Shimano denies design problem with Hollowtech cranks despite reports of cracked arms

anke wrote:

Right. So Mr ex black belt would design bike-parts to a six sigma level of reliability, so there might be just shy of 4 failures in one million parts, over many, many years, in possible highly corrosive environments? What's the weight penaltiy? Pro-level bicycle parts (or at least very high-level ones) - not the stuff for the daily commute?!

 

You clearly don't understand DSSM, you are way over your head. Despite the name it's not about meeting an absolute figure, it is about meeting a set of key requirements which you define yourself for your product and market, and covering much more than reliability. So no nothing whatever to do with meeting 4 failures per million, unless you think that 4 is the right number for a specific parameter on your product. Being an authority on DfM, I thought you'd know that.

Right. But you didn't even attempt to provide an estimate of what level of reliability, failure frequency, weight, ... are acceptable for components for this market (Amateur racing, pro racing, keen / wealthy riders) - but just attack me for appreciating Shimano's innovative crank design.

You know, people can appreciate a cool, elegant, light and clever design - even if it shows problems, in particular problems that seem to occur after a longer time or in corrosive environments - for which these "fair weather products" might not have been designed.

These Shimano-Cranks are cool, because they provide a high stiffnes at low weight, by achieving large, hollow cross-sections from the pedal through to the four arms holding the chain-wheel. Yes, this could also be made from carbon fibre or perhaps, somehow, by welding. But making it from aluminum, with thin walls, cheap enough for hundreds of thousands of bikes is cool. I also believe that Shimano were the first to question the three-arm or five-arm designs - replacing it with a much more logical four arm design that can be tuned for the way a rider will provide the greatest torque during the down-stroke. And finally, I believe that these designs are cheap to manufacture for Shimano - so again, cool engineering for production -- even if the sales people might have added a huge premium.

But instead of appreciating these clever design points, you can just continue attacking me personally and boast about your credentials if that's what you're here for.

Where do you get 4 from - there are about 30 fails mentioned in a self selecting thread on the problem.

hirsute wrote:

Where do you get 4 from - there are about 30 fails mentioned in a self selecting thread on the problem.

Looking at the number of riders that have experienced crank failures, I'd say the failure rate would be much closer to 4% than 4 in a million. Otherwise, riders like Bubba Tex would be incredibly unlucky to have broken three of them and that doesn't seem that unusual.

Ask Grif500, who introduced his six-sigma credentials, and that he used to be a black belt ...

(Six sigma, originally, means that a production step should be managed so tightly that its standard deviation (in a critical outcome) is so small that a single part will only be a failure if it deviates from the specification by six standard deviations (sigma) - which corresponds to approx. four failures in a million parts, assuming a Gaussian distribution.)

Why would I ask anyone other you who wrote

"so there might be just shy of 4 failures in one million parts, over many, many years, in possible highly corrosive environments?"

Given the number of posts and photos of the problem prior to you writing it does seem a rather odd thing to claim.

"Why would I ask anyone other" (than) "you who wrote..."

...you would ask Griff500 because my post is a response to Griff500 (as clearly indicated above the post). And it was Griff500 who argued based on "six sigma" (which, originally, means four in one million). Why he did so? I do not know.

Edit: The black belt and six sigma stuff have disappeared now from the posts of Griff500. That's fair enough - in that case, please ignore my comments on black belts and six sigma...

anke wrote:

"Why would I ask anyone other" (than) "you who wrote..."

...you would ask Griff500 because my post is a response to Griff500 (as clearly indicated above the post). And it was Griff500 who argued based on "six sigma" (which, originally, means four in one million). Why he did so? I do not know.

Indeed, note that word "originally" in your definition?  Good, nothing to do with the Design For Six Sigma Design Methodology then, just a snippet of ancient history you decided to throw in.  "Six Sigma Methodology is a proven tool set for driving and achieving transformational change within an organization. Six Sigma is a continuous improvement process focusing and organization on: Customer Requirements, Process Alignment, Analytical Rigor, Timely Execution"

Apart from which, you will note that your hastily Googled 4ppm definition refers to reliability of a process step, which is very different to reliability of a completed component, made potentially from hundreds of process steps.

You say you don't know why I raised this? You were the first to mention Design for Manufacture, so I gave you the credit of understanding the subject. Because that is what Design for Manufacture is, the control of many process steps used in the manufacture of a component such as a crank, and the summation of small, possibly 4ppm errors and failures at each step to result in an overall figure. It was you who after a quick Google, applied 4ppm to a complete crank, not me!

Interesting: "Because that is what Design for Manufacture is, the control of many process steps used in the manufacture of a component such as a crank, and the summation of small, possibly 4ppm errors and failures at each step to result in an overall figure."

You better go and correct wikipedia then, their definition must be as wrong as my understanding of the term: "Design for manufacturability (also sometimes known as design for manufacturing or DFM) is the general engineering practice of designing products in such a way that they are easy to manufacture."

anke wrote:

Interesting: "Because that is what Design for Manufacture is, the control of many process steps used in the manufacture of a component such as a crank, and the summation of small, possibly 4ppm errors and failures at each step to result in an overall figure."

You better go and correct wikipedia then, their definition must be as wrong as my understanding of the term: "Design for manufacturability (also sometimes known as design for manufacturing or DFM) is the general engineering practice of designing products in such a way that they are easy to manufacture."

Nothing wrong with the wikipedia deinition as a supeficial intro, but totally ignores fundamentals like how to achieve ease of manufacture, and equally important, minimisation of waste. Fundamentally, it is all about control of processes from supply chain through to in service, so that when components come together they fit, and work, as designed, for the period of their design life, as a result of which there is minimal rework and waste and high reliability . Design for Six Sigma is just one way to achieve that, there are others, but the principle is the same 

anke wrote:

Ask Grif500, who introduced his six-sigma credentials, and that he used to be a black belt ...

Just for the record, NOWHERE did I state that bike parts should meet a reliability of 4 parts per million. Absolutely nowhere!  

Anke, seriously? You accuse me of attacking you?  Take a look at my posts and you will see that all I set out to do was correct your suggestion that Shimano was "as light" (it isn't) and cheaper (it isn't), than the opposition and enumerate your exaggerated non recyclable concerns as a mere 10g per year in 240kg of average non recyclable waste per person. No personal attacks whatsoever (until now!). Read your own posts and see who is attacking who, and misquoting me is really scraping the bottom of the barrel. I am sure you knew that coming onto a thread populated predominantly by people with broken shimano cranks in their garage, with talks of elegant design would not gain you much support (just take a read through the other responses). Don't take that out on me for merely introducing some real examples in place of your anecdotes.  Or maybe you are just here to me constrovercial?

This is a much more positive tone now, thank you, and I'm happy to respond in detail. 

You wrote the following:

• "This post has no more fact than your previous post."
• "people get hurt, have you not been listening?"
• "As for the nonsense about affordability and cheap materials". 
• "You clearly don't understand DSSM, you are way over your head."
• "Being an authority on DfM, I thought you'd know that." (I never claimed expertise.)

To me, these statements do read like a fairly aggressive attack.

Also, did you not write about the six sigma black belt - text that was removed by now?

You also claim that I was just giving anecdotes - can you provide any evidence, please?

Finally, you ask whether I'm just here to be controversial. No, I'm not. But when writing about a likely flaw in an otherwise successful product, it's not fair to keep bashing that product, entirely ignoring its strengthes. This is why I combined my observations and criticism with an appreciation of the crank arm design.

anke wrote:

Also, did you not write about the six sigma black belt - text that was removed by now?

Griff> The ONLY reference I made to six sigma black ebelt was that I mentioned that I was qualified, I made no refrenece to DSSM in relation to bike manufacture. The post was there for all of 60 seconds before I removed that line, as I believe resorting to quoting qualification shoudl be unnecessary if your agument is strong enough.

You also claim that I was just giving anecdotes - can you provide any evidence, please?

Griff> Yes, your anecdotal reference to weight and price, which a quick comparison of Red and DA has already shown to not always true. Of course there are examples of some aluminium cranks being lighter or cheaper than some carbon cranks, which suggests there is more than just crank material at play.  

Also anecdotal comment about corrosive environments. My bike spent its life in a warm dry garage in warm dry Provence, 50 miles from the sea.

Finally, you ask whether I'm just here to be controversial. No, I'm not. But when writing about a likely flaw in an otherwise successful product, it's not fair to keep bashing that product, entirely ignoring its strengthes. This is why I combined my observations and criticism with an appreciation of the crank arm design.

Griff> There is nothing elegant about glueing two bits of aluminium together, we've been doing it for years. Some (eg Lotus) do it very successfully. If you, like I, and many others here, had experienced one of the many failures, the "otherwise successful" phrase sounds a bit hollow (pardon the pun) and if you look at the pictures of the many failures on thanksshimano or whetever its called, there is little elegant design on display. A crank failure is not funny, and should not happen without outside intervention. In my case, 25000km, a bike only ever ridden by me, a 70kg rider, and never crashed, the crank should have outlasted the bike. I've already put my money where my mouth is and am changing over to SRAM, as the R8000 cranks also seem to be failing

 

Thank you very much for not responding to my summary of your aggressive statements. It seems you like to attack, rather than to respond to a discussion...

Personally, I find it peculiar that you mention Lotus as a positive example - the very company that has become (in-)famous for it's super-light, dangerously frail and frequently failing Colin Chapman designed F1-cars... 

anke wrote:

I find it peculiar that you mention Lotus as a positive example - the very company that has become (in-)famous for it's super-light, dangerously frail and frequently failing Colin Chapman designed F1-cars... 

Colin Chapman designed F1cars never used bonded aluminium in stressed components. Chapman died in 1982. Lotus didn't introduce bonded aluminium into road chassis construction until 1996., a technology which they largely pioneered, and for which they have become world renowned, transferring in to other manufacturers from Aston and Jaguar at home, to Hyundai, Isuzu in the East, and Chevrolet and Dodge in the West. Not to mention a shedload of awards, such as "The European Aluminium Awards jury recognised that Lotus used its knowledge in lightweight materials, such as aluminium, to its advantage to build the APX, and that Lotus satisfied the jury's key criteria of originality, functionality, design, durability and recyclability."    So yes, I'd say they are pretty good at bonding aluminium. Nothing peculiar there at all.

Why did you remove the final sentence from the quote:

"But as I wrote before, I totally agree that Shimano did not seem to handle this problem (cracked crank arms) in a graceful way. "

To spin my statement in the direction that you like?

Rigobear wrote:

Yes my slightly under three year old Ultegra crank failed in the same way (funny Shimano only give them a two year warranty) . Luckily mine wasn't a full detachment so I didn't crash. Shimano UK didn't want to hear about it telling me I should take it to a Shimano dealer who would return it via the dealer network. It is almost impossible to contact Shimano now. Shaming them on social media is about the only approach. Granted the crank was out of warranty but I would want every failure back to investigate.

Hmm, it may be out of warranty, but an item must be fit for purpose. Wearing out a chainring is one thing but having a major chainset structural failure after 3 years suggests not fit for purpose.

Two years warranty is a very short time for something like cranks that you'd expect to last ten years or so. Shimano must know of the issue, but presumably they're happy with the failure rate (4%?).

Shimano can say whatever they like about their warranty, none of which impacts your statutory rights - against the retailer. My undertsanding is broadly that these rights last six years, although the available remedies taper off with time. How long an item can be expected to last does matter, so the retailer can not simply point to Shimano's 2 year warranty and send you on your way.

Sriracha wrote:

Shimano can say whatever they like about their warranty, none of which impacts your statutory rights - against the retailer. My undertsanding is broadly that these rights last six years, although the available remedies taper off with time. How long an item can be expected to last does matter, so the retailer can not simply point to Shimano's 2 year warranty and send you on your way.

Unfortunately, my bike retailer closed shop a few months after I bought my bike, so that wasn't an option.

You may still pursue a claim with whoever you have the contract with probably the shop or Shimano direct via the small claims court under a fit for purpose clause under the consumers act or against the manufacturer for latent defects which is where the 6 years derives from mentioned by sriracha  (I should google them for proper references / titles etc). If a common design defect is known which this seems to show then in simple terms it makes it easier to demostarte the manufacture knows about a fault and has a duty to replace or correct. The warrenty period is merely to enable a quick and easy means of getting goods replaced or monies returned by making them part of the contract.

still try to go fo warranty. I had my failed just within two years and had it replaced no questions asked. Shimano knows, but doesn't communicate about it. I think they probably replace most people even if out of warranty to make it not become a huge media circus.

"We've never seen it before..."

To which the standard retort should be...

"Well, you have now!"