Potential Catastrophic Frame Failer Averted ….sort of.
- Thread starter e-beach
- Start date
e-beach
10 MW
Buk___ said:
Hummm......Looks like I will have to dig deep into the ole tool box and look for my pop rivet gun.
speedmd
10 MW
e-beach said:Buk___ said:
Hummm......Looks like I will have to dig deep into the ole tool box and look for my pop rivet gun.
Thanks for posting the crack photos. It lived it's life. Looks to be much too small a fillet on the lower weld on the down tube side. The sharp angle at the welds edge certainly did it no favors. Now that it is apart, take a deep breath and kick it to the curb. It did you a favor. Buy the best-strongest frame you can afford. Don't look back.
Chalo
100 TW
You know that steel work hardens. It's why we cold forge things. And we all know that steel fatigues because we see bikes and other objects that have cracked that way.Buk___ said:
To the nearest approximation, bike frames are not built that heavy. Ever. That's the problem with your assumption about steel being immortal. If you built a steel frame that heavy, the bike would suck. And you could build an aluminum frame two-thirds as heavy that would be stiffer and do just as good a job of outliving its rider.
And yet we have some aluminum bicycle frames, and aluminum airplanes, and all sorts of other aluminum machines from the 1930s still in service. Lots and lots of aluminum bikes and airplanes from the '80s that haven't had a break yet, in either sense of the word. So the implications can be lived with.
.
That's just not how things work in the real life of bikes. Weight for weight, aluminum bikes display greater longevity-- which is not an especially fair comparison because even bog standard aluminum frames tend to be substantially lighter than steel ones. Intelligently traditional aluminum frames like '80s straight-gauge Cannondales, Kleins, and Treks outlast their heavier steel contemporaries. They're overrepresented among the long term survivors of their time period.
Aluminum has proved to be so suitable and long-lived for bicycle frames that manufacturers now capitalize on its strengths and do stupid things like curved truss tubes, that result in heavier and weaker frames than necessary.
Steel's awesome for DIY, and more than satisfactory for bikes, but for commercial bicycle frame production it doesn't have any noteworthy engineering advantage over aluminum. The reason I like most commercial steel frames better than most commercial aluminum frames is because they still use straight round tubes, and aluminum frames have gotten silly.
The reason that the pedicabs I design and build are made from chromoly is because heat treating an aluminum frame four feet wide and seven feet long is difficult and expensive, and because there is a strong likelihood that these trikes will need some kind of repair requiring welding eventually.
Consider this: most quality suspension forks are made of aluminum. They're cantilevered and unbraced. How often do you hear of them fatiguing and cracking? I don't think I've seen that yet.
liveforphysics
100 TW
Buk___ said:
It would perhaps work fine that way. It would still be unlikely to be as strong as casting tape.
speedmd said:
Respectfully, that would be a shitty solution at best. Everywhere you add a new welding bead just makes new failure points. As a kid who grew up living at a home welding shop, I played that game with my own broken bike frames and it only shifts where the aluminum fails.
If you're re-welding aluminum that's already been welded and heat-treated, you've already lost the game.
I ended up having my father help me re-build the broken sections out of cromoly steel, and slip-fitting and clamping them to the aluminum tubing after no amount of added aluminum scab-plates and gussets offered more than another few days of riding before it re-broke.
liveforphysics
100 TW
e-beach said:
I'm actually doing grid-tie battery development for a client in LA right now and have a premium TIG welder available here in LA I could weld your frame with anytime. However, it wouldn't be doing you any favors to weld it because it would just fail faster and more impressively after re-welding (I've learned this the hard way).
https://www.amazon.com/82003U-ScotchCast-Plus-Casting-Yards/dp/B01BCREML2/ref=sr_1_9?ie=UTF8&qid=1516244507&sr=8-9&keywords=casting+tape
I just ordered 6 rolls of that tape (should be delivered friday), and I will wrap it for you if you want. Just a couple rolls would be more than adquate to make the head tube joint tougher than it ever was when new. I realize it seems hard to believe, but after you play with the stuff you will understand its freaky tough stuff, and it won't matter if drill-stopping the crack stops it's growth or not, because you could build the frame from just a few rolls wrapped properly around all the joints as a complete welding alternative. It looks goofy because it's bulky compared to using metals, but I would trust my life to it for aluminum over any welded solutions.
e-beach
10 MW
liveforphysics said:
speedmd
10 MW
Is that frame even heat treated? May not be. Do we even know what alloy it is. Certainly a can of worms.
Regardless, welding requires proper bead formation, weld prep and correct amounts of penetration. Not something that can be done by a amateur and expect expert results. I agree, it takes time and understanding to fix something like this and make it reliable. I would junk it if it was mine without giving it a second thought. Have the know how, tools and skills to do it, but would not put the work into it. It has lived a good life. May make a good trade school practice part.
Chalo
100 TW
Buk___ said:
Do you mean built to withstand foreseeable working stresses while remaining under 250MPa throughout? Remember all steels aren't equal in strength, and their fatigue limits are a function of their yield stresses. Very thick or very large diameter steel tubes that could keep maximum stress levels lower than 250MPa in all normal use are a feature of frames made from soft mild steel, which means they have lower fatigue limits stresses than that.
A36 and 1018 mild steel have yield stresses in the range of 250MPa, so their fatigue limits are about half that much.
I'm reminded of the Chicago Schwinn Super Sport, which had an unusual combination of thick walled chromoly tubing and huge brass fillet-brazed joints. If there's a steel bike that stays inside its fatigue limits everywhere, during all normal use, it would be like that-- with unusually thick or unusually large diameter tubing (or both), plus high strength alloy and comprehensive stress reducing joint geometry. As it is, I doubt the dropouts and seat cluster on the old Super Sport are smoothed out enough to prevent high-cycle fatigue.
Edit:
I see you rephrased your question. And there are two reasons I can't give a definitive answer to that question. First is, even though a frame can be characterized by stress analysis, the stresses it will see out in the world are not easy to characterize or quantify. What are normal forces for one rider are abnormal for another. So what constitutes "normal forces" pretty much has to include most of the ugly end of the bell curve.
Second is, I don't know what the stress levels are in any units, by any form of direct measurement. All I know about stress levels is what I can infer by comparing the materials' properties with whatever failures I see in my own bikes and the thousands of bikes I've inspected to one degree or another at work. I can safely assume, based on the cracks I have found and the bikes that have had them, that almost any steel bike sees enough stress in its points of highest stress concentration to tally up an HCF cycle count for some part of the frame. Most don't count up enough cycles to hit the jackpot.
I've seen many steel bikes cracked in the same place as the OP's bike. Most of them were not lightweight sporting bikes. Other common spots for steel frames to crack are the seat tube to bottom bracket joint, seat stay to seat tube joints, and chainstay to dropout joints.
Chalo said:
A Raliegh R700 50cm road racing frame made of Reynolds 853 tube. 1" 0.7/0.4/0.7 top tube; 1 1/8" 0.7/0.4/0.7 down tube; single butted seat tube; 0.7 constant diameter stays. 1.5mm head tube and BB shell. Total frame weight 1.525gms.
Elite rider generating a little over 1000N down strokes, through a 50/15 ratio, at a cadence of 120rpm for 1 minute producing 620W.
Maximum stresses occur in the down tube just down from the head tube; in the chain stays just in front of the dropouts; at the back of the seat tube a bit up from the BB; the outer edges of the BB shell; in the rear lower and front upper rims of the head tube.
The maximum stress seen is < 220MPa. This below the 240MPa fatigue limit of mild steel and well below the 480Mpa fatigue limit of normalised ChroMoly.
Elite rider and very light frame and it is well within -- by a safety factor of 2.something -- the fatigue limits of ChroMoly steel.
Use 0.9/0.6/0.9 or plain 1mm steel tube everywhere for ~double the weight, and cheap steel, and it will never fatigue!
The case for the defense rests.
liveforphysics
100 TW
I've personally never felt like pedaling related forces were significant compared to riding impact related forces.
I've met e-beach, and something tells me his frame didn't crack from pedaling (no offense to e-beach who is a nice guy).
Reynolds 953 was my favorite steel on paper for a while due to better strength to weight than any aluminum or titanium alloys.
http://www.reynoldstechnology.biz/materials/steel/s-953/
But it turns out, it's as rough as working with fussy alloys of Ti to get it to keep its strength through heat-based joining processes. Brazed lugs would be about the only option to keep it's properties, and maybe the ultimate strength and high-fatigue solution for making any performance bike frame short of just using carbon.
Things built to last from aluminum you generally don't find welds on (like an airplane), that said, if you have the right alloy and you weld it in the right environment with the right processes and put it through the right heat-treatment cycles afterwards, it can be maybe 90-95% of it's original strength. Lots of bike mfg's mastered that art, but even then it only works for the first time building the frame.
I've met e-beach, and something tells me his frame didn't crack from pedaling (no offense to e-beach who is a nice guy).
Reynolds 953 was my favorite steel on paper for a while due to better strength to weight than any aluminum or titanium alloys.
http://www.reynoldstechnology.biz/materials/steel/s-953/
But it turns out, it's as rough as working with fussy alloys of Ti to get it to keep its strength through heat-based joining processes. Brazed lugs would be about the only option to keep it's properties, and maybe the ultimate strength and high-fatigue solution for making any performance bike frame short of just using carbon.
Things built to last from aluminum you generally don't find welds on (like an airplane), that said, if you have the right alloy and you weld it in the right environment with the right processes and put it through the right heat-treatment cycles afterwards, it can be maybe 90-95% of it's original strength. Lots of bike mfg's mastered that art, but even then it only works for the first time building the frame.
liveforphysics said:
The point is, it isn't the big impact that cause fatigue failure. They can exacerbate it, and if they're big enough to go over the yield point, bend or even snap.
But fatigue is repetition of 'normal' working stresses. Pedal cadence and road chatter.
Smaller forces that, in steel, because they are below the fatigue limit, do not have any permanent or accumulative affect.
But in aluminium, because it doesn't have that minimum below which it is unaffected -- no fatigue limit -- those constant, repetitious, smaller forces each have a small effect, and they accumulate.
Take a nylon zip tie, lay it on your vice, and hit it with a hammer as hard as you like, pretty much nothing will happen. Clamp a cm or two at one end in the vice jaws, clap a set of mole grips on the other end and pull it all the way over one side hard and back to the other hard; and if you start with a small enough zip tie and are strong and persistent, you'll eventually fatigue it enough to cause it to break.
You can fatigue steel, but only by exceeding its fatigue limits many, many times. Or you can do the Yuri Geller pre-weakening a spoon handle trick by repeatedly bending it back and forth past its yield point.
Once again I say, it doesn't mean that aluminium structures -- including bike frames -- cannot be designed to be strong and last a long time; but the design criteria are much different to steel; the fabrication requirements are far more stringent.
The smallest particle of steel (or brass) from a tool say, or a tiny bit of grit, gets embedded in the Al when pressing or folding or hammering -- instant stress riser. Use a scribe, same again. (In the link above they say even using a normal pencil causes problems because the carbon combines with and weakens the Al. I hadn't heard that one before.)
Stop before completing a weld, and not dress it back to remove any hidden voids and oxide pockets, and then puddle over to continue; stress riser.
Do an iffy weld first time and then puddle weld over and even if you start with O temper Al, you've now double heated the surrounding metal and no amount of heat treatment with restore it.
So you have the triple wammy of: a) it requires different design criteria that are not well disseminated; b) fabrication requires a new set of rules regarding cleanliness and accuracy -- no re-working; c) no matter how low the stresses, or beefy the part, even very small flexure causes changes to the temper; enbrittlement is inevitable; how long only depends on how many cycles, and their frequency.
The downfall of the Comet, besides square corners, was high frequency vibration. Tiny movements and forces, but at 20,000 rpm or whatever speed the jet engines ran at back then, for hours on end. Add in the heat/cold cycle and bye bye.
Having seen the latest set of close ups of the fracture, I'm even more convinced that my original assessment is correct. There is no sign of impact damage, and the position of the start of the crack and the way it has propagated, point to the classic void or oxide inclusion; probably with over-welding.
There's no way to prove that without hacking out a cross-section and doing a micro-graph or ultrasound testing; but I'd put a pint on it, if anyone in the right place and has an ultrasound tester, they'd see the void/inclusion.
liveforphysics
100 TW
I agree with all of that and it matches my own experiences in working with aluminum.
When Aluminum framed bikes were a few grand to buy, they might have budgeted in some/all of the process control and QC needed to weld them right. When aluminum framed bikes started retailing at walmarts for $249, I bet the folks making them only got good at making a bunch for as cheap as possible, and any added processes or cleaning steps don't make it cheaper to build.
When Aluminum framed bikes were a few grand to buy, they might have budgeted in some/all of the process control and QC needed to weld them right. When aluminum framed bikes started retailing at walmarts for $249, I bet the folks making them only got good at making a bunch for as cheap as possible, and any added processes or cleaning steps don't make it cheaper to build.
speedmd
10 MW
Now that we are deep into the armchair failure analysis :lol:, I am not too sure it is a weld void or over weld type defect. To me it looks to be a classic hi weld bead angle causing stress concentration at welds edge in the tube. A bit of finishing with a hand file or grinder as done on the old schwinn may have helped quite a bit here IMO. I agree that aluminum is a whole other animal when coming to the rework side of things.
Any of the steel bike tubes above the 531 level are terrible in welded joints. Even silver solder is borderline in developing large grain growth if you keep it hot a bit too long. If your gas welding, as done in many ultralight air frames, you want to start with normalized temper steel. In general, the higher the temper, the more unsuitable a metal is to welding, regardless of the material.
Back to the OP. Is the head set tight? I forgot to ask.
Any of the steel bike tubes above the 531 level are terrible in welded joints. Even silver solder is borderline in developing large grain growth if you keep it hot a bit too long. If your gas welding, as done in many ultralight air frames, you want to start with normalized temper steel. In general, the higher the temper, the more unsuitable a metal is to welding, regardless of the material.
Back to the OP. Is the head set tight? I forgot to ask.
Chalo
100 TW
liveforphysics said:
Right. Potholes and even little bumps can have much higher instantaneous peaks (even if the durations are short), and hard braking winds up a frame much more than hard pedaling. Make the chainring 22t instead of 50t, and the situation between BB and rear axle changes pretty substantially when pedaling, too. Lots of us here weigh significantly more than 1000N. That means bigger forces than in the von Mises diagram every time we stand to pedal, even if we're not elite. All that is the ugly end of the bell curve of forces that I was talking about.
I remember one time when I was working on bearing mountings for a flywheel motor-generator unit with a 700lb rotor and an evacuated enclosure. The engineer and tribologist who was in charge of that assembly was talking about a vibration on the order of 10G magnitude that we were struggling with at the time. I asked him, when did we get that kind of vibration? He pointed out that the running machine was doing it at that moment. I put my hand on it and only felt a slight hum. The relative suppleness of a bicycle prevents it from hiding large forces as effectively as that, but the same principle-- shock and vibration can easily dwarf other mechanical forces-- does apply to bicycles. Basically, the vibration you feel when riding that comes across as annoying or painful is telling you about force peaks that are much higher than static forces.
Chalo
100 TW
Buk___ said:
That's close to the yield stress of normalized 4130, so the fatigue limit would be not much more than half that.
Chalo said:
Still missing the point. You don't generally ride from one pothole to the next. Most of us try to avoid them.
A commuter: 1 hr a day at an average cadence of say 45rpm.
45*60*5*48*2 == 1.3 million flex cycles/year. This is what kills Al frames!
Maybe he cannot avoid 1 pothole a week. 48/year. So long as they don't exceed the yield strength, no harm done.
liveforphysics
100 TW
Buk___ said:
Everyone has a different kind of riding preference.
Even on my road bike, I look for the biggest sets of stairs to drop or climb, and ride over log piles and aggressive downhill trails, and it still feels like inadquate training for cyclocross racing.
For that reason my roadbike is cromoly, and the frame has not yet broken (though I go through wheels faster than most folks go through tires).
liveforphysics
100 TW
Buk___ said:
It's a 2002 Lemond Zurich using Reynolds 853 (the predecessor of the amazing 953) running a mix of dura-ace components with a cable pull ratio modified modern XTR rear derailleur (for it's auto-chain tensioner function that is awesome for avoiding chain drama on big impacts).
The tubes are butted to leave just the ends where the weld joints happen to be thicker material, and I bet the mid sections of the tubes gives egg-shell a run for it's money at some point, as it's lighter than my friends carbon road bikes.
liveforphysics
100 TW
My bike is 18.5lbs fully assembled with cyclocross tires. I don't know the bare frame weight.
Similar threads
