Dura Ace 9000 11 speed cassette- rubish product design

[4914]

11.4
I don't have anywhere near the expertise regarding manufacturing, materials or failure rates which by the was is very impressive- is say this sincerely and not with sarcasm.
So am asking you this question, if just one of the rivets lost even the smallest amount of tension then the integrity of the structure changes, is it not possible for this change to cause the carbon body to fracture if it was not designed to deal with the change in stress/load?
And considering this is it not possible that the body could fracture whilst leaving the rivet in tact?

[4914]

11.4
Also can you explain to me why you say there are three different failure modes?
I'm not sure how you could come to this conclusion when its only an assumption with no definitive explanation for the undue loads on the cassette causing failure.?

[4914]

11.4
I don't see my self as a track sprinter, I was a track sprinter and I have two uci world cup gold medals and a world championship silver to say so.

I'm not being funny, but do you really think I wouldn't notice if the was anything that wasn't operating perfectly?
The chain is fine and still is, the gears shift smoothly and precisely and still running perfectly....
There was and is absolutely nothing else wrong with my bike then and now and I was not doing anything out of the ordinary, there was no chain hoping or gear poping.... nothing.
can you understand why from my perspective I even started this post in the first place... I couldn't believe it myself.

[104]

Another data point: I've been riding DA9000 since it first came out. I broke my first cassette just like the OP stated, a rivet popped loose from the carbon spider, two of the cogs then lodged together.

I spoke with Shimano who quickly replaced the cassette under warranty as well as my chain which took a beating since I had to nurse my bike home with a broken cassette. They also gave me some bar tape as an apology. The customer service/warranty guy I talked to said "we've seen this a few times so far".

That was in early July.

Just today I was out on a ride. I heard a pop and thought maybe the pawls in my hub jumped or something. No more than 2 min later a loud POP and I was left with another broken carbon carrier in my cassette. Exact same thing as before.

This second broken cassette has abut 900 miles on it.

I am meticulous with cleaning my bikes, and thoroughly inspect them at least after every third ride. I did a pretty complete overhaul on my bike just yesterday and there were no signs of anything wrong.

I installed the first cassette that broke, a top-tier mechanic (who has traveled with numerous professional CX racers in Europe as well as the Czech national CX team to CX Worlds this year) installed the second.

[104]

I don't have the picture of the first one i broke in July, but here's the one from today. The breaks were virtually identical.

[SMITHERS]

104,

Just curious, what wheelset are you using with these cassettes?

The OP stated that he was using the 2013 Mavic Ksyrium SLS.

Could these failures be due to the casette body? Maybe the carbon cassette splines are not strong enough under load? I am dumbfounded...

[104]

@SMITHERS: Handbuilt King R45 hubs to Open Pro rims with CX-Ray spokes (3x all around) this time, and DT Swiss 240s to ENVE 3.4's the first time.

Go ahead and take that freehub body theory out of it.

Also, take the talk of 1700w out of it as well. My best 3s effort ever is something like 1200w.

[104]

I was really pissed the first time the cassette exploded because it nearly caused me to crash immediately in front a bus. No sarcasm, I was within about 15' of being run over by a 40,000lb vehicle.

This time when it broke it also snapped my chain. Luckily I was able to patch the chain on the road and limp home. Nearly caused me to crash again, but luckily there was no traffic this time.

[pdlpsher1]

Here's a theory. Riveting takes a lot of force at a very high speed. Riveting into carbon fiber could causes cracks that would later lead to catastrophic failure.

Another weird thing from the photo is the way the carbon failed- it almost looks like the carbon is brittle. The cassette spider requires a material that has reasonable toughness. Perhaps it's just a bad batch of materials.

[PSM]

I will keep my 7900 group.

[pdlpsher1]

I will keep my 7900 group.

9000 = 11 speeds - 3 speeds = 8 speeds

[mattr]

Didn't shimano stuff up the design of the xt 9 speed cassette as well?
The one with the 3 large sprockets on a 4 arm spider, sure I saw a good few of them folded up. They went to 5 arm shortly after.

[11.4]

Here's a theory. Riveting takes a lot of force at a very high speed. Riveting into carbon fiber could causes cracks that would later lead to catastrophic failure.

Another weird thing from the photo is the way the carbon failed- it almost looks like the carbon is brittle. The cassette spider requires a material that has reasonable toughness. Perhaps it's just a bad batch of materials.

Here's a theory. Riveting takes a lot of force at a very high speed. Riveting into carbon fiber could causes cracks that would later lead to catastrophic failure.

Another weird thing from the photo is the way the carbon failed- it almost looks like the carbon is brittle. The cassette spider requires a material that has reasonable toughness. Perhaps it's just a bad batch of materials.

As far as I can tell, these are internally peened rivets -- this keeps them from putting pressure directly on the carbon fiber spider.

I keep coming back to the fact that when cassette failures happen, they happen in clusters to a very few riders. If three riders got together and had ... what is it, 8 broken cassettes? ... then how many individual cassette failures would we be hearing about from everyone else? I'm not at all denying that the cassettes may be designed so they are reliable in some parameters but weren't designed to cope with certain abnormal or unpredictable stresses. It could even be a bad shift that twisted one cog -- a combination perhaps of a cross-chain load plus a rear derailleur that hiccups plus a connecting link or rivet in the wrong place. And combinations like this may happen because of bike setup so they keep occurring. And no, they don't nuke a SRAM cassette because it doesn't have a spider that is subject to the same peculiar combination of issues.

These kinds of failures are almost impossible to diagnose fully and almost impossible to design around unless you overbuild significantly and in cycling, give away both weight and the tight spacing of eleven-speed cassettes. Shimano shifting geometry and technology is quite different from Campy or SRAM, which is part of why it works so well when it works well. But again, it's why one unique situation may occur that only affects a 9000 cassette. I've also seen current Campy setups that on occasion cause incessant chain breakage for isolated riders -- analogous to what is going on here with 9000 cassettes.

I'd honestly suggest that you have the bike hangar and drive train alignment checked by someone who really knows what they're doing (preferably a good framebuilder). Go to a different chain and intentionally change the number of links. Change the b-screw setting. And so on. Be sure the free hub isn't sticking at all, nor the rear derailleur jockey wheels -- fastest way to get this kind of failure is if a couple links fold up and jam on the cassette right above the upper jockey wheel -- you'll have a failure essentially without warning even if the cause has been there all along or since a change induced the scenario. My point is that there's a way to tear a cassette spider apart, and it can come about from a combination of issues that is almost impossible to diagnose or predict. The cassette is generally reliable and riders shouldn't be hesitant about using them, but if you do start to have problems with one, I'd look at the whole drivetrain and its geometry to see what is causing this.

As for the suggestion above that the carbon looked brittle and may have been a bad batch: If anything, I'm seeing too much shredding, which would suggest the opposite. But I don't really think this is bad carbon. And it's the nature of this design that, in the absence of the kind of abnormal forces such as I discussed above, the actual carbon structural parameters don't matter too much. The design relies on the bulk of the spider and on how the two sets of rivets reinforce the spider with the strength of the cogs themselves. You could almost make that spider out of polycarbonate or lexan and get the same performance. I think this is just a resin loaded with short shopped fiber, which is how you make a composite that has to bear strong compressive loads, and the spider is designed so the loads are basically compressive. This is completely different from how a frame tube would be designed, but not unlike how a carbon fiber stay end is created.

Again, just looking at what was actually happening in photos, and adding a little experience with composite design and failure. I wish I had a more definite answer for you, but this is the kind of failure that frustrates component designers. These kinds of failures will often happen -- every chain, every fork, every stem is subject to it. You just have to allow that it's going to happen somewhere, to somebody.

[pdlpsher1]

As a tandem owner I believe all bicycle parts should withstand the forces experienced on a tandem. I'm running XT cassette and KMC chain on my tandem with zero failures (over 12,000 miles). I've also used older Dura Ace cassettes on a tandem with no issues. I would not dare put a 9000 cassette on my tandem.

[eric]

104 and 4914, what cog were you in when the carrier let go, and were you shifting or had you been in that cog for a while?

I can't see how an out of alignment (but still rideable- these bike were being ridden) drivetrain would cause forces that much different than normal. And in any case, the cassette should be able to handle a poorly adjusted drivetrain the same way it should be able to handle
a 250lb rider and 1500+ watts in a sprint.

If maladjusted drivetrains could easily destroy a cassette we'd be seeing a lot more failures of regular non carbon cassettes. I've been around cycling a long time and that's one component failure that I have not seen, either in person or posted on the internets.

I'm surprised at the wear on the splines on the cassettes. And 104's looks to have broken at the spline (lower left corner) in addition to the rivet. That might have happened after the rivet broke rather than before.