I would like to know what the strength of chains / cranks / bearings etc are. It isn’t a scientific thing, but years of ‘yeah, I broke that with my AWESOME POWAH’ stories and ‘I’m a big guy and I keep snapping cranks’ tales don’t ring true.
I put it down to poor technique, manufacturing fault, accident or bad spannering. I’m a big guy and I don’t destroy stuff. I know that fatigue will eventually play a part, but not that much on cranks etc.
Anyone got any good numbers?
It looks like Wippermann Connex joining links have a breaking strain of 1000kg, which is more than anyones ‘AWESOME POWAH’.

they are probably about in measurements that make no sense, there was an extra on Roam where they showed the destructive testing of a Demo (it was a lot more than use would give it)

Yes, that geek of a person over there in the corner of the internerd whose wearing a cardigan, works in extreme algebra, lives with his mum, drives an allegro and is called Graham.

Generally chains that snap are worn and have been pulled about by loaded shifting. i am 16 stone and only ever broken very worn geared chains. the only ss chain that broke was hit be a rock that also broke my toe

There is an industry standard measurement. It’s a fairly broad scale.
Halfords square taper cranks on £99 full sus bikes = margarine
FSA afterburner cranks from the 90s = butter on a winters day
White industry cranks = if chuck norris made Parmesan and kept it in the fridge.

White industry cranks = if chuck norris made Parmesan and kept it in the fridge.

Chilled vomit, you say? Interesting….

As an aside, i’ve also only every broken one chain, and that was my fault for not joining it up properly.

Halfords square taper cranks on £99 full sus bikes = margarine

Yet the cycling world existed for many years with such technology and wheels kept turning.

If anyone has a chain and a vernier to hand and wants to tell me the rough dimensions and thickness of a side plate I’ll do a quick guestimate calculation for you. I do those kind of calcs all day for work on much bigger things so it’s only a 2 minute thing to get an idea of the breaking load for a chain.

you could work out some basic strengths using cross sections and either making a guess on the material to get a strength value or using a spread to give you a range. Strength varies according to alloy, processing etc which makes it a bit tricky compared to modulus but it should give you an idea.

The thing about a lot of bike components is they’re designed to be strong in use and also light. So they’re strong in one direction, but stressed another way and they fail easily.

E.g. wheels. You can land heavily on most wheels no bother, but land with a little bit of sideways and as there’s less tension in the lower spokes it’s easy to fold the wheel over (if the rim’s not that stiff etc.) Similarly chains, you’ll not pull one apart, but it doesn’t take much if something’s pulling it sideways apart whilst you pedal – cos it’s worn, or the gears are poorly adjusted etc.

Side plates don’t break in the middle, though, they break at the rivet hole.

UTS is never an issue with bike parts, though – a single spoke can easily hold your body weight, yet spokes break all the time. It’s fatigue wot does it…

I’ll play with cranks when I get home for a figure. It’s a fairly simply calculation which o studied not too long ago!

Virtually any component can be analysed, the Aero industry did this for years, by hand. CAE simulations can be carried out. This type of development, done properly, should get you into a good position, at the design stage, even before you go to tooling, manufacture, etc.
In addition to CAE work, where applicable or desired. Engineers then apply a “Saftey factor” into their calcs. My materials lecturer use to refer to it as the factor of ignorance, which I feel was a bit harsh. Its more along the lines of “belt and braces” in my opinion.
🙂
So, once you’ve gone to tooling and you are producing your components / assemblies. Next up should be dynamic testing of the real components think of test riders. Which is where you might find issues. If you’ve done a good job, then you might hit the target first time, but if a bit of tweaking is required, then real world testing should highlight this and of course, testing of real components also covers issues that may arise from “cyclic loading“.

A properly designed and spec’d component or assembly, once its gone through the correct design and testing, should be good to go.
However, as the OP points out, user riding styles and component installation can vary with less than preferred results.
And yes, there are the AWESOME brigade who seem to try very hard to break stuff as some sort of demonstration of their ultra, mega, riding awesomeness.
🙂

I’d Echo Luminous comments…

As an assembly of Structures and components most MTBs made in the last 20-30 years will have been designed to the best of the engineers/designers knowledge and ability to be adequate for their intended use plus a good margin (Safety factor, Reserve factor, etc).
When you design anything for use in a load path you can use a variety of methods to substantiate that part against its envisaged loading long before any metal is cut.
Basic stress calculations, FEA is widespread (but not always trusted implicitly in all fields), and of course there’s a couple of decades of previously designed parts that have probably done the same job and can simply be looked at with a bit of engineering judgement to yield improvements…

The cost of physically making something to destruction test is relatively high, so really you only want to use physical testing as a way of confirming analysis your results at the end of development, you may not even bother if your analysis gives enough confidence, and go straight to manufacture.

Chains are an interesting one, the basic concept hasn’t changed fundamentally in 130 odd years, the bicycle chain has essentially had over a century of refinement.
As a design it’s a very good one, transferring tensile loads along it’s length for transmission and yet robust enough to also deal with lateral movements and misalignment, being covered in water and debris and dragged across mating components and still functioning… Given it’s working environment failures are inevitable.

Very few people actually “snap” a chain, they normally manage to separate a side plate from a connecting pin, in doing so instantly wiping out half that tensile load path…
The causes of these sorts of failure are as much down to use and maintenance, as they are likely to be manufacturing defects.

Fundamentally though yes someone, somewhere will know or be able to lay their hands on a document that tells them the strength of every component fitted to your bicycle against a set of load cases… this will of course be proprietary data, and I doubt any manufacturer would volunteer or publish such information; Essentially it’s there in case they ever end up in court and need to demonstrate they took all practicable measures to ensure the products safety…

Does anyone know actual strength of components?

167.5 +/- 3
depending on the phase of the moon

Yes, that geek of a person over there in the corner of the internerd whose wearing a cardigan, works in extreme algebra, lives with his mum, drives an allegro and is called Graham.

According to a certain thread yesterday, the last bit is apparently true.

Aren’t most structural failures down to fatigue or stress over time though?

And often invole things like cracks, poor welds etc?

I think the OP’s question is too simple – how often does anything break because one single incident exceeded its tensile strength?

how often does anything break because one single incident exceeded its tensile strength?

Set of RF XC ISIS cranks fitted to a DH bike and ridden once in the mega bent in 1 hit.

Like most things wrong component in the wrong place (in this case on a pile of logs at 30mph)

Aren’t most structural failures down to fatigue or stress over time though?

Fatigue is repeated stressing over time…

Fatigue can be estimated (you simply need to define expected loading of a given frequency over time) and a components suceptability to fatigue can be calculated, this can then be factored into other analysis, the question is how much time / effort does a company really want to spend analysing such things, realistically what is the expected design life for any given part, and is it reasonably to expect that the fatigue life of the part will be less than it’s design life?

it’s quite easy to exceed the design loads for a bicycle component in an accident, there has to be a limit to the fault load cases that an engineer is forced to design to or else your bike will become very robust but weigh too much to actually move…

The simplest approach (taken by many) is to assume that if you carry out a basic set of stress analysis’ based on a conservative load case and have a high enough reserve factor, fatigue is less likely to be an issue as you are not regularly stressing the material close to to it’s limits…

Welds are a good example I suppose, The key thing is always conservatism in any analysis.

When calculating weld stresses it’s pretty common practice is to apply a reduction to the actual area of the weld (say 10%) to accommodate any weld imperfections and to also add a factor to the input loading, you are looking for a “Pass” against a set of unrealistically high inputs and poor manufacturing to give you additional confidence…

The other thing is that Stress permissible for analysis are always way below real world figures, there is inherent conservatism in all steps of a weld stress calculation.

Weld stress permissibles are normally taken as a proportion of the parent material or filler material’s minimum tensile strength, the Crane Code essentially applies a factor of 0.3 x the lowest tensile strength material used in the joint.

I’m not sure how this translates to bicycle frame construction I’d imagine the welded joints in many bike frames have an appropriate stress permissible for them to be analysed to, you could use one from another industry (such as the crane code) based on it being pretty good RGP…

All of that conservatism is meant to help rule out excessive fatigue of the joint to a certain extent, however any analysis is only as good as the input data, if your bounding case is based on an 80kg rider and you regularly sell bikes to 110kg rider your calculations are effectively invalid, always better to go high.

bencooper – Member

Side plates don’t break in the middle, though, they break at the rivet hole.

They do, though. I don’t break many chains but I reckon I’m about a 50/50 split, hole or middle.

Tucked away in the internet somewhere is an old Tour magazine test of various road frames which were put on a jig and loaded over several thousand cycles. The loading was more than you would expect from normal peddling. The frames were a mix of Ti, Steel and Alu. Contrary to perceived wisdom, in general the steel frames broke first, then Ti, and then Alu with some Alu frames not breaking.

So are Alu frames the best for strength/longevity? Perhaps, perhaps not. The authors concluded that in general, the Alu frames had been tested more and designed as a whole whereas the steel frames often had very strong areas such a BB shells and fail where weaker parts (stays) met them, as they would have to cope with all the stress. Ti broke or cracked often at areas such as welded bottle cage holes.

But perhaps steel has a better reputation as it copes with day to day knocks a little better.

The authors did come to the conclusion that even the weakest frames were actually strong enough for normal use and unlikely to fail.

Perhaps in my mind it is over-simplified or, as I suspect, the engineers are trying to over-complicate the answer.

To get the ‘stress’ answer, there is going to be a manufacturing point that is optimal to make the components at – something like ‘if it can take a 500kg load before breaking, it’ll last 20,000 cycles of fatigue’. There hast to be a correlation between 2 numbers like that.

To some extent particularly for a simple structure (eg a tube) but that’s the whole issue with fatigue and why it’s much more complex than just UTS. The tests largely focus on fatigue strength – or at least that’s the complex part of it.

Fatigue is really complicated. For some materials there is a stress level below which it will last for forever, this is the fatigue limit. Some, but not all steels are like this whereas aluminium always breaks in the end. For things that are working beyond their fatigue limit, there is no easy way to tie calculated stress to fatigue life.
It is very highly dependent on the material, the nature of the stress cycles, the shape of the part, the number of defects in the material, the surface finish, the corrosive substances in the enviroment that the part is working etc… You can make a decent estimate by doing some pretty complicated stress analysis and looking up a lot of tables but it’ll always be an estimate, and for thing like bike where the loading is highly irregular and impossible to predict, it;ll be a very rough estimate. The only way to be sure is to design things so that the worst possible working conditions will give rise to stresses so low as to either be below the fatigue limit or to give a predicted life of millions of cycles.

The tests largely focus on fatigue strength

You say that, but nobody has referenced any of the testing standards, so it sounds like a hunch. Given that some components and some frames are for ‘XXXkg max’, and you tend not to get that for most components, there must be a large factor of safety in there.
Sod it – I’ll save up and do it myself.

Just on the basis of the equipment I work with and the loads that you can carry with similar sized sections to those seen on bikes, made from similar or weaker materials, there is a very large factor of safety against the simple weight of the rider. I’d guess typically more like 10 than 5.

You say that, but nobody has referenced any of the testing standards, so it sounds like a hunch

Not in my experience. We design to pass our own predetermined, internally set, test standards which can be based to meet a balance between cost and performance, or to comply with legally enforced testing as dictated by the governing agencies of the markets we sell into.

So, for example, a customer of mine has determined to design and test their I/C (engine) for a service life of 150K miles. But the vehicle still has to pass safety regs as stipulated and enforced by the governing agencies who grant license to sell into their markets.
Its not a clear cut thing. I expect there would be safety regs for bikes to have to comply with in order to be offered for sale in the UK ? ….
But you’d expect your XTR crank arm to be designed with more than this fundamental compliance in mind. Components aimed at the upper end of the market will be under pressure to be strong, yet light weight and hard wearing. These boxes can be ticked, but at a price which stems from good, sometimes exhaustive design, manufacturing tech and material spec.

[EDIT]
As someone above has pointed out, most good manufacturers will have internal records of simulation, if they do that, and testing. Especially if they sell into a highly litigious market such as the U.S.
But, they are very unlikely to divulge this data, test procedures or results, as its not needed by the customer and may assist competitors. Also, it could assist those who would replicate parts without license or proper controls.

I read an interesting interview with the then MD of Brompton. They specifically do not patent many components as these can be expensive to maintain and can, effectively, become a manual for people in parts of the world where the copyright law may not be respected or enforced. To copy parts and build illegal replicas.