Ask the experts

[J. Barron DeJong]

I'm seeing roughly 520 lb/mm for that model (maybe more). Been a while since I've used Ansys outside of the standard small deflection assumption, so I'm not sure how confident I'd be of a linear model for this. But if that holds it would mean over a ton of force is shown by the end. Does that seem right?

Would you mind checking my estimates? I'm looking at almost 3 seconds (call it 40% on the slider, so roughly 2 mm, maybe 2.5 mm) and I see the tension in the bottom 6 spokes changing by about 4500 N (1000 x 4 + 250 x 2) while the top 6 change by about 1500 N and the rest are not quite symmetric about the horizontal plane through the hub, but close. So I'm calling that 1350 lb net and saying those 12 spokes are within about 15 degrees of vertical gives a conservative estimate of 1300 lb at 2.5 mm.

I think this is the same video I've seen criticized for using a defined deflection instead of a force, but I don't see why that would necessarily matter. Unless the beads crack at 1000 lb, which would make the majority of the illustration a little bit moot (kinda--I'd certainly prefer the tire blows out before total structural failure, but then again this model doesn't seem capable of predicting a taco anyway).

I’m going to preface this: I have not worked with FEA at all in almost 20 years, and never in a professional capacity, so I’m going to try to be cautious to not get out in front of my skis, and even so, I may still say something wrong (maybe even embarrassingly so!)

If I look at the sim at just before three seconds, I see almost all spokes being yellow, then at the bottom we have four brownish ones with ~0 load, and two green with lower load than the rest of the wheel. From the way the spoke colors change, it looks like a discreet jump between load ranges, not a continuum. To simplify, I’m just going to assume that the yellow spokes have a load right in the middle of the 1250-1500 N range, and the green in the middle of the 750-1000 N range. I’m going to assume all the loads in the yellow spokes in the bottom half of the wheel are balanced by the yellow in the top half. I’m going to assume the brown spokes have no load at all. And to simplify, I’m just going to assume that all the loading in the spokes that remain to be considered is acting straight up and down – ignore the angles.

So, on the bottom of the wheel, you have four spokes with zero load, and two spokes with ~875 N load. On top, you have 6 spokes with ~1375 N load opposite those being considered on the bottom. Difference between them is 6500 N, so I’ll take that to be the rough approximation of the load being applied to the wheel at that point. (6500 N is about 1460 pounds).

Is that high?

DT specs 1200 N as a spoke tension for their rims, I believe (and 1200 N falls in the range for the green spoke color that we see at the start of the sim). So four unloaded spokes plus two more lightly loaded will get you into the 5000 N plus range, and if the tension in all the other spokes has been rising above the static tension (which we see from them changing to yellow), then that’s adding onto that 5000 N, so my 6500 N guestimate seems plausible?

What I don’t know is the stress that would be in the rim itself, and at what point the whole wheel would have just collapsed. My gut feel says that would have happened well before the end of the simulation, possibly by the time we reach the load we’re looking at. 1460 lb is like a wheel loaded at 146 pounds having a 10g impact right? That seems to me like a pretty big hit that would likely wreck a rim, but I haven’t seen any bike wheel load/acceleration data, so don’t actually know.

(Apologies to everyone who is not Jono)

[toast2266]

So, on the bottom of the wheel, you have four spokes with zero load, and two spokes with ~875 N load. On top, you have 6 spokes with ~1375 N load opposite those being considered on the bottom. Difference between them is 6500 N, so I’ll take that to be the rough approximation of the load being applied to the wheel at that point. (6500 N is about 1460 pounds).

Is that accurate? I follow your math and line of thinking, but you're assuming that the reduction in tension on the spokes = the load being applied to the rim. I'm no engineer, but is that necessarily the case? If the rim is deforming by a distance greater than the tensioned spokes have stretched, then the spoke tension will drop to zero. And it doesn't take that much force to deform a rim by a millimeter or two.

Mostly I'm just thinking anecdotally. If a not-super-burly wheel is at 1/2 tension (so let's call in 600N), it's not that hard to push on it by hand and flex it enough that the bottom 2 or 4 spokes go slack. And I'm definitely not putting ~550 lbs of force into the wheel when I do that.

(sorry that a math averse luddite is crashing your engineer party)

[J. Barron DeJong]

Is that accurate? I follow your math and line of thinking, but you're assuming that the reduction in tension on the spokes = the load being applied to the rim. I'm no engineer, but is that necessarily the case? If the rim is deforming by a distance greater than the tensioned spokes have stretched, then the spoke tension will drop to zero. And it doesn't take that much force to deform a rim by a millimeter or two.

Mostly I'm just thinking anecdotally. If a not-super-burly wheel is at 1/2 tension (so let's call in 600N), it's not that hard to push on it by hand and flex it enough that the bottom 2 or 4 spokes go slack. And I'm definitely not putting ~550 lbs of force into the wheel when I do that.

(sorry that a math averse luddite is crashing your engineer party)

Hell no it’s not accurate!

I’m not taking into account any rim stiffness/internal stresses. And the loads in individual spokes are really rough guesses.

All I’d say is that I don’t see anything glaringly wrong with how the model is working (but again, I have extremely limited experience with FEA).

I think the video gives a good visualization of how when a wheel is loaded it’s the bottom spokes that take up that load by losing tension and not the upper spokes increasing tension. I wouldn’t put much more weight into than that without knowing all the parameters that went into the model.

Thinking a bit more about this, if you picture the load on the wheel as being applied through the hub, the only thing constraining the hub is the spokes that are attached to it, so to keep everything in balance the loading in the spokes need to change such that the change in tension balances the force applied to the hub (doing my best to avoid saying Free Body Diagram). Any stresses in the rim itself are going to need to be transmitted through the spokes to the hub, so I’m pretty sure that you know the applied load if even if you only know the tension in each spoke and not necessarily the stresses in the rim itself. Assuming that’s correct, the wheel you’re able to get de-tensioned wouldn’t have had a starting tension of 600N in each spoke since taking the tension off of just two spokes (ignoring that all the adjacent spokes would be taking some of the load as well) requires 1200N which is 270 lb.

Just thinking out loud here, so could very well be missing something.

Edit: I think I know the something I was missing - when you were applying the load to the rim you were probably pushing down on the top of the rim, not the hub axle, right? In that case it’s not change in spoke tension between the upper/lower half of the wheel that’s going to be taking up the load (I think if you checked the spoke tension you’d see that both the upper and lower spokes were losing tension when you pushed on the top of the rim) so mostly what you’re achieving is rim deformation, so will be much easier to get spokes to unload in that case I believe. (Clear as mud explanation!)

[toast2266]

Hell no it’s not accurate!

I’m not taking into account any rim stiffness/internal stresses. And the loads in individual spokes are really rough guesses.

All I’d say is that I don’t see anything glaringly wrong with how the model is working (but again, I have extremely limited experience with FEA).

I think the video gives a good visualization of how when a wheel is loaded it’s the bottom spokes that take up that load by losing tension and not the upper spokes increasing tension. I wouldn’t put much more weight into than that without knowing all the parameters that went into the model.

Thinking a bit more about this, if you picture the load on the wheel as being applied through the hub, the only thing constraining the hub is the spokes that are attached to it, so to keep everything in balance the loading in the spokes need to change such that the change in tension balances the force applied to the hub (doing my best to avoid saying Free Body Diagram). Any stresses in the rim itself are going to need to be transmitted through the spokes to the hub, so I’m pretty sure that you know the applied load if even if you only know the tension in each spoke and not necessarily the stresses in the rim itself. Assuming that’s correct, the wheel you’re able to get de-tensioned wouldn’t have had a starting tension of 600N in each spoke since taking the tension off of just two spokes (ignoring that all the adjacent spokes would be taking some of the load as well) requires 1200N which is 270 lb.

Just thinking out loud here, so could very well be missing something.

Edit: I think I know the something I was missing - when you were applying the load to the rim you were probably pushing down on the top of the rim, not the hub axle, right? In that case it’s not change in spoke tension between the upper/lower half of the wheel that’s going to be taking up the load (I think if you checked the spoke tension you’d see that both the upper and lower spokes were losing tension when you pushed on the top of the rim) so mostly what you’re achieving is rim deformation, so will be much easier to get spokes to unload in that case I believe. (Clear as mud explanation!)

Yeah, that all makes sense. I guess my point is mostly that I don't think you can discount rim stiffness. I think your numbers (or really, the thought process, since the numbers are just guesstimates) only work if the rim doesn't flex or bend at all.

[J. Barron DeJong]

Yeah, that all makes sense. I guess my point is mostly that I don't think you can discount rim stiffness. I think your numbers (or really, the thought process, since the numbers are just guesstimates) only work if the rim doesn't flex or bend at all.

I guess it depends on what you are actually trying to figure out, but I think I’ve pretty much convinced myself that if you want to know what external load is being applied, you only need to know the tension in the spokes, don’t need to know anything about what’s happening in the rim specifically. (That’s assuming the load is transmitted through the axle. I think that’s a safe assumption for normal riding, but not the hand test). But that doesn’t mean the rim has no bearing on how the wheel behaves - a stiff rim will distribute the tension across all the spokes differently than a soft rim would, so applying the same load through the hub is going leave you with a different spoke tension pattern depending on the rim. Stiff rim should distribute the load more evenly across more spokes instead of having a more localized area of spokes taking up the majority of it (I think)

[sethschmautz]

Any carbon fiber experts in here? I am considering making a modification to a frame and would really like somebody who is educated about such things to tell me I'm an idiot (or not). Would be great if they were an engineer. :-)

Sent from my SM-N986U using Tapatalk

[toast2266]

I guess it depends on what you are actually trying to figure out, but I think I’ve pretty much convinced myself that if you want to know what external load is being applied, you only need to know the tension in the spokes, don’t need to know anything about what’s happening in the rim specifically. (That’s assuming the load is transmitted through the axle. I think that’s a safe assumption for normal riding, but not the hand test). But that doesn’t mean the rim has no bearing on how the wheel behaves - a stiff rim will distribute the tension across all the spokes differently than a soft rim would, so applying the same load through the hub is going leave you with a different spoke tension pattern depending on the rim. Stiff rim should distribute the load more evenly across more spokes instead of having a more localized area of spokes taking up the majority of it (I think)

Definitely agreed that a stiffer rim will more evenly distribute the load across all spokes.

But let's say you put 1000 lbs of force into one specific point on a regular aluminum rim. The rim is going to buckle at that specific point. And it's going to buckle regardless of what tension the spokes are at; you're simply exceeding the strength of the aluminum. But the spokes are going to go slack in that area. Which is why I don't think you can directly correlate spoke tension loss with input force.

[J. Barron DeJong]

Definitely agreed that a stiffer rim will more evenly distribute the load across all spokes.

But let's say you put 1000 lbs of force into one specific point on a regular aluminum rim. The rim is going to buckle at that specific point. And it's going to buckle regardless of what tension the spokes are at; you're simply exceeding the strength of the aluminum. But the spokes are going to go slack in that area. Which is why I don't think you can directly correlate spoke tension loss with input force.

I get what you’re saying, but I still think that when things go sideways, like a rim buckling, ultimately all the loading still gets transmitted through the spokes and that you’d still be able to figure out what load was being applied if you knew all the spoke tensions.

That said, we’re kind of off in theoretical land here (which can be a certain type of fun!) but in practice I doubt designers/engineers would care much beyond maybe at what load spokes are going to become de-tensioned, and what load causes rim damage. And they’d be applying a known load to do that, not trying to figure out what load was applied by measuring tension in each individual spoke.

Anyway, I could be wrong… But not convinced that I am yet

[XXX-er]

Any carbon fiber experts in here? I am considering making a modification to a frame and would really like somebody who is educated about such things to tell me I'm an idiot (or not). Would be great if they were an engineer. :-)

Sent from my SM-N986U using Tapatalk

I'm in the " you are trying to be an idiot camp"

but wait till I make some popcorn

[rideit]

Any carbon fiber experts in here? I am considering making a modification to a frame and would really like somebody who is educated about such things to tell me I'm an idiot (or not). Would be great if they were an engineer. :-)

Sent from my SM-N986U using Tapatalk

Just for fun, can you share what you are thinking? Color me intrigued. Hole for a dropper post cable? Rivnuts for water bottle mounts? Or maybe a machine gun mount for grizzlies?

[sethschmautz]

I think I've talked myself out of it. I was hoping that I could figure out a way to use a shock I love (Mara Pro) on this frame I got. Front what I can see the shock would fit if I were to remove some material (2mm is my guess) on the end of the frame mount. That mount sees the vast majority of it's high stresses/forces on compression, not extension, but there are certainly extension forces at play.

I have a department of 11 mechanical engineers at my workplace and have run it by a few of them. In general, they think it would be fine, but they recognize that there is risk. And none of them are composites experts from the Internet. ;-)

At the end of the day, though, there is the potential of ruining a good frame and certainly destroying the resale value. I feel pretty loyal to Manitou for the way that they have supported me over the years and I'm finding this one strangely hard to let go of. :-) I can't decide if it's that I don't like hearing the something can't (or shouldn't) be done or if it's that loyalty playing the biggest part in that.

Sent from my SM-N986U using Tapatalk

[dannynoonan]

^^what if you just flip the shock around the other way?

[jono]

I get what you’re saying, but I still think that when things go sideways, like a rim buckling, ultimately all the loading still gets transmitted through the spokes and that you’d still be able to figure out what load was being applied if you knew all the spoke tensions.

That said, we’re kind of off in theoretical land here (which can be a certain type of fun!) but in practice I doubt designers/engineers would care much beyond maybe at what load spokes are going to become de-tensioned, and what load causes rim damage. And they’d be applying a known load to do that, not trying to figure out what load was applied by measuring tension in each individual spoke.

Anyway, I could be wrong… But not convinced that I am yet

I don't think you're seeing this wrong, either. I was looking at a point a little earlier than you were, because it was convenient to see the change in tension there. I think I was looking at about 2 mm, actually, but our numbers seem roughly proportional.

The rim stiffness plays a role, but less (maybe much less) than we would intuitively expect because the pre-loaded compression makes it so the rim and spokes are unloading for a while near the point of "impact," which makes the rim act much stiffer.

On net, it all seems to add up to the idea that probably most wheels taco before getting as much load as the defined deflection in that model (defining the rim's deflection one-dimensionally instead of just applying a load artificially forestalls rim buckling, at best). And hopefully a bead dent/failure precedes a taco on a highly tensioned wheel, since the local stress right at the bead is usually more concentrated when a pinch flat occurs.

Initial spoke tension in that sim is easier to pick off at the start, of course, but just how much you start with is critical to the whole thing.

The dynamics of using a little rim flex to your advantage isn't totally insane if your rims are free. Like lower tire pressures. But it's riskier and I'm still pretty firmly on the side of keeping the spokes highly tensioned for us mortals. At least until I see something like this that's a lot more detailed (and specific to a rim/spoke combo).

[sethschmautz]

^^what if you just flip the shock around the other way?

Good question. That was my original thought but it's a yoke on the other side in an orientation that would put the piggyback off to one side or another (and probably contact my thigh on every pedalstroke). I thought about reengineeing the yoke to fix this (have a friend cut it on his mill) but it introduces another pivot at that location. And would render the shock ineffectual. :-/

It's possible that a yoke could be designed and cut that prevents it from rotating but it would have to have hard contact with the shock top cap and I'm not sure that's a good idea.

The red piece in the picture is just a shock spacer but the bolt sleeve comes in vertically.

Sent from my SM-N986U using Tapatalk

[Andeh]

Why don't you at least try it with the shock reversed (& res on side) to see if it actually bugs you? I'm guessing your stance is wider than you think.

[sethschmautz]

Why don't you at least try it with the shock reversed (& res on side) to see if it actually bugs you? I'm guessing your stance is wider than you think.

That's a good call - I should. I sometimes ride with an extra water bottle cage and bottle on the top tube when I need extra water or bear spray handy so yeah - it's worth a shot!

Sent from my SM-N986U using Tapatalk

[jono]

Definitely agreed that a stiffer rim will more evenly distribute the load across all spokes.

But let's say you put 1000 lbs of force into one specific point on a regular aluminum rim. The rim is going to buckle at that specific point. And it's going to buckle regardless of what tension the spokes are at; you're simply exceeding the strength of the aluminum. But the spokes are going to go slack in that area. Which is why I don't think you can directly correlate spoke tension loss with input force.

Consider it this way: the only forces acting on the rim are the spokes and the ground. In that simulation there is no acceleration, so the total vertical force balances out and the difference in tension on (roughly) vertical spokes is (roughly) the ground force. (How's that for triggering the FBD-sense, J. Barron?)

In reality it's not static, it's dynamic: the rim does accelerate, both locally and overall (though its mass is very small so the force change is negligible in the simulation). So a little rim flex gets it out of the rock's way a little bit and slightly reduces the force (like ~10 lbs, say), which seems to be the temptation for building more flex into the wheel--not to change the force, but to look for more traction.

It may help to clarify terms slightly, too: a sharp impact can lead to local buckling of the bead(s), but tacoing the wheel (and similar but cooler looking failures) is also buckling--overall rim buckling, which is not local to the applied force (+/- ~90 degrees from force to yield point in the case of a taco). That overall buckling is very dependant on spoke tension.

The relevance of all this is a little esoteric, but obviously comes back to this:

Stories from Snowshoe WC of French teams experimenting with loose spoke tension. Then there’s Daprela’s exploding wheel …..

[radam]

Would an offset bushing in the top eyelet give you enough clearance?

[jono]

Would an offset bushing in the top eyelet give you enough clearance?

If I read it right, that's where he gets the choice of an extra pivot or contact with the shock top (or similar).

[sethschmautz]

I wondered this also. I have meant to reach out to some manufacturers and see if that's an option. That would require no modification, I would think.

[El Chupacabra]

If you're considering machining a new mount for one end of the shock, couldn't you do that and rotate the shock mount 90* at the yoke pictured above? (the yoke between the seat stays and the shock)

[toast2266]

Consider it this way: the only forces acting on the rim are the spokes and the ground. In that simulation there is no acceleration, so the total vertical force balances out and the difference in tension on (roughly) vertical spokes is (roughly) the ground force. (How's that for triggering the FBD-sense, J. Barron?)

In reality it's not static, it's dynamic: the rim does accelerate, both locally and overall (though its mass is very small so the force change is negligible in the simulation). So a little rim flex gets it out of the rock's way a little bit and slightly reduces the force (like ~10 lbs, say), which seems to be the temptation for building more flex into the wheel--not to change the force, but to look for more traction.

It may help to clarify terms slightly, too: a sharp impact can lead to local buckling of the bead(s), but tacoing the wheel (and similar but cooler looking failures) is also buckling--overall rim buckling, which is not local to the applied force (+/- ~90 degrees from force to yield point in the case of a taco). That overall buckling is very dependant on spoke tension.

The relevance of all this is a little esoteric, but obviously comes back to this:

Yeah, that all makes sense.

And yeah, teams have been playing around with soft spoking wheels for a while. The syndicate was doing it ~10 years ago (also with enve rims, interestingly. I wonder if someone at enve is behind that).

And then there's those zipp rims that have flex intentionally engineered into them. Haven't heard much about how that plays out in the real world though.

[sethschmautz]

If you're considering machining a new mount for one end of the shock, couldn't you do that and rotate the shock mount 90* at the yoke pictured above? (the yoke between the seat stays and the shock)

That was another idea I had. However this introduces a pivot at that location. A compression would simply pivot there instead of compressing the shock.

To compress the shock and not rotate at that location the yoke would need to contact the top cap of the shock so it wouldn't be allowed to rotate. This would transfer some of that rotational force to the top cap. I have no idea if that would be bad or not.

Sent from my SM-N986U using Tapatalk

[jono]

Yeah, that all makes sense.

And yeah, teams have been playing around with soft spoking wheels for a while. The syndicate was doing it ~10 years ago (also with enve rims, interestingly. I wonder if someone at enve is behind that).

And then there's those zipp rims that have flex intentionally engineered into them. Haven't heard much about how that plays out in the real world though.

Would those be these?

"Zipp’s 3ZERO MOTO carbon 27.5 and 29 enduro/trail wheelset is inspired by moto to provide riders with the control and durability required for pure trail speed. Already this wheelset and its MOTO Technology are proven with multiple victories in the Enduro World Series with our pioneering partnership with Lapierre Zipp Collective.

While other wheel brands limit their focus to traditional box-section rim designs, Zipp engineers—uninhibited by legacy technology—took a different approach. They looked to motorsports as a model. Zipp’s Moto Technology’s single-wall rim is the optimal approach for an enduro/trail wheelset. The resulting rim provides something we call “ankle compliance.” Imagine a runner rounding a sharp turn, the ankle naturally flexing to maintain grip as the runner leans. The 3ZERO rim can locally flex to stay parallel to the ground during cornering, increasing traction like a human ankle adjusts while running. This ability to twist locally allows it to deflect during single bead impacts without the rider getting bounced offline. 3ZERO MOTO helps you take the most direct route, saving seconds when and where they count most."

Yeah, that sounds just great, doesn't it? MOTO weight and you have to change the lean angle to get the side knobs to bite thanks to "ankle flex." I'm gonna go out on a real short limb here and say SRAM figured out they could make single wall rims cheaper than box section.

[El Chupacabra]

That was another idea I had. However this introduces a pivot at that location. A compression would simply pivot there instead of compressing the shock.

To compress the shock and not rotate at that location the yoke would need to contact the top cap of the shock so it wouldn't be allowed to rotate. This would transfer some of that rotational force to the top cap. I have no idea if that would be bad or not.

Sent from my SM-N986U using Tapatalk

Duh, you're right. I wasn't thinking about how the shock would move in that orientation.