Hollowtech crank failure - cause ?

[Brucey]

re corrosion-assisted cracking;

a) even (apparently slight) corrosion can help to form a crack-initiating defect and

b) as I mentioned earlier conditions inside the cavity can become anerobic with elevated levels of hydrogen. This can greatly lower the fracture toughness of some steels and this may help to both start and propagate a crack.

Regarding shimano's manufacturing and testing regimes; they are pretty good but

1) they are not immune to error; e.g. I have recently had to reject some upper-midrange shimano hubs because they were not machined concentrically... (something I have not seen on even the cheapest and most horrible hubs from other manufacturers), and
2) they certainly don't test their parts in the conditions that prevail in a typical British winter. If they did, they would presumably fit better seals, better grease etc, to their parts.

cheers

[Buk]

I think that a manufacturer like Shimano, who has in-house facilities for both metal production and testing, needs no suggestion about how to do their work, at least from me.

First you provide anecdotal evidence that failures are common place -- unless the bike shop where you work is patronned exclusively by elite riders, you having "seen several" implies this must be a common occurrence statistically -- then you wash your hands of it with your (frequently repeated) "trust the manufacturer" stance.

Wannabes and pretenders. They get everywhere.

[Buk]

The material doesn't always react to completion initially

Impossible. (Go read the document I linked and understand how Expanding PU foam works to understand why.)

But even if it were, then as PU foam is set by moisture; it would simply act to deal with any ingress of moisture by setting.

and nor is it impervious to water after it is cured.

Some aren't; some are. So buy a waterproof one.

If you pull anything steel out of typical expanding foam (after it has seen any kind of weather for a few years) it is usually horribly corroded. In part this is because you have made a nasty water-trapping crevice where no crevice would otherwise be and in part it is because claims of true 'inertness' are in most cases not entirely correct.

Brucey anecdotal evidence. Therefore it MUST be true. Despite the manufacturers specifications, legal requirements, and years of testing.

'kay Brucey. Sorry to have stepped on your toes. See ya.

[Des49]

Going back to the idea of filling the spindle with expanding foam - my opinion is that this a bad move. It can trap water.

I have relevant experience with an especially nasty job where I had to remove such foam from the locker from a boat. It was originally used by the owner as a sound proofing test, but it also eventually got soaked through with water and was a devil to scrape and chip out bit by bit while lying in the bilges for many hours with an arm and various bits of metal hooks/scrapers stuck through a 4" hatch. For good measure a fragment got stuck under a finger nail and caused a painful infection that needed piercing to drain.

[Brucey]

GORILLA FILLER WEATHERPROOF EXPANDING FOAM FILLER

Ah, so we have gone from 'an expanding foam' to a specific product. That is (IIRC) just another moisture-curing PU foam. It may be better than some but it will still suffer the usual problems.

Think about it; you either have too little moisture or too much to start with; the first leaves you with unreacted monomer (which is usually corrosive to steel) and the second leaves you with an excess of moisture in the foam, which, er, causes corrosion. Eventually you end up in the second condition anyway; claims of 'waterproofness' are invariably not entirely accurate.

Corrosion beneath PU foams (of various types) is a well-known problem. There are many mechanisms that turn conditions corrosive within the foam, including that the foam degrades and releases acids, or that (necessary) fire retarding chemicals in the foam cause it to become corrosive when damp.

If it were a great idea that (say) corrosion-proofed the insides of steel structures, you would undoubtedly find folk merrily putting it into steel bicycle frames and inside the sills of motor cars and so forth. The reason you don't see this happen is that it is just not a sensible thing to do.

cheers

[Gattonero]

I think that a manufacturer like Shimano, who has in-house facilities for both metal production and testing, needs no suggestion about how to do their work, at least from me.

First you provide anecdotal evidence that failures are common place -- unless the bike shop where you work is patronned exclusively by elite riders, you having "seen several" implies this must be a common occurrence statistically -- then you wash your hands of it with your (frequently repeated) "trust the manufacturer" stance.

Wannabes and pretenders. They get everywhere.

...while people who can't read English, or do steer a conversation in a way beyond what is about, apparently are even more common .
This is what I said, precisely

Having seen a few failing in the same way, all after a lot of miles and powerful riders (Cat 1 and Elites)

Please explain where do you read "seen several" instead of "a few".
I have not said this is a "common failure", indeed is a failure that may happen in particular circumstances, which are related to fatigue: lots (lots) of miles and powerful riders.
Nothing lasts forever, and those cranks do seem pretty solid if failures like this are not that common.

[Brucey]

I think I said 'several' upthread.

FWIW

'a few' means 'one or more but not many'

and
'several' means 'two or more but not many'

-more or less....

In any event I'd have said that it is a pretty rare failure, and that for many (most) users the stresses may well be below the fatigue limit. If so this leaves three possibilities

1) abnormal service loads (setup, rider strength)
2) manufacturing defect ( there will be the occasional one that will slip through the net)
3) abnormal service conditions (eg corrosion)

I doubt that we'll ever know which is the most important factor; even a proper failure investigation (which would cost thousands BTW) might not resolve the exact cause. It could also be a combination of the three; eg strong cat1 riders often train in all weathers, all year round, and their bikes often suffer corrosion-related failures even in parts that wouldn't fail due to mileage or rider strength etc.; small variations in manufacturing may tip the balance sometimes...

cheers

[Buk]

GORILLA FILLER WEATHERPROOF EXPANDING FOAM FILLER

Ah, so we have gone from 'an expanding foam' to a specific product. That is (IIRC) just another moisture-curing PU foam. It may be better than some but it will still suffer the usual problems.

{sigh} You really don't like to be seen as anything other than the all knowing oracle do you.

Given that the purpose described was to exclude moister to prevent corrosion, of course I would choose a product designed to do that.

Think about it; you either have too little moisture or too much to start with; the first leaves you with unreacted monomer (which is usually corrosive to steel) and the second leaves you with an excess of moisture in the foam, which, er, causes corrosion. Eventually you end up in the second condition anyway; claims of 'waterproofness' are invariably not entirely accurate.

You're just making this up as you go along aren't you. How do you get "unreacted monomer" from the polymer? (hint: you cannot.)

Corrosion beneath PU foams (of various types) is a well-known problem. There are many mechanisms that turn conditions corrosive within the foam, including that the foam degrades and releases acids,

You really have literally no understanding of the mechanisms of corrosion do you.

You keep stating that "this is a well known problem", but offer nothing but anecdote as evidence.

If it were a great idea that (say) corrosion-proofed the insides of steel structures, you would undoubtedly find folk merrily putting it into steel bicycle frames and inside the sills of motor cars and so forth. The reason you don't see this happen is that it is just not a sensible thing to do.

Try reading a little science. LMGTFY:
Abstract

The use of closed-cell plastic polymer foams to control corrosion in hollow, irregular cavities of preexisting metallic structures or assemblages was explored. The method involved blocking the access of corrosive species to the interior surfaces by filling the empty spaces with an expanding, polymeric, rigid foam material. The fact that the foam could be injected from a remote location and then expanded to fill the void solved the problem of accessibility, which precluded the use of traditional corrosion preventive coatings in these areas. The approach differed from prior studies in that it focused on postdesign applications of corrosion inhibition within confined regions. Results of laboratory seawater exposure tests using structural steel tubular specimens with interior surfaces that were as-manufactured, precorroded, or protected with a paraffin-based corrosion preventative and filled with a one-part polyurethane foam were discussed in the context of the foam's ability to prevent water penetration to the metal interface. Effectiveness of the method for corrosion control was demonstrated through visual comparisons of metal deterioration in treated and untreated specimens.:

Abstract

Recent studies show that rigid, closed cell polymer foam blocks seawater penetration and provides corrosion protection to the internal surfaces of metal cavities exposed to aqueous salt environments. The present study explored properties of rigid polyurethane foam that would impact its durability and effectiveness as an agent for corrosion protection. In particular, the bonding strength and cure morphology of polyurethane foam prepared from single-part injectable and two-part pourable formulations applied to steel tubular structures were investigated. Water absorption and retention characteristics of the cured foam also are presented. Tests showed that upon curing, the foam provided excellent coverage and adhered well to as-manufactured and precorroded metal surfaces. For the one-part formulation, no significant water was retained after 8 weeks of total seawater immersion. In the same test, the greatest amount of moisture retained by the cured foam from the two-part formulation was ∼ 1 vol%. Results suggested that within a confined metal cavity, rigid polyurethane foam acts as a well-sealed barrier coating or as an entry plug to block the intrusion of detrimental solid or corrosive fluid material. Two distinct advantages to using self-expanding, rigid polyurethane foam for retrofit corrosion control are its excellent adherence to precorroded metal surfaces and convenient application by remote injection to regions that are difficult or impossible to access for surface preparation and the application of protective paints.

PU foam can also be sprayed onto surfaces, and it is widely used to insulate ships and boats. PU foam acts as both insulation and a corrosion inhibitor by filling in water traps and preventing condensation in gaps within the hull.

Polyurethane solidifying foam is a chemically produced liquid mixture, that forms a cell structure. It is the most advanced insulation material available today.... The effectiveness of Polyurethane foam is achieved by the trillions of air pockets, which make up this insulation material.
...
Watertight: The extremely low permeability to humidity and steam render PU-foam waterproof.
...
Non-deteriorating: In laboratory tests conducted by BASF simulating the accelerated passage of time, it has been proven that PU-foam even after 80 years does not show any significant deterioration.
...
Floating objects
Due to its low weight PU-foam en-hances buoyancy and stability while insulating boats, buoys and other floating objects. A properly insulated object of this nature becomes unsinkable. For this reason the Lloyd's insurance company has issued a certificate for PU-foam to be used on all lifeboats.
...
Protection against corrosion:PU-foam, when sprayed on to iron beams, forms an airtight casing thus preventing corrosion.
...
The insulation capacity of PU-foam is several times that of other insulating materials, as is shown in the following diagram. Please bear in mind that the insulation capacity diminishes rapidly in all the compared materials when these absorb humidity, something which PU-foam never does.
...

I could go on, but there would obviously be no point.

You've decided that you know something -- when all the evidence of what you've said on the subject demonstrates very clearly, you do not -- and you aren't going to let a little evidence get in the way; and thus endanger your big-fish-in-a-small-pond status. Good luck with that.

Thank you for all the good stuff I learnt from reading you on the subject on which you are expert. For the other stuff, I'll look elsewhere.

[Brucey]

I'm very sorry but I am disinclined to indulge your ramblings any more. You basically don't appear to know what you are talking about. Try looking up 'monomer' , 'polymer synthesis' and 'metal corrosion in PU foams' for starters. BTW pretty much all such foams are permeable to water vapour (they have to be, to cure properly) and don't offer a long term moisture barrier, despite what the manufacturers might have you believe.

[Buk]

I'm very sorry but I am disinclined to indulge your ramblings any more. You basically don't appear to know what you are talking about. Try looking up 'monomer' , 'polymer synthesis' and 'metal corrosion in PU foams' for starters. BTW pretty much all such foams are permeable to water vapour (they have to be, to cure properly) and don't offer a long term moisture barrier, despite what the manufacturers might have you believe.

Brucey knows best. Better than science. Better than the European standards bodies. Better than Lloyd's of London. Better than RNLI.

No evidence. No references. Just Brucey's word. That might wash here with some; but other will recognise your attempt to ignore the evidence provided and argument by dismissal.

Right on Brucey. Keep on kidding yourself.

[Farrina]

Thanks for all the contributions and preventative suggestions which I shall carefully consider.

Given the further speculation surrounding the failure, just to provide a bit more context.

I'm not a Pro or 1st Cat (although naturally I consider myself powerful )

My riding style (especially uphill) is to grind rather than spin.

I'm north of 13 stone.

I would estimate failed chainset mileage at about 12,000 miles

The reason for my interest is I have an identical chainset on my other bike bought about the same time. I have closely inspected it (well as best you can) to see whether it is showing any sign of corrosion et al.

I am keen to ensure that the replacement (Ultegra 6750) does not go the same way (one of my cycling friends is also nervously following this discussion!)

[Gattonero]

Use a Campagnolo chainset then

[Pneumant]

Don't fill with cavity foam .
To me this post represents all that is good about this forum - a genuine problem highlighted, examined and discussed in depth.
I am going to check my 105 H2 chain set axle tomorrow and liberally coat with thick 'used' car engine oil. I'll be sensible and wear protective gloves when applying. This is much better than spray on grease products as it won't dry out and can be easily wiped clean for future inspections.

[Brucey]

NB not all HT-II spindles are steel and it is possible that the steel ones vary between models in ways we don't understand too.

As suggested above I think some kind of oil or wax treatment is probably best not least because it can be removed and the surface inspected.

BTW to work as intended any foam material would have to adhere to the steel very well in order not to leave a very nasty crevice at the interface. If the surface is indeed electroless Ni plate as suspected, I don't expect very many adhesives to bond to it at all well.

PU foams vary quite a lot; you would have to know exactly what you were doing to both choose one and source one. Even then you could easily come unstuck (ahem) for a variety of reasons. Note that polymers can be 'practically airtight' whilst still being appreciably permeable to water vapour in the long term.

One take-home here is that if you get a new creaking sound, it is probably a bad idea to just ignore it!

cheers

[alexnharvey]

Pure fatigue is an interesting idea but I'm not sure how useful. Fatigue is always searching for a flaw and developing it into a failure. The flaw could be corrosion, a crystal grain boundary or dislocation, an inclusion, a bubble, a scratch, a sharp step, peak or valley, heat affected zone or others. We can only seek to minimise these by design, manufacturing and prevention in use, not eliminate them.

The interesting thing to me is has something preventable or foreseeable, at any of the stages (design, manufacture, use/maintenance) reduced life below the expected design life, or led to an unexpected failure, particularly if in a dangerous mode.