Page 1 of 2 12 LastLast
Results 1 to 50 of 63
  1. #1
    mtbr member
    Reputation:
    Join Date
    Jan 2011
    Posts
    860

    Nothing Lasts Forever

    I posted this on my blog but also want to share it here.



    Yes, my frame broke. This is the first time I have had a frame that I built break. It is sad but did not fail catastrophically plus I was not hurt so I don't really see it as a bad thing. It's life and we grow from our experiences.

    What do you guys think? Are there things that I should look for in terms of understanding and learning from this?

    Hopefully we can have a good discussion.

  2. #2
    mtbr member
    Reputation: adarn's Avatar
    Join Date
    Aug 2009
    Posts
    375
    Yikes! glad you are OK.

    I'm not really a TIG guy, but it seems like probably Inconsistent welding? Maybe where the crack started was over heated and more brittle?
    -Adam Sklar
    My blog
    Instagram

  3. #3
    mtbr member
    Reputation:
    Join Date
    Jan 2011
    Posts
    860

    What to look for?

    It was one of my better welds but I won't claim it is perfect.

    One thing I am looking for is a discussion of what to look for. I am reviewing the material that I can find such as the following:
    Welding defect - Wikipedia, the free encyclopedia

    When pushed to the limit anything will fail. The question is how should a frame fail and what should I be looking for.

    In my case, the crack went across the tube rather than following the line of the weld.


    The following example from google shows a case where a crack followed the edge of the weld.


    The key question is what should I be looking for and what would the signs be of a weld defect vs. simply too much stress?
    Last edited by febikes; 02-05-2014 at 04:43 AM.

  4. #4
    mtbr member
    Reputation: adarn's Avatar
    Join Date
    Aug 2009
    Posts
    375
    Well the fact that your weld didn't break would imply that it was not an imperfection in the weld itself.

    I'm in engineering school right now and I actually had a lecture on fatigue and stress concentrations today, funny enough.

    The concentration could be caused by a manufacturing flaw, impurity in the steel, a dent, a bad weld (cold or undercut), a hardened section of material (HAZ), maybe a stray file mark... a lot of different things. A frame experiences cyclical loading in its life and a ton of it too. When metal is stressed it cold works - so it gets "stronger" but becomes more brittle. when an area is more brittle than the surrounding material, it works beyond the yield strength faster and in a positive feedback loop sort of way it quickly leads to failure of that area and causes things like cracks and failure.

    If it were my bike, I would probably do a TON of practice joints think a lot about heat control, and cut each one open to look at the penetration.

    That picture is of a Ti frame. I know when Ti was first coming on the scene a lot of frames failed like that because the welds were contaminated.
    -Adam Sklar
    My blog
    Instagram

  5. #5
    Moderator Moderator
    Reputation: Walt's Avatar
    Join Date
    Jan 2004
    Posts
    5,792
    Did you stop/start at the centerline of the frame (ie at the middle/bottom of the downtube)? That will often create a stress riser that will crack the tube. Don't ask how I learned this...

    -Walt
    Waltworks Custom Bicycles
    Park City, UT USA
    www.waltworks.com
    waltworks.com/blog/
    instagram.com/waltworks/

  6. #6
    mtbr member
    Reputation:
    Join Date
    Jan 2011
    Posts
    860

    Not sure if starts/stops are the root cause

    Thanks Walt,

    In terms of welding I have gotten better on every frame that I built and continue to do practice joints. I am still improving but based on my very limited knowledge I am not sure that I see any weld fault in this case. In particular it does make me feel good that the crack did not follow the line of the weld.

    I used back purge, NOVA 969 tubing, 045 wire, 70 amps, and the pulse feature of my Miller 150 STH. The crack did form near a stop/start area but I am not sure that was the cause. Are there specific things I should be looking for in the broken weld?

    No part of the actual welds cracked. It is clear that the crack started just past the weld on the downtube. The crack did not follow the edge of the weld.

    As luck will have it I was having problems with my digital camera when I built this frame so I don't have good records of the welding.

    I can't stress enough how much this frame has seen in terms of riding, past crashes, and general abuse. Just before this happened I had gone into a ditch at speed. In addition, after the crack formed I heard it creaking and kept riding because it was cold and dark. I knew something was wrong but ignored it.

    It is possible that this is a case of "Underbead_cracking". If so, what are things I should be looking for to be sure?
    Welding defect - Wikipedia, the free encyclopedia

    Would adding a gusset be a good idea?

    It is also possible that the frame was correctly built, if this is the case, what should I look for to be sure?

  7. #7
    Nemophilist
    Reputation: TrailMaker's Avatar
    Join Date
    Sep 2007
    Posts
    1,873
    Quote Originally Posted by febikes View Post
    In particular it does make me feel good that the crack did not follow the line of the weld.
    NO!

    I am no engineer or metalurgist at all, but I have a bit of experience with metal of all sorts and in many applications. I would walk right back from this statement of yours, and actually flip it right on its head. Look at it 180* from there. 1) It is clear in this case that the weld was a "decent one," in so far as it was not the point of failure. Unless they are pretty atrocious, welds themselves are rarely the problem in general steel welding. 2) It is clear to me that the initiation point was right at the base of the HT junction, in the root of the weld, in the DT itself. There may have been any of the attributes present that Adarn mentioned to begin this chain of events, and I do believe it was a chain, not one distinct issue. The fact that the crack did not follow the weld suggests that the tube was not crystallized or otherwise overheated to any great degree, as if the weld had drastically overheated the tube at the root throughout the HAZ, it would have broken there all the way around. The fact that the path of least resistance traveled up through the tube beyond the HAZ should make you feel good where your welding is concerned, in my opinion.

    While the HAZ was clearly the point where the failure began, I'm not at all sure that welding itself was the issue, based on this somewhat slim evidence here. I would surmise that based on the length of the HT, the number of miles, and perhaps hard ones at that, the point at where it began, and that this exact point represents the weakest piece of material in the area, that it was a simple fatigue failure from too many stress cycles. If any of the physical points that Adarn raised were contributors is hard to say, but they certainly were not alone, and in my opinion not the lead cause.

    It is always interesting in such cases to very closely scrutinize the edge of the tube for clues. Quite often you can clearly see the evidence of a slow and steady failure, where it had seemed to you that it failed suddenly. The point at where the crack began is often corroded slightly, rust in this case, having been open and exposed for some time prior to it being noticed. Even if the failure point is not corroded, it might often appear darker or dirtier than the rest. This area can often also appear to be polished variously smoother than the rest of the tube edge if it is a spot that sees bending or twisting loads. The creaking you heard being the sign of this friction. The rest of the edge would appear more jagged as if it had been ripped clean away, suddenly.

    This is arguably the point of greatest stress in any frame, and the number of solutions out there to combat this potential is legion. I might imagine that not only might your frames be sporting one of these reinforcements in the future, but that we will all be watching our head tubes more closely!
    Most people ply the Well Trodden Path. A few seek a different way, and leave a Trail behind.
    - John Hajny, a.k.a. TrailMaker

  8. #8
    WIGGLER
    Reputation: todwil's Avatar
    Join Date
    Feb 2007
    Posts
    896
    Though the picture is not a good one it looks to me that the crack didn't start at the weld but some where else you can tell by rusted portion being the oldest and the area close to the weld is shiny being the last piece to fail. I could be wrong we need a better pic
    PAYASO 36er.....Live the Circus

    MR. 36er TROLL


  9. #9
    mtbr member
    Reputation:
    Join Date
    Dec 2009
    Posts
    21
    Glad your OK and thanks for shairing!

    It looks like a toe crack to me which is a welding defect. The bottom of the down tube can give bad torch angle which can do a few things.. undercut the tube, bad penetration or by sticking the electrode out to far not shield the weld properly. I like to slow down in this area and come at it more from the side. I also like to add more filler in this area just as long the toes fuse in I think it helps blend the joint better.

  10. #10
    Most Delicious
    Reputation: dr.welby's Avatar
    Join Date
    Jan 2004
    Posts
    1,102
    Fatigue crack initiated at the edge of the weld. Could be a bad joint, or just the 'weakest link'. Once you have the crack it will propagate where the stress is highest, which means with this joint geometry it will travel away from the weld and across the tube at some point.

    This is a common failure, especially on a 29er with a shortish head tube and big miles.

  11. #11
    mtbr member
    Reputation:
    Join Date
    Jan 2011
    Posts
    860
    Quote Originally Posted by DEFCON4130 View Post
    It looks like a toe crack to me which is a welding defect. The bottom of the down tube can give bad torch angle which can do a few things.. undercut the tube, bad penetration or by sticking the electrode out to far not shield the weld properly. I like to slow down in this area and come at it more from the side. I also like to add more filler in this area just as long the toes fuse in I think it helps blend the joint better.
    After reading about a toe crack, I am inclined to agree with this although I am still not 100% sure. I will be doing a lot of practice welds this year.
    Welding defect - Wikipedia, the free encyclopedia

    Quote Originally Posted by TrailMaker View Post
    The fact that the crack did not follow the weld suggests that the tube was not crystallized or otherwise overheated to any great degree, as if the weld had drastically overheated the tube at the root throughout the HAZ, it would have broken there all the way around. The fact that the path of least resistance traveled up through the tube beyond the HAZ should make you feel good where your welding is concerned, in my opinion.
    Thanks, although I am still pretty nervous about it. I guess my concern is not that it failed but rather I want to be sure it was a "good death" where the failure does not indicate any HAZ issues.

    Is the theoretically perfect failure occur at the butt transition zone or is failure at the toe of the weld considered normal?

    Quote Originally Posted by TrailMaker View Post
    This is arguably the point of greatest stress in any frame, and the number of solutions out there to combat this potential is legion. I might imagine that not only might your frames be sporting one of these reinforcements in the future, but that we will all be watching our head tubes more closely!
    Thanks, I am thinking about moving up to 171 down tubes and adding a small gusset on future builds in addition to doing a lot of practice for smoother welds.

  12. #12
    Picture Unrelated
    Reputation: zebrahum's Avatar
    Join Date
    Jun 2005
    Posts
    5,118
    It looks to me, and I will admit that the pictures do not show the failure line well, that the metal in the weld is the cleanest which would tell me that it was probably the last to break in the line of failure while the opposite end of the tube is quite dark which could mean it's been open for a while or it could mean that you hit the dirt when things went south.

    You said you stuffed the thing into a ditch before this failure, that's your problem among others. Your frame was going to fail because you caused damage. Even with a standard cyclical failure there isn't much you can do apart from normal caution and the use of best practices during welding. Metal which is cyclically loaded will fail, full stop. Metal which is loaded beyond the yield strength has failed and altered the strength of the material. If you put those two things together you get a failure which is what happened here.

    It is less important that you identify the exact point of failure and more important that you take this as a lesson in dialing in your technique. I don't think that anyone can point to any specific place where you went wrong, especially without a lab doing failure analysis on that tube, so I don't think there is anything you need to change before you build again in particular. All the comments above are best practice sorts of comments about controlling your HAZ, preventing undercutting, getting good penetration, and so on; these are fantastic points. Your problem isn't that you necessarily messed something up, your problem is that you broke a frame which saw its fair share of use. This exposed some questions which needed to be asked about the techniques you employ whether or not they apply to you or what you did.

    Sometimes we need to experience a failure to remind us of all the stuff we could have been doing better even if what we've been doing has been functional. We can all get better, right?
    Don't you hate it when a sentence doesn't end the way you think it octopus?

  13. #13
    Most Delicious
    Reputation: dr.welby's Avatar
    Join Date
    Jan 2004
    Posts
    1,102
    Quote Originally Posted by zebrahum View Post
    It looks to me, and I will admit that the pictures do not show the failure line well, that the metal in the weld is the cleanest which would tell me that it was probably the last to break in the line of failure
    In this case it's the opposite - the joint cycles between tension and compression and the oldest part of the crack will polish.

    I've literally seen pallets of frames with this exact same failure in varying levels of destruction. The solution was to make sure stop/starts were never near the bottom of the tube, and switching to a heavier-duty down tube.

  14. #14
    mtbr member
    Reputation:
    Join Date
    Jan 2011
    Posts
    860
    Quote Originally Posted by zebrahum View Post
    It is less important that you identify the exact point of failure and more important that you take this as a lesson in dialing in your technique. I don't think that anyone can point to any specific place where you went wrong, especially without a lab doing failure analysis on that tube, so I don't think there is anything you need to change before you build again in particular.
    Thanks! it's great to have this sort of support on the forum. I am going to really work this year to get every aspect of my welding and process 100% and am not really worried about the next build.

    Right now, I am thinking about the various gusset designs and will be planning on doing one on the next bike. It seems most if not all of the big brands like Niner, Surly, Voodoo, Salsa, etc use gussets so I will be designing something similar for my next frame.

  15. #15
    mtbr member
    Reputation:
    Join Date
    Jan 2011
    Posts
    860
    Quote Originally Posted by dr.welby View Post
    In this case it's the opposite - the joint cycles between tension and compression and the oldest part of the crack will polish.

    I've literally seen pallets of frames with this exact same failure in varying levels of destruction. The solution was to make sure stop/starts were never near the bottom of the tube, and switching to a heavier-duty down tube.
    I am sure the crack did not exist prior to the start of the ride. The bike was in good shape and while it looks dirty please remember that it was raining that night and the day before. I am sure that it was not cracked on Sunday and after going in the ditch I immediately heard the creaking sound. Between the ditch and the point where the tube became a full break was around five or ten minutes. After the downtube broke I was able to ride a little bit as the top tube was bending/tearing. The reason the entire thing came apart was that it was cold so I tried to ride even when I knew it was tearing then tried to push then carry the bike until the top tube snapped fully.

  16. #16
    Picture Unrelated
    Reputation: zebrahum's Avatar
    Join Date
    Jun 2005
    Posts
    5,118
    Quote Originally Posted by dr.welby View Post
    In this case it's the opposite - the joint cycles between tension and compression and the oldest part of the crack will polish.

    I've literally seen pallets of frames with this exact same failure in varying levels of destruction. The solution was to make sure stop/starts were never near the bottom of the tube, and switching to a heavier-duty down tube.
    Certainly an interesting viewpoint, not sure I'm 100% in agreement though. That's ok, we're trying to diagnose metal failure from not great pictures in an unknown environment of both what happened at failure and what has happened to the frame since failure. Freshly broken metal can be pretty reactive so who knows what the corrosion pattern would look like in this exact case and who knows if the picture is actually showing what I said it was showing? I'm no tube failure expert, that much I can assure you.

    For the start/stop point, would you recommend that the start/stop be moved to the top if you were welding a top tube instead of a downtube? Or a seat cluster? Is the idea to prevent the start/stop point from being in a location loaded in compression?
    Don't you hate it when a sentence doesn't end the way you think it octopus?

  17. #17
    Most Delicious
    Reputation: dr.welby's Avatar
    Join Date
    Jan 2004
    Posts
    1,102
    Quote Originally Posted by zebrahum View Post
    Certainly an interesting viewpoint...
    Let me expand my viewpoint then - I have never, ever, seen any tube crack on the inside-facing portion of a tube without some point of initiation like a braze-on.

    You're essentially arguing that the portion of the failure that is at an obvious stress riser failed only after a failure in a portion of unadulterated tube. That seems unlikely, barring better photos.

  18. #18
    mtbr member
    Reputation:
    Join Date
    Jan 2011
    Posts
    860
    Quote Originally Posted by dr.welby View Post
    You're essentially arguing that the portion of the failure that is at an obvious stress riser failed only after a failure in a portion of unadulterated tube. That seems unlikely, barring better photos.
    I will try to get some better photos.

    From looking at the tube in person, what I think happened was that the crack started at the toe of the weld on the bottom of the down tube. When I looked at the crack it was more jagged at the toe of the weld where the bead creates a stress riser. The crack was smoother as it "travelled" around the tube.

    Starting and stopping beads on the sides makes sense and will be part of my process refinements.

    I am pretty sure that the crack progressed around the tube while I was riding. I heard creaking that in hindsight was progressive. It was raining so everything was wet and dirty.

  19. #19
    WIGGLER
    Reputation: todwil's Avatar
    Join Date
    Feb 2007
    Posts
    896
    Quote Originally Posted by dr.welby View Post
    Let me expand my viewpoint then - I have never, ever, seen any tube crack on the inside-facing portion of a tube without some point of initiation like a braze-on.

    You're essentially arguing that the portion of the failure that is at an obvious stress riser failed only after a failure in a portion of unadulterated tube. That seems unlikely, barring better photos.
    I have! It was on a famous ti maker rode frame and the crack started
    About 1 1/2" north of the seat tube at the 9 o'clock position and traveled south straight thru the seat tube and seat stay and welds almost a perfectly straight line and ended in the seat stay no where near the brake bridge or cable holders just in the middle of the tubes both ends.

    I Was thinking it started from the internal butting but I would think The butting was further up the tube not 1 1/2-2" from ST
    PAYASO 36er.....Live the Circus

    MR. 36er TROLL


  20. #20
    mtbr member
    Reputation: Feldybikes's Avatar
    Join Date
    Feb 2004
    Posts
    695
    Dumb question: did you double check the butts? Could the crack be in the thin section? You can probably take some calipers to it now.

    I had a braced frame break in a similar place, beyond the fillet. Also not catastrophic. I, rightly or wrongly, have attributed it to cooking the tube too much (as well as being an overal badass). That was a 1.5" 9/6/9 TT Verus HT. The next bike has a 1/7/1 supertherm, and hasn't broken so far, so we'll see how it goes. Since the previous frame lasted for 4 years, I'm not holding my breath.

  21. #21
    Nemophilist
    Reputation: TrailMaker's Avatar
    Join Date
    Sep 2007
    Posts
    1,873
    Hey Mark;

    It is my opinion that this was a slow failure, and that a crack existed there for some time. You would have had to specifically look for it to see it. It could have been nothing more than a hairline crack in the finish, and virtually undetectable. I've had cracks like that in frames that I knew were prone to failing in a certain spot. I had to use a magnifying glass and a bright light to see them, but there they were, looking like a simple scratch in the paint. Being in the root of a weld told me other wise, and I watched them open and spread over time until I no longer felt comfortable riding the thing. Next bike!

    It seems very unlikely to me that stuffing it into a ditch would cause a rapid failure like this. Although this may have hastened the failure, I believe it was only contributory, not causal. I've had any number of crap generic cheapo bikes as a kid that had been ridden into curbs, walls, parked cars, etc.,, forks bent back, both in the legs, crown, and steerer, DTs kinked at the HT joint, you name it. All good studies of the concept of rake and trail! None of them ever failed, and I usually rode the crap out of anything I had, as likely to be in the air as on the ground.

    Perhaps this is why we see so many bikes with a bend in the DT just aft the HT. It makes for an easier weld joint. I guess everyone has decided that it also looks good too!
    Last edited by TrailMaker; 02-06-2014 at 05:33 AM.
    Most people ply the Well Trodden Path. A few seek a different way, and leave a Trail behind.
    - John Hajny, a.k.a. TrailMaker

  22. #22
    Most Delicious
    Reputation: dr.welby's Avatar
    Join Date
    Jan 2004
    Posts
    1,102
    Quote Originally Posted by TrailMaker View Post
    Perhaps this is why we see so many bikes with a bend in the DT just aft the HT. It makes for an easier weld joint.
    That's usually for clearance so the adjustment knobs on the tops of suspension forks don't smack into the down tube when the wheel rotates more than 90 degrees.

  23. #23
    mtbr member
    Reputation:
    Join Date
    Jan 2011
    Posts
    860
    I took more photos...
    Failure Analysis | Farnsworth Elemental

    My camera is not really good at close up shots but looking at the way the tube failed it is clear that the weld did not actually fail. The crack started at the crotch then moved around the tube.

    Big thanks to everyone who contributed to the discussion. I feel a lot better about the failure and think my craft will ultimately improve as a result.

  24. #24
    Most Delicious
    Reputation: dr.welby's Avatar
    Join Date
    Jan 2004
    Posts
    1,102
    Yeah, you can see the polished/peened spots around where the crack left the edge of the weld. Then traveling up the tube up towards the 3 and 9 o'clock points you can see the characteristic scalloping (though on a thin tube it looks more sawtooth-y) of the crack propagating upwards. As the crack moves up the cross-section of the remaining tube starts to drop quickly so the crack moves faster, hence the change in texture from polished to scalloped. Finally at some point the pattern should just turn uniformly rough where the tube gave up.

  25. #25
    Harrumph
    Reputation: G-reg's Avatar
    Join Date
    Jan 2004
    Posts
    1,600
    It's part of learning the craft, one of those not 'if' but 'when' things. It's what you do with that experience that matters for subsequent frames. Think of all you've learned since that frame, well before this analysis/post-mortem.

    What Walt said about where you start and stop the weld is a new consideration to my ears.
    Slowly slipping to retrogrouchyness

  26. #26
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    293
    Wonder if the tacks make any difference? I've been putting my tacks for the dt 2 oclock, 4 7 and 10. I start my dt weld on one of the tacks but really not paying too much attention to the stop start points. I do as many as I have to so I don't get the torch out of position too much. I do the starts on the last bit of welding slowly bringing up the heat to get the puddle as wide as the last weld, then move on, so as to make the start-stop as incongruous as possible. I would have guessed we would all start the dt weld in the middle like at 6 oclock to minimize the amount the tube is pulled out of alignment, but this is a good discussion. Thanks for starting it.
    cheers
    andy walker
    Walker Bicycle Company | | Walker Bicycle Company

  27. #27
    Nemophilist
    Reputation: TrailMaker's Avatar
    Join Date
    Sep 2007
    Posts
    1,873
    Hey;

    I do not weld in the most acute angle area of the joint first, as this will cause the most draw. Other than that, my welding sucks anyways, so I'm not sure it matters that much.

    Indeed, the highly random and jagged pattern of the crack at the base of the HT shows it slowly meandering along as it propagates. Drip, drip, drip from a frozen faucet. The last 80% is the sudden rush as the ice plug shoots out and the flow is free!
    Most people ply the Well Trodden Path. A few seek a different way, and leave a Trail behind.
    - John Hajny, a.k.a. TrailMaker

  28. #28
    mtbr member
    Reputation:
    Join Date
    Jan 2011
    Posts
    860
    For this joint, I think the tacking was at clock points 11, 1, 5, 7.

    When welding, I back purged with argon flowing using a tube inserted through the down tube from the BB side. For this frame, I did this joint prior to any other joint. The spans from tack to tack were completed with the top (11..1) followed by right side (1..5) then left side (11..7). The final bit in the tight crotch of 7..5 was done in two approaches with (7..6) followed by (5..6).

    No part of the actual weld broke and the while the crack did form at the "toe" of the weld it did not follow the weld so to me it seems the welding and the area right near the weld was stronger than the tube it's self. The bike saw an a lot of usage with many crashes and harsh trails. My feeling is that this was a fatigue crack more then anything else.

    The tube broke and not the weld so I am thinking that a stronger tube and/or a gusset would have helped out.

    For next time around, I plan to increase from 36mm 969 down tube to a 38mm tube and may use 171 supertherm. I am looking at gusset designs and will likely add one. I am also thinking about a bent down tube because they do help with the tight crotch and I think it would help create a stronger joint.

    I am also planning a lot of time to work my welding technique; I don't think my welding was "bad" but everyone can always improve.

  29. #29
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    293
    +1 supertherm. Your weld was fine, probably just the hard riding. I get the vibe that people are ambivalent about gussets. It just makes a thinner part of the tube stressed.
    Here's a snippet from a bmx builder Why we don?t use plate gussets | fbom bikes I've tried one gusset but probably won't do any more. Does anyone think Reynolds 853 actually gets stronger with welding? They say it does, and make long straight and bent ones, BX3038L, 30L, 50,52,54. The headtube butt thickness is 1.1 to 1.2mms also fyi.
    cheers
    andy walker

  30. #30
    Moderator Moderator
    Reputation: Walt's Avatar
    Join Date
    Jan 2004
    Posts
    5,792
    Your weld sequence in the crotch of the joint did it IMO. You put it through 2 heat cycles and there are 3 stop/start points under there. No matter how careful you are with starts and stops it's pretty much impossible to get the exact same penetration, shape, etc - so stress risers will be inevitable. You have 3 of them where the tube broke.

    In my book if the break starts at or on the weld, it's builder error (and yes, I've made plenty of errors like that of my own over the years.) If the tube buckles away from the joint, it can *still* be builder error (ie picked the wrong tube, put a big can-opener-pointed lug or gusset somewhere it shouldn't be, etc) but it is usually just "rode into a ditch" rider's fault.

    -Walt
    Waltworks Custom Bicycles
    Park City, UT USA
    www.waltworks.com
    waltworks.com/blog/
    instagram.com/waltworks/

  31. #31
    Moderator Moderator
    Reputation: Walt's Avatar
    Join Date
    Jan 2004
    Posts
    5,792
    Do not, do not start the welds on anything at the bike centerline (ie 6 or 12 o'clock). It will be fine for a year or two, and then you'll have the bike back with a fatigue failure from a stress riser you created. If you are building for yourself, whatever - nothing is going to fail suddenly on you. If you are building for others, you're going to have to fix a lot of frames down the road.

    -Walt
    Waltworks Custom Bicycles
    Park City, UT USA
    www.waltworks.com
    waltworks.com/blog/
    instagram.com/waltworks/

  32. #32
    Nemophilist
    Reputation: TrailMaker's Avatar
    Join Date
    Sep 2007
    Posts
    1,873
    Yeh;

    I like your idea on heat cycles, Walt. The hardest thing for a TIG welder to do is do the joint in one pass. I'm not sure how... well... I know how, but I just don't have those kind of chops, let alone the setup. Not even CLOSE. I'd say it is also very much more important on hardened tubing, as I imagine it has a much narrower window of stress it can handle before it crystallizes and becomes brittle. I'm guessing that generic old ChroMo is a bit more forgiving in that regard, and being a hack, that's another reason I use it.

    Weld sequencing is kind of a black art, and it can have many parameters influencing it. If you weld a bike wrong, you might have to tweak the crap out of it after you are done. Oh... sorry... Cold Set. The smarter you are at sequencing (assuming good miters) the less tweaking you need to do. With the stuff I weld the most (below), you weld NOTHING to the chassis (other than tacks) until the cage itself is completely welded to itself, and then within that you play the same game of deciding which way you want to draw which tube or assemblage of tubes. The more tubes or complex the assemblies, the harder it is to decide the sequence.

    It is REALLY critical. You get used to not jumping through the rafters when the base tacks break, usually in the middle of intense concentration, and it is LOUD. Do it all right, and everything is as it should be. If you get the sequence wrong, THE DOORS WON'T SHUT! You have no idea the difficulty involved in fixing that.... It can get ugly.
    Attached Thumbnails Attached Thumbnails Nothing Lasts Forever-alexdoorbars4.jpg  

    Last edited by TrailMaker; 02-06-2014 at 03:09 PM. Reason: Spellung
    Most people ply the Well Trodden Path. A few seek a different way, and leave a Trail behind.
    - John Hajny, a.k.a. TrailMaker

  33. #33
    mtbr member
    Reputation:
    Join Date
    Jan 2011
    Posts
    860

    When fatigue cracks occur in perfect welds what does it look like?

    Quote Originally Posted by TrailMaker View Post
    Weld sequencing is kind of a black art, and it can have many parameters influencing it.
    Agreed; I also think fatigue analysis in bicycles is a bit of a black art as well.

    The frame had a lot of hard miles (most of them off road and all of them rough). For example, when I commute, my route includes two flights of stairs. In addition, I am a regular user of gap jumps on our local single track plus I do all sorts of log crossings and often crash because my eyes are bigger than my bike handling skills.

    The tube is where the fatigue crack occurred. Although as we discussed the point where the crack started is on the edge of the weld bead. No part of the actual weld broke and the crack did not follow the weld bead line very far. It seems possible that the tube simply fatigued beyond what that tube it's self could handle.

    What I think happened was that the tube was too small for the forces involved. A 36mm diameter tube will flex more than a 38mm diameter tube. Each flex event is a fatigue cycle and when a fatigue crack appears it seems that it almost has to start near the weld because that is where the flex point is. If the tube was indeed too small for my riding style then even "perfect" weld technique would not have stopped the crack.

    As a question for Walt and others; if the weld was perfect and it was indeed a fatigue crack how would things "look" different from a bad stop/start point on the weld?

    The reason that I thing a gusset and larger diameter down tube will help is because it will reduce the fatigue flex cycles and allow me to continue to ride hard and aggressive with much more head room.

  34. #34
    Most Delicious
    Reputation: dr.welby's Avatar
    Join Date
    Jan 2004
    Posts
    1,102
    Quote Originally Posted by febikes View Post
    What I think happened was that the tube was too small for the forces involved. A 36mm diameter tube will flex more than a 38mm diameter tube. Each flex event is a fatigue cycle and when a fatigue crack appears it seems that it almost has to start near the weld because that is where the flex point is. It the tube was indeed too small for my riding style then even "perfect" weld technique would not have stopped the crack.

    .....

    The reason I thing a gusset and larger diameter down tube will help is because it will reduce the fatigue flex cycles and allow me to continue to ride hard and aggressive with much more head room.
    That's not quite how it works - 'flex' is merely displacement from a load. A smaller tube will see higher peak stresses for a given load since it has a smaller cross-sectional moment of inertia. This will also cause it to flex more, but the amount of displacement has nothing to with fatigue. You can't reduce cycles unless you ride your bike less.

    Also we can't get mixed up with flexing a paper clip back and forth since we're not causing large plastic deformations here.

    Things get additionally confusing because in dynamic situations a more flexible member can see lower peak stresses. Look at the wings of a 747 - they can flex up and down an incredible amount by design.

    A gusset is basically just adding butting in a high stress spot. If you weld around the end of the gusset then you've added another stress riser. Hopefully it's somewhere that has lower stresses...

    Quote Originally Posted by febikes View Post
    As a question for Walt and others; if the weld was perfect and it was indeed a fatigue crack how would things "look" different from a bad stop/start point on the weld?
    The edge of a weld will always be a stress riser. I think all you can do is try to figure out where the crack started and look for a something that magnified the stress riser - stop/start, undercut, etc.

  35. #35
    Moderator Moderator
    Reputation: Walt's Avatar
    Join Date
    Jan 2004
    Posts
    5,792
    I think maybe you misunderstood me - you *created* a stress riser (or several) with your weld sequence. That stress riser is what broke the tube and your frame. It is true in a technical sense that the "weld" did not break here but a better weld sequence probably would have prevented the problem. The joint was the problem, not the tube.

    The stress risers are bad from a fatigue standpoint, and also from a peak load standpoint - in both cases they shorten the life of the tube at that spot.

    As others have noted, the weld itself is not generally going to break unless you use way too little filler or do something really weird.

    The failure mode you should see here (in the "ran into a ditch/brick wall/shorted a double) is a buckled tube a few inches from the joint, usually at the butt transition. If the joint is the starting point for the failure you screwed up, regardless of whether it failed in a crash or just due to fatigue over time (or both).

    The bottom line is that you should do your utmost to avoid making stress risers at highly loaded areas. How you want to go about doing that is up to you, of course, but the low hanging fruit is not starting and stopping at the centerline.

    -Walt
    Waltworks Custom Bicycles
    Park City, UT USA
    www.waltworks.com
    waltworks.com/blog/
    instagram.com/waltworks/

  36. #36
    mtbr member
    Reputation: Eric Malcolm's Avatar
    Join Date
    Dec 2011
    Posts
    870
    Mark, this has been very interesting so far.

    I will give another completely different angle to consider.

    As a fresh faced kid many years ago, watching a TV documentary of a steel manufacturing works making steel beams. The molten steel poured through a die and cooled down progressively along its length. The key element that has stuck with me was seeing the steel cool in various colours. Initially white, it turned red, then patches of black could be seen before the whole length became black. There was in action here variations in steel molecular 'Texture' (there is probably a correct term for this).

    Later in life I now take students through Powder Actuated tool use with the firing of Nails into steel. New Zealand is one of the few developed countries still using Direct acting 'guns' for this task. When we fire into steel along the length, the nails penetrate with a lot of inconsistency, some will be at the desired depth, others, randomly will have penetrated deeper and others barely got in at all. The first reactive thought is that the charges are inconsistent, but this is not so. Firing into concrete yields even penetrations out of the same packet. Those patches I remember in the cooling steel have been explained to me as density patches. That is, there are fine dense grains of steel and coarse soft grains. When on a construction site, these variations in the steel cannot be identified by the naked eye. So, this is a risky process. We also train people to be aware weld area's as these are too hard to fire into.

    We also sell Carbide drill bits for Rotary drilling into concrete. A drill bit of this nature is a forged item, and a very small percentage of these break because of the 'hard metal' stress riser caused at the junction of dense/coarse grains in the steel. I have seen many bits over the years that have a smooth polished look on part of the break and a tear throughout the rest. The tear is the softer coarse metal. Rotary Hammer drills have an impact rate of over 4000blows per minute for a machine suited to say, drilling a 12mm hole. The blow force is likely to be 3+joules. Plenty of cyclic loading to expose a steel fault.

    Steel tubing used for bikes are drawn cold. They are no less susceptible to the grain fault than any other steel. 4130 is of better quality, specialised tubes that have vanadium, niobium added have the ability to strengthen the grain structure when heat is added (welding) so as not to loose structural integrity. Reynolds 853, TT S3, Columbus Life/MAX/Spirit have variations of additives in their steel, as well as being heat treated.

    I have seen frames that have inexplicable breakages away from welds, and conclude that at times, steel can throw out a random mystery.

    Your welds look good, no breakage within the weld 'rainbow'. Plenty of advise to review your methods, but in this instance, you may have been caught by the random hard metal break exposed by your hard riding style.

    Up-scaling to a larger diameter down tube is a good idea, but with double-butted tubing, using a gusset and putting the stress into a thinner wall section of the tube might require some revision of thought.

    Don't beat yourself up for the breakage, the inquiring nature of your question is healthy.
    Go build another frame, better for the soul.

    Eric
    If I don't make an attempt, how will I know if it will work?

  37. #37
    mtbr member
    Reputation:
    Join Date
    Jan 2011
    Posts
    860
    Thanks Eric and Walt and the good Doctor.

    All suggestions will be used to improve my process.

  38. #38
    Nemophilist
    Reputation: TrailMaker's Avatar
    Join Date
    Sep 2007
    Posts
    1,873
    Quote Originally Posted by dr.welby View Post
    Things get additionally confusing because in dynamic situations a more flexible member can see lower peak stresses. Look at the wings of a 747 - they can flex up and down an incredible amount by design.
    Abso-freakin-lutely;

    Defaulting to increasing the tube diameter can make the situation worse if the circumstances are right, sending greater stresses into spots that can't handle them. I build and sell a bolt in roll bar for a certain model Porsche. Every now and then, I get a guy who wants to know if he can substitute grade 10.9 fasteners for the 8s that I supply with the bar. I say "sure"... but I wouldn't. I've gotten into endless arguments with guys that just don't understand that everything must be thought of as a system. Increase the fastener strength and you increase the load on that which it threads into. Can the receiver of said bolt handle the increased load, or would it be better to have a fastener that works in better harmony with its partners?

    I think the jury will remain out on this. Some improvement in welding might help, for sure, but it seems to me that you'll have to be satisfied that you jus dun broked it, brother! Eyes bigger than your talent? You'll age out of that one... eventually!!!

    I applaud you for dragging this out here. Good show, Old Bean!
    Most people ply the Well Trodden Path. A few seek a different way, and leave a Trail behind.
    - John Hajny, a.k.a. TrailMaker

  39. #39
    mtbr member
    Reputation:
    Join Date
    Dec 2013
    Posts
    29
    Thank you all for enriching my evening.

  40. #40
    Moderator Moderator
    Reputation: Walt's Avatar
    Join Date
    Jan 2004
    Posts
    5,792
    Let me be clear - I think it is *awesome* that you posted this and I have made the same mistake myself (along with LOTS of others!) Did not mean to be anything but constructive but rereading my post I think I came across as being pretty harsh. This is the kind of thread where everyone learns something, I think, so kudos for putting it out there!

    I'm actually fixing (replacing) the toptube on an older frame of mine next week where I did exactly the same thing.

    -Walt
    Waltworks Custom Bicycles
    Park City, UT USA
    www.waltworks.com
    waltworks.com/blog/
    instagram.com/waltworks/

  41. #41
    mtbr member
    Reputation:
    Join Date
    Jan 2011
    Posts
    860
    Quote Originally Posted by Walt View Post
    Did not mean to be anything but constructive but rereading my post I think I came across as being pretty harsh. This is the kind of thread where everyone learns something, I think, so kudos for putting it out there!
    It's all good, I enjoyed and will follow your advice.

  42. #42
    mtbr member
    Reputation:
    Join Date
    Jan 2011
    Posts
    860

    This turned up in some research that I did online

    I found this great material shared by Worcester Polytechnic Institute.
    http://www.wpi.edu/Pubs/E-project/Av...cycle_Fork.pdf

    There is a lot of great information in the article. It is specific to forks but the general themes can be applied to frames as well.

  43. #43
    mtbr member
    Reputation: Testmule's Avatar
    Join Date
    Jul 2013
    Posts
    126
    Quote Originally Posted by febikes View Post
    I found this great material shared by Worcester Polytechnic Institute.
    http://www.wpi.edu/Pubs/E-project/Av...cycle_Fork.pdf

    There is a lot of great information in the article. It is specific to forks but the general themes can be applied to frames as well.
    Great read, Mark, and also a great thread for all. Striving to understand and make better is what we all want, I think.

    In my hobby frame building over the last twenty years I have been given lots of broken frames from my owner friends of LBSs. Breaks at the head tubes are common in all frames, even from what looked to me to be really well designed frames from prominent builders.

    Mountain biking is about a tough on materials as it gets, and there are no easy answers to making frames both tough and light.

  44. #44
    mtbr member
    Reputation: Meriwether's Avatar
    Join Date
    Jul 2007
    Posts
    420
    how many of us are now wondering when our frames are going to break at the 12 or 6'oclock spots on the tube?! damn...

    This has been a very illuminating post. Thanks Mark for sharing this frame failure, it helps us all in the long run and will make all our frames safer in the near future. And thanks to all you guys for the knowledge and opinions, this has been one of the best threads in awhile IMO.

    I think the hardest thing about welding *through* the acute angles in one pass is that it's, well...hard! I've also seen or read many top builders tack along the centerline of the frame to keep the frame in vertical alignment before taking it out of the fixture to weld. I've been told to tack in the acute angles first to draw the tube miter tighter around the other tube.
    Walt and other's words make me want to try again and again is tack around the 10/2, 4/8 positions and weld the vertical passes IN the fixture to try and retain that frame alignment.

    Very very few professional builders share their weld sequence. It's effectively a secret since it took so long for them to develop to get proper alignment. But the frame's structural integrity hasn't been discussed as much in terms of weld sequence (to my limited knowledge). It seems to be one of those things that they want you to learn the hard way. Luckily Mark wasn't going down those stairs when this occurred...

  45. #45
    Eric the Red
    Reputation: edoz's Avatar
    Join Date
    Jan 2004
    Posts
    438
    I can't add anything past what Walt did, but props to you for sharing this failure and using it as a learning experience for yourself and others. I salute you.

  46. #46
    RCP Fabrication
    Reputation:
    Join Date
    Jun 2011
    Posts
    389
    I tack at 12 and 6. When it comes time to weld, I weld 12 and 6 at least 1/2", but in reality as much as possible. IMO the LOCATION of the start/stop points is not nearly as important as the EXECUTION of the tie-in.

    I'm not sure what you meant by welding though acute angles in one pass. If you mean welding from one side to the other with out stopping, yes it takes some practice. If you mean welding it in two passes you are BEGGING for a failure.

  47. #47
    mtbr member
    Reputation:
    Join Date
    Jan 2011
    Posts
    860
    Quote Originally Posted by Meriwether View Post
    how many of us are now wondering when our frames are going to break at the 12 or 6'oclock spots on the tube?! damn...
    Frames break; Even when the welding is perfection they still fail in fatiuge cycle testing. Everyone is right on the money with comments here and we will all focus on our "tie-ins" and starts stop locations but even if we do things perfectly metal fatigue still happens at some point.

    It is really hard to stress how much and how hard my frame was ridden prior to the failure. If there is one thing other "hobby" builders should take away is that even sub-optimal welds are pretty strong. I road that bike many cycles and had many crashes plus on just about every ride I would hit stairs, gap jumps, rock piles, etc. The bike was also occasionally in serious off road conditions with up to 60 pounds of bike packing gear strapped to the frame as dead weight.

    Quote Originally Posted by Meriwether View Post
    Very very few professional builders share their weld sequence. It's effectively a secret since it took so long for them to develop to get proper alignment. But the frame's structural integrity hasn't been discussed as much in terms of weld sequence (to my limited knowledge). It seems to be one of those things that they want you to learn the hard way.
    As RCP pointed out really perfect tie ins are also key. The good thing about doing tacks outside of the stress plane is that we can avoid the start point and aim for the smoothest possible weld for high stress areas.

    Quote Originally Posted by Meriwether View Post
    Luckily Mark wasn't going down those stairs when this occurred...
    On thing that makes me feel good about steel is that even in this case the bike gave a bit of warning to me in terms of creaking and such prior to the failure. I never felt it was unsafe and as you can see in the following photo even after the down tube broke the top tube was doing it's job until it finally tore.



    In general, steel bikes are quite safe because they give a bit of warning before they fail.

  48. #48
    The cat's name is jake
    Reputation: BungedUP's Avatar
    Join Date
    Aug 2003
    Posts
    274
    I fabricate bikes in a production setting, and have done so since 2007. In the process of welding several thousand frames, I have gone from thinking that the exact sequence of welding a frame is important, to believing that welding sequence is actually not very important to frame alignment. Neither is tacking sequence. I've changed my personal opinion about this, and the more frames I've made, the more I have started seeing other factors appear to make the most difference. The variations I see in frame alignment come down more to:

    1) joint fit
    2) properly clamped and positioned fixtures and frame parts
    3) tubing straightness (depending on types of jigs - I use both plate fixtures with v-blocks, and fixtures with centering cones)

    After that, assuming that one doesn't weld the entirety of a BB area, or not take into consideration BB warping, or weld the entire right side of the frame, then the left, you can weld the whole thing up any old way, as far as I'm concerned. If you don't properly address items 1-3 listed above, you might need a welding strategy to address those issues. The sequence does have an effect, but it's minor compared to the items I listed above, in my experience, working with the tools that I do. I check and align if needed every frame that I weld.

    Regarding the frame breakage, I have a few thoughts which may or may not pertain to that specific bike.

    The first is, tubing can often contain defects which are clearly visible on the inside as pock marking. Some of the worst that I see is Fairing tubing, as well as high end Reynolds tubing (853). If you aren't looking for it, you might not see it, but it can come in batches where it's visible in every single tube in a box. The pock marks can be deep.

    Second, the underside of the downtube/HT joint is a highly stressed area. Many frames, even if welded perfectly, will fail over time. That's why the structure should be designed to resist that failure, whatever way that may be. Reynolds now produces a 853 "DZB" tube to prevent just this problem, which has a thicker butting in that area - it makes sense. You can read a story about a builder having trouble passing EN standards that used this tubing to solve that issue. The EN tests are extremely rigorous. Ultimately, I would guess this is where the real problem in the frame was - not construction, but tube selection in the first place. This is something that manufacturers can have trouble with as well. Many places do not do the necessary testing to really ferret out these types of problems. It's too bad, because you'd learn a huge amount, without risking injuries to people. I'd actually warrant that most small to medium sized professional bicycle builders do not do the type of testing comparable to EN. It's costly though, so it's no wonder that individuals don't do it. I think it's unfortunate that medium sized outfits don't do it however, and that knowledge gained from testing doesn't get made available to all. I understand why it doesn't, but as a science minded person, I'm sad that scientific discoveries are not made available to all.

    In the fore-aft cycling of the HT/DT region, I would consider building some flexibility into the headtube by either thinning it down, or raising the DT intersection, so that the reinforced headset area doesn't become a super stiff member, sitting right next to a fairly thin member. The brazed equivalent is having a lug be thick all the way to the tips, causing fractures where all of a sudden there is no more lug support. If the whole area can flex (safely) and distribute that movement, as opposed to one small point operating as a fulcrum, I wonder if that might help. I don't know how FeBikes constructed his bike, but it's a strategy I'd personally consider. In the end, that's just a nasty little area, which doesn't have too many easy solutions.

  49. #49
    Nemophilist
    Reputation: TrailMaker's Avatar
    Join Date
    Sep 2007
    Posts
    1,873
    Thanks Peter;

    Always most worthwhile. More gold for the mining, from "The Professor."
    Most people ply the Well Trodden Path. A few seek a different way, and leave a Trail behind.
    - John Hajny, a.k.a. TrailMaker

  50. #50
    The cat's name is jake
    Reputation: BungedUP's Avatar
    Join Date
    Aug 2003
    Posts
    274
    HA! Oh Jeez - if that moniker sticks, I'm in real trouble...

    I was thinking about the testing thing more, and really, how great would it be to have a database for framebuilders AND consumers where incidents of failures and the specific details were listed? Talk about a goldmine of information...

    Quote Originally Posted by TrailMaker View Post
    Thanks Peter;

    Always most worthwhile. More gold for the mining, from "The Professor."

Page 1 of 2 12 LastLast

Similar Threads

  1. Seems like forever since I've been out !
    By onbelaydave in forum Colorado - Front Range
    Replies: 6
    Last Post: 11-12-2013, 09:34 PM
  2. Cassette off, spins forever. Cassette on, not forever
    By westin in forum Wheels and Tires
    Replies: 2
    Last Post: 08-05-2012, 04:14 PM
  3. How long does jockey wheels lasts?
    By bratz in forum Drivetrain - shifters, derailleurs, cranks
    Replies: 4
    Last Post: 10-04-2011, 02:03 PM
  4. Full finger gloves-what lasts, what doesn't?
    By TwistedBars in forum Apparel and Protection
    Replies: 18
    Last Post: 07-14-2011, 11:24 AM
  5. Haven't been on in forever, but I got a new toy....
    By esundell90 in forum 29er Bikes
    Replies: 0
    Last Post: 01-17-2011, 09:35 AM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •