This is a technical video that shows the mechanism by which aluminum bicycle rims develop fatigue cracking, and shows how spoke tension influences fatigue propagation according to the magnitude of changes of stress during loading cycles.
Very interesting, I'm here today because my MTB disc rims from 2002 just started developing cracks. I'll take this information into consideration when I rebuild them with new rims.
Regarding fatigue life and low vs high stress range, the plot you show is likely assuming a zero mean stress. Having a non-zero mean stress decreases stress life, hence it is not obvious if the highly tensioned spokes are better of or not (it has lower stress range but higher mean stress, hence they might cancel each other). Search for Goodman-Soderberg equation for example of how to use both the mean stress and stress range to calculate fatigue life.
Tension changes under load ... exactly as predicted and explained by Jobst Brandt. Many ppl still believe in the old "hangs from the top spokes" fallacy. This video is very useful.
Hertog, thanks for your comment. You're right that the wheel doesn't "hang from the top spokes" but it doesn't "stand" on the bottom ones either. It's considerably more involved than that.
I've recently developed these very same cracks in an Exile+ 35mm aluminum rim being used on my emtn bike (Jam2). I don't know the tension spec but at this point I've switched over to a back up rim and I'm looking at DT Swiss Hybrid wheels since they are designed for ebike use. I'm a light rider, 145 lbs and the bike is a relatively light ebike at 44.5 lbs but obviously after 2 years of use this particular rim isn't up to the task and is probably dangerous at this point. A side note is that it seems like as the spoke tension is reduced by the crack, the cyclic stresses that caused the cracks to form go up and the process is accelerated. I think I might have dodged a bullet and it was just chance I saw these cracks last week and found your video today. I counted 8 out of 32 spoke holes with multiple cracks today. Thanks, very informative video!
15 years ago these kind of cracks didn’t happen even on walmart bike rims, reality is that it’s absurd for a rim to crack this way after two years usage. They make atrociously garbage rims these days. The same thing happened to me on a Boardman comp bike, after two years rims are cracked up around nipple holes, spokes coming out, meanwhile this never happened to a cheap walmart bike owned for 15 years, I ran into a few cars with a front wheel, took it off and hammered it back into shape with an axe, adjusted spokes, there is some permanent minor "limp" but no cracks, it's still ridable for years. It's absurd quality of rims today, back in the day rims were made for life, not 2 years of commuting, never buy wheels from that joke manufacturer again, it's not your riding, it's their disposable utter abomination rims.
Hi Bill, I am a structural engineer and my rims cracked and I kinda knew why, but you have filled in the blanks, especially about why it cracks on both sides of the nipple. Many thanks, well done!
Ive been cycling for 50 years and I find these types of crack a rare event. I have a lot of wheels and have only had this type of cracking a handfull of times. I also build my own wheels as a hobby and in the past have caused a couple of rims to crack due to overtensioning. Since I bought a spoke tension meter, I feel that Ive "cracked" the problem with wheel builds, it's getting the tension as even as possible thats the key and staying within the limits, for the rim/spoke combination. The only cracks I keep a close eye out for are the ones which develop along the wall, as the rim wears with braking. Im talking of rim brakes of course.
Bill, the chart you show at ~12min does not represent the wheels in the way you labeled it. You labeled "Low tensioned wheel" as "high stress range" but the caption for this chart in the Janssen book specifically states Stress Ratio R is the same. You can't have a constant R where the one that starts at a higher stress also has the lower range - so it is in fact the opposite of what you measured on the two wheels. Stress Ratio R is min stress divided by max stress. So for an example, min stress 10, max stress 20 - has an R of 0.5 and stress range of 10. If you look at a different case where min stress is 16, for R to be the same 0.5, max stress would be 32 and provide a stress range of 16. This is also why the "high stress range" plot ends at a lower crack length - because the increased mean stress inherent for that case takes a shorter crack to cause full failure. I think you should do your tension measurement with a mounted pneumatic tire since the measurements you took without a tire focus the load on the rim's material instead of spreading it out. You can see this presented in the paper titled "Bicycle Wheel as Prestressed Structure" by C.J. Burgoyne and R Dilmaghanian. In that paper, you can see a peak strain in the measured spokes of ~650 for the bare rim vs a peak strain of ~450 for a wheel with a tire at 90psi. I believe a good test to demonstrate the effects of high vs low tension on fatigue life would be to build 6 identical wheels, 2 each at varying average tensions from 'too low' (80kgf) to 'normal' (110kgf) to 'too high' (140kgf) and run these loaded on the drum tester until you encounter rim cracks.
Very thorough and understandable, Bill! When you got to the "saddle" analogy (~21:00) I thought you were going to say "hyperbolic paraboloid"! My calculus is reeeeeally rusty, but you explained this perfectly. I'll be interested in your follow-up with the washers. Meanwhile, my wider rear rim has not developed the "sagittal" fracture in the tire bed, going on 3,000 miles. Stay safe! Steve Weeks
Bill, I wonder how your experiment is going. You omitted in this video the effect of anodising in crack propagation. It would be interesting to put non-anodised rims into the mix but I understand that doing that and the controls required, is a full time job, nevermind a little side-line.
Great techical video , some carbon rims top brands dont drill spoke holes drilling weakens the structure of the spoke bed and washers help spread the load and not leave it on one small shoulder of a nip
Could you remove the bar from the bb shell, install a drivetrain, and hang the weight from a pedal with the crank horizontal? You would need to install a tire and perhaps lean the bike to balance it. This would load the wheel with the torque and side load of sprinting.
Enjoyed learning these details..... Question: What kgf spoke tension should i design for using SAPIM Race spokes on rear wheel 274 mm disc brake side and 276 mm drive side using a carbon fiber 700C road bike rim designed for a max of 180 Kgf. Spokes are installed with brass nipples c/w spoke rim washers? This is my very first wheel build.
This is interesting. My old Velocity rim has developed cracks at most of the nipples on the drive side but the cracks are perpendicular (cracks propagate toward the edge of the rim) to what you show in this video.
Makes me glad I was wise enough to keep all my steel rims. Aluminum and carbon fiber just can't handle the stress of riding. Especially where there are a bunch of bumps in the road.
I suspect that an inflated tyre would increase the degree of uniformity of tension among the spokes of the loaded wheel, mainly because it will redistribute the ground load more evenly around the rim, instead of just pushing at one point on the bottom of the rim, as in the laboratory test.
Great experimental work. Even though I’ve seen tyres every day forever I’ never thought about how load is transferred when you put a tyre on the wheel. I’m guessing that the side wall of the tyre can not transmit a great deal of load in compression to the rim. Air pressure would be equal everywhere in the tube so, is the load is transferred to the rim by tension/shear in the side wall of the tyre lifting the top of the rim upwards?
I'd say 'no' to the sidewall transferring any significant force on the rim in any place other than the (almost) point load. Rubber is so pliable that the force of this type is so small compared to the stiffness of metal that it is not a factor. The rider depends on the stiffness of the alloy to handle load by - as much as possible - not bending, whether the contact point is at a spoke point or not. The force of the *tire to the rim* in any significance is only at the point the tire is making contact with the ground. It's about the same as having a tireless rim sitting on a small rubber ball. That is what and how the rim supports the load right at that point, and the metal is doing most of the work there (work being defined as resisting bending). All other places of the rim is just being held in place by the spokes. As an aside, LL below referenced a paper done on the difference in spoke tensions between a rim with a tire and one without. I think this is interesting but can't help but think that a point load on the metal rim without the tire is approaching (or even exceeding) the modulus of elasticity of the alloy and causing odd behavior in the rim. (If this is the case, I don't think this is technically very useful.) It'd be interesting to know.
BIll, I built wheels for our tandem using Ryde Andra 35 rims (559), 36 Saprim butted spokes, and Saprim polyax nipples. Sadly….After 7000 miles, much of which was heavily loaded, 4 of the spoke holes in the rear wheel have developed 5 mm long hairline cracks in the rim. The spokes are not next to each other, but in very different parts. The wheel remains perfectly true with average tension of 105 kgf (both sides, Rohloff hub) with 15% variation. Ryde Andra rims have radial hole drilling of 8 degrees and the spokes do point straight to hub from nipple. The polyax nipple presumably helps with additional angulation. Your video shows that the MG washer is not a simple flat washer. Would this washer, rather than the conical HM washer, be best for pressure distribution to help stop cracking?
Michael, Sapim MG washers might work well if they can sit relatively flat on the inside of the rim. It's hard to say for sure without knowing the cross-sectional dimensions of the rim.
@@BillMouldWheels Hi, I have a Duroc 30 - 32H rim with 6 holes with small cracks from 2 to 4mm. Using Sapim MG washer in these holes can prevent the crack from progressing?
Whenever I analyze a rim with some miles on it, I invariably see that the nipple (polyaxial or not) has set into the hole where the sharp edge is gone because the aluminum has been peened blunt (or worn blunt, but probably peened), so I'm not certain the saddle theory holds in practice because of the 360 degree contact. Knowing the peening effect, I think one should lean more to what really is happening than the mathematical explanation, which, by the way, is brilliant. Considering the peening, isn't the cracking at the nipple more because of the large side to side forces caused by braking and frisky acceleration? (My front rims never cause problems, but my back ones... I can't seem to keep them going very long no matter how well tuned I keep them, but then again I mean business when pedaling from a stop - which puts those lateral loads on the rim because of the alternate loads from a left spoke to a right spoke to a left spoke to... etc. and each time these alternate loads are put on a rim the rim bends ever so slightly. Lateral - or perpendicular to radial - loads aren't so much caused by side loads since a bike is ridden in a gyroscopically balanced method that guarantees minimal side loads unless you hit a root or something, but... are caused by the hub either trying to accelerate or retard the spokes, rim, and tire.) Braking probably doesn't do as much damage as the pulsating action of pedaling. Isn't this more of a factor than loads? Less? I know the experiments were only exploring loads - totally respect not mixing other things into that and mussing everything up - but maybe it's food for thought?
Paul, thanks for your thoughtful comments. There is a lot going on for sure. I was trying to explain why the rim cracks occur in the direction they do, but the proximate cause of the cracking is due to a pulsating change in spoke stress, and that pulsating effect will be greater on the rear wheel because that's where the rider's weight is concentrated.
@@BillMouldWheels I'm inclined to agree. The more I think about it, the more I think the majority of the cause of rim fatigue is the concentrated weight if for nothing more than this is "working on the rim" 100% of the time, whereas the things I mentioned is a much smaller percent of the time. (Both things can overlap in time. The back rim really gets abused with all that going on!) Thanks
Does the extrusion process give a directional grain to the aluminum, in the same way that forging does? If so, wouldn't the lateral weakness of the aluminum also contribute to what you describe? Early on, you say that your analysis doesn't apply to carbon rims. Is that because the overlap of plies at varying angles reduces the weakness in one direction?
While it does result in some level of difference from extrusion direction to 90° from extrusion direction, most metal rims are heat treated after extrusion and the directionality is nearly all gone.
A friend bought a pair of Hope w35 wheels ,spoke tension was too tight ,they split between every spoke .The same thing happened to my friends Bmw R80 motorbike front wheel
Richard, thanks for your comment. How do you know that the spoke tension was too high? What was the tension on the Hope wheels? There are other reasons that the rims my have cracked.
The change in tension percentages is less for a tightly strung wheel as compared to a looser one. This increased range (not increased tension) is precisely what causes fatigue. Less variation, less fatigue. It is all about that delta.
I’m about to build a set of aluminum, 36mm outer, 32mm inner rim width mountain bike wheels and was considering spike nipple washers. Seems like they could be a good idea, but maybe only if the nipple interface of the rim is flat. Any results or recommendations to share Bill?
Possibly you could use a washer of the type mentioned, but at least use spokes of the right length and with lots of tension. A goal of 120 KGF on the high tension side of a wheel is a good target.
Very interesting stuff, I wonder if the stresses on the wheel might also come from pedalling. In particular with ebikes increasing effort, noticed my rear wheel has cracks appearing on some of the spokes. Would a steel rim be better for ebikes? Thanks for the video, found it very informative
Yes there are cyclical stresses from pedalling. The pushing spokes lose tension and the pulling spokes increase tension during pedalling. However, pedalling stresser are far, far smaller than load stresses, so we consider those to be neglible in the greater scheme of things.
I had Mavic X-717 26" rims with straight gauged spokes, 32 hole. After a few years of city commuting, and light duty trails my rims developed crack at most of the nipples. What causes this? I never did any down hilling or jumps. Was it a bad wheel build or is the Mavic rim cheap junk?
Rick, it could have been many things that caused the cracking. Without being able to examine the wheel, it is impossible to determine the most likely cause.
This is as often misquoted as the bible, aluminum alloys ae not all the same any more than iron and its alloys are, and all are subjected to fatigue, it is a matter of scale only.
As a note, your loading and expectations for % increase/decrease in spoke tension is not perfect. You're assuming a 100% vertical force, but that is not the case with your setup - there is a horizontal component as well which helps to explain some of your results more clearly.
You can not have any horizontal force in back or forward direction. And it is not setup dependent at all. It's trivially simple why - the bearings will just spin under the torque of this force until the force disappears. I mean, if the frame flexes, the wheel will move back or forward, depending on how this particular frame flexes. And that's it - no force transfer is possible trough the bearings in this direction. The only way to transfer force will be to somehow lock the bearings. Than, with locked bearings, the transferred force will be equivalent to ether spinning up the pedals, or braking with a hub mounted brake, depending on how this particular frame flexes. Horizontal force transfer trough the bearings is only possible in left or right direction. And this force is again, "almost" in this case, independent of the setup. Because if you have any significant left to right imbalance of the weight applied, than the resulting torque will make the bike fall to the left or to the right. All these directions are obviously relative to the normal way the bike moves. Boldly accusing someone of being wrong, but using extremely vague and unclear statement as a "proof" is not polite at best...
@@2L40K I don't have time for a lengthy reply so I will just say that my feedback was meant to be constructive as bill asked for at the end of his video as an attempt to help bridge the gap between experimental and theoretical results. I was not trying to be boldly accusatory. I believe your comment was meant to be far more malicious than mine. "Let ye without sin cast the first stone."
@@smithe53 I have no idea what your answer it's meant to be, but it is nowhere near "constructive". Go to Wikipedia and check what a constructive proof in mathematics is. en.wikipedia.org/wiki/Constructive_proof "In mathematics, a constructive proof is a method of proof that demonstrates the existence of a mathematical object by creating or providing a method for creating the object." Nowhere you demonstrate where and why this force exists, and nowhere you show how do we find this force. You just boldly assert that the force definitely is somewhere. That's not what you do if you are not sure. If you are not sure, you say something like: "I think, there might be a horizontal force, that probably can explain... (this and that)..."
@@2L40K Bill is a smart guy, he doesn't need other engineers doing research for him; what he needs are ideas. Please find something else to do with your day than write back; I certainly have more important things to do.
not sure why the rim sees a stress concentration at 18:41. I see that the lower edge sees a compressive force but dont see any increased tension on upper edge
unless youre proposing that the wheel is bending in the perpendicular to the radial direction and the bottom surface sees increased compression which bends the cross section upward putting increased tension on upper surface?
@@BillMouldWheels altho I would call it increased tensile stress rather than a stress concentration which is caused by a configuration rather than an increase of loading
15 years ago these kind of cracks didn’t happen even on walmart bike rims, reality is that it’s absurd for a rim to crack this way after two years usage. They make atrociously garbage rims these days. The same thing happened to me on a Boardman comp bike, after two years rims are cracked up around nipple holes, spokes coming out, meanwhile this never happened to a cheap walmart bike owned for 15 years, I ran into a few cars with a front wheel, took it off and hammered it back into shape with an axe, adjusted spokes, there is some permanent minor "limp" but no cracks, it's still ridable for years. It's absurd quality of rims today, back in the day rims were made for life, not 2 years of commuting, never buy wheels from these joke manufacturers again, it's not your riding, it's their disposable utter abomination rims.