Belts and Chains: Tension, Wrap, and Why Too Tight Kills the Bearings

Measure the span between the two pulley centres. Press the belt at the midpoint with a known force and measure how far it moves. The specification is a deflection of roughly 1.6 mm per 100 mm of span, at a force the belt maker states. That gives a number instead of a feeling, and it is repeatable by somebody else — which is the same argument as weighing salt rather than measuring it in spoons. Re-tension after the first few hours of running. A new belt seats into the pulley grooves and stretches, so it loses a noticeable fraction of its tension almost immediately. Most belt drives that slip were tensioned once, when new, and never again.

A roller chain transmits force through the engagement of rollers and teeth, so it needs no tension to grip. Leave roughly 2% of the span as sag on the slack side — for a 500 mm span, about 10 mm. A tight chain loads the bearings exactly as a tight belt does, and it also forces the rollers hard into the sprocket teeth, so both wear faster. A chain that 'stretches' has not stretched: the pins and bushes have worn, and the pitch has grown. Oil is what prevents that wear, and it has to reach INSIDE the rollers where the pins turn. Oil applied to the outside of a chain lubricates nothing that matters; it goes on the inner edge of the slack run, where the links open slightly and draw it in.

Two pulleys that are not in the same plane make the belt run onto the flange, rub, heat up and wear on one edge. The same misalignment on a chain drive makes it climb the sprocket teeth and eventually jump. Check with a straightedge across both pulley faces, or a string line. On a V-belt drive the tolerance is about half a degree; on a timing belt or chain it is tighter still. A belt that wears on one side only, or a chain that is bright on one face of its plates, is reporting misalignment rather than bad luck. That is the diagnostic habit the last rung of this batch is about — the wear pattern names the fault.