close
Pinkbike is now part of Outside! As of December 3, 2021, please refer to the Outside Terms of Use and Privacy Policy which govern your use of the Pinkbike website and services.

SRAM Produces Generative Design Prototype Cranks in Partnership with Autodesk

May 20, 2021 at 3:20
by James Smurthwaite  

SRAM has created a number of new crankarm prototypes in conjunction with Autodesk that use generative design to attempt to maximize their properties.

What is generative design? Well, rather than starting with a drawing of what they believe a product should look like, the designer instead defines the parameters needed for the final result - for a crank arm that might be its length, strength, weight, Q factor and construction method (forged vs 3D printed). These parameters are then entered into a computer software, in this case Fusion 360, and thousands of potential designs are generated in just a few hours. SRAM isn't the first mountain bike brand we've seen use this technique and there have been similar experiments in design optimization from Robot and Starling in the past few years.


Autodesk claims these designs are unlike any a human could come up with and SRAM has now selected a number of them to prototype for real-world testing. Forbes reports that, "The end results from the design team employ differing materials of construction, and vary from something that still looks quite a bit like any other crankarm, but with significant portions of material removed, to a futuristic design that resembles a structural truss."

SRAM chose to focus on the crankarm because, unlike other areas of the drivetrain, it has remained generally identical to the version we were riding when the sport was invented. It's a structural component, so it has to be safe, but SRAM believed there was an opportunity to reduce the weight if it was approached in a new way. The titanium truss pedals in the video above certainly look lighter than conventional cranks but we don't currently have any concrete numbers on them, other than that they are apparently able to handle up to 10 g of force.


The method doesn't just remove material but it removes time from the process as well. By reducing the steps involved, its claimed an engineer can go from the design phase to the final product more quickly, and that ultimately means a cost-saving over the course of the R&D process.


At the moment, this is all early prototyping and there's no guarantee that it will be taken forward into a real-world product. Instead, SRAM is using it as a learning experience for its future product developments. Will King, Senior Design Engineer at SRAM, said, "We design products to get to the podium in the Tour de France, or to hope for medals in the Olympics. This helps us get to the finish line faster and more educated—to see where material needs to be, and where it can be removed. Generative design has a great app for our traditional construction materials, helps us understand loads and constraints, and pushes ideas we can evaluate.”

Basically, while the more unorthodox designs the software has generated are unlikely to ever be stocked on shelves, SRAM may be able to produce lighter, stronger and cheaper cranks in the future because of its experimentation. We look forward to seeing what real-world results come from this very cool partnership.


237 Comments

  • 393 3
 Computer designed to meet all design loadings in a more efficient way.
Pinkbike: nah mate looks weak dont trust it.
  • 54 19
 Well, you now have to trust both the data set, the algorithm, the structural knowledge about your alloy (well this one doesn't count), the reliability of the machine...
So it would be reasonable not to trust it without more proof of the viability of the final product.
  • 37 6
 @AyJayDoubleyou: The first is a marketing claim. The second is an opinion. Nothing inherently wrong with either.
  • 26 0
 And a year after enough people test it and if works out ok fine. Pinkbike: Looks cool actually...Take my money.
  • 25 2
 Since when was Pinkbike Australian?
  • 10 3
 It’s not just about design loads. But also stiffness. Old school cranks were “strong” enough but flexed and wasted energy.
  • 26 5
 e*thirteen has tried that some time ago and Aaron Gwin found the first sample in his garage...Smile
  • 62 3
 Until it's available in oil slick, it will never be 5% faster
  • 7 7
 @2pi: the part you're referring to was a several years old crank that should not have been on the bike
  • 6 4
 whatever.99 - SRAM forever!
  • 3 3
 @mi-bike:

The first is an engineering claim, I’d claim.
  • 16 0
 @AyJayDoubleyou we don't truss it
  • 9 1
 @IluvRIDING: I mean really letting stuff prove out a while before spending your own money isn't really a bad thing is it?
  • 2 0
 @NEKbiker: ALIEN TECH!! Nuclear powered bikes next!!
  • 7 2
 @Bhaack: and now,stiff cranks transfer more load to tiny bearings,hurray
  • 5 0
 No kidding, software never has a bug in it.
  • 9 0
 Yes and no. I would not trust autodesk in a million years when it comes to the plastic deformation of materials. Leave this kind of design to the big dogs (I.e. Siemens, ANSYS or dassault)
  • 8 0
 Having had a crank catastrophically fail...not a place I am in a rush to save grams
  • 3 0
 @faul: cause who needs experts, right?
  • 2 0
 @monkeybizz: If you were going to just let Autodesk make a crank and run it on your bike then yes I would agree, however I would think they would print a crank and then actually test it on the same rig as their current cranks first.
  • 1 0
 This is where pinkbike needs to allow gifs! So many posts would consist of the “terminator” nucular fall out scene where she melts to a skeleton…..
  • 1 0
 @2pi: savage :-)
  • 8 13
flag RedRedRe (May 20, 2021 at 11:51) (Below Threshold)
 I feel sorry for kids growing up nowdays and falling for every single BS feed to them. Enjoy your BETA life
  • 5 0
 @monkeybizz: I mean, technically, your cranks shouldn't plastically deform. So, sounds like this is in Autodesk's wheelhouse then! Razz
  • 6 0
 "Let's see here....I just plug in the numbers, press render and.....SHIP IT!"
  • 3 1
 @RedRedRe: Big alpha energy. Keep on complaining about the good old days red. Sorry technology and thinking outside of the box scares you.
  • 1 0
 @monkeybizz: Why do you believe that's an issue here? I'm fairly sure cranks are designed such that stresses are within the elastic region.
  • 5 2
 Less material+less manufacturing time= higher cost to the consumer
  • 1 0
 @ReformedRoadie: Have Josh Bryceland test em!!
  • 2 1
 I smell another standard coming our way........hold on! It’s degenerate design
  • 2 0
 Aaron Gwin's mechanic should give these a look
  • 1 0
 @curendero: that would make it an e-bike.
  • 1 0
 @scottlink: Sad but true.
  • 5 0
 @monkeybizz: This made me laugh as I work for one of the 3 you've just mentioned.

Generative Design has been around for a while now and while it will prove in time that the results produced can be better than a manual design the additive manufacturing process used to produce them still has a long way to go. Porosity in the material build up can be an issue and I certainly wouldn't trust them for something as key as a pair of cranks for a number of years yet!
  • 3 1
 @scottlink: just like ladies underwear.
  • 1 0
 @EnduroriderPL: they are 0.99 / 1.0 this is a better way.
  • 2 0
 @scottlink: that they are 1.0 /1.0 certain of.
  • 2 0
 @NEKbiker: any AI worth a crap would have oil slicked it already during prototyping.
  • 1 0
 @MattyBoyR6: Condoms as well....
  • 2 0
 @faul: We are sorry, your crank broke by being loaded in a direction it was not designed for and therefore not covered by warranty.
  • 1 0
 @novajimmer: haha I used to sell one of them. Got to visit some pretty cool places doing additive manufacturing Smile I agree there still is a long ways to go in that regard
  • 1 0
 @curendero: I'll take your word for it.
Ha ha Wink
  • 1 1
 @allmtn303: How is this "outside of the box thinking" this is more "inside" then ever, if you don't like individual creativity (I think we know why that might be) let others do. Not everybody wants to be surounded (or hurt) by waste, fails and advancement of technology.
  • 97 5
 The thing with these structures is, that they can only work well if your input data matches the real world well. For the main stresses coming from eg pedalling and jumping, that might work very well. But for e.g. crashes, where the stresses differ a lot from normal use, it is very hard to model that. And in theses cases, the crank might be way more likely to fail with a lot of optimisation for certain stresses
  • 23 4
 OK, so without taking real world stresses into account, the crank 'might' be more likely to fail. That is allways true. Is there anything in these designs that make you think that they did not do that or just your assumption that it's hard to model?
  • 26 2
 you might me right. but keep in mind:
"We design products to get to the podium in the Tour de France, or to hope for medals in the Olympics." - Will King, Senior Design Engineer at SRAM

If this is the product of this goal only, it works. because with crashes you most likely not gonna win any medals at Olympics and TDF.
"Manufacture cranks that will hold up in a crash" therefore is another goal that needs to be considered. Frankly it bites the main goal: winning Olympics. Wink
  • 8 0
 I would be worried especially about pedal (crank) strikes. It might be ok on road bikes, where it's not such an issue.
  • 21 0
 This is essentially automated guess and check FEA, right? i.e. start with a design and loading data, run FEA, remove material in low stress areas and add in high stress areas, then repeat over and over. FEA is a great tool. It can reduce, but can never replace real world testing.
  • 8 1
 Furthermore, 10g acceleration might sound like a lot, and it is if sustained for any length of time. But the impact of even a small hammer blow will be far more than that, even if it only lasts milliseconds.
  • 8 0
 technically this is correct, but it's weird to mention this in the context of the article. of course you will need to test it in the real world, exacltly the way you test every product or new technology. so this is supposed to be a con of the technology? let's just use a piece of tried and true wood instead.

btw. the x shape shape you see on the rf evolve and sram gx cranks is created with a similiar algorithm. seems to work alright?
  • 5 1
 I don't fully agree with you. It all depends on the input and I am sure there was an input on e.g. lateral forces which are usual when crashing. I very much doubt they only considered radial forces (pedaling/jumping).
  • 2 0
 @ciechan: that's why I said it only works well if your input data fits the application/real world. Of course they didn't only consider radial forces. (Btw, one normally jumps with level pedals. This won't cause mainly radial forces)
  • 17 0
 It’s almost like you guys didn’t read the part about them now doing real-world test riding on the prototypes.
  • 2 0
 Actually if you take an other perspective: With that you are able to test your «virtual world stresses» and compare it to the real world. In the end you are able to opt the digital stresses and having much better data for the real product development.
  • 2 0
 @kcy4130: Yes, from what I've been shown you give the program a volume. Say a rough outer crank shape, and then give it some areas it can't change, say pedal threads and crank axle interface. Give it some loads/cases to check, press ok, and get some coffee. Kinda.
  • 1 0
 @bashhard, I'd certainly be wanting to feed in some real-world loads from a variety of riders and terrains into the optimiser (rather than guess, or estimate based on suspension loads etc). Strain gauges on the pedal spindle ought to do it. Bit of a PITA to log, as it's rotating. For that few channels, for a short period, slip rings would probably work.
  • 1 0
 Yep. Garbage in, garbage out.
  • 1 1
 @kcy4130: Nah this is the opposite direction of FEA optimization. You start with load conditions and manufacturing methods and the software adds material where necessary.
  • 1 0
 @rickybobby18: Yes. But you really need to do real world testing.
You were saying? Smile
  • 3 0
 @kcy4130: No it's topology optimization which is guess and check product shaping. The load is not varied you input them and it starts solving. The material volume is varied based on your intended production material, allowable factor of safety, specified loads, fixed features, and your intended manufacturing method's capability for producing certain types of shapes. You can take it one step further and optimize the generated shapes with FEA after the shape optimization is done for additional load cases to continue the iteration of the design.
  • 2 0
 @sack-zement: you mean winning the olympics or tdf with the UCI arbitrary 15lb minimum weight limit where teams used to stuff lead in the hollow part of the crank?

Brilliant, now there is more space for more lead since the crank is lighter...
  • 1 0
 Tell me about it I’ve got a wonky bent crank from hitting a tree stump that I never saw. I think the little fella broke my toe.
  • 1 0
 @kcy4130: The hammer bit is true, but not as relevant here. The peak force (or acceleration) in an impact will be strongly influenced by the effective stiffness of the impact. A solid steel impact is very stiff and generates very high forces - for very short periods of time. A person landing on a bike is very compliant by comparison.

Imagine riding off a three foot drop on your full suspension MTB - not a problem. Now imagine dropping off of that and landing on your heels with your back straight and knees locked - this is going to be a whole different experience.
  • 1 0
 @kcy4130: FEA is passive.....human makes design, FEA tests it....human edits, FEA tests again..... This autodesk tech is reversing that flow. Human sets design goals, computer figures out configurations that would work and then they're tested. The human ain't actually designing it, merely approving the renditions, and editing post design.
  • 2 0
 @mlangestrom: Even better than that, if I remember correctly, you can specify manufacturing methods too. The big benefits come with additive manufacturing, but I think this can also optimize for structures that are machined or cast as well.
  • 1 0
 @SLOCALBIKES: Yeah the hammer bit is more representative of impact forces from rock strikes. I do a little FEA but am not an expert. Modeling loads from pedaling, hard landings etc would be relatively easy. But modeling impact loading from a rock getting kicked up by the front wheel (or when the crank arm hits an embedded rock) is more challenging, and require lots of assumptions usually. Not sure if they did this or just left it to lab/real world testing. Intuitively the ti arms look like pretty susceptible to fatal rock strike damage.
  • 1 0
 @st-lupo:
Sweet, then you/I can drink more coffee! Big Grin
We've talked about using this at work, but for structural/non visible sheetmetal parts (automotive stuff).
And now that I write this, I accualy have time and a part to try this with.
  • 2 0
 @mlangestrom: we've done it for whole cylinder blocks. Some of the optimisations came out kind of interesting, but by the time we added all the real world manufacture stuff, it didn't end up saving much weight. For items with fewer constraints, it works pretty well generally
  • 1 0
 @kcy4130: right, good point. Real world testing.

Look, it might be all good, and maybe they have done a thorough job (and continue to do so) with this project, but the problem is I have a tiny caveman brain.

The visuals make me nervous, even if they end up being just as strong. If I saw engineers removing material from the wing of an airplane that I was about to fly 1000 miles in and was supposed to be just as strong after testing, I still couldn’t get on that plane.

And the comment about “cost savings” makes me snicker because it’s not like the money saved in R&D would be passed on to us. They’d probably come up with a new slogan like “Vulture “ and actually raise the price.
  • 72 0
 Artificial intelligence has discovered drillium. Or speed holes. Whatever it's called.
  • 13 0
 AI needs to be fed the Danny Mac dataset and the Sam Hill motofoam dataset
  • 2 0
 yeah, those are-um.... speed holes!
  • 7 0
 Just wait until AI discovers oil slick !!!
  • 65 0
 So this is why alien technology in movies always looks so weird, they're just using autodesk...
  • 6 0
 I'd hate it when my cranks would come alive.

(That said: If it's not in a Terminator movie kind of way, but helping me pedal, that would be OK)
  • 5 0
 @jeroenk: But what if they won't let you stop, ever?
  • 1 0
 @jeroenk: I immediately thought of Terminator as well...
  • 10 0
 This technology is pretty common and widely available - and by no means limited to Autodesk. The shapes it tends to produce are often weirdly organic/alien in appearance. The reason you don't notice them in products very often is that the output has to be greatly simplified and adapted to traditional manufacturing methods. With additive manufacturing, though, possibilities get interesting!
  • 12 3
 @ABhardtail: it's not really the manufacturing that stops these designs. The first ti 3d printed one could be machined easily.

The problem with topology optimization in general is that it's extremely sensitive to the boundary conditions and selected loads of your FE model.

Most of the optimization routines also do a horrible job of exploring the possible solution set. GE had a competition obviously targeted at topology optimization---a simple design from a human that maximized MOI ended up being far superior compared to the computer generated solutions.

www.ge.com/news/reports/jet-engine-bracket-from-indonesia-wins-3d-printing#:~:text=A%20jet%20engine%20bracket%20designed,receive%20%247%2C000%20in%20prize%20money.

Engineers haven't needed computers to design parts for 100 years. It's not going to change any time soon.
  • 5 0
 @ABhardtail: It shouldn't be a surprise that the shapes often appear organic. All aspects of living organisms including structures have been "optimizing" thru evolution for hundreds of thousands of years (millions? depending on the species).
  • 6 0
 @thelibrarybiker: True, you can achieve a lot of complex shapes with CNC machining but that gets expensive. The structures I design are usually made from flat steel plates and my team has had occasional success with topology optimization even in that context (with our proprietary software), though you're absolutely right about loads and boundary conditions. The last time I considered using the technology I ultimately decided against it because the loading was too variable and hard to define. Cool link btw, that's an elegant little bracket
  • 1 0
 @thelibrarybiker: This its not building it that's the problem, the software needed to optimize a design this far is astoundingly complex and has mind boggling amounts of variables.

Also I feel like my Autodesk app management suite on my work computer has been listening to me now after this page came out....
  • 6 0
 @thelibrarybiker: „ Engineers haven't needed computers to design parts for 100 years. It's not going to change any time soon. „

Sarcasm?
  • 3 0
 @danielson-01: Not really. I'm a mechanical design engineer with experience in product design, structural design, extensive FEA work, and failure analysis. The computer is a tool that is massively, massively useful obviously, and my work would be near impossible without it, but overall, layouts/concepts/geometry of parts is all from the head - a mixture of repeating known optimal solutions, using theory to develop new optimal solutions, or simply intuition. The computer helps with materials, sizing etc, which is all just calculations, but no actual creation at all is done by it.

I can see generation/optimisation becoming more useful very soon, but currently and in the immediate future it's pretty worthless. The crank above is a good example of this. It's an awful design.
  • 1 0
 @AgrAde: Engineers don't need computers
Also @AgrAde: My job is impossible without computers
  • 1 0
 @AgrAde: "It's an awful design". Seriously interested: Based on what?
  • 3 0
 @jeroenk: There are a few things that with a 5 second look over it, don't appear well engineered. The first is not using a closed section to improve the torsional stiffness and strength of the crank. The second is the section near the BB where it splits completely into two near-parallel flat sections. This will be inducing a lot of bending into the actual members of the truss, raising stress. Normally in a truss you'd use triangles so that all the loads can be resolved into pure tension or compression of the members, and no members are being bent. The generator has added stabilisation to this part of the crank with the extra members attached to the bb interface part (and thickening this area), and the random weird internal member just in front of the chain in the pic, as well as thickening the left side of the X part on the side of the crank, and I feel that they're all crutches to aid the highly loaded flat sections. A lot of the general shape seems odd to me. I don't like where it goes onto the bb spindle and I don't like the lumps between the X and the sram logo. I don't like how much it necks down before the pedal.

Stress flows through a part in a way that encourages engineers to use simple, broad shapes that don't allow stresses to be concentrated in edges or bends. Using weird shapes like this can work on a part with very well controlled loads, but even a crank is too variable for this imo. Changes in the ratio between bending and torsion in the crank (say, a heavy person with narrow cleats vs a light person standing on the edges of their flat pedals) can mean large changes in the way stresses manifest in the material and mean high stresses in under-built areas in the crank. A simple broad shape (like a hollowtech XTR crank) will experience much smaller changes in the peak stresses across a range of load scenarios. And that's not even considering pedal strike or crashes.

I'm prepared to eat my words if I could see how they've run the simulations and it's very comprehensive, and strong compared to other cranks for its weight across all scenarios and crash/impact loads, but my thoughts are currently not. I'd love to see it compared to a tubular 3d printed ti crank with well-optimised geometry.

I do think that a very good looking, good performing crank could be designed using this one as a basis.
  • 2 0
 @mtskibum16: near impossible* Smile

My point is that the computer is currently still used purely as a calculator. It doesn't "do" the design work.
  • 54 1
 wait until it's full of mud Big Grin
  • 14 1
 Could easily have a sheath over the top to keep the lines clean, but structurally within it’s all wibbly wobbly.
  • 15 0
 A carbon fiber sheath would be cool...
  • 30 0
 @dingus: sounds itchy...
  • 4 23
flag ciechan (May 20, 2021 at 4:55) (Below Threshold)
 @dingus: what would be the point of using carbon fibre sheath? It wouldn't transfer any forces anyway...
  • 37 1
 @ciechan: to keep the mud out? Did you even read the first comment?
  • 7 2
 Wrap it in cling film. Then you can see the sketchyness aswell as benefit from the low weight. I don't know the density of mud but it can't be far off aluminium.


Does aluminium float in mud?


No, my bike sinks when I ride through mud.


That's how it works, right?
  • 5 0
 @deadlyhifi: then you would lose all the feeling!
  • 11 0
 it is always the brits bringing this up
  • 1 0
 @TannerValhouli: why does it need to be carbon fiber? Simple plastic sheath will be lighter and cheaper and still will keep mud out.
  • 3 0
 @deadlyhifi: “wibbly wobbly” are engineering terms I am unfamiliar with.
  • 1 0
 @ReformedRoadie: oh thats usually associated with the bigger, most efficient and smartest engineers, who make the greatest inventions like german cars
  • 2 0
 @carrottread: Everybody knows that carbon fiber is faster
  • 1 0
 "You never said it would be used in dirt or mud in the design phase.
  • 3 0
 @inked-up-metalhead: if it weighs the same as a duck...
  • 30 0
 "Up to 10 g of force".


Is that 10 grams or 10G as in g-force?

Cos my 10G is probably higher force than yours, without a weight for the 10G of force to act upon it means nothing, its just 10x gravity, or 10x your normal weight.
  • 27 0
 Engineering probably did a 10G static force simulation using the weight of a ~95th percentile rider as an approximation of the dynamic forces a crank would see. And then when marketing asked them "how strong is it?" they tried to explain this and the resulting interpretation is what we see in the article, haha.
  • 8 2
 @ABhardtail: oh I imagine so, but doesn't mean it doesn't stink of bs. If it said 10G for a 85kg rider, that'd be absolutely fine, no qualms here, but without the weight that was tested/rated at for the 10G, it means nothing, cos that could have been for their lightest team rider at 55kg, which would only be 4G for me.
  • 4 10
flag rickybobby18 (May 20, 2021 at 5:50) (Below Threshold)
 10 grams is like 1/10th the weight of an apple. I’ll let you use your noggin to figure out which unit they’re talking about here, given that pedaling forces are involved.
  • 5 1
 @inked-up-metalhead:

Bike stuff is usually designed for a 300 pound rider. If otherwise, the manufacture will make the weight limit prominently known.
  • 3 0
 For the record, I am grateful for marketing folks and happily acknowledge that I would misunderstand your job just as well as you would mine (as an engineer). Thank you for helping the stuff I design actually sell so I can have a job!
  • 1 0
 @rickybobby18: Like a Spinal Tap Stonehenge moment?
  • 4 0
 Hopefully they used earths gravity too
  • 2 1
 But how many Joules of Force can it take during a 100 Watt pedal strike? That's the real test my grand pappy always told me boy.
  • 1 0
 @JohanG: you are mixing work with force. Though in reality it is probably the yank that would truly break a crank or the change in force over time. I don’t know that for certain but that is my gut feel.
  • 33 1
 I see they've set the Attractiveness parameter to 0
  • 9 0
 I came here to say this in a less funny way.
  • 19 0
 At 1:28 we can see the engineer performing the full SRAM testing and QC cycle.
  • 16 0
 In other news, Trek partners with Autodesk to create a bike frame using generative design and AI optimization. You'll never guess what it looks like...
  • 10 0
 Gigers session?
  • 13 0
 "handle up to 10 g of force"

10 grams! Not a force. That's a mass, needs acceleration to make a force.

You probably meant G, but that's not a force either. G is acceleration, needs mass to make a force.

And if it did means grams, well, 10 grams at an implied 1 G is not impressive.
  • 1 0
 In the bike industry, everything is a force.
  • 15 0
 I feel a pair of these should be tested by Brage Vestavik on his Real MTB hits!
  • 15 1
 SRAM just bought a pedal company so this software will probably help them develop a new spindle standard
  • 1 1
 Embrace for SRAM's new spindle standard people.
Going metric to 15.099mm offering 8.3% more rigidity!
  • 10 0
 maybe those big brains at autodesk can help sram make a competent crank preload system and main bolt design.
  • 6 0
 What, you don’t like fixing your bike in a vise and pulling on the cheater bar with a tow strap attached to a vehicle? Weird.
  • 13 1
 ha nerd cranks
  • 3 0
 Underrated comment right here. I say this as someone who has done the exact project this article is about (generative design and 3D printed Ti mountain bike cranks). Been riding them for 2 years or so. I do prefer 'geek' though.
  • 10 0
 The sight of those 3d ti printed arms in the in a grid in printer and the CNC'd al arm gave me an engineering boner.
  • 10 1
 the t800 could probably shit out a better crank design
  • 9 0
 Citation Needed
  • 10 4
 All that time making "better" crank arms, shame they couldn't spend a little engineering time to sort out their bottom brackets. Whichever designer dreamt up the GXP axle/bearing layout needs to go back to school.
  • 6 0
 What‘s frightening me is the idea that the same guy might have done the crank arms…
  • 1 0
 @FuzzyL: Heard his name is Johnny Five.
  • 4 0
 Has DUB not fixed your problems?
  • 5 0
 "By reducing the steps involved, its claimed an engineer can go from the design phase to the final product more quickly, and that ultimately means a cost-saving over the course of the R&D process."

... and these savings will be passed on to the consumer, riiiiiiiiiiiiiiiiiight? Cheap high end cranks for all!
  • 4 0
 The big problem really is the cost of the renishaw titanium feedstock. It’s in huge demand in the aerospace sector and the price isn’t coming down anytime soon. There are heaps of parts we would love to use it for on our UAV’s at work, but 90% of the time, the cost of traditional machining for small batches (1-10 pieces) outweighs any marginal benefits from printed parts. Our printer is madly churning out landing gear prototypes at present.
  • 6 0
 Will these cranks be sold as a yearly subscription like all AutoDesk products? The consumer will be offered a yearly update with features that they do not want or need?
  • 1 0
 Hahahahaha...
  • 3 0
 Best AI is Nature.
Bones are complex strutures, but yet, simple and beautifuly "design"
The concept is interesting for Adding Manufacturing design.
All our parts were designed for tradicional (cost effective) machines.

I for once thurst Shimano and their new cranks, and I know I'll maintain my way of living and don't need to sell my house to continue with my sport.
Do this cranks will cost less or equal to existing ones?
What are the major advantages, besides some grams less, and a real pain to clean up?
Humans are always looking for improvement, but sometimes, what's needed, is just saddle time!
  • 1 0
 Had a brand new set of Shimano XT cranks catastrophically fail at the splined interface landing a biggish drop. All splines on the spindle sheared right off. Completely smooth, all the way around. Both pedals in ‘down’ position like they’d been mounted 180deg off. 230lb rider. Cranks installed correctly, torqued to spec, nothing loose.
  • 3 0
 It’s interesting to see prototyping is action. Sure, the programming used may not account for all of the stresses the cranks will undergo, but field testing prototypes should help refine that testing protocol.

My only concern is that SRAM has a track record of selling items that don’t seem out of beta testing. Some stuff works great (35mm chassis forks, current Code brakes). Some have worked poorly for years (3 generations of Avid/SRAM brakes, Reverbs).

Nobody bats 1000, but I sincerely hope that any products derived from this process are field tested to a degree of reliability that matches or exceeds the product they replace.
  • 1 2
 Not really true is it? Sram probably sell more products than everyone else combined, off course your going to have issues when everyone’s riding sram products. Just due to the number of people using and abusing sram. Like shimano have way more issues than sram especially brakes and far fewer people are running Shimano brakes than their is running sram.
  • 2 0
 @thenotoriousmic: having worked at shops for a lot of years, SRAM has had more issues than the other S brand. Shimano isn't perfect (remember dual control MTB brake levers?? Dura Ace 7400 bottom brackets??) but the rate of problems has been a lot lower.

At one point, every.....single....pair of OEM SRAM brakes I saw had the master cylinders fail. Likewise, every Reverb from the first 3 generations would inevitably go soft. To SRAM's credit, their current brakes have proven reliable. The Reverb has an IFP bleed that keeps it working right between (still hateful) rebuilds. But these weren't manufacturing problems, they were engineering problems. That persisted for years.

If SRAM is pushing new crankarm designs, I really, really, really hope they're ready to go from day 1. Neat looking underdeveloped cranks (in pretty colors) were a thing back in the 1990's. A bad thing.
  • 2 0
 I work with 3D prototyping and it‘s very common in most industries. While AI sounds „strange“ for many mtb riders, it has been used to design/test frames etc. since years.

This sure will be the future, as the human brain will never be capable to „guess“ such complex structures. Only nature and maybe AI can do that.
  • 2 0
 Let's hope AI has algoritms for beauty Because these cranks are ugly as f.........
  • 5 0
 I just turned on my old copy of ACAD14 and couldnt find the generative design option.
Anyone else have this problem?

lol.
  • 2 0
 I'm all for light weight and strong, but not for something that will be pebble, mud, and stick trap. These truss-like designs although visually cool, are incredibly impractical for mountain biking. They would all need a sheath of some kind, but then they wouldn't look nearly as cool. You could do a clear (plastic) sheath but that wouldn't likely survive or look good for very long. You could use polycarbonate, but then the crank gets heavier defeating the point. I think we're better off focusing not on how to make the bike lighter, but rather on how to make the rider's lighter, but that doesn't do anything for the profits of bike companies. C'est la vie! At this point I think bikes are about as light as they'll ever be and I do not think innovation in the crank arm design is a worthwhile pursuit. Just change the aesthetics every couple of years and we'll gladly keep buying. Innovate elsewhere on the bike.
  • 2 0
 The lack of true technical details here that riders care about makes it seem like a bit of a marketing piece set up by AutoDesk & SRAM to get some publicity during the olympic events.

Commentator:
"World Champ (insert name here) is riding new parts that were designed using AI(!), in a collaboration w/ AutoDesk & SRAM. That is crazy, it is the future! Look at them go! Amazing what those engineers have done!"

Totally otherwise uninformed Forbes reader who only sees cycling during olympics, saying to self:
"Oh wow, AutoDesk and SRAM are really at the forefront on technology. I'd love to have those parts on a new bike in my garage. Aannnd, I wonder if this AutoDesk company could have a technology that would make me a hero at work!"

Cut to commercial: "AutoDesk....revolutionizing the world we live in with AI(!)"

Totally otherwise uninformed Forbes reader who only sees cycling during olympics, saying to self:
"I'm calling this AutoDesk comapany when I get into the office on Monday! Imma gonna be a HERO AT WORK! IT WAS IN FORBES!"
  • 4 0
 Might as well integrate a "little mouth" to get the full Alien / HR Giger looks
  • 5 0
 It wouldn't be HR Giger without loads of phalluses
  • 2 0
 But no human could come up with that design.
  • 1 0
 Looks like the brake caliper designs Bugatti came up with using 3D printed titanium. Those were the strangest brake calipers I’d ever seen but they tested them on basically an aircraft test bed and they didn’t blow up. Cranks should be a walk in the park.
  • 5 0
 These appear to be the hole-y grail of cranks.
  • 4 0
 I am waiting for another company that makes tools that fit in the crankarm now.
  • 1 0
 Booooring. I was hoping for Pierce Brosnan to steal the crank out of a submarine before it was used to power a dangerous new reactor. Way too much real information about the product here. Hard pass. Also next time I think the crank should either be shaped more like a veiny flexing arm if possible. Thanks!
  • 3 1
 As someone who uses autodesk products every day (Maya) my trust in the working is -10.

"It broke? Don't worry we will fix it in the next update... that you will have to pay for."
  • 1 0
 Why I own SolidWorks. They don't own me.
  • 1 0
 YES. I want more parts that look like they were made by the Borg.

Also stoked to see a Stache hangin up as a "featured" bike. One of the best bikes that Trek killed. Hoping to eventually see a spiritual replacement now they only make the 1120. 29+ was what it was, really I just want them to keep making a slack XC bike with sliding dropouts.
  • 1 0
 I'd be curious to see the metalurgical aspect of this study. How does the shape resulted from this process look like if they'd be using forged or printed metal or whatever vs. diecasted and machined aluminium, would it be significatnly different, or the raw material is not significant? Interesting concept though.
  • 1 0
 Generative design like this has been around for a while. But designers have generally been intelligent enough to know it's limitations.

This is a case where they weren't. Those crank arms have lost more stiffness and strength than they have lost weight. They've gone backwards. A hollow tube is the stiffest and lightest shape, then you modify that to fit a pedal and chainring/axle and you're done.
Removing the walls like they've done loses all integrity.
  • 1 0
 This is awesome. Generative design is the eventual future of mountain biking (amongst many other things). The technology that went into designing this crank has been used in much more high-risk environments for some time now (such as high performance car parts). If anything, it's more advanced than the technology used to create whatever crank you're riding now, the amount of risk in the comments is quite overstated.
  • 1 0
 This is total crap and looks like weird organic alien space tech as well. Honestly. “Let’s engineer it for 10G forces.” That’s great for leg pressure. Blow through a rock garden at even 15-20mph and have the edge of a rock strike this thing. Newton says the pressure at that point in that moment will be waaaaay higher than 10G. Snap! crackle pop goes your fancy generative design crank. Do the folks coming up with these ideas even ride bikes? Plug in some numbers and let the computer do all the work??? C’mon give me a break. I’ll take a solid aluminum Hopetech EVO crank any day of the week. www.ted.com/talks/tim_huenke_sustainable_humanity_artisanship_in_an_age_of_automation/up-next
  • 2 1
 this is how Skynet takes us all out! first, a pandemic....then a bike boom from the resulting pandemic. then, bike components made from AI software(Skynet underlings), then.....those parts implode and kill us all!!!!!
  • 2 0
 Oh god, SRAM parts designed by autodesk, that sounds like a true nightmare. Any parts will be outdated by the time they get put in packaging
  • 3 0
 What no negative comments? I can’t be the only one with complete disdain for AutoDesk
  • 1 0
 I mean...it beats freecad.
  • 2 1
 @slabhardcheese: the product isn’t that bad, the subscription model though is a small business and independent consultant killer
  • 1 1
 I prefer SolidWorks/HSM works because you own your seats. NX is the most complicated and better than Fusion 360. MasterCam is cool too. There's no do all end all CAD generative design. I will say Fusion 360 is by far the most affordable subscription available for CAD/CAM. Just my 2 cents.
  • 3 0
 I look forward to buying a left-foot-forward crankset. Super optimized. This is what I NEED!
  • 2 0
 What does the PB comment AI bot think about this. The humans need to know.

Note: This comment was not generated by an AI bot
  • 1 0
 I can't be the only one who saw that guy snap his "cost no object" race bike in half the other day in the Giro. Just sayin, lighter isn't always better. Especially for us joto's that buy our own stuff.
  • 2 1
 Why not remove the cranks all together? If you add an electric motor, you won't have to worry about excess weight either. Then, just add a throttle and voila!
  • 2 0
 If it means using less resources to make the same or better product I'm all for it
  • 1 0
 Second picture looks like the bloated head of the bad guy im Terminator 2... I see where it gets the inspiration the machine...
  • 1 0
 Save weight on the computer but doesn’t take into consideration the 5lbs of mud stuck in the crank arm. They now all ship with Saran Wrap/RideWrap to keep the mud out.
  • 1 0
 AMAZING! Literally computers will make better design than men. Just put the right input and the output will be something that's imposible for a human to come up with.
  • 1 0
 IS THE 3D PRINTER going to come to my house and clean the horse manure, mud, dead stuff, debris and my DNA from the spiderweb of metal HAL created?
  • 2 0
 I read "generative" like " oh wow Sram is making cranks that recharge Ebike batteries. ha. need coffee.
  • 1 0
 Make Sam Pilgrim test those by dropping on the set of stairs that's always seen in his videos. If it survives underneath him, then I'll consider.
  • 1 0
 Bikes will be out here looking like ass, but at least they'll be lighter, stronger, and safer.
  • 3 1
 I need this. Bored of all cranks that look all the same.
  • 2 0
 No clearance in these cranks to accept water bottles??? Dig it!
  • 2 0
 They should've just dug up H. R. Geiger for these. (W/ all respect to him)
  • 2 0
 H.R. Giger was his name Wink
  • 2 0
 Bet they didn't model heel rub
  • 2 1
 "...SRAM may be able to produce lighter, stronger and cheaper cranks..." Pick two, and only two !
  • 2 0
 they look like they are made of playdoh theyre so ugly
  • 1 0
 Anyone who's familiar with F360 and this function knows how it's not exactly perfect.
  • 1 0
 Autodesk you say... I hope SRAM doesn't start doing "Groupset as a Service".
  • 1 0
 Interesting how nobody is chatting about the reduced Q-factor. I guess feet can just keep going wider, and wider, and wider.
  • 1 0
 Gee..trust something that has not been properly tested... where have I heard that before?
  • 2 0
 Autodesk...sound like a ...Fatal error! For the cad users.
  • 1 0
 I'll swallow the red pill, but only if you wrap it in thermoplastic so as not to look like some alien muck eating beast.
  • 1 0
 Function over form in its purest here. Interesting but also disgusting looking.
  • 1 0
 Raceface did this first with Altiar's Optistruct/inspire topology optimization...
  • 1 0
 they should do the same for their brakes... so maybe one day Sram brakes will really brake?
  • 1 0
 It looks cool but it also looks like it ups the ante on efficiency of stabbing in the event of a break
  • 1 0
 we haven't figured out to create a useful product yet, so we made this cool video talking about artificial intelligence.
  • 1 0
 "they are apparently able to handle up to 10 g of force." really sunshine, what are the units on that? m/s^2 of force?
  • 1 0
 This is good for the environment! Less material, less natural resources! Hope to see more of this.
  • 1 0
 So shit expensive cranks for uber elites to save less grams than taking a shit. Pointless application
  • 1 0
 Shimano crank, alloy, and still kicking after rock strikes and bad pedalling decision.
  • 2 2
 Could have been even lighter if drive side was mounted to chain ring, but computer was not told to do that?
  • 1 0
 Thanks for sharing this! Pretty incredible example that answers a fair share of the reservations brought up in the comments. At the scale of the frame (rather than a smaller part like a crank arm) the resulting structural shapes are fewer/simpler and could be easily refined for aesthetics/pragmatics, as well as optimized for reductive CNC fabrication. The latter negates the environmental benefits of additive mfr in ti but makes it more realistic from a cost perspective. Super interesting!
  • 1 0
 Looks like the back of my GX cranks. Maybe I should do some cutouts!!
  • 1 0
 It's one thing to design it...it's quite another to make it, economically.
  • 1 0
 How well does it work after an average crash into a rock?
  • 1 1
 The only innovation I can see here is they are riding with AUTODESK FUSION 360 JERSEYS
  • 2 0
 okay
  • 2 0
 Already sold out
  • 1 0
 I think they look amazing... you can take my money!!
  • 2 1
 New crank length standard coming. 168.37mm
  • 1 0
 I want this as replacementpart in my shoulder.
  • 1 0
 Where’s the power meter
  • 1 0
 So FEA is being used with AI. Interesting.
  • 1 1
 Let's throw everything we have at the least problematic and most simple part of the bike! Pure innovative GOLD!
  • 1 1
 most of you must admit you're too dumb to get why this is game changing tech....
  • 1 0
 must be made of elastic bands
  • 1 1
 the second image generates in me a feeling of urgence to eject...
  • 2 2
 Autodesk... Ptsh what gobble funk they talk in the marketing department.
  • 1 1
 And now go back to 24mm axle please.
  • 1 0
 Looks so alien...
  • 1 0
 that looks wild!
  • 1 1
 NX >>>> Autodesk, btw.
  • 1 0
 Take my money
  • 2 3
 As long as they are not made of plastic it's all good.
  • 1 3
 NOPE!
  • 1 4
 i dont like it
Below threshold threads are hidden

Post a Comment



You must login to Pinkbike.
Don't have an account? Sign up

Join Pinkbike  Login
Copyright © 2000 - 2021. Pinkbike.com. All rights reserved.
dv56 0.023571
Mobile Version of Website