How to Measure Roller Chain Stretch: Why 3% Is Not Your Replacement Point

Key Specifications

Types & Variations

Maintenance Steps

1. Isolate all energy and expose a straight span
   Stop the drive. Mechanically restrain the take-up carriage, counterweight, and any raised load, then apply lockout/tagout (LOTO), release stored spring energy under the approved procedure, and bleed hydraulic or pneumatic pressure to zero. Verify isolation with the normal start control and return it to off before removing the guard.

   A stopped chain can still move when a suspended load drops, a take-up releases, or trapped pressure shifts an actuator. Restrain those first. A start button or selector switch is not an energy-isolating device and does not replace the lock.

2. Identify pitch and mark the measurement span
   Confirm the chain series and pitch from its marking, drawing, or manufacturer data. Select at least 6 to 10 pitches on a straight run and paint-mark both end rollers.

   A one-pitch counting error invalidates the result. Do not infer pitch from sprocket appearance.

3. Pull out slack and take two readings
   Apply light tension in the working direction. Measure across the outside faces of the two marked rollers, then across their inside faces. Average the readings to obtain the span between roller centers.

   Keep caliper jaws square and fingers clear of the sprocket pinch point. Never jog the motor or use powered movement while the guard is open.

4. Calculate and repeat around the chain
   Subtract nominal span from measured span, divide by nominal span, and multiply by 100. Repeat at three marked locations and retain the highest repeatable result. Log the reading against operating hours so the next inspection yields a wear rate.

   Recover tools, reinstall the guard, clear people from the drive, and remove LOTO before indexing an endless chain to the next location. Stop, isolate, and verify again before touching it.

5. Inspect sprockets and document the decision
   Compare the result with the lowest applicable manufacturer limit. Inspect for hooked teeth, poor seating, tight joints, cracked plates, loose pins, and seized rollers. Record the reading and replacement decision.

   Replace damaged chain regardless of measured elongation. A percentage cannot prove that plates, pins, or a connecting link are safe.

Tools, records, and a safe setup

Bring a steel rule long enough to cover the chosen span, or calipers for a shorter one, plus a paint marker and the chain drawing or pitch data. Check the tool against the decision you have to make. Ten pitches of ANSI 50 chain measure 158.75 mm (6.25 in) new and 161.13 mm (6.344 in) at 1.5%, so the whole judgement lives inside about 2.4 mm (0.09 in). A rule graduated in millimetres barely resolves that; over six pitches it does not. A chain wear gauge is faster for routine rounds, but it must match the chain series and pitch, and it reads a band rather than a number.

Stop the machine with a straight, accessible section of chain under light tension. Mechanically restrain the take-up carriage, counterweight, and any raised load first, because those move on their own once the drive is free. Then lock out/tagout (LOTO) every energy source and bleed hydraulic or pneumatic pressure to zero. Try the normal start control afterwards and return it to off. That control is a test, not an isolating device; the guard comes off only after the lock is on and the test comes up dead.

Clean enough dirt from the roller faces to seat the tool. Do not solvent-wash the joints before measuring; that carries grit into the pin and bushing area and strips out the lubricant. Record the asset tag, chain designation, operating hours, measurement location, number of pitches, nominal span, and measured span. One percentage with no history behind it cannot tell you how fast the chain is going.

Measure a lightly tensioned span, never the slack side

Choose a straight run and mark the first and last roller. Draw the span lightly taut in its normal load direction so every pin and bushing clearance seats the same way. Tsubaki states the requirement plainly: measure the chain while stretching it slightly to eliminate any slack. Hand force is enough on a small isolated drive; use the machine maker's approved take-up method on a large conveyor. Never bump the motor to tighten the span.

Measure at least 6 to 10 pitches, and use more when access and tool length allow. Tsubaki publishes its reference lengths over 6 and 10 links for that reason, and tells you to cover as many links as possible when a tape measure is the only option. One pitch is too sensitive to roller clearance, dirt, and jaw seating. Take one caliper reading across the outside faces of the end rollers and a second across their inside faces, then average the two. That average cancels one roller diameter and gives the span between roller centres without hunting for the centres.

Count pitches, not visible side plates. Nominal span is chain pitch multiplied by the number of pitches between the marked rollers, so one miscounted pitch swamps the wear you are trying to read. Marking both end rollers is what prevents it.

Calculate wear elongation and make the call

Wear elongation (%) = (measured span - nominal span) / nominal span x 100. Ten pitches of 15.875 mm (0.625 in) chain are 158.75 mm (6.25 in) new; measure 161.13 mm (6.344 in) and you are at Tsubaki's 1.5% limit.

Which limit applies depends on the drive, and the two most-quoted numbers come from different places. Tsubaki's 1.5% is a smooth-transmission figure for a large sprocket of up to 60 teeth, and it falls to 1.2% at 61 to 80 teeth, 1.0% at 81 to 100, and 0.8% at 101 to 110 — big sprockets simply run out of tolerance for mispitched rollers sooner. Renold's 3% is the point at which its wear guide calls for immediate replacement, with 2% as the point to start planning one. Read 3% as the end of the road rather than the target, and take the lowest limit that applies. Indexing, reversing, shock loads, fixed centre distances, and chains running in parallel all pull the decision earlier.

Damage overrules the percentage. Stop there when plates are cracked, pins have turned or walked, rollers are missing or seized, the chain has tight joints, or it rides high and snaps into the sprocket. Elongation measures accumulated joint wear; it does not clear the chain for fatigue damage, corrosion, or a bad connecting link.

Common mistakes that spoil the reading

Measuring the slack strand without pulling out joint clearance gives a low and inconsistent result. Measuring over one or two pitches lets roller clearance dominate the number. Holding caliper jaws at an angle adds length, while seating them on grease or burrs can move the result either way.

Another common error is checking only the easiest spot. Wear runs uneven after poor lubrication, contamination, or a stiff link cycling through the small sprocket. Measure at least three marked locations around the chain and keep the highest repeatable reading. Moving an endless chain to the next location means recovering tools, reinstalling the guard, clearing people from the drive, removing LOTO under the site procedure, indexing the chain, then stopping and re-isolating and re-verifying before anyone touches it again.

Temperature quietly moves the number too. A steel span read hot off a long run is not comparable with a cold baseline taken at the last shutdown, and over a long tape measurement that difference is the same size as a few tenths of a percent of wear. Take the trend under similar thermal conditions, and note the state on the record.

Do not use take-up travel as chain elongation. Sprocket wear, center-distance movement, temperature, and adjustment history all change take-up position. Do not remove one pitch to recover adjustment range either. The worn pin and bushing joints remain worn, and the shortened chain still seats badly.

Verify the result before returning the drive to service

Repeat the highest reading with the tool reseated. Two measurements should agree within the resolution of the rule or caliper; a larger spread means the span was not held consistently or the jaws were not square. Compare the result with the chain maker's limit and the machine manual, then inspect every accessible plate, pin, roller, and connecting link.

Check both sprockets before fitting a new chain. Hooked teeth, thinned tooth flanks, uneven wear, runout, or a new chain that will not drop fully into the gullets are reasons to replace the sprocket as well. Tsubaki is direct about it: a new chain run on a worn sprocket wears faster than normal, so replace the sprocket with the chain. Then find out why the old chain wore — lubrication method, shaft parallelism, sprocket alignment and runout, and an over-tightened take-up all wear a fresh chain the same way in the same number of hours.

Set slack or take-up to the equipment specification, lubricate the pin and bushing joints, recover every tool and marker, reinstall the guard and its fasteners, move people clear, and remove LOTO under the site procedure. Follow the manufacturer's guarded test-run procedure and watch from outside the guard for smooth seating, stable slack, noise, and vibration. Recheck tension after the run-in interval.

Keep the baseline, because the rate carries the decision. Two readings at the same location, same pitch count and similar temperature give you a slope: 0.4% gained over 2,000 operating hours is 0.2% per 1,000 hours, so a chain sitting at 0.9% against a 1.5% limit has roughly 3,000 hours left. Compare that against the shutdown calendar and the chain gets ordered before it strands the line, not after.

FAQ

References

• Tsubaki, Installation and Maintenance Guide
• Tsubaki, Chain Wear Indicator
• Renold, Chain Wear Guide Instructions
• iwis, Wear Resistance Testing
• iwis, CCM-S Chain Condition Monitoring
• OSHA, Control of Hazardous Energy (Lockout/Tagout), OSHA 29 CFR 1910.147