Shimano Crankset recall: bonded Hollowtech ones

Page may contain affiliate links. Please see terms for details.

Jameshow

Veteran
My wife knows a bit about galvanic corrosion, as she is a materials scientist. She confirms that corrosion happens where the two dissimilar materials are in contact, and welding would be worse than bonding, because the glue stops the electrical contact. So, if the issue was anything to do with galvanic corrosion what you would expect to see is the crank arm detaching from the spindle, not the sort of debonding reported.

The more I watch of his videos, the less impressed I am with his knowledge.

But the glue/weld is between two similar metals surely. The corrosion is caused by the steel axle alu crank interface???
 

C R

Guru
Location
Worcester
But the glue/weld is between two similar metals surely. The corrosion is caused by the steel axle alu crank interface???

That's the point. Hambini says galvanic corrosion damages the box section surfaces causing the debonding. If there was galvanic corrosion the damage would be at the steel/aluminium interface.
 

Jameshow

Veteran
Agreed!

Better to weld metals rather than bond them.
How many gms did they save doing that on ultegra/DA? 10gms?!
 

C R

Guru
Location
Worcester
Agreed!

Better to weld metals rather than bond them.
How many gms did they save doing that on ultegra/DA? 10gms?!

The difference between ultegra and the rest is that ultegra arms are a box section formed from two bonded aluminium u sections. The rest use a solid aluminium arm. The welding/bonding I was referring to is the method of joining the steel spindle to the arm, but we aren't seen any failures due to separation from the spindle, as I understand it the failures are due to the two u sections separating.
 
OP
OP
Ajax Bay

Ajax Bay

Guru
Location
East Devon
. . corrosion happens where the two dissimilar materials are in contact, and welding would be worse than bonding, because the glue stops the electrical contact. So, if the issue was anything to do with galvanic corrosion what you would expect to see is the crank arm detaching from the spindle, not the sort of debonding reported.
The more I watch of his videos, the less impressed I am with his knowledge.
Think his preferred hypothesis is that it is galvanic corrosion is adversely affecting the glue (whereas the welding is unaffected, even is as your wife notes, there maybe more corrosion in a welding join).
Fact is the 105 cranks are not failing to any SM extent, and these glued ones have been, for over a decade.
So there has to be a primary failure mechanism. If not galvanic corrosion then what? Link please
https://www.bikeradar.com/features/shimano-crank-failure/
road.cc"an essential preliminary stage in the failure of the cranksets is the partial debonding of the inner and outer U-shaped channels that together make up the rectangular tube section of the crank arm."
Peak torque (2020):
View: https://www.youtube.com/watch?v=Rj__lexd_BI

Candidates:
Moisture ingress
Trapped moisture (during manufacture) and thermal cycling (cool indoors/garage condensation, hot outdoors)
Meniscus effect at the joins (spider arms)
Galvanic corrosion
1698135747324.png
 

silva

Über Member
Location
Belgium
Galvanic corrosion is like a battery. If I assume both U crosssectioned crank parts are alu, where is the other pole?
 

Twilkes

Guru
Not read the whole thread but I'm sure Mark Beaumont snapped a crank on his Africa trip a few years ago, any idea what cranks he would have been riding? I remember the pedal and part of the crank still clipped in to the bottom of his shoe.
 

silva

Über Member
Location
Belgium
That's the point. Hambini says galvanic corrosion damages the box section surfaces causing the debonding. If there was galvanic corrosion the damage would be at the steel/aluminium interface.
The problem with that is that that "damage" is actually alu of the part that disattaches as loose white powder, which at a mechanical connection, causes disconnection. So, that 'box section' should then just come loose from the alu crank parts, without the latter (glued together) separating from eachother.
Yet; that is apparently what does happen here. What is left: the glue itself, regardless the parts. Maybe they chosed a glue that didn't suit the usage.
 
  • Like
Reactions: C R

Jameshow

Veteran
The difference between ultegra and the rest is that ultegra arms are a box section formed from two bonded aluminium u sections. The rest use a solid aluminium arm. The welding/bonding I was referring to is the method of joining the steel spindle to the arm, but we aren't seen any failures due to separation from the spindle, as I understand it the failures are due to the two u sections separating.

Not so....

"The crank arms are forged around a steel block which is later removed, to create a hollow crank arm.
But don’t think for a second that this equates to anything which feels less than rock solid; you’d need specialist equipment to register the slightest amount of flex on a set of Shimano Hollowtech crank arms."

https://www.cyclist.co.uk/reviews/shimano-105-r7000-groupset-review
 
  • Like
Reactions: C R

C R

Guru
Location
Worcester
Think his preferred hypothesis is that it is galvanic corrosion is adversely affecting the glue (whereas the welding is unaffected, even is as your wife notes, there maybe more corrosion in a welding join).
Fact is the 105 cranks are not failing to any SM extent, and these glued ones have been, for over a decade.
So there has to be a primary failure mechanism. If not galvanic corrosion then what? Link please
https://www.bikeradar.com/features/shimano-crank-failure/
road.cc"an essential preliminary stage in the failure of the cranksets is the partial debonding of the inner and outer U-shaped channels that together make up the rectangular tube section of the crank arm."
Peak torque (2020):
View: https://www.youtube.com/watch?v=Rj__lexd_BI

Candidates:
Moisture ingress
Trapped moisture (during manufacture) and thermal cycling (cool indoors/garage condensation, hot outdoors)
Meniscus effect at the joins (spider arms)
Galvanic corrosion
View attachment 711101


The galvanic corrosion mechanism would produce a different failure, which would be seen in 105, Tiagra and Sora as well.

I've been involved in projects where we were using similar processes, and there are two main reasons why it is difficult to get right in mass production.

1. The adhesive is usually a two part system, which requires care to ensure that the correct proportions are used. Some systems are very demanding, and sometimes people get sloppy and end with substandard adhesive because of that. It can also happen that the factory decides that something cheaper and less demanding will be fine instead, and by the time you find out a lot of almost good enough but not quite parts are out there. DAMHIKT.

2. Surface preparation. A good bond usually requires particular surface features and cleanliness. A particular surface finish might be expensive, and again, the factory may decide that a visually similar but cheaper finish will do, until it doesn't.

As I said before, there's probably nothing wrong with the design, other than probably being quite demanding in manufacturing control.
 
OP
OP
Ajax Bay

Ajax Bay

Guru
Location
East Devon
The inside of aluminium spider and crank deteriorates and that adversely affects the glue between the two aluminium parts which then separate under stress because the crank's strength depends on its box section. The weld is a weld.
I have no figures but SM reporting suggests that this failure is rarely reported in the left crank (which has the same section etc but no steel to provide the cathode). Of course it hasn't got all the spider interfaces that the right hand crank has either, so other failure mechanisms (see above) could be in play.
 
Last edited:

silva

Über Member
Location
Belgium
The spindle is steel; the spider/crank material is aluminium. Add electrolyte (water with some salts) and the higher metal will corrode.
The "least noble" metal corrodes, alu is one of the least of all metal, the corrosion manifests as dissolving the alu, so one would expect the spindle to start having play inside the crank hole, but the crank, constructed of two longitudal parts, glued together, rips apart. If it was galvanic corrosion inside the two crank parts, that could serve as explanation, but not the spindle. I had had galvanic corrosion in my bike stand, an alu block with two bolt holes in it, between the frame tube and the stand, the bolts were A2-70 stainless, and my stand started to get play, needing further retensions all the time, until I threw the thing open, and saw the bolt holes full of white powder.
Nothing broke or so, just play, because that powder is "lost" material form the parts.
So it's like hard to see galvanic corrosion here as an explanation for the U parts separating.
The lost material / play actually makes the electric current stop flowing, or less contact time, decreasing or halting further corrosion, but play is not acceptable on a drive train. For ex, a tyre valve is also stainless, and if you look good, alot show some white dust traces around their basis, that's alu from the rim, but it does the hole grow in diameter so the problem stops, and the play of the valve in the hole is not a problem.
 

fossyant

Ride It Like You Stole It!
Location
South Manchester
Blimey folks, they were a bad design. I'm sure if you looked after them they'd be fine, but most folk get them covered in crap, salty road water etc etc and don't clean them, and the bonded design didn't hold out. It's a silly design for a 100g saving or so from a forged crank. The MTB cranks aren't hollow, nor are the older ones. I'd take older or generic cranks over these as they are much easier to keep clean, for a slight weight penalty that's not in a crucial area. I really object to being even more limited on chain rings - you need the OEM fancy ones per chainset or they look crap.

I don't like being limited when replacing wearable parts.
 
Top Bottom