SurferRosa

Better be on the safe side and list all the parts that have ever failed, even once. We could start the list with my '80s Davidson frameset in big bold letters. Better safe than sorry. Even one life is worth it!

steelbikeguy

good point, and I was just too lazy to bring the subject up.

Not too long ago, someone mentioned that (some?) GB stems have the same arrangement, so I checked my '74 Raleigh International.... and it did have the slots with the square corners! I did a coarse job of drilling the end of the slot to get rid of the stress risers. In my days working on aircraft, this was also a common technique to stop a crack in the aircraft skin, although they also put a patch over the skin too, IIRC. Disclaimer: my job was electronics, and I was told about this technique.

This is what my GB stem looks like now:





As much as Campy gets a lot grief for this type of failure, other manufacturers produced the same sort of thin web of aluminum at the junction of the spider and the crank arm. My beloved Sugino AT does it too, mostly because the machined the back side of the arm flat. The thin web of aluminum ends up located at the point where the most flex occurs and the thin web is easy to start a crack.

I've got a Sugino AT that I bought used and put about 50k miles on. I filed out that web once, but upon inspection not long ago, I thought I saw a tiny crack....




I don't like unpleasant surprises, so I filed it out some more, until I didn't see a crack. There are dye penetrant tests that can be used to do a proper job of inspecting for cracks, I've been told. This is the "after" photo of the second filing.....




Steve in Peoria
(maybe I should start an ebay search for "Sugino AT 175mm" just to be safe??)

La Brea Bike

Ofmega Mistral crank arms are prone to cracking. It's unfortunate because I really love that gaudy script.

Chuck M 

This is a good enough idea for a thread. But what I have noticed quite often is when folks buy old bikes they want to share with the BF community the people that are aware of potential problems with anything are good about pointing said problems out. Even things that look suspect and not necessarily because they have a reputation or history for failure. For example someone posted a picture of some scratched rims today and it was quickly pointed out that the poster should check the spoke length. The C&V gurus do a good job helping with practically anything.

Road Fan

If you really want to be safe, I think you need to be able to verify that cracks have been removed and demonstrate that they won't form later on with added stress cycling. Like any case of test engineering, the more margin you put into the specification of the stress levels/cycling requirements, either for physical evaluation or for model-based evaluation (i.e. FEA) So some means of evaluating stress risers that are present after doing stress relief, and of evaluating whether the safety measures performed will be adequate under foreseeable stress/cycling patterns.

Aluminum is not elastic like steel is. All these stress cycles can be seen as cold-setting: stressing the auminum beyond its elastic deformation regime, for enough cycles that the mateial breaks down and ultimately, breaks. These problems can be engineered, but I think nobody here can afford to do that. If anyone here has an analytical model if not one based on physical or model-based cycling, that would be a major step forward.

Buy a new one? Sure, but how do you ensure no fatigue failures are in its future?

steelbikeguy

Thomas Jefferson is reputed to have said "Eternal vigilance is the price of freedom". I suppose that applies to the freedom from having a crank break and dump you on the ground.

It doesn't hurt to look for a crank that appears fairly untouched by the usual wear and tear. I don't know if you can rely on a seller to know how to look for these cracks and notify potential buyers if it exists.

Steve in Peoria

EddyR

All Cannondale bikes as everyone knows how fragile aluminum is. Just kidding as older Cannondale’s are my favorite bike. I have seen comments on here almost this bad as things get repeated over and over as if it was a proven thing.
Ed

dmark 

https://bike.shimano.com/en-US/infor...12-cranks.html
https://www.cpsc.gov/Recalls/1997/CP...cle-Components
I have a bike waiting on parts due in April; this is still honored 25 years later.

Erzulis Boat 

Much ado about nothing. The only cycling catastrophes that I have seen have involved cars.

Road Fan

Yes, so vigilance implies monitoring, and monitorinng implies trigger thresholds, and if the process is to be efficient, we should think about optimizing the thresholds. So back to the modeling and analysis I was talking about earlier.

steelbikeguy

So your plan is to purchase a statistically significant number of new Sugino AT cranks, devise various tests that will allow modeling of the cranks' failure characteristics (allowing for the various rider weights, strengths, etc) and then be able to predict the expected lifetime, assuming that we could buy new Sugino AT cranks?

My only experience with this sort of thing, at least in terms of testing parts to failure under known stess levels, is military aircraft. My experience is somewhat second hand, combining experience working on aircraft and then working at an aircraft manufacturer and watching aircraft undergoing load testing. Side note: watching an F-18 Hornet airframe being slowly loaded and unloaded, with all of the assorted creaks and groans, is surprisingly entertaining.

The load testing does provide data on how many cycles of a particular load pattern that the aircraft structure can survive. I'm not sure how often they inspect for early crack development during the tests. With that sort of data, aircraft are monitored for excess stress, as well the general level of stress and number of hours on the airframe. At certain intervals, aircraft were pulled out of the squadron, sent off to NARF (Naval Air Rework Facility) and totally stripped down. They got new wiring, hydraulics (probably), and were thoroughly inspected for cracks.

Some aircraft types get patches and reinforcements added, based on whether cracks were appearing. In fact, the early F-18's had a problem with cracks at the base of the rudders, due to pilots flying at high angles of attack much more than expected. Vortices from the leading edge of the wing were impinging on the rudders and causing the cracks. The solution was to add external reinforcements at the base of the rudders, as well as a strake ahead of the wing to help break up the vortices.

Anyway... for bikes, I suspect that we'll never have data on the expected lifetime of any parts, so we have to fall back upon routine inspections for damage or cracks.

Steve in Peoria

Road Fan

Actually I would focus on simulation in favor of testing, but they're both pretty impractical out here in the wild.

I agree with the impracticality. If a lot of riders, techs, and shop owners become extra careful about radiusing slots and rounding edges (buy some decent files, folks!) to minimize observable stress points, I think whatever the current rate of faiure is, it will be reduced. But no records are kept, so nobody will ever know. No tracability.

2flit

I would add this advice;
Take your vintage Lyotard peddles apart and carefully examine the root of the spindle. Mine snapped off during a modestly hard sprint. But it showed dark staining of the metal that clearing indicates that there was a crack forming there well in advance of the catastrophic breakage.


Lyotard breakage on a 1963 French vintage frame

steelbikeguy

apologies in advance, but is there any chance that you've got photos showing the the surfaces at the crack, as well as anything showing whether there was some sort of stress riser on that part of the axle? i.e. a sharp edge or change in axle cross section? Retired engineers and material scientists are waiting with bated breath.

Steve in Peoria

2flit

Hi Steve,
See this thread (It happens and worth checking you peddls)
https://www.bikeforums.net/classic-v...-danger-2.html

steelbikeguy

that is a fascinating thread!
a bit frightening.. but fascinating!
plus, it makes me feel a bit better about avoiding most French components. I do have faith in my Stronglight A9 and Delta headsets, though. I like my Mavic rims too. ... and T.A, water bottle cages.
Maybe it's just French pedals that I don't like?

Steve in Peoria

JohnDThompson 

FWIW, it's not just French pedals that break:

steelbikeguy

given that it's a track pedal, I'm going to guess that it belonged to someone like Nelson Vails or Marty Northstein??

I've had the flip tabs break on Campy SL pedals, but that is quite surprising!

Steve in Peoria

JohnDThompson 

Not sure. I picked it up used, dirt cheap for the spindle and cone.

JohnDThompson 

The swaging is the weak spot; it can break loose so the chainring is no longer attached to the crank arm. IMO, the hierarchy is swaged crank < "melt forged" crank < cold forged crank.

Wildwood 

but I so want to put the swaged crank on a city bike

eeuuugh

Swaging is a method of attaching a spider to a crankarm, forging is a method of shaping crankarms. I would guess that most swaged cranks have melt-forged arms, but I've seen some I thought were cold-forged... Don't remember which or why though CNC should be in the hierarchy too. Early Kooka cranks are notorious for cracking. Most CNC cranks now have a bulky look (IMO), since the manufacturing process imparts no grain to the metal.

Not all cold-forged cranks are created equal, either. Cheaper cranks are forged from blanks in as few as three dies iirc. Safety-wise this is probably plenty, but the crank resembles a simple bar of aluminum. Nicer cranks are made in five or six dies and can be more slender and have nicer curves. There will always be some final machining after forging. Then there's whether every length of a particular model is made in the same die and then cut to length, in which case the grain won't run all the way around the pedal hole, or whether each length is made with separate dies. The latter process is called net-shape forging; Jan Heine wrote a blog post about it last year. I believe other high end cranks especially in the past used net-shape forging too.

Drillium Dude 

Interesting break, almost like the rider didn't get his foot out during a crash. I imagine abnormal wrenching forces subjected by a stuck-in cleat would overcome the design strength parameters. Never seen anything like this on any pedal, so it's prolly not something to worry about in normal operating circumstances. Note there is no wear to the anodizing of the forward portion of the cage, suggesting low mileage.

DD

JohnDThompson 

Well, it's the rear cage plate that takes most of the wear, especially if you use metal cleats on an aluminum cage. Here's a Zeus pedal of mine that eroded to the point of cage failure in that way:

zukahn1 

+ One