Monday, September 06, 2010

29 Safety Moment : Speed Wobble and Jaw Fracture

Will Cheng is an electrical engineering Ph.D. student at Stony Brook University here in NY. In their free time, he and his sister ride their bikes with a group of retiree friends in Long Island.

Early in August, Will met with a nasty bike accident during a pace-lined group ride that left him with a fractured jaw. He was rushed to the ER where doctors had to perform an 8 hour surgery on him to patch up the severe injury. Some who witnessed the accident had advised him to contact an attorney. Meanwhile, the medical and dental costs for the operation had been tallying up and taking uncomfortable proportions.

He contacted an attorney who recommended him to get in touch with a mechanical engineer who could look into the background of his cycling equipment.

At the time of the accident, Will happened to be riding a 2006 Orbea with Mavic Cosmos wheels fitted with Schwalbe Blizzard tires. I was informed by his sister that Mavic no longer makes the wheelset.  Will's own account of the accident later to her was as follows :

"I was on Clay Pitts road [in East Northport] and I moved up to second wheel in the pace line at the light on Elwood road. After pedaling for a while [at 20mph], I noticed that I was a little to the left of the shoulder. I corrected by moving the wheel a hair to the right so I was heading towards the white line slowly. I then turned the wheel back to the left to straighten out.

That's when I felt no resistance or feedback from the wheel and handle bars. I assumed that I went over a dip in the road and recall crying "WHOA" and thinking that the leader should have warned us. This was followed  by a gross turn of the wheel to the left. I panicked, simultaneously turning the wheel straight ahead and clipping out my left foot.

I felt the bike wobble a little, after which it diminished and stopped. At that point I thought I was OK but a split second later I felt something was wrong. Before I could do anything, I was falling. I do remember that when the wobble disappeared, I was staring at my handle bars and saw that it was straight without signs of the wobble. I thought I was safe and I looked back up at the road. I don't really remember much, but I think I still had my hands on the handle bar right when I hit the ground with my chin. I didn't have a death grip on the bars but my grip was firm and my hands were always on the top of the bars."

As far as I have looked into the wheelset through some internet searching, I haven't found any design related issues and its performance limiter really depends upon who built it. In more cases than not, a crash is what causes a wheel failure. At other times, it is fatigue failure or some very high external load not expected in normal usage of a bicycle that causes spokes to pull through. Wheel experts say something in excess of 2000 N of force is required to pull a spoke out of the rim.

From the attached pictures (see below, and more here), it seems that about 5-6 of the straight pull spokes in total had pulled out and that more spokes pulled out on nut side of the front wheel skewer than on the lever side. This corresponds with weakening and rupture of the wheel rim on the left side, when viewed from the front. Also take note that the rupture occurred right underneath a sticker on the rim so its hard to tell whether there was a hidden crack formation well before the accident.

From the description of the actions of the rider before the accident, I don't see anything particularly out of the ordinary. Steering motions such as this is absolutely normal and is to be expected. I perform more wilder maneuvers on my bike path in order to avoid sharp twigs and bumps.

What may be significant though, is the faint evidence of a speed wobble before the crash. Could a rapid left-right steering correction at 20 mph together with a sketchy road surface amplify an unwanted oscillation? Check out the image of the site of the accident.

Moreover, what gave away first - the spokes or the rim? Another bit of interesting testimony is the loss of "feedback" just before the crash, which almost wants me to question whether Will had remembered tightening his skewer that day.

Unfortunately, these questions are really hard to answer through images. I would check the tension on the spokes with a tensionometer, consult with a metallurgist who would be able to analyze the sample of broken aluminum rim (Stony Brook should have a professor who may help) and try as much as possible to take a similar wheel with the same tire, attach it to the same bike and perform some maneuvers at the speed in question.

Here's wishing Will the best of luck in recovery. Meanwhile, if any of you have had similar experiences, do share some of your thoughts.

*  *  *


  1. Is proper speed wobble on a road bike even possible at 30-35kph?

  2. Its a good question Jason. I would answer that by saying its the resonant frequency of the bike-rider combination that determines the precise timing of the wobble. But you don't hear of wobbles at slow speed that much, do you?

  3. Anonymous11:56 AM

    Whenever I encounter anything like a speed wobble, the first thing I do is check for loose bearings and such in the headset. The next thing I do is to take a different wheelset I have in my basement, that costs me next to nothing, and substitute that instead of the wheel-fork combination that led to the wobble. Sometimes, changing your moment of inertia of the front end in this way may change the frequency at which the wobble's modes are introduced.

  4. Marco3:07 PM

    I love your posts. I was wondering if the presumption that speed wobble caused the accident is accurate by your analysis? Or correct me if I'm wrong, but I read the OP's quote and found wobble could also be as a result of whatever led to a "loss in feedback". Besides, like Jason asked, how easy is it to induce speed wobble at 20mph?

  5. Anonymous4:04 PM

    I'm pretty sure the crack is where the rim is welded together, opposite the valve stem with a sticker covering it. It's difficult to conclude what caused Mr. Cheng to crash, but this serves as a reminder for all of us to frequently inspect critical components of our bike, especially right before a ride. Calfee Design has an interesting article about the cause of speed wobble, basically it is caused by fork asymmetry, which seems unlikely with modern production and quality control standards.

  6. There is a paper out there that has a model for speed wobble on a pedal bike - I can't find it right now again but from what I remember it seemed to indicate that significant speed was required for bikes with "big wheels" (ie 700c), high pressure tires and typical geometry. IIRC as well tire pressure was quite important.

    Wobble is not only an issue of resonance frequency it also requires some sort of in phase forcing to get it going and for the rider to not be "clamped in" with a proper riding position at speed.

    If you're talking about loose bearings/headsets it has nothing to do with speed wobble, that's just shitty worn out kit or a lousy setup.

    I'm more inclined to think something happened to the wheel and it lost it's dish for whatever reason and then things went even worse.

    I've never lost dish on a front wheel, but I have on a rear wheel (debris flew into the spokes and cleared out about 5 split between two sides and then locked up the rear wheel) and it feels very much like the bike no longer responds to steering and wobbles all over the place. Had I been going faster and that been the front I'm sure I would have gone down hard instead of swerving all over and then just toppling.

  7. @Anon

    "Calfee Design has an interesting article about the cause of speed wobble, basically it is caused by fork asymmetry, which seems unlikely with modern production and quality control standards."

    Sorry but that's been shown to be patently untrue by numerous sources in numerous locations.

    Colour me sceptical about their motivations too, but I bet that statement was followed by claims how amazingly accurately they align their outrageously priced custom frames.

  8. You can get wobble at lower speeds, but it's usually not a huge concern because you can actually control a wobble - the wheel isn't turning very fast so it takes less effort to hold onto the bars. I know this because I experience this when riding on the Cane Creek Speed Bars and I put a lot of weight on them. Typical speeds of 20 to 35 mph, on flat roads. If I ease up on them weight-wise the bike is fine, but I can feel the "potential wobble".

    Holding the regular parts of the bars (tops, drops) at high speeds (50+ mph) is fine. It's the very forward and low position of the Cane Creeks that seem to cause the wobble.

    I've used the bars (Cane Creek as well as Scott Rakes) on many other bikes and never had a problem. But the other frames put my bars closer to the saddle so I had less weight on the bars. This bike is much longer, 3 cm longer TT, much steeper seat tube angle (another 2 cm or so), and shorter head tube.

    On Will's crash, the wheel seems strangely straight for a crash, meaning it doesn't look like it broke on impact. The high number of broken spokes, with no bends etc that I noticed, seems to reinforce the idea that the wheel wasn't broken because of the crash.

    Interesting. In a bad way. I hope that Will recovers completely.

  9. Trackasaurus4:56 PM

    Ouch, whats a jaw fracture like? Will be be able to chew.

  10. Anonymous5:00 PM

    Why must he instantly contact an attorney, can't an accident just be an accident? The equipment is older and used and i'll bet nobody will admit to being the last person too touch that wheel with a spoke wrench.

  11. Andrew6:07 PM

    Jason, maybe you could cite some of the sources which refute Calfee's article. On the face of it it does seem like having the front wheel mounted off-center would cause steering and stability problems, but I'm no expert. I don't think it's fair to dismiss the idea just because it came from a custom frame builder who you think charges too much for his work.

    To make an extreme example, imagine if the front wheel were mounted several inches off center. In order to stay upright while riding a straight line you would have to constantly lean toward one side, which would cause you to have to steer a little bit opposite the way you are leaning in order to keep going straight, which seems like the perfect setup for a rider-induced oscillation.

  12. Tripelt6:49 PM

    The break does appear to have happened at the spot where the rim is joined, I think in the case of this rim by welding.

    Here is a link to the Calfee paper:

    He offers the service of checking your fork for symmetry and claims his forks are more symmetrical than other manufacturers.

  13. that rim failure looks like it occurred at the part where they join the rims. the failure looks too straight and clean. and they usually put a sticker over that part to hide the line.

    Sometimes, that joint is pinned. sometimes welded.

    I call factory defect on that part

  14. It looks like the rim just failed where it is joined together - yes very commonly where the decals are on most rims - opposite the valve stem.

    Does not look damaged from striking anything on the road as the rim would be dented.

    This brings back memories when I was a Junior on the US Cycling Team and my two team mates and I were racing in Poland for the summer.

    I had Mavic G-4 tubular wheels (later renamed to G40) and hit a couple of nasty bumps in training causing a big dent in the rim. The club's mechanic took out a hammer and a block of wood and fixed the rim and later trued it to just as good as new. I later even raced the Nationals with the same rim.

    In this incident - it looks like the rim cracked causing it to break the spokes throwing the rim radically out of true which hit the brakes and caused a dramatic speed reduction in the front wheel - throwing the rider over the bars.

    Not sure why all the talk of a speed wobble as that is not even a factor at speeds over 50 MPH on most bikes.

  15. I'd question a crack across the rim at the joint (or anywhere else) as the only cause of the failure. The force from the spokes should be pulling any crack like that closed & I'd guess you'd need a pretty strong (read: noticeable) outside force, like a pothole or sudden hard steering input, to make the crack open up enough to cause a failure.

  16. @Andrew

    There are numerous sources which describe speed wobble to varying degrees. At least one journal paper and plenty of technical works related to motorcycles.

    Not one really seems to talk about frame alignment with that issue.

    I'd definitely trust Sheldon Brown and his legacy and Jobst Brandt over Calfree any day of the week including holidays.

    No bike is perfectly aligned anyway. Bikes are built to tolerance, not absolute perfection.

  17. The rim could be defective - thus the reason why Mavic is no longer making them. If the rim cracks (especially at it's weakest point - where it is joined) then it's going to break spokes - particularly on this wheel design (with maybe 16 spokes) - if spokes break rapidly they will cause a sudden and dramatic wheel misaligment causing the rim to hit the brake pads on one or both sides - thus causing a sudden change in velocity on the front wheel = going over the bars.

    That is my theory after seeing many road accidents as a member of the US Team.

  18. The cosmos are built with open pro rims. I have a pair a year or two older and the are pretty bomb proof. I've commuted and trained on them for 6 years withou tproblem. Nor sure what happened here but they have been good to me

  19. Nate : Observe the name of these rims though. Its labelled "SUP". Any significance? Are they really open pros?

  20. Dear Marco,

    To make my stance clear, I did not at any point claim that wobble led to this crash. However, wobble can occur at some low speeds and frequencies and depends on a number of factors. Most people agree that there is a component from bike geometry and then a component from a combination of tire and steering influences. For a while, I have been holding this theory that wobbles are something that low wheelbase, flexible bikes with low amount of trail see. Perhaps there is something to do with the moment of inertia of the front end of the bike, as someone mentioned before. In other words, if the mass of the front end components of the bike multiplied with the square of its distance from the COG is low, the bike may not be as stable as one whose front end moment of inertia is higher. However, after chatting with bicycle dynamics experts on this topic, all maintain that we're all neglecting a big factor in shimmy - whats happening at the tire-ground interface. I unfortunately have little expertise with that but would love to learn how that parameter plays into the equation.

    Either-way, I do agree - it seems a bit pointless now talking about shimmy. That rim crack seems much more interesting!

  21. Ron-

    SUP is mavic's marketing term for a welded seam as opposed to pinned. The wheelset, as far as I know, has always been an open pro rim straight spoke laced to a pair of mavic's generic hubs. They are the same dimension/shape and I've been told by high end bike shop owners that I've worked at that they are the same rim.

  22. Is it possible one or more spokes failed prior to the crash. To which the wobble and loss of feedback in the steering could be attributed. It's hard to reconstruct the actual events when things happen so quickly in a crash.

    It sounds to me that the spokes might have failed, the wheel went floppy (the most technical term I can think of). Things go wrong pretty quickly from there when it is the front wheel.

    It appears some of them have pulled out of the hub, not snapped as is common in spoke failure.

    Hard to say what events prior to the crash might have led up to whatever failure occured.

  23. @Shlepzig

    Actually, it looks to me like all of the spokes pulled out of the hub with the flare to hold into the hub basically intact. Notice the plate on the side is bashed out of contact too from when they slipped out.

    So was it in fact the hub that failed first?

  24. In the chicken & egg world ...It looks like the rim failed at the joint...thus allowing the straight pull spokes to disengage from the spoke seats in the hub. This is my complaint with this hub design. It is extremely easy for the spokes to disengage from the hub during a rim incident. The plastic spoke retention ring on the hub flange is more decoration than functional
    highlighted by there is barely a bend on the rim let alone a broken or bent spoke

  25. Anonymous11:11 PM

    it looks like bad alloy...not a clean break
    any one know if mavic controls there own alloy prior to extruding?

  26. Anonymous11:13 PM

    and by bad it looks like pot metal...something was wrong w the composition

  27. Anonymous4:59 PM

    those aren't broken spokes - they're straight-pull spokes that are no longer attached to the hub. It appears to me the problem is related to the spokes coming loose from their slot in the hub. The failure of the rim (at the joint) is most likely an after effect of the crash.

  28. Did the tire/tube remain inflated? I've seen failure like this at the rim joint due to a tube exploding. In one case I remembered immediately upon reading the above, it wasn't clear whether perhaps movement in the rim joint itself as the result of weakness or an impact had pinched/punctured the tube just prior to the blow out. This wheel was a standard but high quality road wheel with typical road pressures (120psi or thereabouts.) I cannot recall the brand of the rim, but at the time, there were a lot of Mavic Open4CD rims on the bikes we were selling. It was 10+ years ago, so that may make it not applicable to this situation. At any rate, spokes were broken, and the rider crashed when the wheel stopped turning. But again, it wasn't really clear which part of what came first. These were just my suspicions at the time.

  29. Anonymous8:15 PM

    speculation here, but maybe he fell on top of the wheel? the rim braking surface is barely worn for a 4 yr old wheel; the spokes are popped out of the hub but not broken, nor is the hub flange broken. having broken spokes (drive side rear revolutions, 10 at once, rim fine; front vs derailleur at laguna seca circuit race, again rim fine) and hub flanges (rear, spokes and rim fine)it's hard to imagine anything other than force directly on top of the wheel causing this mess. feel for the guy, but lawyers making shit up with expert witnesses (metallurgy for jurors..right) is a poor response.


Thank you. I read every single comment.