What if you never changed your motorcycle chain?

A chain does not stretch like rubber. It wears at its 118 tiny pins until it no longer fits the sprocket, and then it either jumps a tooth or breaks.

Close-up of a motorcycle roller chain made by Enuma Chain, showing plates, pins and rollers
Photo: Ohgud.kibn7ewyu2, CC BY-SA 4.0, via Wikimedia Commons

If you never changed your motorcycle chain, it would not stretch the way rubber does; it would wear. Each of its 118 pins (on a KTM 390 Adventure's 520 chain) rubs inside its bushing every time the chain bends onto a sprocket, the clearance grows, and the chain gets longer. At about 1.5 to 2% elongation it passes a typical service limit, sits higher on the sprocket teeth and starts hooking them; by about 3%, Japanese chain researchers note, a chain begins skipping teeth. After that it can jump off, jam the rear wheel or snap. Worked out below: about 35,000 km at best, 9,000 to 17,000 if ignored.

Key takeaways
  • A chain does not stretch; its pins and bushes wear, and each pitch grows a few micrometres.
  • 317.5 mm is 20 links of new 520 chain; Kawasaki's service limit is 323 mm, about 1.7% longer.
  • At about 3% elongation, chains begin skipping sprocket teeth.
  • A worn chain rides about 1.8 mm higher on a 41-tooth sprocket and hooks its teeth, so change sprockets with the chain.
  • A 390 single at full power in a low gear pulls about 5.9 kN, 16% of a 36.5 kN chain (worked out below); rust and shock loads eat that margin.

So how many kilometres? Here is the worked estimate, with these example inputs. A maintained final-drive chain lasts about 20,000 km on average (Ducati service bulletin, 2019); take that as the distance to a 1.73% service limit, and assume the wear keeps the same pace up to the 3% where chains start skipping (Noguchi et al.).

Then assume a never-cleaned, never-oiled chain wears 2 to 4 times faster: that multiplier is my assumption, not a measured figure, because Tsubaki only says a chain stripped of lubricant wears along a nearly straight, shorter line.

\[ d_{skip} = 20{,}000 \times \frac{3}{1.73} \approx 35{,}000\ \text{km}, \qquad \frac{35{,}000}{4} \text{ to } \frac{35{,}000}{2} \approx 9{,}000 \text{ to } 17{,}000\ \text{km} \]

Calculated in this post from Ducati's 20,000 km average chain life, Kawasaki's 1.73% service limit (worked below) and the 3% skipping point from Noguchi et al., with these example inputs (steady wear, a 2 to 4 times neglect penalty). Real wear speeds up past the limit, so treat 35,000 km as a ceiling, not a promise.

Here is the engineering of each stage, with the measurements you can take yourself, and why the end comes suddenly after a long, quiet decline.

317.5 mmnew length of 20 links of 520 chain
323 mmKawasaki's 20-link service limit (about 1.7% longer)
about 3%elongation at which chains begin skipping teeth
36.5 kNrated tensile strength of a DID 520 VX3 chain

Why does a chain "stretch" if steel does not?

A chain lengthens because its joints wear, not because its plates stretch. A roller chain is a loop of links, and every link is a pin turning inside a bushing.

When a link straightens or bends around a sprocket, the pin rotates a few degrees inside the bush under the full chain pull. Metal comes off both surfaces, the clearance grows, and each pitch gets a few micrometres longer. Multiply that by more than a hundred joints and the whole chain grows by millimetres.

The story

On 7 February 2024, Suzuki Motor USA issued a stop-sale notice and safety recall for its brand-new GSX-8R. About 360 bikes built between 1 and 11 December 2023, roughly a fifth of the model's US population, might have a drive-chain link that its supplier, RK Japan, had never riveted. Such a chain might not hold together, and it could come apart while the bike was being ridden.

No crashes or injuries were reported. Dealers inspected each chain and, where the link was faulty, replaced the chain and both sprockets. It took one unriveted pin out of more than a hundred to turn a new motorcycle into a recall, which is a fair measure of how much every single joint in a chain matters.

RideApart: Suzuki USA issues stop-sale order on 2024 GSX-8Rs due to chain issue

A 520 chain, the size KTM lists for its 390 Duke and 390 Adventure, has a pitch of 5/8 inch, 15.875 mm, so a length of 20 links measures 317.5 mm new. Kawasaki's Z750 manual gives the new range as 317.5 to 318.2 mm and the service limit as 323 mm, measured from the centre of the 1st pin to the centre of the 21st under a 98 N (10 kg) pull. That limit is 1.7% longer than new:

\[ e = \frac{L_{measured} - L_{new}}{L_{new}} = \frac{323 - 317.5}{317.5} = 1.73\% \]

Spread over 20 pitches, that is 0.28 mm of extra length per joint, made of wear on two surfaces, each one thinner than a sheet of paper.

Where the "stretch" comes fromNew joint: pin fits the bushBush (outer ring)Pin (filled), tiny clearanceWorn joint: both surfaces lose metalPin shifts in the gap:+0.28 mm per pitchx 20 pitches = +5.5 mm317.5 mm new → 323 mm limit (Kawasaki, 520 chain)
The chain's plates hardly change. Length comes from clearance opening up between pin and bush, joint after joint; Noguchi's tests found the wear concentrated near the pin ends.
A labelled rendering of a roller chain showing the outer plate, inner plate, pin, bushing and roller
The parts of a roller chain. The pin turns inside the bushing every time a link bends, and that sliding contact is where nearly all the "stretch" comes from.Photo: Niabot, CC BY-SA 3.0, via Wikimedia Commons
Evaluation of Wear between Pin and Bush in Roller ChainNoguchi, Yoshiba, Nakayama and Kanada · Journal of Advanced Mechanical Design, Systems, and Manufacturing 3(4) · 2009

Because a whole chain takes too long to test, the team built a wear tester that loads just two links in tension and rocks them through the bending angle a chain sees on a sprocket, then measured wear with a roundness tester and modelled the contact with finite elements.

They ran four greases for 317 minutes at tensile loads of 1,000 and 2,000 N. Wear depth roughly halved when the grease's base oil was more viscous, and the wear was not spread along the pin: it concentrated at the pin's two ends, because tension bends the pin so it touches the bush only near the plates.

The paper states the industry rule of thumb that a chain starts skipping cogs at about 3% elongation. The limit you should know: two links in a lab, with grease and no road grit, so real chains with dirt and water wear faster and less predictably.

Read the paper (free)

What happens to the chain, year by year, if you ignore it?

An ignored chain goes through five stages, and the rider notices only the last two. The kilometre figures swing with lubrication, dirt, rain and riding style, so the table tracks elongation, which you can check with a steel rule; the kilometre estimate is worked out at the top and in the verdict.

ElongationWhat is happeningWhat you notice
0 to 0.5%Running-in; pins and bushes bed in; slack needs adjusting once or twiceNothing, apart from the occasional adjustment
0.5 to 1.5%Steady wear; on a sealed chain, the O- or X-rings hold factory grease inside the joints as long as they surviveSlack returns faster between adjustments
1.5 to 2%Past typical service limits; rollers sit higher on the teeth; sprocket teeth start to hookChain lifts away from the rear sprocket when pulled; tight and loose spots
2 to 3%Seals dry and crack; rust stiffens links; load concentrates on the last tooth in contactJerky drive at low speed, clunk on throttle changes, kinked links
Around 3% and beyondChain can climb over a tooth under load; fatigue cracks grow at plate holesSkipping under hard acceleration, then a jump-off or a snap

Why the seals decide how long it lasts

The seals deserve a word. Most larger motorcycles now use a sealed chain: rubber O-rings or X-rings sit between the plates and keep grease packed around each pin. Bajaj's NS200 manual lists a "sealed drive chain" and asks for it to be cleaned, lubricated and adjusted at every service.

Exterior lube on a sealed chain mostly protects the seals and the rollers. Once a seal cracks, water and grit get into that joint, and it wears much faster than its neighbours. That is where tight spots come from.

Why does a worn chain destroy the sprockets too?

Because a longer pitch no longer fits the sprocket's tooth spacing, so the rollers ride further out on each tooth and the whole load lands on fewer teeth. A sprocket's pitch circle is set by its pitch and tooth count:

\[ D = \frac{p}{\sin(180^\circ / N)} \]

For a 41-tooth rear sprocket and a 520 chain, D = 15.875 / sin(4.39°) = 207.4 mm. A chain worn by 1.73% behaves as if it wants a pitch circle 1.73% bigger, 211.0 mm, so each roller sits about 1.8 mm further out on the tooth flank.

Roller on a 41-tooth sprocketSolid: new chain seats at the rootDashed: worn chain sits about 1.8 mm outHooked tooth("shark fin")Load onthe tip
Not to scale: the 1.8 mm lift is exaggerated so it shows. Pitch circle 207.4 mm new against 211.0 mm "wanted" by a chain worn 1.73%.

Every tooth then takes its load on a narrower strip near the tip, wears there, and develops the hooked "shark fin" shape mechanics look for. A hooked sprocket then chews up a new chain quickly, which is why Kawasaki's manual, like most, says to replace both sprockets whenever the chain is replaced.

Big sprockets suffer first in one sense: for the same elongation, the roller climbs further in millimetres on a large sprocket than on a small one, so a large rear sprocket tolerates less stretch before skipping.

My NS200 got its chain and sprockets as a kit, a Rolon brass chain with an RS200 41-tooth rear sprocket, at around 39,000 to 40,000 km.

A chain sprocket with its teeth visible, showing the tooth profile that the rollers seat into
A sprocket's teeth are shaped for one exact pitch. As the chain lengthens, the rollers climb the flanks and wear them into hooks.Photo: Arc1977, CC BY-SA 3.0, via Wikimedia Commons

Does a worn or dirty chain waste power?

A little, and less than people think, but a dry or rusted one wastes more. A clean, lubricated roller chain is one of the most efficient ways ever devised to move power, which is why bicycles, which have only a human engine to spare, still use one.

Lubricant Performance in Bicycle Roller ChainsDowd, Cavanaugh and Månsson · Purdue University; preprint on engrXiv 2024, published in Sports Engineering 2025

The team stripped new Shimano Ultegra 11-speed chains of factory grease, applied commercial and custom wax and dry lubricants, and measured efficiency on a rig at 200 to 500 W and a range of cadences, then repeated after wear cycles at 250 W and 100 rpm.

Under high-torque uphill conditions, efficiency peaked above 98%; in low-torque downhill conditions it peaked at only 96.6%. The best wax lubricants held their efficiency for more than 60 hours of running, while dry lubricants fell off in under 30 hours. The limit you should know: bicycle chains carry a few hundred watts, not the 30-plus kW of a motorcycle, and the tests were clean, with no road grit or water.

Read the preprint (free)

An older Johns Hopkins study, reported in August 1999, found bicycle chain efficiency between about 81% in the worst test and 98.6% in the best, and found that in a clean lab even removing the lubricant made little difference; the researchers concluded that lube matters mainly by keeping dirt out.

Larger sprockets and higher tension gave better efficiency. For a motorcycle the lesson is the same. A clean, oiled chain loses a percent or two; one packed with grit and rust loses more, and that lost power is heat in the joints, which speeds up the wear.

What is the load on a motorcycle chain, and when does it snap?

A chain's pull is the power it carries divided by how fast it moves. At full power in a low gear, a 400 cc single puts several kilonewtons through a 520 chain, a fraction of its rated breaking strength. Chains rarely break from one pull. They break because wear, rust and shock loads grow small cracks in the plates.

\[ F_{chain} = \frac{P}{v_{chain}}, \qquad v_{chain} = v_{road} \times \frac{R_{sprocket}}{R_{tyre}} \]

Take 33 kW, the KTM 390 Adventure X's rated output at 8,500 rpm, at 60 km/h in a low gear, with a 41-tooth rear sprocket (pitch radius 103.7 mm) and a rolling radius of about 0.31 m. The chain moves at 16.7 × 0.1037 / 0.31 = 5.6 m/s, so the pull is 33,000 / 5.6 ≈ 5.9 kN. That is 16% of the 36.5 kN that retailers list as the tensile strength of a DID 520 VX3 chain. Try your own numbers:

Calculated in this post from the KTM 390 Adventure X's rated 33 kW at 8,500 rpm (moto-data.net) and the 36.5 kN tensile strength of a DID 520 VX3 chain (Parts Denmark), with these example inputs (60 km/h in a low gear, a 41-tooth rear sprocket, a rolling radius of about 0.31 m).

How hard is the engine pulling on your chain?

The margin looks generous until you add the things neglect brings:

  • A rusted link that will not bend jerks the chain at every sprocket entry.
  • A loose chain slaps and snatches on every throttle change, multiplying peak load.
  • A tight chain (my KTM's chain came back from a service too tight or misaligned) loads the gearbox output bearing and the chain at full suspension compression.

Each extra peak grows the cracks that start, as fatigue studies of roller chains show, at the edges of the plate holes and inside the bushes.

Real-world failures are rare but not imaginary, as the GSX-8R recall in the story above shows: in February 2024 Suzuki stopped sales of about 360 early 2024 GSX-8Rs in the US because some chains had not been riveted properly and could come apart while riding.

A new chain that fails is a recall. An old one that fails is maintenance.The working above

What happens when the chain finally breaks?

When a chain snaps at speed, the engine suddenly drives nothing and the rear wheel is free, which sounds harmless. The danger is where the loose end goes.

Thrown by the gearbox sprocket, it can wrap around that sprocket and smash the crankcase cover, or jam between the rear sprocket and swingarm and lock the rear wheel. A locked rear at highway speed slides the bike sideways. This is what the chain guard and the case saver behind the front sprocket are for.

If you feel the drive go and hear a clatter, pull the clutch, keep the bike upright and use the front brake gently to stop in a straight line. The bike will not drive again until the chain is replaced, and the cases need checking for cracks before it does.

The rear hub and sprocket of a Royal Enfield Crusader with the chain running onto the sprocket
A classic rear sprocket and chain run. On older bikes with open chains and no seals, a neglected chain wore out in a fraction of today's distances.Photo: Motacilla, CC BY 4.0, via Wikimedia Commons

How do you check your chain in five minutes?

Measure it, look at it, and compare with your manual. You need a steel rule and five minutes.

  1. Engine off, neutral, bike stable. Put it on a centre or paddock stand if you have one.
  2. Measure 20 links. Pull the lower run taut, measure from the centre of pin 1 to the centre of pin 21, and repeat at two or three places. For a 520 chain, new is 317.5 mm; check your own manual's limit (Kawasaki's is 323 mm).
  3. Pull it off the sprocket. At the back of the rear sprocket, pull the chain rearward. If you can expose half a tooth or more, the chain or sprocket is worn.
  4. Look for tight spots and kinks. Rotate the wheel slowly and check that the slack stays the same all the way round.
  5. Inspect the teeth. Symmetric teeth are fine; hooked or pointed teeth mean the sprockets go with the chain.

For the routine that keeps you far from the end of this what-if, see the chain cleaning, lubing and adjusting guide, and for what a sprocket change does to your gearing, the sprocket gearing calculator. The chain is also one of the first things to inspect when you buy a used bike.

Verdict: about 35,000 km at best, about 9,000 km if ignored

With these example inputs (a maintained chain reaching a 1.73% limit at Ducati's 20,000 km average, steady wear, skipping at 3%), a chain that is cleaned and oiled but never replaced starts skipping teeth at about 35,000 km. Never cleaned or oiled, on my assumed 2 to 4 times faster wear, it gets there in roughly 9,000 to 17,000 km, and passes the service limit at about 5,000 to 10,000 km. After that it jumps or snaps, and it takes the sprockets with it.

Calculated in this post: 20,000 × 3 ÷ 1.73 ≈ 35,000 km; divided by 4 and by 2 for neglect; 20,000 ÷ 4 = 5,000 and 20,000 ÷ 2 = 10,000 km to the service limit.

What if next?

Questions people ask

How do I know when my motorcycle chain needs replacing?

Measure 20 links: if they are past your manual's service limit (for a 520 chain, Kawasaki's limit is 323 mm against 317.5 mm new), replace it. Other signs are tight spots that adjustment cannot remove, kinked or rusted links, and a chain you can pull half a tooth clear of the rear sprocket.

Should I replace the sprockets with the chain?

Yes, in almost every case. A worn chain hooks the sprocket teeth, and hooked teeth wear a new chain quickly. Manufacturers' manuals, including Kawasaki's, tell you to replace the front and rear sprockets whenever the chain is replaced.

Can I fix a stretched chain by tightening it?

You can take up the slack, but not the wear. Adjusting moves the wheel back to restore tension; the pitch is still longer than the sprocket's, so the rollers still ride high on the teeth. Once the adjuster runs out or the chain passes its limit, it has to go.

Does chain lube matter on an O-ring or X-ring chain?

Yes, but for different reasons than on an open chain. The grease that protects the pins is sealed inside; the outside lube keeps the rubber seals supple, protects rollers and plates from rust, and keeps grit away from the seals. Bajaj's NS200 manual asks for the chain to be cleaned and lubricated at every service.

Sources

  1. Noguchi et al.: Evaluation of Wear between Pin and Bush in Roller Chain, Journal of Advanced Mechanical Design, Systems, and Manufacturing (2009)
  2. Dowd, Cavanaugh and Månsson: Lubricant Performance in Bicycle Roller Chains, engrXiv preprint; published in Sports Engineering (2024, 2025)
  3. ScienceDaily / Johns Hopkins University: Bicycle chain efficiency study (1999)
  4. Kawasaki Z750 service manual: drive chain wear inspection
  5. Parts Denmark: DID 520 VX3 X-ring chain specifications
  6. KTM: 390 Duke technical specifications (520 X-ring chain)
  7. moto-data.net: KTM 390 Adventure X (2025) technical specifications
  8. Bajaj Auto: Pulsar NS200 owner's manual
  9. RideApart: Suzuki USA issues stop-sale order on 2024 GSX-8Rs due to chain issue (2024)
  10. Saito et al.: Fatigue Strength Analysis and Fatigue Damage Evaluation of Roller Chain, Metals (2018)
  11. Lanaspeze et al.: Identification of key parameter influence on efficiency of chain drive for track cycling, Mechanics & Industry (2025)
motorcycle chainchain stretchchain wear limitsprocket wear520 chainX-ring chainchain lubricationdrivetrain
Sapta

Sapta (Saptarshi Sarkar) is a mechanical engineer turned marketer in Bengaluru and a solo touring rider. He writes the seven Saptaverse blogs. Work with him or read more of RoadHook Machines.

Keep reading in RoadHook Machines

From the other blogs