Shimano Crankset recall: bonded Hollowtech ones

[Chislenko]

My replacement Ultegra unit finally arrived today and contrary to initial information from the dealer I have a complete new Boxed Chainset with brand new rings.

When I dropped them off three months ago I was informed it would be just new crank arms with your old rings bolted on to them.

Consequently I am now in possession of a brand new unused 105 Chainset which will probably sit on a shelf in the garage for evermore!

 

[silva]

The labor cost to unmount/remount at dealer individually probably exceeds the cost of new rings+bolts in China, Malaysia or Singapore as mass production.

 

[Chislenko]

The labor cost to unmount/remount at dealer individually probably exceeds the cost of new rings+bolts in China, Malaysia or Singapore as mass production.

They are stripped down to be examined and only the cranks are sent back to Madison. The rings etc are retained at the dealer. Like a motor car recall the dealer is given a set fee for each inspection.

 

[silva]

They are stripped down to be examined and only the cranks are sent back to Madison. The rings etc are retained at the dealer. Like a motor car recall the dealer is given a set fee for each inspection.

Rather surprising that crank inspection requires removal of the rings?

 

[Ajax Bay]

@silva Have you got one of these Hollowtech cranks, or even taken off/replaced the rings? A proper inspection requires the removal of the rings given where the mode of failure typically initiates.
What is your take on Shimano's 24mm spindles?
This is Hambini analysis from 2022, a year before Shimano's recall (volume down and swearing threshold set to high for first minute) - recommend going to 6:00 in the video:

View: https://youtu.be/-uK7FSZ6OFM

 

[Chislenko]

@silva Have you got one of these Hollowtech cranks, or even taken off/replaced the rings? A proper inspection requires the removal of the rings given where the mode of failure typically initiates.
What is your take on Shimano's 24mm spindles?
This is Hambini analysis from 2022 (volume down and swearing threshold set to high for first minute) - recommend going to 6:00 in the video:

View: https://youtu.be/-uK7FSZ6OFM

Must admit I am one of the people who he talks about on the video. Tried everything to eliminate my "creak" new bb, stripped and serviced my Spd-sl pedals, new cleats and so on.

Haven't had chance to fit and test my replacement Chainset yet, time will tell if it was the cause of the creak but the evidence looks damming.

 

[Ajax Bay]

"Look Mum, no crank!"
A friend's experience, after his chainset had been checked and passed 'no problem':
"If your cranks are subject to the Shimano recall do keep an eye on them. Mine had been checked by me and a bike shop and no issues were identified. Today it failed catastrophically though fortunately no major issues other than a ruined ride.
Before it went I thought my cleat was loose as I could feel a bit of flex. I was going to check at the top of the hill but didn’t get that far. Lesson here is that if something doesn’t feel right do check." [especially if you know you have these dodgy cranks]

 

[Chislenko]

Finally had a dry enough day to go out on the "good bike" and give the replacement Ultegra crankset a test.

Repeatedly throwing the chain on upshift. A few in ride tweeks on the in line barrel adjuster has got it pretty much sorted but will give it another coat of looking at in the stand when I warm up!!

The good news, the creaking I have lived with for years has gone so perhaps my cranks were an accident waiting to happen.

 

[silva]

Watched that video, good points there, but I don't think galvanic corrsion plays the major role here.
It dissolves material, so that's a dump of the theory of expansion of material due to corrosion, breaking metal parts away from eachother, overcoming glue (and any) binding?
That contradicts, since most glues are electrically insulating - they need a filler to make them conducting, and why would Shimano chose such a glue - they cost more and have a specialty usage (electronics).
I think it's the glue binding that was already poor at fabrication time, and gradually disattaches due to mechanical stresses and also moisture.
Look at how the guy puts a screwdriver in the hole and after some mmmpff forth and backs and then screwdriver inserting in the appearing sleeve to get them come apart completely. I'd say that was a fastforward of realtime due to leverage and uncommon force directions.
Glue is never a reliable connection for things that suffer a leveraged force under an angle let alone more angles, as soon as clamping gets lost at some point, the leverage causes an exponential going rest of failure process.
Why did Shimano go for a hollow crank?
The given explanation is weight reduction, but apart from in racing where splits of seconds make THE difference, who cares?
A cookie more in the pocket or a last haircut longer ago makes the same difference regarding speed and regarding inertia, it's not like that a bike starts and stops or goes faster and slower all the time, rather the contrary.
I think the root of the choice lies in cost saving / profit reservation.
The processing of massive alu pieces to a crank form probably costs more than stamp-bending plate material over a mould then glue them together.
Look at how glue (and silicone) is dominant in all the cheap crap in the world.
Moisture is a big enemy of all glues. a slow but surely moving enemy.
Look at all the light (= one layer textile) raincoats out there. "Taped seams". After some rain the tape disattaches, and if you go to the shop they show you a silicone sprayer to make the coat again a raincoat. Recommended to repeat after every usage.

Designed by Shimano, demanding a cheapskate price, then the order goes to the company that is boldly willing and also allowed by the countries state to do whatever it takes to achieve that price. And Shimano must surely have been aware because of their rather long in time denial.
If you smell nothing, see nothing, hear nothing, but your neighbor yells Fire! then you gonna take a look anyway eh?
Shimano didn't, when forced to look, ignored, when forced by alot to look, denied, and only when forced to look by the many (big media) finally admitted.

Any info about how the replacements solved any claimed cause?

 

[classic33]

Watched that video, good points there, but I don't think galvanic corrsion plays the major role here.
It dissolves material, so that's a dump of the theory of expansion of material due to corrosion, breaking metal parts away from eachother, overcoming glue (and any) binding?
That contradicts, since most glues are electrically insulating - they need a filler to make them conducting, and why would Shimano chose such a glue - they cost more and have a specialty usage (electronics).
I think it's the glue binding that was already poor at fabrication time, and gradually disattaches due to mechanical stresses and also moisture.
Look at how the guy puts a screwdriver in the hole and after some mmmpff forth and backs and then screwdriver inserting in the appearing sleeve to get them come apart completely. I'd say that was a fastforward of realtime due to leverage and uncommon force directions.
Glue is never a reliable connection for things that suffer a leveraged force under an angle let alone more angles, as soon as clamping gets lost at some point, the leverage causes an exponential going rest of failure process.
Why did Shimano go for a hollow crank?
The given explanation is weight reduction, but apart from in racing where splits of seconds make THE difference, who cares?
A cookie more in the pocket or a last haircut longer ago makes the same difference regarding speed and regarding inertia, it's not like that a bike starts and stops or goes faster and slower all the time, rather the contrary.
I think the root of the choice lies in cost saving / profit reservation.
The processing of massive alu pieces to a crank form probably costs more than stamp-bending plate material over a mould then glue them together.
Look at how glue (and silicone) is dominant in all the cheap crap in the world.
Moisture is a big enemy of all glues. a slow but surely moving enemy.
Look at all the light (= one layer textile) raincoats out there. "Taped seams". After some rain the tape disattaches, and if you go to the shop they show you a silicone sprayer to make the coat again a raincoat. Recommended to repeat after every usage.

Designed by Shimano, demanding a cheapskate price, then the order goes to the company that is boldly willing and also allowed by the countries state to do whatever it takes to achieve that price. And Shimano must surely have been aware because of their rather long in time denial.
If you smell nothing, see nothing, hear nothing, but your neighbor yells Fire! then you gonna take a look anyway eh?
Shimano didn't, when forced to look, ignored, when forced by alot to look, denied, and only when forced to look by the many (big media) finally admitted.

Any info about how the replacements solved any claimed cause?

What are the "uncommon force directions", given what it was built for.

If shops are selling you waterproofing sprays, telling you to used every week, they've seen you coming.

 

[Ajax Bay]

Watched that video

I don't think galvanic corrosion plays the major role here.
It dissolves material, so that's a dump of the theory of expansion of material due to corrosion, breaking metal parts away from each other, overcoming glue

Why did Shimano go for a hollow crank? The given explanation is weight reduction, but apart from in racing where splits of seconds make THE difference, who cares?

Shimano must surely have been aware because of their rather long in time denial.

Any info about how the replacements solved any claimed cause?

Assume you mean Hambini's one from 16 months ago.
Do you mind me asking if you have a background in materials science?
Can you suggest why it's (almost) always the RH crank/spider which fails, and not the LH crank? Is the rain/spray always coming from the right?
At the 'top' end, racing cyclists ARE interested in weight reduction and seconds DO count. They care, and spectators care who wins.
Your 'who cares' argument fails over many decades. Also see @Sallar55 comment (p7) suggesting that Shimano felt pressured to compete with carbon crankset offerings which started to appear (almost all of which withered on the MTBF vine).
It's quite clear Shimano were aware for several years and the wholesale replacement (USA) and worldwide inspection was eventually forced on them. They did not handle this well, certainly in retrospect and probably elements of the company were arguing strongly for this versus those (?marketing/sales) who wanted to maintain the line: 'a few failures, nothing to see, move on'.
105 cranks don't fail (like this) because though they use two-piece construction, the difference is that welding is used as the joining process as opposed to chemical bonding with glue or epoxy. So they're stronger but heavier.
@faster explained this on Page 2 https://www.cyclechat.net/threads/shimano-crankset-recall-bonded-hollowtech-ii-ones.294152/page-2
It's been suggested that the later Dura-Ace and Ultegra cranksets made post-2019 though still 'hollow' have had some modifications to reinforce them or changes to the manufacturing process/environmental conditions (in factories) - don't know what these are.
An analysis, rather irritatingly with no quality conclusions :
"An essential preliminary stage in the failure is the partial debonding of the inner and outer U-shaped channels that together make up the rectangular tube section of the crank arm. When satisfactorily bonded the section is of adequate strength to resist any loads that it commonly sees. However, if partial debonding occurs between the parts the upper/inner channel is vulnerable to the size of loads that may commonly be applied during use.
"No conclusions were possible as to how debonding between the two U-shaped channels initiated. Flaws resulting from the manufacturing process, environmental deterioration and fatigue were all considered as possibilities. Whichever was at fault, such partial debonding of the two channels is a notable weakness that might eventually lead to gross failure."

Off (thread) topic but associated failure incoming, we're going to have fun for good while with the "How safe are 'hookless' rims?"

View: https://youtu.be/rgra7i7i_-I

 

[silva]

What are the "uncommon force directions", given what it was built for.

If shops are selling you waterproofing sprays, telling you to used every week, they've seen you coming.

See the video I talked about, the screwdrivers here and there, pointing here and there direction, until the crank became 2 separate parts, a matter of a minute.
Mounted on a bike, the crank parts may have stayed together a great magnitude minutes longer, depending on a variety of conditions. Uncommon conditions occur, a hit against a wall at home, a pedal hitting an elevated road section, a bike falling on its side, a splash water over it, and where it is build for, that's for being sold of course, the rest is secondary.

... because they knew beforme me that their raincoats weren't holding out rain long enough to justify their sticker "raincoat".
Ofcourse I didn't buy that spray, and neither another raincoat there, instead, I learnt how to recognize fake raincoats as fake,know when to not buy, regardless price.

 

[silva]

Can you suggest why it's (almost) always the RH crank/spider which fails, and not the LH crank?

Maybe the spider eats the glue?

 

[silva]

@Ajax Bay

A bit more serious this morning after yesterdays humour before on the road again: is left/right, regarding forces, comparable in the 1st place? One can compare with the other only when they are mechanically seen symmetrical, and also all force vectors.
Also, this is hollowtech, the right crank is an integration with a mount for chainring(s), the left one is mounted by a clamp over a splined cilinder.
Judging right cranks construction based on left crank cases is not really justified.

You said that an inspection resulted in an OK, and yet it failed afterwards.
I don't really understand that inspection round, no external signs of coming loose doesn't imply not any surface disattached.
To be sure, one would need to try to break the crank, a sacrificing inspection - it doesn't make sense.
The cranks construction isn't reliable, nothing less than that. You can't trust it, including after an external inspection.
Which is a rather unpretty situation, you ride around knowing that a crank may suddenly fail, and you have to inspect it frequently enough to avoid stuck along the road.

https://www.cpsc.gov/Recalls/2023/Shimano-Recalls-Cranksets-for-Bicycles-Due-to-Crash-Hazard

Remedy:

Consumers should immediately stop using the cranksets manufactured before July 1, 2019, and contact an authorized Shimano dealer to schedule a free crankset inspection. Only consumers whose cranksets show signs of bonding separation or delamination during the inspection will be provided a free replacement crankset and installation.

... but during the first ride after the inspection the first external sign may appear.
Maybe even more likely than before, since the unmounting handling and inspection themselves may have given the failure process a boost.

https://www.bikeradar.com/news/shimano-crankset-recall
It links a video showing the inspection:

The process covers three steps, as below. Shimano has also published a video running through the process.

View: https://youtu.be/mKMKsbts3t8?si=YcvYmjI1OC9dB56z

... but it is flagged as "private video", you need permission, a login, to view it?
Also there, about their replacement cranks:

How will replacement cranks differ?

Because these models have now been discontinued, Shimano says an affected crankset will be replaced with a new one that “uses the latest construction and bonding techniques

... so they had previous construction and bonding techniques.
... meaning that they changed these.
... meaning that they had reasons to do so.
... meaning that they were aware of the being faulty of their product.
Just logic, that is, no?
And even upto that statements time, they still don't want to explain people those reasons, the fault in their product.
Guys we didn't make our parts thick enough.
Guys we chosed a glue whoms bound is vulnerable to bicycle usage conditions.
Guys our spider frets away our glue.
... things like that, I would say, inflicting so many people such a hassle, deserves a technical explanation, but no, the latest of the latest fixed the fault people, trust us?

 

[Yellow Saddle]

Having never examined one of these failed cranks, I don't know the answer as to why this happened. I have some observations though. A U-shaped structure is tortionally very weak. But put a cap on it and make a D or a square, and you have a pretty good tortion transmitter. Think plastic rain gutter vs a plastic downpipe. Same amount of material in both, but handling the gutter is like playing with an extended Stanley retractable tape measure.

Shimano's crank is a gutter with a cap on. The cap is the crucial part and as long as that remains stuck to the U, the who thing is pretty strong, light and easy to manufacture. Welding would have been ideal but glue was chosen. Glued aluminium is perfectly OK if you understand aluminium and organise your assembly line accordingly. Aluminium oxidises within hours of being exposed to air. That first layer of oxide then protects it against further oxidation. However, that outside layer is not strongly attached to the base aluminium. If you glue over it, the glue will stock to the oxide and the latter will separte from the aluminium.

The aerospace industry uses a lot of glued aluminium, but on structural parts they either glue immediately after the alu was finished or, they anodise the aluminium. If the parts are made outside the factory, they are almost always anodised first since there is a time-lapse they cannot control.

I would like to see how Shimano addressed the problem. Maybe they already have, but I wonder if they won't bite the bullet and go carbon.

One thing I got out of this disaster is to finally find out how Shimano's magical hollow forging worked. Turned out it wasn't forged hollow, it was built hollow from solid forged pieces. It was true magic when I first saw it. Now the trick is just an illusion.