Shimano cranks failing. Time to do something about it?

I have zero inside knowledge re: this particular crank issue…but I do know that Shimano is a very conservative / cautious company. If they believed / knew there was a large scale issue with these cranks, I would think they would have issued a recall.

What we do know is that they have changed their manufacturing processes along the way (TBH, I forget the exact details, but there were changes made).

Are there a large number of anecdotal incidents of crank failures? Yes, with a heavy emphasis on “anecdotal”. There is no doubt there are cranks that are failing…to what percentage they are failing no one knows (except Shimano).

My gut reaction is that it is very likely a very small percentage of cranks…and a question that needs to be asked is to what percentage they are failing vs. previous crank models. Again, that is a question to which only Shimano knows the answer to.

The fact that Shimano is continuing with the design concept in some fashion to the 12 spd groupsets also indicates that they are confident in the design overall, albeit it updated as noted above.

Finally, it is worth noting that not every incident of the purported failures is the result of a potentially flawed design. We assume it is when we see these reports, but we have no idea what the history of the individual cranks are.

I’m not defending Shimano necessarily, just providing additional context to the discussion.


My current 9100P crankset survived another ride today, 102 days and counting… but I did hear a creak I need to look into, perhaps the end is nigh.


105 cranks are 1-piece not bonded 2-piece like DA/Ultegra. They are not prone to the fail mode being discussed. I have both and no problems so far. But 20-30 yrs later, who knows. Given a choice, I’d choose 105.

1 Like

Any factory is going to have some failure rate, that’s a given. Think about how many of these cranks are actually out in the world and how many have actually failed. The percentage failing is minuscule. Thanks to one IG account the problem is blown out of proportion. Maybe I’ll start an IG account that focuses on every time a Roval rim cracks? If it was a problem “worth” fixing, Shimano would fix it. I’m sorry if this sounds gruff to anyone who has had a crank snap on them mid-ride, that would suck…


You’re being way too logical and pragmatic. You need to go scrolling through IG and get woke. Shimano are the devil, and need to issue an immediate recall for every bike that has Ultegra or DA OEM cranks. Pull the bikes out of the shipping containers they are stuck in if they have to. This lack of action will not stand!


Close to 60,000 km on my R8000 Ultegra cranks and no issues.
I do have an acquaintance in Sydney who did break theirs, so it’s something I’m conscious of and occasionally take a longer look to see if there’s anything developing.

1 Like

I personally only knew 2 persons that have that problems. Both with the early R8000 cranks.


Road cc has an article about the issue. The usual problem around quality issues is that you can’t really know how large percentage that is affected, and you can’t know how many of these failures that are genuine manufacturing/design issues and how many that are caused by say crash damage.

I think it’s great that actually reaches out to Shimano and try to get some answers even if it is sort of a hopeless venture.

1 Like

Shimano made the choice to start gluing cranks together, instead of melt-forge or otherwise non-gluing technique. Why glue? It’s probably cheaper. Could they make a crank of the same weight without gluing the cranks together. Of course they could and they have.

The fact that they have made lightweight non-glued cranks previously but no longer, says something about Shimano.

The solution is simple. Don’t buy or use their cranks. Shimano will adapt accordingly.

1 Like

True but won’t happen as the vast majority of them are OEM, and the end-user just use what’s on the bike they bought.

1 Like

Just had to reorder a new R8000 crank yesterday. The bonding/weld of the right crank arm began to separate causing noise. Took my shop about 12 hours to find it.
No corrosion on the crank at all.
Ultegra has a 2 year warranty and my cranks are about 2 yrs and 10 months since new.

1 Like

My understanding is that this perhaps relates to the different way the R8000 and R9000 cranksets are “forged” compared to older Ultegra and DuraAce and current lower tier items. This is purportedly why 105 and below seem not to have the same issues.

Does anyone know what the threshold for percentage of catastrophic failures is before we say shimano must do something?

This is a fairly dangerous mode of failure. Let’s say there was a car whose engine exploded, killing everyone inside, in 0.1% of units. Would you say “that’s statistically irrelevant”? It seems like there are more shimano crank failures than SL7 steerer failures. Definitely in absolute terms, but also as a proportion. Of course, anecdata blah blah blah, but we have maybe one image on the entire internet of a snapped SL7 steerer. There aren’t even any anecdotes to turn into data.

Everyone is fine with saying “screw the minority” until they end up in the minority.


What? Crank failure while riding isn’t nearly as likely to produce a crash as catastrophic steerer tube failure is. It isn’t even comparable.

Furthermore, cranks will see the greatest forces acting on them typically at low speed when torque is large, while steerers will see the greatest forces acting on them at high speed. I’d much sooner crash due to crank failure while grinding out of the saddle up a steep hill than crash due to steerer failure hitting a road irregularity going down it at 80 km/hr.

To make the comparison even worse, gradual crank failure can be observed before catastrophic failure happens and cranks replaced because cranks are external, while steerers are internal.


no-points-idiotic (1)

1 Like

Ok cool. So as long as we’re on the same page… the issue is not the frequency of the problem, but the severity of it. This entire thread so far has focused on the frequency of the issue, but as soon as I ask for a hard number, you shift the goalposts to the severity of the failure.

I agree that the severity of this failure isn’t that bad. But people have gotten hurt over this. So again, I ask you, what is the threshold? Give me hard numbers. How many cranks have to fail (absolute, as a percentage, whatever) or how many people have to get hurt before we decide that shimano must do something? Is there a legal guideline?


Yeah, that is isn’t even remotely comparable.

Link to the data that you are basing this statement on?

There is no clearly defined number for this. There are a multitude of factors that need to be considered. Frequency, likelihood of injury, etc.

Yeah it seems that the CPSC doesn’t have a good definition for it either. From Title 15, chapter 47, section 2064(a)(2), a “substantial product hazard” is defined as “a product defect which (because of the pattern of defect, the number of defective products distributed in commerce, the severity of the risk, or otherwise) creates a substantial risk of injury to the public.” No other insight there, it seems.

I’m not saying that the law should have a hard and fast number for this. But it’s clear to me that these crank failures aren’t as obvious of a hazard as the sl7 steerer but not a non-issue, either. It’s a gray zone, and the only productive line of discussion is to determine how many failures or injuries we’re willing to tolerate.

1 Like

OK, but if you want to have a “productive line of discussion” you shouldn’t use false analogies (car engines blowing up) or make up your own claims re: the failure rate.

Again, Shimano has historically been a very cautious and conservative company. They aren’t likely to just roll the dice and hope a serious issue just vanishes.

We know they adjusted their manufacturing processes for the cranks since the design was introduced…what the failure rate difference is after the change, only Shimano knows. What overall percentage of cranks are failing, again, only Shimano knows.

The car thing was just to assert that frequency alone is not enough to determine whether something is worth a recall. The CPSC agrees with me, as do you.