, and expensive. They demand tight shaft alignment. Every transmission choice is a vote for a particular bundle of advantages and a particular failure mode, and the honest selection logic begins by asking which failure the rest of the system can survive.[fn::The classical selection literature frames this as a cost-or-efficiency optimization, but in practice the binding constraint is usually the failure mode. A timing belt that strips its teeth under shock leaves you coasting; a chain that breaks at speed can flail and damage adjacent structure; a gear tooth that pits can shed metal into the lubricant and propagate damage through the train. The question "which is cheapest?" is almost never the question that matters; the question is "which fails the way I can afford?"]

Belts, Chains, Sprockets and Gears

A robot must transmit motor power to wheels or mechanisms. Speed and torque often need altering. Belts, chains, sprockets, and gears accomplish this. This module provides insights into mechanical techniques for transmitting mechanical power as well as how to calculate gear (and sprocket) ratios:

Belt Types

Belts transmit power by friction (flat, V) or positive engagement (timing/synchronous). This distinction governs most design decisions.

Power capacity scales with belt width, cord tensile strength, and speed; speed limits come from centrifugal tension (the belt's own mass pulls it off the pulley as speed rises) and from flex fatigue at the pulley entry. A typical timing belt tops out around 30–50 m/s surface speed; flat belts can exceed this because they flex more easily over small pulleys.

Chain Types

Chain speed limits come from the polygon effect. Sprockets are not circles, so the chain rides up and down each tooth. This produces velocity ripple inversely proportional to tooth count. Impact at engagement also limits speed. High-tooth-count sprockets reduce both. A sprocket with 12 teeth produces roughly 3× the speed ripple of one with 36 teeth, and small sprockets are the usual culprit when a chain drive is noisy or short-lived.[fn::The polygon effect is why drives almost never use sprockets below ~17 teeth if they can avoid it, and why the old "hunting tooth" rule exists: choose sprocket and chain tooth counts with no common factor so that the same chain link does not hit the same sprocket tooth every revolution, distributing wear. A 17:25 ratio hunts; a 16:32 ratio does not and wears the same teeth into early failure.]

Efficiency Comparisons

The literature reports wide ranges; representative numbers under typical, well-maintained conditions:

Drive typeEfficiency (typical)Notes
Timing belt, well-tensioned96–98%drops with small pulleys, high speed
V-belt, properly tensioned93–97%flex losses dominate; slip if under-tensioned
Flat belt95–99%best of the belts; low flex, high speed
Roller chain, lubricated96–98%unlubricated drops to ~90% fast
Spur gear, oil-bath98–99%per mesh; a train multiplies losses
Worm gear50–90%high ratio, low efficiency; self-locking often a feature

The steelman of the belt against the gear: at the same ratio and power, a belt drive is cheaper, isolates shock, tolerates shaft misalignment, and is often more efficient than a multi-stage gear train because it does the ratio in one step. The steelman of the gear: it is compact, handles reversible and bidirectional loads cleanly, and has the best efficiency per unit size at high reduction. Belts win on cost and forgiveness; gears win on density and precision; chains sit in the middle, carrying torque like a gear but running like a belt.

Tensioning

Belt and chain drives both need controlled slack-side tension to engage the teeth (timing belt) or maintain wrap and prevent skipping (V-belt, chain).

The rule: a belt or chain that is "just right" today will be loose in 100 hours as it beds in and stretches. Design for retensioning, not for a set-and-forget assembly — unless you have chosen a center distance that takes up the initial stretch elastically (which is what fixed-center printer builds do, at the cost of a stiffer frame requirement).

Shaft Alignment

Power transmission assumes the input and output shafts are where the design put them. In practice they are not — tolerances, frame flex, thermal growth, and assembly error all conspire. The consequences differ by drive type:

The alignment discipline scales with the precision of the drive: a V-belt drive verified with a straightedge is fine; a gear drive demands dial-indicator or laser alignment, and a high-speed gear train demands it under hot and loaded conditions because the cold, static alignment is not the operating alignment.[fn::Thermal growth is the silent alignment killer. A steel gearbox case grows ~12 µm per meter per °C; a 50 °C temperature rise across a 1 m center distance shifts the shaft by 0.6 mm — enough to destroy a precision gear mesh. The cure is to either align hot (run, shut down, re-align immediately) or to compute the expected growth and pre-offset the cold alignment. Most industrial alignment standards (ANSI/ASA) now specify hot-alignment checks as a separate procedure.]

Selection Logic

A rough decision sequence, with the caveat that every rule below has exceptions driven by specific constraints (space, cost, environment, duty cycle):

1. Is position synchrony required? Yes → timing belt or gear. No → V-belt or chain. 2. Is the reduction ratio high (>~8:1)? Single-stage gear is hard above 8:1 for spur; worm or multi-stage gear or belt-chain combo. Chain and V-belt single-stage ratios typically top out around 6:1 for reasonable sprocket sizes. 3. Center distance? Large center distance favors belts and chains (the length is cheap); short center distance favors gears. A 2 m center-distance gear train is absurd; a 2 m timing belt is routine. 4. Shock or reversal? Belts slip (good for shock, bad for precision); chains and gears transmit it. If the load reverses, timing belts and gears are bidirectional; V-belts and flat belts are not (slip on reversal). 5. Environment? Heat, oil, abrasion favor chain or gear (metal survives) over belts (polymer degrades). Clean, dry, quiet favors timing belt. 6. Cost? For equivalent power, generally belt < chain < gear, though gear cost is dominated by the gear itself while belt/chain cost is dominated by the sprockets and the tensioning hardware.

Lifting Mechanisms

Robotics competitions often involve lifting objects, which is a power-transmission problem: converting motor torque into vertical force via screw jacks, winches, or linkages. Merged from the former Lifting Mechanisms note.

Lifting is the high-torque, low-speed corner of the power-transmission space, and the selection logic inverts: here the screw jack (high ratio, self-locking, slow) and the winch (chain or cable on a drum) dominate, and the belt's slip is a liability (a slipping belt under load drops the load) rather than a protective feature. The governing tradeoff in lifting is not efficiency but holding — does the mechanism hold the load with power off? A worm gear or a self-locking screw does; a timing-belt drive does not, and a brake must be added, which is itself a failure-prone component.

Related