I had some intermittent clanking noise on yesterday's ride, thought the chain just needed cleaning and derailleur aligning. But got home to find that one side of one chain link had broken, so only the other side was holding the chain together. The broken side was occasionally clanking against some of the drivetrain components.

I had pushed my 110kg system weight uphill and out of the saddle, which got me wondering how much strain the broken link could have taken.

Are chains designed to withstand a certain amount of force along their length, and how much would a half-broken link compromise that? Or is it when a chain bends around sprockets that it's more vulnerable, not just straight-line force?

Edit - I suppose the half link question is whether a chain actually loses any 'strength' if only one side panel is present on a link, or whether the dual side panel design is solely to locate the chain over the teeth. I'm thinking it would lose some strength, but nowhere near what I would imagine, and probably much less than half its original strength.

  • The larger danger, if the chain splits on one side of one of the "outside" links, is that the pins through the "inside" links will bend. This will make the pins want to slide out, possibly pulling them out of the plate and leading to complete failure. (But I have never seen a chain fail in the manner you describe -- it was probably just a fluke.) Mar 1, 2021 at 19:30
  • Did the side plate crack, or did it simply come off the rivet? I've never seen the former happen, but the latter is common enough.
    – Adam Rice
    Mar 1, 2021 at 19:34
  • 1
    Side plate split in two, each half of it was still attached to each of the adjacent pins.
    – Wilskt
    Mar 1, 2021 at 19:38
  • Did you push a normal pin back in to close the chain? Did it break at a closing pin? How worn was the chain?
    – Michael
    Mar 1, 2021 at 19:42
  • It's getting off-topic, it was an old chain.
    – Wilskt
    Mar 1, 2021 at 19:45

3 Answers 3


The design tensile strength of bicycle chains is probably in excess of what almost all humans can reasonably hope to output. I'm not aware of publicly available figures for this parameter, and I'd assume that the manufacturer specifications are their trade secrets. Adam Kerin of Zero Friction Cycling is in the process of testing tensile strength, so if you are interested in this, you could watch this page. *

I'm not an engineer, but once the chain fails on one side, the force on the other side presumably increases by up to double. Plus the whole structure would get pulled sideways and fail as the pin gets pulled out of the other roller and link plate. That chain is done. Don't ride it.

Notice that the first paragraph said the design tensile strength. As with any manufacturing process, you will occasionally get dud items. Alternatively, if the chain was joined by a master pin, the mechanic in question might have mis-installed it. This is not so much a veiled swipe at you as it is a general statement of human fallibility. I haven't broken a master pin, but that's only because I don't use pins. I quite recently incorrectly connected a master link by failing to engage one side. The link failed in much the same way I'd expect your chain to fail if you keep pedaling it.

* -In summary, the page insists a chain can take up to 9kN before failing. This is equivalent to 900kg static load (standing still onto the pedals). Or perhaps 450kg when sprinting.

  • 2
    The tension will double almost exactly, but the effect will be worse as there's now a bending force, that means the pins are no longer experiencing pure sheer forces (another way of putting what you said; I'd expect failure to be fairly quick, or the bent link could jam in the rear derailleur). Master pins can drive in to make a stiff link; freeing that up could leave a weak link, where one end of the pin isn't well-supported, but I wouldn't expect that to cause plate failure. It would take creative clumsiness to damage the plate significantly
    – Chris H
    Mar 2, 2021 at 16:40
  • Thought experiment (real if you really want but you can sort out the precautions): close a chain over a chin-up bar or branch, and lift yourself up on it. Add a friend, then more until it breaks. You'd probably struggle to get enough people hanging off it to snap a chain. Or use sandbags instead of people, add them gently and count up the weight (force) needed to break it
    – Chris H
    Mar 2, 2021 at 16:44
  • "design tensile strength of bicycle chains is probably in excess of what almost all humans can reasonably hope to output" has to be a huge understatement... Mar 2, 2021 at 17:58
  • @user2705196 that's also what I'm getting at with my thought experiment. The idea of towing a car with one crossed my mind, but that's a dumb idea from a legal point of view, and I don't have anyone to help me.
    – Chris H
    Mar 3, 2021 at 8:56

Chains do withstand all reasonable forces you can put at them. Current chains are bushingless design where the punched side plates have a half-bushing on each side. Not having a press fit bushing but integral bushing-halves, the bushing doesn't weaken the chain and the chain withstands a huge amount of force.

There are people weighing over 150 kg riding bikes standing on the pedals. The first thing that fails for them isn't the chain, it's poorly built wheels.

There are mountain bikers using chainrings as ridiculously small as 20 teeth. This gives more mechanical advantage for the crank chainring system, thus increasing chain loads. Chains don't regularly fail for them.

There are e-bikes that have pedaling assist at the cranks (mid-drive), thus increasing chain loads. Those e-bikes don't require e-bike-specific chains (except sometimes you see e-bike specific components; that's just marketers wanting to sell more expensive stuff -- normal bike components work just fine for e-bikes)

It's the straight upper run of the chain that bears the load. Once the chain turns around the sprocket, the load immediately reduces because the first tooth carries most of the load. The chain links flexing cause chains to wear, but the greatest forces are encountered in the straight upper run.

Thus, I'd say your chain had a manufacturing defect that caused at least one link to be weaker than it's supposed to be and thus caused the chain to fail. This isn't entirely uncommon.

  • It's more likely lack of cleaning and maintenance but aye we'll go with manufacturing defect. :)
    – Wilskt
    Mar 1, 2021 at 18:37
  • @Wilskt I think it's much more likely to be manufacturing defect unless you let the chain go an extremely long time without replacement, such that the pins and rollers were very worn.
    – Weiwen Ng
    Mar 1, 2021 at 18:59
  • The outer plate split halfway along, half of it was still attached to each adjacent pin, so it's not quite like the cases linked above. It's possibly 9+ months and a good few thousand Scottish miles old, and the last two rides I had some noisy slippage on the sprockets or chainring so I'm thinking some mechanical stress too. I don't think it owed me anything, put it that way!
    – Wilskt
    Mar 1, 2021 at 19:08
  • 20 teeth isn't “ridiculously small”. Trials bikes use 18 or even 16t on the front, and they put massive force on the pedals in gap jumps. Mar 2, 2021 at 16:06
  • Wouldn't a larger chainring put more strain on the chain? Using a small chain ring allows you to use a small amount of force to do a large amount of work, by doing more revolutions. Using a large chain ring allows you to actually put a lot of power through the chain. The amount of force on the chain is multiplied by the distance from the center of rotation to get the amount of torque. The larger the chainring, the more torque applied.
    – Kibbee
    Mar 3, 2021 at 20:12

Bicycle chain plates are typically made of steel. Steel has very high tensile strength. For chain tension to permanently deform a chain link plate (permanently stretch but not break), the tensile force would likely have to be at least 200 lbf per plate. The ultimate strength of the plate (causing a break, not just permanent stretch) would typically be 2-3 times as much. In actual use, bicycle chain plates don't stretch (the observed chain stretch is primarily the wearing down of the pin-bushing interface surfaces).

  • To paraphrase - the material is strong, the failures are more likely to be pins and joints in the chain rather than the plates ?
    – Criggie
    Mar 8, 2021 at 20:42
  • This answer touches on an important aspect the others don't cover: chains in use must not significantly stretch, a chain that stretched a mere 0.5% would be considered toast in modern drive trains. This means the amount of plastic deformation permitted is extremely small, and all of the expected force a rider will deliver has to be within the plastic deformation regime. Exceeding that, to ultimate tensile failure, is going to require a good deal of additional force. Mar 9, 2021 at 6:42

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.