This needs a test: power meter on the cranks and power meter in the rear hub on a flexy and on a stiff frame to see if there's any difference in the power delivered to the cranks vs. that delivered to the rear wheel. Use the same crankset, and wheels. Test with a sprinter and a climber to see if rider weight has an effect. Maybe throw in a flat (no climbs to minimize any effects from weight difference) time-trial with a rider putting out a steady power output on the crank-based power meter to spice things up.
There is indeed power lost between the cranks and the rear hub. Most of it in the chain. Stiff is better because it's more stable. Not because you actually lose energy by loading a spring. You lose a lot more energy in places that you don't even look. If you're getting chain or brake rub that's an entirely different story. If you're not getting obvious performance problems from the flex you should go search for "lost watts" in the tires and the chain. Once you're close to optimizing that (for what you can afford) then go hunting for "lost watts" in your bearings. You lose a lot more power in a dirty chain than you do in a "flexy" frame (as long as you're not causing friction from bad alignment). When you lose energy in the chain you can not get it back. As illustrated in the video, most of the energy used to load the spring (frame) is returned to you and the actual values are trivial.
@@indonesiaamerica7050 doesn't this suggest that a smoother ride (which could come from a softer frame) is more stable? Also, that clutch derailleurs could help too
the test only shows how much those frames store energy like a spring would it has nothing to do with efficient transfer of energy that would make a bike fast. a bike frame cant be stiff enough. most comfort comes from seat post deflection and tyres anyway.
Hello, from an old and experienced rider. I believe I can add to this discussion quite a bit. I’m heavy, 220lbs. I love steel, and my favorite bike was a Basso Columbus SL set. However, it was too flexible. It had the best ride, as I’ve said, and I could even feel the energy stored pushing me forward. It was fun! Still, when I started to climb the short steep hills one finds in Southwest Missouri, the frame would flex so much the chain would rub. This equates to drag when least needed or wanted. Flex is great but needs to be managed. This is the first time I’ve ever responded to anything, and would like tell you guys at GNC, thanks for everything you do. Also, I’ve ridden all over America, and for road biking, it doesn’t get better than Missouri, especially the Ozark’s. If you are ever out this way, please get in touch. I have rides you’ll never forget.
very well said, John. I would also add to your comments that not only do I believe the riders weight can make a difference, but correspondingly their strength. Somebody who can really apply torque to the pedals can benefit as well (for similar reasons) to the stiffer frame. Whereas, someone lighter &/or who doesn't peak too high on power would actually benefit from an appropriate amount of flex. This is similar to a golf club. Someone who has a real fast/strong swing benefits greatly from a stiffer club by it's ability to keep the club head more in line with the hands & gets distance on the ball from pure swing/power. However, a person that has more of a slower/modest swing, actually won't hit the ball as far with the stiffer club than they would with a regular or softer flexing club (whichever is most appropriate for their swing/strength), by using the "whip" action of the flexing club shaft to snap the ball & create more energy on contact than the stiff club would.
Taking this to the extreme, a bike frame made of wet ramen noodle would also yield fairly poor results due to low stiffness (among other poor properties).
For every action there is a reaction, part of the energy goes to the wheel and part of it goes back to your leg, if the the frame flex too much, like a spring, it also can delay the reaction to your leg by the time you have less power, the dead zone of pedal stroke.
This is subjective. Your personal observations are meaningless - we need *REAL SCIENCE* - and everyone knows red bikes are faster.. That's just common knowledge. :D
I really enjoyed this video and (as an engineer) it makes perfect sense. What it made me think is that while there is little energy loss due to frame flex (there must be some as materials aren’t perfectly elastic and have internal friction) the difference between a stiff and flexible frame has to do with responsiveness. In other words, the stiff frame won’t flex much and will transfer the energy to the drive train right away (ie isn’t storing the energy to return later) and so the bike “jumps” when you put on a large load like a sprint. On the other hand, a flexible frame will store the energy and return it to the drive train later so the bike will feel “sluggish” when a large load is applied even if the net energy applied is almost the same. A stiffer bike will win in a sprint.
Physics! Even if you get that energy back, I would bet that some of it is used by a flexible frame. It could generate minimal amounts of heat for example. That energy then being transferred in the the air rather than the drive train. We know that when things flex they warm up at least a small amount. Therefore a stiffer frame that can't absorb as much energy doesn't conduct it to other places like heat and air. That would lead us to believe that the energy is more efficiently transferred to the drive train. Result: Stiffer frames are indeed fast, just by less than we originally thought. It all boils dow to what feels better though. If you don't like how a bike feels you won't ride it to your capacity. We look at tire size and pressure, bike weight, frame stiffness, etc. and the issue always will fall to the engine of the machine, the rider.
Michael Albany You're exactly right. Energy conversions from mechanical, electrical, whatever are never 100% efficient. A stiffer bike will lose less energy in this conversion process. Either way great point and great Gcn video.
I think that perception of how a bike feels and equating that to a rider not riding it to his/her potential is pretty much rubbish. It certainly is a narrative but it has no basis in technical fact.
Michael Albany as a general rule stiffer frames are not faster you actually experience more energy loss on a stiffer frame then you do on a frame that Flexes in the right way and at the right time. #planing
Ive own and ridden for thousands of miles both "stiffer" and "flexible" frames. Its something a rider can actually feel, if you dont feel such changes going trough your arms and legs you're dead. I always comeback to my stiffer frame, I like that when you push your pedals it reacts extremely fast and it feels nimble, i just point and shoot and there she goes. Having flexible rims and frames its not of my liking. And im not even debating between frame materials, I own both steel frames that are stiffer than other carbon bijes I have, and also stiffer carbon frames that are by far much stiffer than other steel frames i have. I enjoy both, but for everything I prefer the stiffer frames. I might not lose much energy but its a thing of feel that no experiment can take away from me :)
Quite a while back Jan Heine from Bicycle Quarterly has written about frame flex characteristics and what he calls "planing". The gist of it was that the stored energy can help you be more efficient, if the frame flexes the right amount.
Indeed, was reading his articles with alot of joy. And to my surprise I confirmed for myself that the flexy frame I have (Look kg86, first carbon frame from 1986, La Vie Claire) was faster for me than a new, stiff frame from Felt (F1).
I would also add that Jan Heine was also way early touting the performance & comfort benefits of wider tires. He's been quite the "prophet" on these topics.
I recently built a bike highly influenced by the Bicycle Quarterly crew - 650b x 42 tires, lightweight steel tubing that planes, low trail geometry, friction shift 3x10 drivetrain. It’s the funnest, most comfortable, useful bike I have ever had.
I think the chain is under tension and the frame is being puled by that tension on chain so when you let the brake of, the frame acts like a spring on the chain, puling it making the wheel turn
The most acurate test I can think of. And one I’d love to see performed. Would be to take two power meters. One pedal based - like Garmin Vector, the other hub based - like powetap. Do some runs on both bikes and compare the results.
My first custom bike in 1988 was a 27” steel frame. Unlike other Clydesdale frames I had seen, there were no stiffening stays or lugs. It flexed liked a flexy, flexy thing. It didn’t take much power to flex the BB. With rear panniers it was like riding an old Porsche 911; I once drifted through a wet roundabout with a bus right behind me! It is now permanently attached to my KICKR where it can do some good. I now ride a Trek Domane, reputed to have one of the stiffest bottoms of any frame, yet supple to ride.
Very interesting...and reminds me of a pendulum effect that once the bike flexes from side to side, that exactly provides the momentum. Very similar to when you're lifting weights, a bit of swing in your body or arms provides that little bit of inertia to make it seem easier to lift the weight compared to a dead lift. However, on the flip side, the flex might also be stretching the chain slightly that when the brake is released, the energy stored in the stretched out chain is released, making the rear wheel spin.
No research needed. The explanation is fairly simple, and it's not damping/thermal loss as a lot of people are suggesting. The bike is "returning" the power, but not at a useful time. It's holding onto the extra energy that you're trying to put into propelling you forward, and giving it back later, through the crank when you release the pressure on the cranks. A good analogy would be tying heavy weights to your feet. Any energy you put into lifting your feet is given back when you put your feet back down. But if you try and run, it's going to be a lot slower.
I’ve been looking forward to this. Yes, you’ve shown frame flex stores energy and gives it back to the drive train. Excellent, this what we want to know. What would also be nice to see is a comparison where you charge up a steep slope at a set wattage with bikes of the same weight but different flex and time the ascent and also get the rider’s impressions. Anyway, good stuff 👍.
Very interesting demonstration, but there might be a different interpretation. With the bike on the stand and the rear brakes applied, pushing the pedal down 1 cm causes frame deformation but it also stretches the portion of the chain running from the top of the chain ring to the cogs on the rear wheel (energy is stored in the stretched portion of the chain). When the brakes are released, the energy stored in the stretched chain pulls on the cog, leading to the dramatic spinning of the rear wheel that we saw. That's the simplest explanation I can think of. As for the benefits of frame flex, maybe the energy of the bottom bracket moving the other way helps lift the rider's leg up on the back part of the pedaling. Just a guess on that one...
I think the energy stored in a flexing frame is released in the form of a slightly slower crank rotation just after the powerful part of the stroke. In other words, when you come off of the most powerful part of your stroke, the frame straightening isn't going to propel you forward, it will just slightly slow the pedal stroke. So the power really is lost because it didn't get translated to forward speed while there was pressure on the pedals. If the pedals were being turned by a machine that never varied it's power at all, then the energy might be released into the wheel, but people don't work that way.
(I hope i can write properly in English....) In addition, any time the frame flex, it causes to lose the straight line between both wheels (more precisely they get away from the same plane, where they mast be), and it results in a mesh of not parallel forces (both tires tend to go in different directions) that generate a high loose of energy. Also, as both tires tend to go in different directions, there is a loose of energy due to increased friction forces between tires and road...
The biggest problem I have with this demonstration is that the load which flexes the frame peaks where his pedal was and drops off towards the bottom of the pedal stroke. This means that the frame bending back to its original position at a 6 o'clock would merely move the pedal out to the right and not cause any additional rotation or therefore put the power back into the drive system.
For me stiffness has always been about handling. I know I’m not strong enough to significantly flex a frame. But stiff bikes always feel more planted in corners and snappier coming out of corners.
We are talking about stored energy here. If rotating the wheel is like pushing an object through a spring, the stiffer the spring the harder to store energy in the spring as we push. I would rather the energy transfer through to the wheel during our force pulses rather than have a portion store in the frame and come back out through my foot on the backstroke or twisting me at the handlebars and dispersing across my body. Remember, energy stored in the frame doesn't necessarily go out through the drive chain. It can come back into us at innopertune moments.
I think it's mainly about ride feel and user preferences what you feel more comfortable riding. But probably a stiffer frame is a little more efficient, as energy you convert and reconvert always involves losses to internal energy and these losses are relative to the time dependent behaviour (and in part thus to the displacement). A stiffer frame with a higher elastic modulus probably also has a more purely elastic behaviour and less viscoelastic parts, which would mean a lower energy loss during hysteresis of the loading/ unloading cycles; also less damping in both loading and unloading due to less pronounced frame flex displacement wise. Which means although the amount of energy stored in frame flex is probably almost identical for different frames, there is more direct conversion in stiffer frames resulting in less time dependent behaviour qnd less energy losses. But I still think it's mainly rider preferences as it also involves comfort and personal compatibility next to pure efficiency.
It's about predictability. Experienced cyclists like having fine control over their power input, and a flexy frame will delay and distort your pedal stroke. It also creates a cyclic distortion to your handling which can cause speed wobbles and reduce overall controllability. the amount of energy stored by the frame in that experiment was on order 10 J. Steel tubes have pretty low hysteresis, and it's probably losing less than 1J per downstroke. Your thermal loss might be up to 3W. For this, stiffer is better assuming constant maximum force: Let the frame obey F=kx. Then W = F*x = F*F/k. This is probably accurate since your pedal stroke is probably force limited by your body, which is much less stiff than the frame.
In this experiment, they had constant pressure on the pedal and then released the brake. In reality, however, you don't always have the same pressure on the pedal if one pedal is at the bottom and the other at the top, for example, no pressure can be exerted. In this case, the energy in the frame can also be released in such a way that a pedal simply moves back a little. In the experiment, it would be like simply getting up from the pedal, then the pedal would return to its original position and the energy would be wasted. In order to get a result from such an experiment, how much energy is released from the frame back into the pedal and how much really goes into the chain, the pedal and brake would have to be relieved of the pressure at exactly the same time. Unfortunately, this experiment was not meaningful, the conditions in reality have not been met.
Yes. This experiment has multiple things wrong with it. the biggest thing would be that they pressed on the pedal and held it down. which just the frame right now is acting like a spring but because you're holding a single pedal down 99% the forces allowed to go back into the drive train. but when were peddling were alternating legs which means we're not constantly holding pressure on one side. when were actually pedaling alternating which side is relaxing and which side is applying Force. So this means that the frame is always flexing in the direction of the side with the least amount of force. so this means that the force is going in the opposite direction of the drivetrain weather be back into the frame and into your leg this equals force/power loss
I suggest you climb a hill with two bikes, with identical weight and identical aerodynamics (or as close as possible to that) - one quite flexible and one quite stiff. Keep same average power with both bikes and see if the stiff one is faster
Davide Archetti I’ve done exactly that with my Colnago C40 vs Tramac SL3 and no difference. I’ve also done varying watts with the same PM. 300 watts and 550.
It's all about timing, or as many others call "planing". When you apply a force to the pedals, there are time delays as the various components stretch. These might be only microseconds but they tend to smooth out the force over time. When you add it all together, (shoes, pedals, cranks, chain, wheels, tires, and frame) you might be looking at a lot of energy. So what does this feel like? Depends on your pedal stroke and how you apply force. If you spin super smooth, the constant force will keep everything loaded and you won't get out of time with your bike. In the steel frame old days this lead to the idea that you had to spin circles to be efficient; you kind of had to with the old flexy bikes. If you mash a flexy bike, some of your force will load the bike, stretch the chain, and wind up the spokes, and then potentially be released at the top and bottom of the pedal stroke, when your feet are no longer applying much pressure. So really its about timing, if you can match the timing of your bike to your personal style of force application, you will find that golden ride. I found this once personally with a 2011 Opus Cresendo w/Mavic Ksyrium wheels, that sweet spot where every push of the pedals matched perfectly by the timing of the bike.
Measure the temp change and then you can calculate the wattage required to cause the increase. Bingo there is the amount of watts you wasted heating up your frame .002 degrees
Great vid. Genuinely thought provoking. My head still thinks that a stiffer B.B. is better from the initial pedal stroke. However, my head believes that a noodle like floppy frame will dissipate the energy back through other areas of the bike and not back through the drivetrain.
The sweet spot is when the frame planes for you. See Bicycle Quarterly for more info. When a frame planes it means that the rider is able to put out more power and feels less fatigue in his or her legs. So it's really a physiological thing!
I could only think of the concept of ressonance to support the frame plane hypothesis. That will probably mean that frame plane will only occur at a certain power/cadence combination, which again will make it pretty useless in practical use.
Stig Aannø If it's resonance that gives the benefit then it would apply to all power outputs as long as your cadence was near the natural frequency that the frame flexes at. You would only need to worry about maintaining cadence, which is easier with more and more gears on the bike, as long as you don't go 1x.
Erik H you are of course right, it will depend on the frequency and not the amplitude. But still, the ressonance frequency will be a product of a combination of the frame and wheels. Maybe we will see people tuning their wheels through spoke tension to set the preferred ressonance frequency in the future? Well I think I myself will go for the stiff bike/wheelset combination. Look up my other post to see why I think this is the best option.
Stig Aannø Bikes can plane. That is proven. Read Bicycle Quarterly! It is very relevant in the real world and it depends on your power and weight, not on your cadence.
Max Sievers you made me read what was available without paying and now I am even more sceptical. The stiffness of the wheels will for example make up a lot of the characteristic of the stiffness of the bike. I do not see it even mentioned. They mention measuring power, but I don't see if they measure at the pedal or at the wheel hub. It is a crucial difference given the thesis they are trying to prove. They tell that this is the secret of covering ultralong distances like Paris - Brest - Paris, which make it really look like quasi science (I have covered a distance of 540km more or less non-stop on a stiff carbon frame and wheelset, my legs were fine, my ass hurt and I needed sleep). Boats plane, frames can theoretically resonate, but I do not buy that it has much to do with your weight and power (due to certain laws of physics). I think the biomecanical argument makes a compelling case for oval chainrings, but not for a wobbly frame.
The static test isn't comparable to a dynamic test for a few reasons but simply put in the dynamic test the load coming back does so through the riders legs which have moved past their optimal stiff position. Having said that and as others have said my 'feel' is that there is a sweet spot for some frame flex for rider comfort which ultimately will cause fatigue and reduce your efficiency anyway.
A flexible frame will have more losses, because that flexing creates heat. It's not a lot, but it is there. Also, flexing will effectively delay the pedal stroke, therefore hurting acceleration and makes the bike feel sluggish. There has been a lot of research into frame stiffness in motorcycle racing. The trend used to be to get the frame as stiff as possible. It only ended when the lean angles became so extreme that the suspension wasn't doing anything. So they strategically introduced flex into the chassis to act as a primitive form of suspension when leaned over. The current consensus is that flex within a certain range is good. So it seems stiffer frames definitely are faster in a straight line. But when taking road surface into account when, you might say a little flex is good. How much? It depends on the road surface and the weight of the combined rider and bike. And then there's the fact that we don't tend to corner that much... TL;DR I'm sticking with my stiffer frame.
StuntpilootStef precisely! The difference between the two frames are the time that the frame use to bounce back. A stiffer bike means less comfort on a crappy road but it'll have a really ready response compared to the flexible one. I've got a steel fixed gear for my commuting and it is super flexy but I always enjoy to ride it even to do some climbs, and still can go hell fast if I want.
Intuitively a stiffer frame makes sense in terms of power transfer, and in most cases stiffer probably is better, however there is an interesting area of study into what has been dubbed "planing" or basically if more flex can lead to better power transfer in certain cases. There definitely isn't a large depth of study in the field to prove either hypothesis, with logic being on the side of a stiffer frame, however it is a really interesting idea to explore. You should take a listen to this podcast, it is very interesting and very thorough. cyclingtips.com/2017/06/cyclingtips-podcast-does-frame-stiffness-matter/
janerney I have just read about that for the first time and that Heine also experimented with polymer bushings. He seems to find quite a difference between different bikes. The differences were there with quite subtle changes in stiffness. I do wonder how you are going to find the right one by any other means than trying out all possibilities. Because it does seem quite easy to get wrong and there are a lot of factors involved. You have the weight of the rider and the bike stiffness, but also what BB you use, maybe which tires, groupset, chain, rims, etc. It does make things bloody difficult and almost impossible to get right for anybody that's not a pro cyclist.
Yeah, i get what you are saying. It is so nuanced and there are so many variables fro scenario to scenario that it is hard to see any quantifiable real world benefits. Interesting to examine and read about, definitely, but is it something that i am going to think about when buying my next bike, definitely not.
A lot of really good comments about the the delay in flex causing a different feel. The next step would be to have a pedal stroke analyzer attached to the hub to see the different plots of force over time. I'm thinking that you would need a pedaling robot to eliminate the human variation. My guess is that the flexible frame will have a smoother response due to the damping of the frame acting as a energy storage spring in the system.
Some significant errors in the argument being followed here. The flex on the frame decreases when the force component in the original direction that caused the flex, decreases. This will be when the pedal in question is at six o'clock. Releasing the flex in the frame only when the pedal is in that position, will result in no wheel drive as seen in the experiment performed in the video. Instead, the rider's weight will counter the force, in other words, the rider's legs will absorb the energy. A stiffer frame will carry less energy to the six o'clock position, pushing more energy into driving the rear wheel. A perfectly stiff frame will carry no energy into the six o'clock position. In other words, a stiffer frame would be faster. Is the performance worth the tradeoff in comfort? Depends on the situation.
Thanks a lot for saving me the typing :) Especially when sprinting (while not being seated) this becomes visible in the 6o'clock position. To visualize the problem let's assume the frame being a perfect spring (that means, it will return all the energy it absorved). And we shall only look at the vertical part of the flexing, so just put a spring vertically on a table. To compress it, you have to apply a downwards force on top of it. If you put a weight on it, the gravity-force will compress the spring to a certain level. On the bike, the compressing force is made up from gravity and the muscles straightening the leg. Back to the spring: Action=reactio, the gravity on the weight is keeping the spring compressed, but the spring is pushing back with it's own force which increases with compression - so that the spring will only be compressed to the level where the spring's extending force equals the gravity force. When is the spring going to release energy? While extending back to it's normal size again, which can only be done, when the weight on top of it is removed (or reduced). On the cyclist that translates to lifting the leg or reducing the power with which the leg was straightened to push down the pedal, which in turn means: The frame can only flex back when the cyclists let's it push it's foot upwards again. In the end, the frame flex is not giving much energy back - you only use your energy to compress it and maybe even need some uncompress it. Anecdote: Cycling home from shopping with a heavy backpack (weight increase ~30%) I tried to sprint. I noticed that while sprinting I usually would "jump" when my foot reached about 6o'clock - with the backpack, I couldn't. I was just compressing the frame and get back to where I was before jumping. At 6o'clock. Compare that to a trampoline. Stand still on the trampoline, compress your legs and then jump. Your legs extend, but because of the springiness of the trampoline you're only raising your body by a few centimetres instead of getting about half a meter off the ground. Wasted energy.
This of course also applies to the 3o'clock position as shown in the video. But in the video, the energy was transfered back to the wheel, because the weight was taken off the pedal. When cycling, you push downwards until you reach 6o'clock - so the frame flex introduced at 3o'clock is not released at 3 but rather after 6 - where you won't have any more gains from it.
Yes to this but, I will add that if the frame at particular points then this would transfer loads into the drive train decreasing the efficiency of the drive train. A better experiment would be to increase the load on the frame and measure the efficiency of the drive train. The drive trains are very quiet so there is likely to be very little energy being lost there, but possibly in heat. Again the principles describe above would be the likely loss, transferring it into the rider by delaying the power transfer. Does anyone know how the manufactures test the frame stiffness in the dynamic sense, eg as a rider analogue?
One of my bikes frame cracked at the seat tube and down tube just above the bottom bracket. It did not break all at once. It was gradual. The first symptoms was a self shifting rear derailleur. Then as the crack got bigger, the frame flexed more and the crank felt soft on the down stroke. It felt soft since the crank was flexing inward and the result was a noticeable loss of power as some of the forces was being directed inward instead of downward. If a frame is too flexy your pedaling force may not be be optimized to forward motion. Kind of like climbing or sprinting out of the saddle and rocking the bike slightly. Rocking the bike may enable to dig in more power, but flex of frame and change of wheel angle will be a little less efficient to forward motion. You can test yourself by cutting slightly into your frame to start a small crack. One note also about the test in video, the wheel is not clamped perfectly tight and still, so that could affect results a bit. Also in real riding, that much crank flex would cause wheel to spin since brake is not holding wheel still. So amount of flex in frame would even be less. I think how frame flexes in regards to the line of travel is important too. If a frame flexes, the energy could cause the head of the bike to change line and produce more road and wind resistance, whereas as stiffer frame would not deviate as much.
It seems to me the wheel spinning is mostly from the potential energy stored in the chain, at around 4 minutes you see the chain jolt/twitch forward upon releasing the brake. IMHO for 99% of us the majority of frames are stiff enough and frame flex (on most bikes) isn't that big a deal and not another thing we have to worry about upgrading. Ride upgrades don't buy 'em.
I recently built a new road bike around a different frame model. At first it felt like I needed some time to get used to it - regarding not only the handling, but also how exactly it transfers power. Thanks for the video to explain the background!
Stiffness is great for accelerations over 1000w but not going to make much difference at a constant 200w, and you get into a better rythm with some flex.
Simon, there is an experiment you guys could perform that could possibly reveal any power loss difference between a "flexy" frame and a "stiff" frame. It would require a PowerTap hub power meter and a crank based power meter. Using the same power meters on the two different bikes, you could see what the offset was between the power reading at the crank and at the hub. Since this offset might be small and possibly not resolvable at lower power outputs, it would likely be best to do this test either in an all out sprint or in all out effort uphill. you could then compare the percent offset with respect to the crank based power, i.e. (offset - crank power)*100/crank power. If you don't maintain the same power output between tests, you might see some uncertainty bias, but that's not necessarily a game killer. You wouldn't be looking for the quantitative difference in power loss between the two bikes but rather the qualitative: is there more power loss in the flexy bike vs. the stiff bike.
With due respect, your experimental setup doesn't represent what happens in real life. You were reacting the frame torsion with the turbo trainer. On the road however, it would be reacted by side load in the tyre. The frame itself will have low hysteresis (steel especially) but the tyre certainly doesn't. So, the cyclic torsional loading of the frame does come at a cost. I would still agree that frame stiffness is a 'feel thing' to a large degree. However, that doesn't mean it's not also beneficial. It is impossible to generate load without something to react it. Imagine a super flexy crank - you'd never be able to push hard against it as it will bend out of the way. If we were perfect engines it wouldn't be a problem - we'd just load it up. But, we provide power with alternate legs and to differing loads round the cycle. So, flex in the system causes a lack of reactive load at the point of peak power generation. The load doesn't simply fall out of phase, the peak drops too. In short, static tests and analysis of an inherently dynamic system, can lead to flawed analysis. Not that I'm saying my hypothesis is correct either though...
Worth noting that tyres generate grip through 'slip' at the macro scale. Because the tyre is rotating, cyclical side loading will cause a pathline that meanders side to side - think of a little sine wave. Accordingly, the losses in such a setup are not tied solely to the lateral stiffness of the tyre and the hysteresis associated with 'non-slip' side loading (in say a non-rotating tyre). In short, the losses aren't the same physics as vertical compliance of larger section tyres and rolling resistance.
Gonna agree on most of this. The efficiency of the frame probably doesn't change a lot in the minute details from frame to frame, BUT this isn't the same to be said for handling. A stiffer frame will handle better and this is where I believe the sensation of stiffer=faster originates: simple human misinterpretation.
Sounds like you're describing a suspension fork that has semi-independent L/R travel. This would affect balance first, and loss of balance (and energy required to correct/maintain it) would lead to less power available to turn the cranks smoothly. If we consider a suspension fork or a FS mountain bike, some of that power goes to compress the suspension, but the return on the shocks (if the bike goes down after a bump) sends power back to the crank the same way this video shows (as long as the cranks are still being turned at that point). If the rider bounces while riding on flat, the energy of the return can be used to move the rider back up, and this feels like lost power, but then the rider regains potential energy as they are higher. Do they use that potential energy to power the cranks? If they go higher than their initial position, yes. If not, it was extra energy output that did *something* but did not power the drivetrain. This sort of "accidental" energy redirection is what we really want to avoid, right? I can't quite tell if this same redirection away from the drivetrain happens when flexing a rigid frame on a road bike... Any thoughts?
Besides, even if everything that happened in the test was accurate, which I agree it's not, according to the 2nd law of Thermodynamics states basically that " In any given exchange of energy, there will always be energy lost", which basically means that with all the exchanges in the experiment the losses will be bigger than with a stiff frame which will produce no such exchanges.
That was great. I had a LOOK back in early 2000. Was a fairly flexible frame. I could easily tell. I loved that bike though. It was pretty fast geometry-wise (thanks for the other video with Tom) and when I got to climbing, I could feel that effect - especially out of the saddle, which felt wonderful. The bike felt alive. There was a give on the one side and a spring on the other as you pedaled - subtle but it was there. I've always been thinking due to the stiffer is better mantra that the bike was inefficient. Really liked how it felt. Turns out, I just rebuilt it yesterday. I'll get to ride it again with this in mind, having since been on really stiff aero frames.
@@Jesus_s_Real For me, in comparing the two, higher stiffness contributes most noticeably to acceleration. In most daily riding and training, the differences are not too large in terms of average speed. When racing, I'd always pick the stiffer bike, other things being equal. There's a little more average speed to be found there, especially as in racing you are always accelerating and always riding hard. I love the feel of the less stiff bike and it's my training bike choice most days.
I think there's a lot to be thought about within this debate. If we examine a damping spring mass oscillator, that is a spring with some weight attached to the end, and we apply a force to the weight - we will find that it oscillates, but the oscillation decays due to damping of the spring. I think that a "flexy" bicycle frame will be analogous to this oscillator in that kinetic energy will be lost by restoring the frame to its original shape before applying some to the drive train. Will this loss be noticeable? Maybe not, maybe it is in the marginal gains territory.
I had a super stiff frame a long time ago, and it put a smile on my face every time I rode it. It was the immediacy of the response that I loved. You pick up the speed, the bike second guesses you and you're off. Surely it depends on where the frame is flexing. If you have a 'ladies' crossbar, the flex is extreme and only releases the front end to come back in line with everything else.
I lost a tremendous amount of energy due flexing when my steel single-speed mtb flexed in a corner while mashing the inside corner pedal. Causing me to drop the chain and the sudden drop of the pedal sent me flying to the ground.
In this experiment the rear wheel is fixed and when the pedal is loaded the distance between the rear wheel axle and the bottom bracket axle is increased so the chain is prolongs and when the pedal is released the chain backs to its normal size and the wheel start to spin.In the flexible frame the distance between the axles under the load changes not so much( the rear triangl bends) so after the load is gone the wheel will spin not so much respectively as with the stiff frame.So the stiff frame has a better efficiency.
So glad you prompted this discussion. In my opinion, frame stiffness is a red herring. Anecdotally, Sean Kelly won a hell of a lot of sprints on a Vitus 979, likely the least stiff frame ever sold. A stiff frame "feels faster" but it's not. The materials bikes are made out of make excellent springs and all of the energy is returned. A slightly flexy frame is also much more comfortable.
imo, in line with the mechanic, stiff frame will feel faster as the flex is lesser so it feels more responsive as you load those drive train. whereas less stiff bike will feel sluggish or lost of power as there is a time delay for the transfer of energy. ideally, you want responsive bike to attack at the right moment so stiffer bike is still the better option. lastly, the flex of bike frame is commonly compensated with the slight twist of handlebar to maintain balance. that may result in a utter different finding.
Theoretically though, according to the video, torque should be returned before the end of the pedal stroke, so power will be returned before 1/2 of a revolution on the pedal. Assuming pedalling at around 80-90 rpm, thats only about 1/3 of a second. I personally don't think (at least at most people's riding level) that will make too much of a difference.
Feeling faster ≠ actually faster Case in point: bigger tires at lower pressures feel slow but are measurably faster while skinny tires feel faster but are comparably slower.
The handlebar twisting as people pedal isn't just from frame flex, but also from: 1-The rider putting uneven hand force on the bars through their pedal stroke. 2-The response of the steering geometry (i.e. trail) to the rider rocking the bike.
You are on the right track. The faction of the 1/3 of a second of your muscle exerting effort while the spring loads and unloads adds up as it happens on every revolution. Think of a stiff frame as your muscle working for a slightly shorter duration every pedal stroke to deliver the same power. If the frame were infinitely stiff, the muscle would have shortest possible duration of effort. I agree people will feel the stiffness, but that little extra effort will also add up. My guess is this can be significant.
I have a stiff bike and a not stiff bike and I’ve ridden both with the same wheels and tires and it’s like running on sand vs running on concrete. I don’t feel power loss in the stiff bike and it’s very responsive and snappy. It’s like making a chain from rubber, it will store that energy like that but by the time you reach your dead spot in the pedal stroke the energy can’t be used to produce forward motion.
I personally am more interested in frame flex during cornering, if it helps the tire track the road better like a rear suspension in mtb or does the change in angle of wheel relative to frame might cause instability like a flexing chassis in a non performance oriented car.
Adrian L Good point.. Usually, espevially with skinny tires, corner performance is underestimated. My empirical evidence suggests stiffer frames give me more confidence in cornering.
Adrian L that is a very keen observation because that's exactly what the Grand Prix in world superbike teams are working on. More torsional less lateral stiffness
I felt instability over the comfort in fast cornering sections with my former steel xc mtb. I’d rather feel the ground below, not a flexi bike frame in between, trying to do some mediation. But if you’re not doing a sports ride, this feature can be comfortable. But than again, sitting on the couch in front TV is even more comfortable.
In my opinion, they both transfer a similar amount of energy. The biggest difference is the reaction time. Stiffer bikes react faster to load switching between handlebar and two pedals (for road bikes). On the other hand, the less stiffer bike reacts slower but is good at absorbing external forces ( for cycle cross or mountain bike).
This would have been a much more accurate test if you had both been wearing your science glasses.. What were you thinking Si.. attempting science without the correct tool.. shame..
I thought frame manufacturers were aware of this and incorporate specific amounts of flex into their designs. I was reading that a prototype Specialized Demo bike was seen with varying thicknesses of carbon which had been crudely adjusted during testing. The explanation was that they were tuning the amount and type of flex. And yes, perhaps akin to tyre pressure, a stiffer harsher ride could fatigue the rider sooner, negating any power gains that may have been achieved.
The experiment has one condition: the leg doesn’t move when the frame springs back. On a road that won’t be true. Basically, the frame’s flex will work towards lifting the leg rather than moving the bike forward.
The feel is a timing or tuning issue. Just like string tension in a tennis racket or golf driver shaft stiffness (pros will use higher/stiffer here too), the frame stiffness affects *when* the stored energy is released back during your pedal stroke. If it comes back at the wrong time the bike will feel slower, even though you are losing very little energy regardless of the frame stiffness. Roughly the higher your cadence the stiffer, and hence faster spring back, you are likely to prefer. Of course there is both mass and geometry of the bike involved in its dynamic response, so it won't be that easy to correlate to just the stiffness. It probably varies with rider strength too.
I suspect this translates to acceleration. I.e. higher frame stiffness = better acceleration. Very small numbers in difference of course. Moreover, a frame with very low stiffness could potentially "lose" energy by the fact that the frame never seems to return to the initial position, this part is more a though experiment since no frames are that bad (the energy would most likely go into heat at this point by bending of the frame). But in my experience the biggest difference i noticed when moving from an old alloy frame to a TCR (acclaimed of very high stiffness), was the acceleration. And with better/sharper acceleration i felt that my momentum didn't suffer as much when the gradient changed, i.e. i was able to keep the speed higher.
I think you nailed it on the head. It’s not necessarily more stiffness that make a bike “faster” because even super stiff frames still flex. I feel it’s stiffness in certain areas of the frames that brings greater acceleration when cornering for example. As a surfer, I can see the similarity with surfboards and especially surfboard fins. Super stiff boards or fins don’t accelerate well in corners, but the “foil” or actually profile of either one will create better performance, having stiffness at high speeds, but also great acceleration in tight corners when making sharp turns. I think this is going to evolve as technology and testing reveals more proof of efficient carbon profiles.
I think the flexing frame moves some of the energy to the dead spot. Stiff frame feels fast because you have almost all the power going in the same space of crank revolution. That might not make much of a difference. Unless the the frame acts as a damper and eats the energy. But flexy frame has a lot to contribute to rider comfort.
what about doing a road test with the two frames both mounted with pedal and hub based power meters. if frame flex matters for actual speed the difference between power meters should be larger on the more flexible bike.
Exactly, I proposed this very test in an earlier GCN video wherein flexy was deemed to be energy lossy. Hell, use power pedals/crank arms/bottom bracket/rear hub meters and test with seated versus standing and so forth. Flat road, up hills, around corners, the whole schmear. Light rider, heavy rider, amateur rider, pro rider. Swap meters to other bike, calibrate, repeat. Science, in other words, not conjecture and thought experiments.
Si, you've just restated exactly what I'm bitching about. 'I'm not sure' 'I think you'd probably need'... How about measuring for meaningful data, publishing the results, and commenting on that? If you can't measure a power difference from noodly frame to uber stiff wonder bike, then it doesn't exist in a meaningful way, and that's important. The 'method' employed in this video is dubious, as many have pointed out. We have actual tools: please use them. Thanks.
Okay then, let me rephrase. There would be absolutely no point in doing the test you suggest as the tools to measure meaningful data don’t exist. The levels of accuracy in current power meters would leave vast statistical errors.
Therefore, you're declaring there's no meaningful difference between frames in terms of the rider's energy reaching the rear wheel and moving the bike forward. Thanks for clarifying your position.
I think that after a frame flexes under load, when less load is being applied (top dead centre?) and the built-up energy is being released that most of that released energy is transfered back to the crank and pedal where it slows down pedal rotation marginally. This would be the flow of least resistance of the energy and if this hypothesis is correct then a flexy frame would indeed be slower than a stiff frame. The best way to test the idea of flexy vs stiff frame may be to use 2 identically equipped bikes that weigh the same up a very steep gradient.
Two things: First, no form of energy transfer can be 100% efficient. While most of the force that went into flexing the frame will be returned to the drive train, some will end up getting lost. Secondly, it isn't the difference in speed that we feel going from a flexible to a stiff bike. It is the difference in responsiveness. When we apply force to the pedals of a flexible bike, that initially goes into flexing the frame instead of moving you foreword. As a result, a stiffer bike will be quicker to jump off the start line as apposed to a flexible one.
also worth noting the importance of stiffness in the front end as well. You want the bike to steer as soon as u point it right or left as oppose to just flex and then bounce back to steer whenever the frame feels like it... which at the end of the say will also make you slower as you corner with less precision and confidence
Nice to see an attempt at showing the principals involved here. I know that its going to be hard to give a perfect experiment when your not in a lab and can test for all sorts of variables. Though it at least shows that perhaps due to conservation of energy we shouldn't be so concerned that if there is some flex in your frame then you are just wasting all of your energy just to make a frame flex instead of making you go forward which is what your looking for. I am sure that some more more testing in a lab would help a little more but good one to get the ball rolling! Nice to also show you don't need to chuck out that steel/aluminium frame because it will flex.
The same thing happens, just mirrored. Therefore, it was a bit odd that he said the effective "clockwise" motion of the BB - on the non drive side it will be an effective counter clockwise motion of the BB that gives you the little kick when the frame returns into its normal position.
This is exactly what i thought. He's testing the "clockwise" side of the bike. (Maybe due to the setup being against the wall and inherently wanting to see the drivetrain during tesing). On the other pedal the return kick would produce anti-clockwise energy?? Would this cancel out the clockwise energy produced in the previous half stroke [hence his IS Wasted energy] ?? Would this if tested similarly, force the drivetrain backwards and no wheel spin created, since it would just spin the freehunb? Or would it still "create" energy since no matter which direction clockwise or anti, still be energy and movement thta the bike has, and that is used for movement...? (I ride a reasonable flexy cheap Allez so would love to know!)
The clockwise/anti-clockwise energy is at the level of the crank, but is applied in different points of time. All is transferred into "clockwise" (forward driving) energy at the level of the drivetrain. It's just a perspective thing. By the way, this doesn't imply the theory of flex-restored-energy is correct, just expanding the theory to the non drive side.
I would say that you see the same amount of flex on both sides. There may be a small variation as they are mechanicaly asymetrical, this may be insignificant, but pro teams with pedal stroke analysis may say different. On a more flexible frame you would see a pendulum effect. As you are on the upstroke of the right pedal the frame flex is unloading to the right, while also being flexed to the right by the downstroke of the left pedal. As the left pedal is on the upstroke the frame flex is unloading to the left, while also being flexed to the left by the downstroke of the right pedal. Even the stiffer frame seemed to store energy though. A stiffer frame may still have same effect, perhaps to a lesser degree, just that the flex isn't so noticably focused around the bottom bracket and spread around the whole of the seat tube and down tube flexing to a lesser degree.
I once had a go at calculating the lost power due to an inch of sideways flex at the bottom bracket. It came out as about half a watt if I recall correctly.
There is a net loss in heat generated by flex, though small, its not insignificant, and also the stored energy released by unloading the frame comes at a different moment in time to the power stroke, so the effect will be (on a bike that flexes more) a more sluggish response, though a more compliant feel.
This experiment is fundamentally flawed and does not answer the question. The flaw: when you apply load to the pedal and the frame flexes, the pedal becomes immobilized between the block and force applied to the pedal. When you release the brake, the frame moves back to center while the pedal is still immobilized; however, as the frame moves while the pedal doesn't, the crank moves. This occurs because the crank axles moves upward with the frame while the pedal remains stationary; thus the crank turns. All you have done is pre-loaded energy into the system and forced it to the rear wheel. If you move the pedal to Bottom-Dead-Center and repeated the experiment, the results would differ because the rear wheel would not move. Most likely frame stiffness contributes little to power gain or loss due the dynamic nature of pedaling. Frame stiffness is more likely to contribute to handling.
Simon has described two possible outcomes. 1. Is the experiment incorrect and stiffness does cost Watts. The solution would therefore be a stiffer bike reducing Watts lost. 2. Energy is transferred to potential in the frame, which in turn gets converted to kinetic energy in the drive train. The problem with 2. is that no system is ideal. There will still be normal resistive losses in the drive train but added to that will be efficiency losses in the flex part of the experiment system. The solution would therefore be a stiffer bike reducing Watts lost. Ultimately though, the Watt losses are tiny and insignificant. And most likely its only the riders in the top 1% of the sport can see benefits. Amatures will affected more by other variables fluctuating (for example, weight or fuel choice) and probably shouldn't worry about minor stiffness improvements.
I have been thinking about this a lot. I agree with this to a point. Most of the energy goes back into your cranks of course, but you load it in a different direction than your pedal stroke is positioned when your legs ease off the power. So your crankset gets it's energy back in the wrong direction, not adding to your speed at all.. I think.
Does the wheel move when this test is done on the non-drive side (left)? I'd argue it doesn't because if you put the force on the right side the flex in the frame would put some tension on the chain making the wheel move when released while putting the force on the other side would actually slacken the chain. So.... what's the deal?
Johnny Cab I disagree in part. The experiment does exaggerate the amount of flex in the frame, but the bottom bracket does move, and that movement would provide force to the chain (or slack when applied to the pedal opposite the drive side).
Flexible frames obey Hooke's Law. When you apply a force to a frame (push down on a pedal) the frame will be displaced, as F=-kx. When you release that force (at the bottom of the pedal stroke), the elastic potential energy, due to the displacement (U=1/2*k*x^2) is released. That energy just goes back into the drivetrain due to energy conservation laws (ignoring the very minor efficiency losses). This is the first time I've seen it demonstrated empirically, and it's a very good demonstration, kudos to Tom Sturdy - never thought about it like that - great video! My only minor dispute would be the focus on the vertical movement, as it's simply a component of the overall displacement. Assuming 100% mechanical efficiency, even the energy lost via the work done to displace the frame horizontally is recovered as the frame flexes back to its starting point, therefore, Work in = Work out so the net change (loss) is zero.
Loss of energy is not zero, there is little loss maybe in the frame but the tranfer of it through all the transmission components is not negligible, and above all there is the tire. Whatever it is the compound the gum suffer form hysteresis which probably wastes all the energy returned from the frame flex that managed to get there. There is also the problem of fatigue, when you repeatedly put the frame under load within the elastic range the material is weakened by a tiny bit due to imperfections. If a frame flexes too much that could shorten the life of it.
So the brake goes on, then the pedal moves down which puts tension on the chain (and flexes the frame as the chain can't move). Of course the wheel will move when you release the brake, you've essentially just pulled the chain a bit. Yes, the frame 'absorbs' the tension and releases it but I don't think that says much about frame stiffness.
Love this vid., and some of the comments below are excellent. Perhaps a refinement of the set up to measure the heat loss/kinetic energy loss. That said, as a guy who recently ended his hard tail hold out status, my new full suspension mountain bike yielded me numerous p/r's on trails I've been measuring on Strava for years. Lot's of heat loss through making shock oil get warm yet the net result was quicker times on all sorts of trails, not just downhill segments. And lastly, fatigue is another component...even on the road, a comfortable (flexy?) bike may prove slower on short, high output situations, but end up being quicker during long days on a bumpy road in spite of the heat loss of bending and releasing the frame tension thousands of times.
I personally think there’s something to this, after three quite high quality frames I moved to a mid priced modern steel frame (Ritchey Logic) and although I’m not some guy who is all old school the frame has a lovely springy and lively feel, despite being 1kg heavier and undoubtedly nowhere near as stiff as the carbon frames, over and over I am quicker on this bike, climbing, riding hard on the flat or to an extent sprinting.
Great video, no real errors in physics I can see, BUT... 2 key aspects are not mentioned. 1. One needs to calculate or measure the mechanical energy lost to heat as the frame is flexed. Your experiment assumes the energy is all stored in the frame material like a spring and all released. Any real material will have dissipation (think of it as internal friction). That energy is gone forever into heat in the environment. Maybe the loss is small, maybe not. WIll depend very much on the material (steel, carbon, Al, Ti). Unfortunately, this loss is VERY hard to measure! 2. One needs a DYNAMIC measurement in addition to a STATIC force measurement. The phasing of how the energy is returned will vary with cadence (frequency of spring). I would be that Al with the same static flex is going to absorb energy different than carbon composite flex while at 100 rpm! In the end, I think this non-linearity is important. Stiffness is not everything. Peter Sagan needs a much stiffer frame for the same effect than I do :-) Still a great video! Now Si needs to write a PhD thesis on it...
Two thoughts: 1.) The resonance frequency of an body depends on its stiffness to weight ratio and resonance phenomena are related to energy absorption. Thefore if the excitation frequency is in the range of the resonance of the frame, the energy "loss" could be higher. 2.) If the deformations of the frame are leaving the linear elastic region of the material, the osscilating stresses could lead to a hystersis loop. This could also mean energy absorbtion.
There is also the matter that more sideways movement has an effect on the rider. If the frame flexes it will cause your body to move slightly to the side. This can be compensated, by pulling on the bars and/or using core muscles to balance. Either way, these muscles are using energy which could be going to your legs. An exaggerated way to exemplify this is how you feel after a road ride compared to a mountain bike ride. For the former, just your legs are tired, for the latter, your whole body is tired from moving around on the bike and levering it into position. I imagine the difference between modern road frames is pretty marginal though. Personally I always ride a steel, moderately flexible frame on road because if I used a super stiff carbon frame on the roads where I live, I would have no spine left after a few weeks.
Heres the answer for you. So the frame flex shifts the position of the bottom bracket on the downstroke. But the pressure gets released on the upstroke, thereby putting the energy into reversed pedaling. Thats a clear loss right there. There is no gain or recycling of energy happening as explained in this video. Im surprised this wasnt discovered earlier, and that this idea reached so far. If anything, this experiment shows that a stiffer frame is indeed faster. Sadly nobody will read my comment anymore... Thumbs up please.
You either dont have a clear understanding of whats going on here(the physics), or you didnt understand the video.. They literally diagram out the opposite of what you are saying.. The BB gets flexed(deflected downward), while the pedal stays in place.. When the system unloads the BB moves back up into its unloaded position while the pedal stays in place. This causes a FORWARD motion in the pedal stroke.. not a backwards one.. Thats is what is returning energy into the system.
Golf clubs are engineered for various amounts of flex in the shaft so that the quantity of energy storage and the timing of its release maximize clubhead speed at impact. A stiffer bike stores more energy for the same amount of flex (E=0.5Kx^2, K is the modulus and x is the deflection), but it also returns the energy faster. Perhaps if flex were engineered (in the crank arms?), the release would occur during the dead spots in the stroke, maximizing acceleration. It would have a negligible effect at steady state, though. In the meantime, a stiff frame feels better and accelerates faster, so all else being equal, is better.
I put my old aluminum TCR and new carbon TCR on the trainer and found huge differences on frame deformation. The softer frame killed me on my KOM event.
Well, bike designers care. And people buying custom bikes care. And people buying bikes off the rack care. But you don’t have to care, because the caring has been done for you.
The energy that move the rear wheel came from stretched chain. The flexed Bottom bracket actually moved the big chain ring forward and stretched the chain in which pulled the rear cassette but wheel movement was stopped by the brake until it released. If you move the front chain ring to small one and rear cassette to the big sprocket then the energy transferred will be way less. In my opinion, flexible frame is not ideal for sprinting which require instant energy transfer.
I would assume a rider would be using energy to counteract the frame and keep the bike steady. Incorporating more stomach and back muscles this would increase rider temperature showing where energy is being converted. Another factor could be aerodynamics tiny flexes in every revolution would cause the rider and bike to cut through the air less efficiently. As a straight line from A to B is always shorter a 2cm flex at the seat would be exaggerated at the riders head, meaning power is used moving more air out the way. Really interesting content once again GCN, keep it coming guys.
Excellent demonstration of the frame flex returning stored power to the drivetrain. This definitely helps smoothen the power delivery. People have mentioned that there will be a loss of energy transfer, but a springy steel frame should have very little loss when returning the energy. Think of it as a steel spring and how it is pretty efficient. I would rather take a slightly flexy frame that works with my pedal strokes over a stiff one that bolts forward instantaneously. PS: i a not a sprinter or a racer.
So, I would like to add when the break was released, the resistance from the wheel from spinning is less than the resistive forces of the frame. However when at top output your application of power matches the resistive forces to moving forward. therefore when at a top sprint you will only lose power rather than getting it back. It is illustrated by the brake in the example in the video, where the force is only released back when the resistance drops.
One thing that was completely neglected is that when the frame yields under load gears and chain are no more aligned properly and this can be draining energy straight into friction and heat. Actually, this may be even more of an issue for belt drive systems where alignment is very important. Mine starts to make noise when I push harder and I have a suspicion that it may be because of the frame flex. Rear triangle tubes are quite thin on my bike.
In my opinion, the energy applied on a stiffer frame is transmitted "directly" to the drive train. In a flexing frame, the energy used to flex the frame (let's discard the fraction that is actually transformed into heat at molecular level) is stored and will "spring" back, BUT the spring action will happen when there is no input from the rider (when the cranks are nearly vertical). At this point, given the bike in moving, the small amount of energy that the frame stored would not have the same effect (would not actually push the weight of the bike + rider) as it would "together" with the rest of the energy on a stiffer frame, all because of the way pedaling happens. I might be wrong tho. In any case, a static test is leaving out many factors in play in a dynamic situation.
Simplistically, the frame etc. act as a spring. It stores energy as you flex it and releases the energy when un-loaded. As long as there are no losses during this cycle, it doesn't hurt. Less simplistically, there is likely some damping in the system which will result in small losses due to friction and such so you don't get back quite what you put into it. Would be interesting to expand this test to accurately measure the energy input and output. Good start and thought provoking!
Ralf Hoenhe had already brought up what I just posted about the energy being stored in the stretched chain in this demonstration (I didn't read all the comments before posting). To see if there's any merit to this idea, they could do the experiment with an old worn-out chain and see if there's any visible stretch in the chain. My guess is that they used a new chain that would only deform very slightly and not be visibly stretched. More energy would be stored in a new chain being deformed by a 1 cm pedal movement (because more force would be needed). An old compliant chain would be more easily deformed by the 1 cm pedal movement and wouldn't store as much energy.
Just to throw another confounding variable into the mix, recent research on the physiology of fatigue in endurance sports such as cycling are pointing to fatigue being largely a mental construct (i.e., as your muscles fatigue you need to recruit more muscle fibers to complete a given task and that it this extra mental effort to recruit additional muscle fibers that your brain interprets as fatigue). This implies small differences in mechanical efficiency between frames could also have a multiplicative interaction with our perception of fatigue and therefore realized performance. As a quick test, started swapping a power meter across three frames that I felt rode either lively or dead. Wheels and body position were the same. I rode to perceived exertion levels without power readings, then compared the average wattage. I tended to put out about 57 watts more (95% Confidence Interval: 39.0 watts, 76.4 watts) than the bikes that did not feel as lively. The power meter was crank based so any energy return from the frame would not have been included. I also compared average heart rate, which tracked the higher wattage suggesting higher effort despite perceived exertion being similar. My interpretation, was that it was mentally "easier" to ride the bike that I felt pedaled more efficiently so my focus was better and my muscle recruitment was better. There of course may be energy return from the frame, but I think how this interacts with the mental component could be where the real big returns come from. Anyway, to do this experiment properly, the rider would need to be blind to the bikes (e.g., built up with different tubing or layups but visually identical)
Tall guy who just got his first carbon frame after riding on an old steel frame. I ride a 62 cm bike, I wonder if the length of the seat tube has any effect on the amount of flex. One of my more seasoned cyclist friends suggested to me that longer tubes flex more, which seems to make sense. 🤔 I know that my new carbon bike certainly feels faster, but I’m sure there are many factors at play other than just frame stiffness. Very interesting vid.
The reason for the wheel turning was due to loading. The loading is possible because of flex in the frame and other bike parts. The flex in this demo isn't exact due to the brakes being engaged which meant force is put onto the break and the part of the frame it connects with as oppose to the hub to spoke to tyre.Flex rebound will be absorbed by your body and tyres, your body may be able to reutilise this energy. The real test is to have a flexible bike vs stiff bike and have the same wattage applied and see which goes further. the chain length between the sprocket and chain ring shortens on a flexible bike which means it will be harder to apply high wattages due to the give? Anyway the energy won't be resonating due to tyres.
Stiffer frames channels your power input directly into the drivetrain, with minimal elaatic energy being stored when you pedal. Flexible frames channels a bit less energy directly into the drivetrain immediately, but gives it back with the elastic return later on. So stiffer frames are essentially more responsive and crisper in feeling whereas flexible frames may feel like its blunting the power input slightly, but the frame can feel like its moving with you. My old aluminum Raleigh was more flexible thanmy Cervelo, and I can feel that I am moving more relatively to the Cervelo's frame because its much stiffer, but it feels much crisper and more responsive, so if I put power down, I instantly feel the bike reacting and going faster. Id say its a bit like the frame geometry argunent - slacker geometry gives more comfort and is more forgiving, but more aggressive frames are more twitchy, more responsive and less forgiving...
Back in 2015 when I was tossing up whether to get a new bespoke steel bike or just an off-the-rack Carbon Fibre bike. At the time I was reading claims from one or two steel bike builders about the benefit of steel frame flex/power transfer. Googled this matter but found nothing else on this issue. Consequently bought my plastic Giant Propel. No regrets purchasing CF. I love the feel of the ride. One thing I am absolutely sure about is CF has much better power transfer. On steel, I literally had to fight to stay with on fast club rides after every traffic stop. (Hated it but it really made me much stronger). CF is so easy to take off with. I am not an Engineer and so dont know the reason, but I can say whatever flexie power transfer a steel bike provides, it is not as efficient as simply having a stiffer CF bike. My guess is that you probably lose more power from the flexing process than what you get back.
heat is lost, also the dead spot in the pedal stroke may absorb the energy, with less or no chain tension i dont think energy will come back to help you go
Sean Kelly won paris nice 7 times and a zillion other races. He won most of his races on a Vitus bonded aluminum bike that was a flexy noodle. Damon Rinard, top cervelo engineer and now at Cannondale says that they don't design the supersix for ultimate stiffness but ideal stiffness.
That way I look at it, there are two related dynamics to consider. Stiffness/amount of flex and flex/ rate of spring to true. It could be that this is a dynamic that can engineered and likely already is with companies that have the budgets for 5hat kind of research. Wanting a stiff bike that allows a combination of transference of power and flex to allow spring to occur and engineering a high speed of recovery springing back to True. For custom builds this could be beneficial for weight and average load could be measured and taken into consideration when calculating rate of stiffness and spring. I think many people confuse actual squat with flex. Meaning, sure too much flex robs you of your power when applying force and the whole system is loaded, but lacking flex at all you Rob yourself of your excess energy that is transferred and returned in the form of spring. Interesting video highlighting that.
Well it's just like a guitar string being struck. If the bridge is not stiff, if the top is flexy, it will dampen the vibration of the string, so the sustain will be shorter and it will sound less bright/dull. For a bike it's kind of irrelevant because you don't expect your pedal stroke to be sent back at you. The effect that could be seen is conservation of momentum. It seems to me that most of it is lost through friction both with the tire rolling resistance and the freehub, unless you're riding a fixie. I'd be ready to sacrifice a little bit of efficiency for more compliance and comfort. However when a frame bounces from left to right when you're putting a lot of power you also lose time because the rear wheel is not travelling as much in a straight line. So while power is conserved it's less used to propel you forward. In any case I've won a few town sign sprints with my old steel bike during group rides, so the effect is probably negligible most of the time.
This needs a test: power meter on the cranks and power meter in the rear hub on a flexy and on a stiff frame to see if there's any difference in the power delivered to the cranks vs. that delivered to the rear wheel.
Use the same crankset, and wheels.
Test with a sprinter and a climber to see if rider weight has an effect.
Maybe throw in a flat (no climbs to minimize any effects from weight difference) time-trial with a rider putting out a steady power output on the crank-based power meter to spice things up.
There is indeed power lost between the cranks and the rear hub. Most of it in the chain. Stiff is better because it's more stable. Not because you actually lose energy by loading a spring. You lose a lot more energy in places that you don't even look. If you're getting chain or brake rub that's an entirely different story. If you're not getting obvious performance problems from the flex you should go search for "lost watts" in the tires and the chain. Once you're close to optimizing that (for what you can afford) then go hunting for "lost watts" in your bearings. You lose a lot more power in a dirty chain than you do in a "flexy" frame (as long as you're not causing friction from bad alignment). When you lose energy in the chain you can not get it back. As illustrated in the video, most of the energy used to load the spring (frame) is returned to you and the actual values are trivial.
@@indonesiaamerica7050 doesn't this suggest that a smoother ride (which could come from a softer frame) is more stable? Also, that clutch derailleurs could help too
the test only shows how much those frames store energy like a spring would it has nothing to do with efficient transfer of energy that would make a bike fast. a bike frame cant be stiff enough. most comfort comes from seat post deflection and tyres anyway.
Hello, from an old and experienced rider. I believe I can add to this discussion quite a bit. I’m heavy, 220lbs. I love steel, and my favorite bike was a Basso Columbus SL set. However, it was too flexible. It had the best ride, as I’ve said, and I could even feel the energy stored pushing me forward. It was fun! Still, when I started to climb the short steep hills one finds in Southwest Missouri, the frame would flex so much the chain would rub. This equates to drag when least needed or wanted. Flex is great but needs to be managed. This is the first time I’ve ever responded to anything, and would like tell you guys at GNC, thanks for everything you do. Also, I’ve ridden all over America, and for road biking, it doesn’t get better than Missouri, especially the Ozark’s. If you are ever out this way, please get in touch. I have rides you’ll never forget.
very well said, John. I would also add to your comments that not only do I believe the riders weight can make a difference, but correspondingly their strength. Somebody who can really apply torque to the pedals can benefit as well (for similar reasons) to the stiffer frame. Whereas, someone lighter &/or who doesn't peak too high on power would actually benefit from an appropriate amount of flex.
This is similar to a golf club. Someone who has a real fast/strong swing benefits greatly from a stiffer club by it's ability to keep the club head more in line with the hands & gets distance on the ball from pure swing/power. However, a person that has more of a slower/modest swing, actually won't hit the ball as far with the stiffer club than they would with a regular or softer flexing club (whichever is most appropriate for their swing/strength), by using the "whip" action of the flexing club shaft to snap the ball & create more energy on contact than the stiff club would.
Taking this to the extreme, a bike frame made of wet ramen noodle would also yield fairly poor results due to low stiffness (among other poor properties).
Are you stiffer on a faster bike?
That's a very personal question. But I'd say it depends on time of day. I'm definitely stiffer in the morning.
definitely
I'm always stiff for a fast bike #bikegasm
Is that the same stiffness I am thinking about?
Oh No, you didn't. 😁
For every action there is a reaction, part of the energy goes to the wheel and part of it goes back to your leg, if the the frame flex too much, like a spring, it also can delay the reaction to your leg by the time you have less power, the dead zone of pedal stroke.
Let's do some *REAL* science. Figure out why red bikes are always faster!
Jim Hansen There are actually already multiple ok studies about the color red i sports :P
they aren't ....Trek Segafrdo
And my Focus is red and it's slower than my Bianchi which is Celeste
That is known as the Bianchi price effect.
This is subjective. Your personal observations are meaningless - we need *REAL SCIENCE* - and everyone knows red bikes are faster.. That's just common knowledge. :D
Ha! I wasn't aware of this scientific "fact." Good to know as my bike is red!
I really enjoyed this video and (as an engineer) it makes perfect sense. What it made me think is that while there is little energy loss due to frame flex (there must be some as materials aren’t perfectly elastic and have internal friction) the difference between a stiff and flexible frame has to do with responsiveness. In other words, the stiff frame won’t flex much and will transfer the energy to the drive train right away (ie isn’t storing the energy to return later) and so the bike “jumps” when you put on a large load like a sprint. On the other hand, a flexible frame will store the energy and return it to the drive train later so the bike will feel “sluggish” when a large load is applied even if the net energy applied is almost the same. A stiffer bike will win in a sprint.
Physics! Even if you get that energy back, I would bet that some of it is used by a flexible frame. It could generate minimal amounts of heat for example. That energy then being transferred in the the air rather than the drive train. We know that when things flex they warm up at least a small amount. Therefore a stiffer frame that can't absorb as much energy doesn't conduct it to other places like heat and air. That would lead us to believe that the energy is more efficiently transferred to the drive train. Result: Stiffer frames are indeed fast, just by less than we originally thought.
It all boils dow to what feels better though. If you don't like how a bike feels you won't ride it to your capacity.
We look at tire size and pressure, bike weight, frame stiffness, etc. and the issue always will fall to the engine of the machine, the rider.
Michael Albany You're exactly right. Energy conversions from mechanical, electrical, whatever are never 100% efficient. A stiffer bike will lose less energy in this conversion process. Either way great point and great Gcn video.
Yes all the minute losses will add up over the course of a ride
I think that perception of how a bike feels and equating that to a rider not riding it to his/her potential is pretty much rubbish. It certainly is a narrative but it has no basis in technical fact.
Michael Albany as a general rule stiffer frames are not faster you actually experience more energy loss on a stiffer frame then you do on a frame that Flexes in the right way and at the right time. #planing
Michael Albany
Yes. How much entropic loss.
Ive own and ridden for thousands of miles both "stiffer" and "flexible" frames. Its something a rider can actually feel, if you dont feel such changes going trough your arms and legs you're dead. I always comeback to my stiffer frame, I like that when you push your pedals it reacts extremely fast and it feels nimble, i just point and shoot and there she goes. Having flexible rims and frames its not of my liking. And im not even debating between frame materials, I own both steel frames that are stiffer than other carbon bijes I have, and also stiffer carbon frames that are by far much stiffer than other steel frames i have. I enjoy both, but for everything I prefer the stiffer frames. I might not lose much energy but its a thing of feel that no experiment can take away from me :)
Quite a while back Jan Heine from Bicycle Quarterly has written about frame flex characteristics and what he calls "planing". The gist of it was that the stored energy can help you be more efficient, if the frame flexes the right amount.
Thanks for the tip! I love reading stuff like this.
Then you might want to have a look at his most recent blog entry on that topic: janheine.wordpress.com/2018/01/28/myth-4-stiffer-frames-are-faster/
Indeed, was reading his articles with alot of joy. And to my surprise I confirmed for myself that the flexy frame I have (Look kg86, first carbon frame from 1986, La Vie Claire) was faster for me than a new, stiff frame from Felt (F1).
I would also add that Jan Heine was also way early touting the performance & comfort benefits of wider tires. He's been quite the "prophet" on these topics.
I recently built a bike highly influenced by the Bicycle Quarterly crew - 650b x 42 tires, lightweight steel tubing that planes, low trail geometry, friction shift 3x10 drivetrain. It’s the funnest, most comfortable, useful bike I have ever had.
I think the chain is under tension and the frame is being puled by that tension on chain so when you let the brake of, the frame acts like a spring on the chain, puling it making the wheel turn
The most acurate test I can think of. And one I’d love to see performed. Would be to take two power meters. One pedal based - like Garmin Vector, the other hub based - like powetap. Do some runs on both bikes and compare the results.
My first custom bike in 1988 was a 27” steel frame. Unlike other Clydesdale frames I had seen, there were no stiffening stays or lugs. It flexed liked a flexy, flexy thing. It didn’t take much power to flex the BB. With rear panniers it was like riding an old Porsche 911; I once drifted through a wet roundabout with a bus right behind me! It is now permanently attached to my KICKR where it can do some good.
I now ride a Trek Domane, reputed to have one of the stiffest bottoms of any frame, yet supple to ride.
Very interesting...and reminds me of a pendulum effect that once the bike flexes from side to side, that exactly provides the momentum. Very similar to when you're lifting weights, a bit of swing in your body or arms provides that little bit of inertia to make it seem easier to lift the weight compared to a dead lift.
However, on the flip side, the flex might also be stretching the chain slightly that when the brake is released, the energy stored in the stretched out chain is released, making the rear wheel spin.
No research needed. The explanation is fairly simple, and it's not damping/thermal loss as a lot of people are suggesting.
The bike is "returning" the power, but not at a useful time. It's holding onto the extra energy that you're trying to put into propelling you forward, and giving it back later, through the crank when you release the pressure on the cranks.
A good analogy would be tying heavy weights to your feet. Any energy you put into lifting your feet is given back when you put your feet back down. But if you try and run, it's going to be a lot slower.
Exactly - if the returned power is at the wrong time it will work against you
I’ve been looking forward to this. Yes, you’ve shown frame flex stores energy and gives it back to the drive train. Excellent, this what we want to know. What would also be nice to see is a comparison where you charge up a steep slope at a set wattage with bikes of the same weight but different flex and time the ascent and also get the rider’s impressions. Anyway, good stuff 👍.
Very interesting demonstration, but there might be a different interpretation. With the bike on the stand and the rear brakes applied, pushing the pedal down 1 cm causes frame deformation but it also stretches the portion of the chain running from the top of the chain ring to the cogs on the rear wheel (energy is stored in the stretched portion of the chain). When the brakes are released, the energy stored in the stretched chain pulls on the cog, leading to the dramatic spinning of the rear wheel that we saw. That's the simplest explanation I can think of. As for the benefits of frame flex, maybe the energy of the bottom bracket moving the other way helps lift the rider's leg up on the back part of the pedaling. Just a guess on that one...
I think the energy stored in a flexing frame is released in the form of a slightly slower crank rotation just after the powerful part of the stroke. In other words, when you come off of the most powerful part of your stroke, the frame straightening isn't going to propel you forward, it will just slightly slow the pedal stroke. So the power really is lost because it didn't get translated to forward speed while there was pressure on the pedals. If the pedals were being turned by a machine that never varied it's power at all, then the energy might be released into the wheel, but people don't work that way.
(I hope i can write properly in English....)
In addition, any time the frame flex, it causes to lose the straight line between both wheels (more precisely they get away from the same plane, where they mast be), and it results in a mesh of not parallel forces (both tires tend to go in different directions) that generate a high loose of energy. Also, as both tires tend to go in different directions, there is a loose of energy due to increased friction forces between tires and road...
The biggest problem I have with this demonstration is that the load which flexes the frame peaks where his pedal was and drops off towards the bottom of the pedal stroke. This means that the frame bending back to its original position at a 6 o'clock would merely move the pedal out to the right and not cause any additional rotation or therefore put the power back into the drive system.
For me stiffness has always been about handling. I know I’m not strong enough to significantly flex a frame. But stiff bikes always feel more planted in corners and snappier coming out of corners.
We are talking about stored energy here. If rotating the wheel is like pushing an object through a spring, the stiffer the spring the harder to store energy in the spring as we push.
I would rather the energy transfer through to the wheel during our force pulses rather than have a portion store in the frame and come back out through my foot on the backstroke or twisting me at the handlebars and dispersing across my body.
Remember, energy stored in the frame doesn't necessarily go out through the drive chain. It can come back into us at innopertune moments.
I think it's mainly about ride feel and user preferences what you feel more comfortable riding. But probably a stiffer frame is a little more efficient, as energy you convert and reconvert always involves losses to internal energy and these losses are relative to the time dependent behaviour (and in part thus to the displacement). A stiffer frame with a higher elastic modulus probably also has a more purely elastic behaviour and less viscoelastic parts, which would mean a lower energy loss during hysteresis of the loading/ unloading cycles; also less damping in both loading and unloading due to less pronounced frame flex displacement wise. Which means although the amount of energy stored in frame flex is probably almost identical for different frames, there is more direct conversion in stiffer frames resulting in less time dependent behaviour qnd less energy losses. But I still think it's mainly rider preferences as it also involves comfort and personal compatibility next to pure efficiency.
It's about predictability. Experienced cyclists like having fine control over their power input, and a flexy frame will delay and distort your pedal stroke. It also creates a cyclic distortion to your handling which can cause speed wobbles and reduce overall controllability.
the amount of energy stored by the frame in that experiment was on order 10 J. Steel tubes have pretty low hysteresis, and it's probably losing less than 1J per downstroke. Your thermal loss might be up to 3W. For this, stiffer is better assuming constant maximum force: Let the frame obey F=kx. Then W = F*x = F*F/k. This is probably accurate since your pedal stroke is probably force limited by your body, which is much less stiff than the frame.
In this experiment, they had constant pressure on the pedal and then released the brake.
In reality, however, you don't always have the same pressure on the pedal if one pedal is at the bottom and the other at the top, for example, no pressure can be exerted.
In this case, the energy in the frame can also be released in such a way that a pedal simply moves back a little.
In the experiment, it would be like simply getting up from the pedal, then the pedal would return to its original position and the energy would be wasted.
In order to get a result from such an experiment, how much energy is released from the frame back into the pedal and how much really goes into the chain, the pedal and brake would have to be relieved of the pressure at exactly the same time.
Unfortunately, this experiment was not meaningful, the conditions in reality have not been met.
Yes. This experiment has multiple things wrong with it. the biggest thing would be that they pressed on the pedal and held it down. which just the frame right now is acting like a spring but because you're holding a single pedal down 99% the forces allowed to go back into the drive train. but when were peddling were alternating legs which means we're not constantly holding pressure on one side.
when were actually pedaling alternating which side is relaxing and which side is applying Force. So this means that the frame is always flexing in the direction of the side with the least amount of force. so this means that the force is going in the opposite direction of the drivetrain weather be back into the frame and into your leg this equals force/power loss
I suggest you climb a hill with two bikes, with identical weight and identical aerodynamics (or as close as possible to that) - one quite flexible and one quite stiff. Keep same average power with both bikes and see if the stiff one is faster
Davide Archetti I’ve done exactly that with my Colnago C40 vs Tramac SL3 and no difference. I’ve also done varying watts with the same PM. 300 watts and 550.
It's all about timing, or as many others call "planing". When you apply a force to the pedals, there are time delays as the various components stretch. These might be only microseconds but they tend to smooth out the force over time. When you add it all together, (shoes, pedals, cranks, chain, wheels, tires, and frame) you might be looking at a lot of energy.
So what does this feel like? Depends on your pedal stroke and how you apply force. If you spin super smooth, the constant force will keep everything loaded and you won't get out of time with your bike. In the steel frame old days this lead to the idea that you had to spin circles to be efficient; you kind of had to with the old flexy bikes. If you mash a flexy bike, some of your force will load the bike, stretch the chain, and wind up the spokes, and then potentially be released at the top and bottom of the pedal stroke, when your feet are no longer applying much pressure.
So really its about timing, if you can match the timing of your bike to your personal style of force application, you will find that golden ride. I found this once personally with a 2011 Opus Cresendo w/Mavic Ksyrium wheels, that sweet spot where every push of the pedals matched perfectly by the timing of the bike.
As a material flexes repeatedly, you are bound to lose some energy in the form of heat. To what degree? Who knows.
Probably to very small degree. So small it's pointless to care about it.
Definitely less than 10% less than 1%? who knows?
Have you ever made your frame warm by flexing it?
Measure the temp change and then you can calculate the wattage required to cause the increase. Bingo there is the amount of watts you wasted heating up your frame .002 degrees
1%? Never... much less than 1% lol. That would be ridiculous.
Great vid. Genuinely thought provoking. My head still thinks that a stiffer B.B. is better from the initial pedal stroke. However, my head believes that a noodle like floppy frame will dissipate the energy back through other areas of the bike and not back through the drivetrain.
The sweet spot is when the frame planes for you. See Bicycle Quarterly for more info. When a frame planes it means that the rider is able to put out more power and feels less fatigue in his or her legs. So it's really a physiological thing!
I could only think of the concept of ressonance to support the frame plane hypothesis. That will probably mean that frame plane will only occur at a certain power/cadence combination, which again will make it pretty useless in practical use.
Stig Aannø If it's resonance that gives the benefit then it would apply to all power outputs as long as your cadence was near the natural frequency that the frame flexes at. You would only need to worry about maintaining cadence, which is easier with more and more gears on the bike, as long as you don't go 1x.
Erik H you are of course right, it will depend on the frequency and not the amplitude. But still, the ressonance frequency will be a product of a combination of the frame and wheels. Maybe we will see people tuning their wheels through spoke tension to set the preferred ressonance frequency in the future? Well I think I myself will go for the stiff bike/wheelset combination. Look up my other post to see why I think this is the best option.
Stig Aannø Bikes can plane. That is proven. Read Bicycle Quarterly! It is very relevant in the real world and it depends on your power and weight, not on your cadence.
Max Sievers you made me read what was available without paying and now I am even more sceptical. The stiffness of the wheels will for example make up a lot of the characteristic of the stiffness of the bike. I do not see it even mentioned. They mention measuring power, but I don't see if they measure at the pedal or at the wheel hub. It is a crucial difference given the thesis they are trying to prove. They tell that this is the secret of covering ultralong distances like Paris - Brest - Paris, which make it really look like quasi science (I have covered a distance of 540km more or less non-stop on a stiff carbon frame and wheelset, my legs were fine, my ass hurt and I needed sleep). Boats plane, frames can theoretically resonate, but I do not buy that it has much to do with your weight and power (due to certain laws of physics). I think the biomecanical argument makes a compelling case for oval chainrings, but not for a wobbly frame.
The static test isn't comparable to a dynamic test for a few reasons but simply put in the dynamic test the load coming back does so through the riders legs which have moved past their optimal stiff position. Having said that and as others have said my 'feel' is that there is a sweet spot for some frame flex for rider comfort which ultimately will cause fatigue and reduce your efficiency anyway.
A flexible frame will have more losses, because that flexing creates heat. It's not a lot, but it is there. Also, flexing will effectively delay the pedal stroke, therefore hurting acceleration and makes the bike feel sluggish.
There has been a lot of research into frame stiffness in motorcycle racing. The trend used to be to get the frame as stiff as possible. It only ended when the lean angles became so extreme that the suspension wasn't doing anything. So they strategically introduced flex into the chassis to act as a primitive form of suspension when leaned over. The current consensus is that flex within a certain range is good.
So it seems stiffer frames definitely are faster in a straight line. But when taking road surface into account when, you might say a little flex is good. How much? It depends on the road surface and the weight of the combined rider and bike. And then there's the fact that we don't tend to corner that much...
TL;DR I'm sticking with my stiffer frame.
StuntpilootStef precisely! The difference between the two frames are the time that the frame use to bounce back.
A stiffer bike means less comfort on a crappy road but it'll have a really ready response compared to the flexible one. I've got a steel fixed gear for my commuting and it is super flexy but I always enjoy to ride it even to do some climbs, and still can go hell fast if I want.
Intuitively a stiffer frame makes sense in terms of power transfer, and in most cases stiffer probably is better, however there is an interesting area of study into what has been dubbed "planing" or basically if more flex can lead to better power transfer in certain cases. There definitely isn't a large depth of study in the field to prove either hypothesis, with logic being on the side of a stiffer frame, however it is a really interesting idea to explore. You should take a listen to this podcast, it is very interesting and very thorough.
cyclingtips.com/2017/06/cyclingtips-podcast-does-frame-stiffness-matter/
janerney I'll fo that! Thanks for the tip;)
janerney I have just read about that for the first time and that Heine also experimented with polymer bushings. He seems to find quite a difference between different bikes. The differences were there with quite subtle changes in stiffness. I do wonder how you are going to find the right one by any other means than trying out all possibilities. Because it does seem quite easy to get wrong and there are a lot of factors involved. You have the weight of the rider and the bike stiffness, but also what BB you use, maybe which tires, groupset, chain, rims, etc. It does make things bloody difficult and almost impossible to get right for anybody that's not a pro cyclist.
Yeah, i get what you are saying. It is so nuanced and there are so many variables fro scenario to scenario that it is hard to see any quantifiable real world benefits. Interesting to examine and read about, definitely, but is it something that i am going to think about when buying my next bike, definitely not.
A lot of really good comments about the the delay in flex causing a different feel. The next step would be to have a pedal stroke analyzer attached to the hub to see the different plots of force over time. I'm thinking that you would need a pedaling robot to eliminate the human variation. My guess is that the flexible frame will have a smoother response due to the damping of the frame acting as a energy storage spring in the system.
Some significant errors in the argument being followed here. The flex on the frame decreases when the force component in the original direction that caused the flex, decreases. This will be when the pedal in question is at six o'clock. Releasing the flex in the frame only when the pedal is in that position, will result in no wheel drive as seen in the experiment performed in the video. Instead, the rider's weight will counter the force, in other words, the rider's legs will absorb the energy.
A stiffer frame will carry less energy to the six o'clock position, pushing more energy into driving the rear wheel. A perfectly stiff frame will carry no energy into the six o'clock position.
In other words, a stiffer frame would be faster. Is the performance worth the tradeoff in comfort? Depends on the situation.
Thanks a lot for saving me the typing :) Especially when sprinting (while not being seated) this becomes visible in the 6o'clock position. To visualize the problem let's assume the frame being a perfect spring (that means, it will return all the energy it absorved). And we shall only look at the vertical part of the flexing, so just put a spring vertically on a table. To compress it, you have to apply a downwards force on top of it. If you put a weight on it, the gravity-force will compress the spring to a certain level. On the bike, the compressing force is made up from gravity and the muscles straightening the leg. Back to the spring: Action=reactio, the gravity on the weight is keeping the spring compressed, but the spring is pushing back with it's own force which increases with compression - so that the spring will only be compressed to the level where the spring's extending force equals the gravity force.
When is the spring going to release energy? While extending back to it's normal size again, which can only be done, when the weight on top of it is removed (or reduced). On the cyclist that translates to lifting the leg or reducing the power with which the leg was straightened to push down the pedal, which in turn means: The frame can only flex back when the cyclists let's it push it's foot upwards again.
In the end, the frame flex is not giving much energy back - you only use your energy to compress it and maybe even need some uncompress it.
Anecdote: Cycling home from shopping with a heavy backpack (weight increase ~30%) I tried to sprint. I noticed that while sprinting I usually would "jump" when my foot reached about 6o'clock - with the backpack, I couldn't. I was just compressing the frame and get back to where I was before jumping. At 6o'clock.
Compare that to a trampoline. Stand still on the trampoline, compress your legs and then jump. Your legs extend, but because of the springiness of the trampoline you're only raising your body by a few centimetres instead of getting about half a meter off the ground. Wasted energy.
This of course also applies to the 3o'clock position as shown in the video. But in the video, the energy was transfered back to the wheel, because the weight was taken off the pedal. When cycling, you push downwards until you reach 6o'clock - so the frame flex introduced at 3o'clock is not released at 3 but rather after 6 - where you won't have any more gains from it.
Perfectly stated. You must also be a mechanical engineer.
Eon du Plessis don't think so... #planing. #janheine
Yes to this but, I will add that if the frame at particular points then this would transfer loads into the drive train decreasing the efficiency of the drive train. A better experiment would be to increase the load on the frame and measure the efficiency of the drive train. The drive trains are very quiet so there is likely to be very little energy being lost there, but possibly in heat. Again the principles describe above would be the likely loss, transferring it into the rider by delaying the power transfer. Does anyone know how the manufactures test the frame stiffness in the dynamic sense, eg as a rider analogue?
One of my bikes frame cracked at the seat tube and down tube just above the bottom bracket. It did not break all at once. It was gradual. The first symptoms was a self shifting rear derailleur. Then as the crack got bigger, the frame flexed more and the crank felt soft on the down stroke. It felt soft since the crank was flexing inward and the result was a noticeable loss of power as some of the forces was being directed inward instead of downward. If a frame is too flexy your pedaling force may not be be optimized to forward motion. Kind of like climbing or sprinting out of the saddle and rocking the bike slightly. Rocking the bike may enable to dig in more power, but flex of frame and change of wheel angle will be a little less efficient to forward motion.
You can test yourself by cutting slightly into your frame to start a small crack. One note also about the test in video, the wheel is not clamped perfectly tight and still, so that could affect results a bit. Also in real riding, that much crank flex would cause wheel to spin since brake is not holding wheel still. So amount of flex in frame would even be less. I think how frame flexes in regards to the line of travel is important too. If a frame flexes, the energy could cause the head of the bike to change line and produce more road and wind resistance, whereas as stiffer frame would not deviate as much.
It seems to me the wheel spinning is mostly from the potential energy stored in the chain, at around 4 minutes you see the chain jolt/twitch forward upon releasing the brake. IMHO for 99% of us the majority of frames are stiff enough and frame flex (on most bikes) isn't that big a deal and not another thing we have to worry about upgrading. Ride upgrades don't buy 'em.
I recently built a new road bike around a different frame model. At first it felt like I needed some time to get used to it - regarding not only the handling, but also how exactly it transfers power. Thanks for the video to explain the background!
Stiffness is great for accelerations over 1000w but not going to make much difference at a constant 200w, and you get into a better rythm with some flex.
Simon, there is an experiment you guys could perform that could possibly reveal any power loss difference between a "flexy" frame and a "stiff" frame. It would require a PowerTap hub power meter and a crank based power meter. Using the same power meters on the two different bikes, you could see what the offset was between the power reading at the crank and at the hub. Since this offset might be small and possibly not resolvable at lower power outputs, it would likely be best to do this test either in an all out sprint or in all out effort uphill. you could then compare the percent offset with respect to the crank based power, i.e. (offset - crank power)*100/crank power. If you don't maintain the same power output between tests, you might see some uncertainty bias, but that's not necessarily a game killer. You wouldn't be looking for the quantitative difference in power loss between the two bikes but rather the qualitative: is there more power loss in the flexy bike vs. the stiff bike.
With due respect, your experimental setup doesn't represent what happens in real life. You were reacting the frame torsion with the turbo trainer. On the road however, it would be reacted by side load in the tyre. The frame itself will have low hysteresis (steel especially) but the tyre certainly doesn't. So, the cyclic torsional loading of the frame does come at a cost.
I would still agree that frame stiffness is a 'feel thing' to a large degree. However, that doesn't mean it's not also beneficial. It is impossible to generate load without something to react it. Imagine a super flexy crank - you'd never be able to push hard against it as it will bend out of the way. If we were perfect engines it wouldn't be a problem - we'd just load it up. But, we provide power with alternate legs and to differing loads round the cycle. So, flex in the system causes a lack of reactive load at the point of peak power generation. The load doesn't simply fall out of phase, the peak drops too.
In short, static tests and analysis of an inherently dynamic system, can lead to flawed analysis. Not that I'm saying my hypothesis is correct either though...
Worth noting that tyres generate grip through 'slip' at the macro scale. Because the tyre is rotating, cyclical side loading will cause a pathline that meanders side to side - think of a little sine wave. Accordingly, the losses in such a setup are not tied solely to the lateral stiffness of the tyre and the hysteresis associated with 'non-slip' side loading (in say a non-rotating tyre). In short, the losses aren't the same physics as vertical compliance of larger section tyres and rolling resistance.
This is the video I want to see
Gonna agree on most of this. The efficiency of the frame probably doesn't change a lot in the minute details from frame to frame, BUT this isn't the same to be said for handling. A stiffer frame will handle better and this is where I believe the sensation of stiffer=faster originates: simple human misinterpretation.
Sounds like you're describing a suspension fork that has semi-independent L/R travel. This would affect balance first, and loss of balance (and energy required to correct/maintain it) would lead to less power available to turn the cranks smoothly.
If we consider a suspension fork or a FS mountain bike, some of that power goes to compress the suspension, but the return on the shocks (if the bike goes down after a bump) sends power back to the crank the same way this video shows (as long as the cranks are still being turned at that point). If the rider bounces while riding on flat, the energy of the return can be used to move the rider back up, and this feels like lost power, but then the rider regains potential energy as they are higher. Do they use that potential energy to power the cranks? If they go higher than their initial position, yes. If not, it was extra energy output that did *something* but did not power the drivetrain. This sort of "accidental" energy redirection is what we really want to avoid, right? I can't quite tell if this same redirection away from the drivetrain happens when flexing a rigid frame on a road bike...
Any thoughts?
Besides, even if everything that happened in the test was accurate, which I agree it's not, according to the 2nd law of Thermodynamics states basically that " In any given exchange of energy, there will always be energy lost", which basically means that with all the exchanges in the experiment the losses will be bigger than with a stiff frame which will produce no such exchanges.
That was great. I had a LOOK back in early 2000. Was a fairly flexible frame. I could easily tell. I loved that bike though. It was pretty fast geometry-wise (thanks for the other video with Tom) and when I got to climbing, I could feel that effect - especially out of the saddle, which felt wonderful. The bike felt alive. There was a give on the one side and a spring on the other as you pedaled - subtle but it was there. I've always been thinking due to the stiffer is better mantra that the bike was inefficient. Really liked how it felt. Turns out, I just rebuilt it yesterday. I'll get to ride it again with this in mind, having since been on really stiff aero frames.
Hi, so what's your opinion on the bike stiffness now? Does higher stiffness contribute to speed?
@@Jesus_s_Real For me, in comparing the two, higher stiffness contributes most noticeably to acceleration. In most daily riding and training, the differences are not too large in terms of average speed. When racing, I'd always pick the stiffer bike, other things being equal. There's a little more average speed to be found there, especially as in racing you are always accelerating and always riding hard. I love the feel of the less stiff bike and it's my training bike choice most days.
I think there's a lot to be thought about within this debate. If we examine a damping spring mass oscillator, that is a spring with some weight attached to the end, and we apply a force to the weight - we will find that it oscillates, but the oscillation decays due to damping of the spring. I think that a "flexy" bicycle frame will be analogous to this oscillator in that kinetic energy will be lost by restoring the frame to its original shape before applying some to the drive train. Will this loss be noticeable? Maybe not, maybe it is in the marginal gains territory.
I had a super stiff frame a long time ago, and it put a smile on my face every time I rode it. It was the immediacy of the response that I loved. You pick up the speed, the bike second guesses you and you're off. Surely it depends on where the frame is flexing. If you have a 'ladies' crossbar, the flex is extreme and only releases the front end to come back in line with everything else.
Did you say 'ladies'???? So 1950's....
I lost a tremendous amount of energy due flexing when my steel single-speed mtb flexed in a corner while mashing the inside corner pedal. Causing me to drop the chain and the sudden drop of the pedal sent me flying to the ground.
In this experiment the rear wheel is fixed and when the pedal is loaded the distance between the rear wheel axle and the bottom bracket axle is increased so the chain is prolongs and when the pedal is released the chain backs to its normal size and the wheel start to spin.In the flexible frame the distance between the axles under the load changes not so much( the rear triangl bends) so after the load is gone the wheel will spin not so much respectively as with the stiff frame.So the stiff frame has a better efficiency.
My conclusion is: ride what you own and stop worrying about other people's bikes and marketing hype
I am quite certain they understand that, but if some would ask they had to put up a video for that.
ride what makes you smile I would say. I'm glad I got rid of my previous aluminium bike
Agreed! My top two bikes that I ride are made out of aluminum! Stiff sometimes but they make me smile so I ride them as much as I can.
The science of cycling is, for some, a large component of the fun. It's pretty easy to trivialize non-essential activities.
Exactly, want a bit faster and smoother ride all around? Get an e-bike, even of low levels of assist.
So glad you prompted this discussion. In my opinion, frame stiffness is a red herring. Anecdotally, Sean Kelly won a hell of a lot of sprints on a Vitus 979, likely the least stiff frame ever sold. A stiff frame "feels faster" but it's not. The materials bikes are made out of make excellent springs and all of the energy is returned. A slightly flexy frame is also much more comfortable.
imo, in line with the mechanic, stiff frame will feel faster as the flex is lesser so it feels more responsive as you load those drive train. whereas less stiff bike will feel sluggish or lost of power as there is a time delay for the transfer of energy.
ideally, you want responsive bike to attack at the right moment so stiffer bike is still the better option.
lastly, the flex of bike frame is commonly compensated with the slight twist of handlebar to maintain balance. that may result in a utter different finding.
Theoretically though, according to the video, torque should be returned before the end of the pedal stroke, so power will be returned before 1/2 of a revolution on the pedal. Assuming pedalling at around 80-90 rpm, thats only about 1/3 of a second. I personally don't think (at least at most people's riding level) that will make too much of a difference.
Feeling faster ≠ actually faster
Case in point: bigger tires at lower pressures feel slow but are measurably faster while skinny tires feel faster but are comparably slower.
The handlebar twisting as people pedal isn't just from frame flex, but also from:
1-The rider putting uneven hand force on the bars through their pedal stroke.
2-The response of the steering geometry (i.e. trail) to the rider rocking the bike.
You are on the right track. The faction of the 1/3 of a second of your muscle exerting effort while the spring loads and unloads adds up as it happens on every revolution. Think of a stiff frame as your muscle working for a slightly shorter duration every pedal stroke to deliver the same power. If the frame were infinitely stiff, the muscle would have shortest possible duration of effort. I agree people will feel the stiffness, but that little extra effort will also add up. My guess is this can be significant.
I have a stiff bike and a not stiff bike and I’ve ridden both with the same wheels and tires and it’s like running on sand vs running on concrete. I don’t feel power loss in the stiff bike and it’s very responsive and snappy. It’s like making a chain from rubber, it will store that energy like that but by the time you reach your dead spot in the pedal stroke the energy can’t be used to produce forward motion.
I personally am more interested in frame flex during cornering, if it helps the tire track the road better like a rear suspension in mtb or does the change in angle of wheel relative to frame might cause instability like a flexing chassis in a non performance oriented car.
Adrian L Good point.. Usually, espevially with skinny tires, corner performance is underestimated. My empirical evidence suggests stiffer frames give me more confidence in cornering.
Adrian L that is a very keen observation because that's exactly what the Grand Prix in world superbike teams are working on. More torsional less lateral stiffness
I felt instability over the comfort in fast cornering sections with my former steel xc mtb. I’d rather feel the ground below, not a flexi bike frame in between, trying to do some mediation. But if you’re not doing a sports ride, this feature can be comfortable. But than again, sitting on the couch in front TV is even more comfortable.
In my opinion, they both transfer a similar amount of energy. The biggest difference is the reaction time. Stiffer bikes react faster to load switching between handlebar and two pedals (for road bikes). On the other hand, the less stiffer bike reacts slower but is good at absorbing external forces ( for cycle cross or mountain bike).
This would have been a much more accurate test if you had both been wearing your science glasses.. What were you thinking Si.. attempting science without the correct tool.. shame..
and a wrinkled lab coat, of course
if i can assume that u're british, so this comment is funny, but otherwise not somuch
I thought frame manufacturers were aware of this and incorporate specific amounts of flex into their designs. I was reading that a prototype Specialized Demo bike was seen with varying thicknesses of carbon which had been crudely adjusted during testing. The explanation was that they were tuning the amount and type of flex. And yes, perhaps akin to tyre pressure, a stiffer harsher ride could fatigue the rider sooner, negating any power gains that may have been achieved.
The experiment has one condition: the leg doesn’t move when the frame springs back. On a road that won’t be true. Basically, the frame’s flex will work towards lifting the leg rather than moving the bike forward.
The feel is a timing or tuning issue. Just like string tension in a tennis racket or golf driver shaft stiffness (pros will use higher/stiffer here too), the frame stiffness affects *when* the stored energy is released back during your pedal stroke. If it comes back at the wrong time the bike will feel slower, even though you are losing very little energy regardless of the frame stiffness. Roughly the higher your cadence the stiffer, and hence faster spring back, you are likely to prefer. Of course there is both mass and geometry of the bike involved in its dynamic response, so it won't be that easy to correlate to just the stiffness. It probably varies with rider strength too.
I suspect this translates to acceleration. I.e. higher frame stiffness = better acceleration. Very small numbers in difference of course. Moreover, a frame with very low stiffness could potentially "lose" energy by the fact that the frame never seems to return to the initial position, this part is more a though experiment since no frames are that bad (the energy would most likely go into heat at this point by bending of the frame).
But in my experience the biggest difference i noticed when moving from an old alloy frame to a TCR (acclaimed of very high stiffness), was the acceleration. And with better/sharper acceleration i felt that my momentum didn't suffer as much when the gradient changed, i.e. i was able to keep the speed higher.
I think you nailed it on the head. It’s not necessarily more stiffness that make a bike “faster” because even super stiff frames still flex. I feel it’s stiffness in certain areas of the frames that brings greater acceleration when cornering for example. As a surfer, I can see the similarity with surfboards and especially surfboard fins. Super stiff boards or fins don’t accelerate well in corners, but the “foil” or actually profile of either one will create better performance, having stiffness at high speeds, but also great acceleration in tight corners when making sharp turns. I think this is going to evolve as technology and testing reveals more proof of efficient carbon profiles.
I think the flexing frame moves some of the energy to the dead spot. Stiff frame feels fast because you have almost all the power going in the same space of crank revolution. That might not make much of a difference. Unless the the frame acts as a damper and eats the energy. But flexy frame has a lot to contribute to rider comfort.
what about doing a road test with the two frames both mounted with pedal and hub based power meters. if frame flex matters for actual speed the difference between power meters should be larger on the more flexible bike.
Exactly, I proposed this very test in an earlier GCN video wherein flexy was deemed to be energy lossy. Hell, use power pedals/crank arms/bottom bracket/rear hub meters and test with seated versus standing and so forth. Flat road, up hills, around corners, the whole schmear. Light rider, heavy rider, amateur rider, pro rider. Swap meters to other bike, calibrate, repeat. Science, in other words, not conjecture and thought experiments.
I'm not sure power meters are accurate enough to generate any meaningful data. 1.5% is about the best you can find, I think you'd probably need 0.01%!
Si, you've just restated exactly what I'm bitching about. 'I'm not sure' 'I think you'd probably need'... How about measuring for meaningful data, publishing the results, and commenting on that? If you can't measure a power difference from noodly frame to uber stiff wonder bike, then it doesn't exist in a meaningful way, and that's important. The 'method' employed in this video is dubious, as many have pointed out. We have actual tools: please use them. Thanks.
Okay then, let me rephrase.
There would be absolutely no point in doing the test you suggest as the tools to measure meaningful data don’t exist. The levels of accuracy in current power meters would leave vast statistical errors.
Therefore, you're declaring there's no meaningful difference between frames in terms of the rider's energy reaching the rear wheel and moving the bike forward. Thanks for clarifying your position.
I think that after a frame flexes under load, when less load is being applied (top dead centre?) and the built-up energy is being released that most of that released energy is transfered back to the crank and pedal where it slows down pedal rotation marginally. This would be the flow of least resistance of the energy and if this hypothesis is correct then a flexy frame would indeed be slower than a stiff frame. The best way to test the idea of flexy vs stiff frame may be to use 2 identically equipped bikes that weigh the same up a very steep gradient.
Two things: First, no form of energy transfer can be 100% efficient. While most of the force that went into flexing the frame will be returned to the drive train, some will end up getting lost. Secondly, it isn't the difference in speed that we feel going from a flexible to a stiff bike. It is the difference in responsiveness. When we apply force to the pedals of a flexible bike, that initially goes into flexing the frame instead of moving you foreword. As a result, a stiffer bike will be quicker to jump off the start line as apposed to a flexible one.
also worth noting the importance of stiffness in the front end as well. You want the bike to steer as soon as u point it right or left as oppose to just flex and then bounce back to steer whenever the frame feels like it... which at the end of the say will also make you slower as you corner with less precision and confidence
That's theory. Makes me wonder if you even ride a bike
Or maybe you are riding a really noodly bike in which case I'm sorry for that comment
Tobias Grätzer Go watch some videos of a racecar hitting a rumblestrip in slow-motion, you will see the science in its truest form.
A frame that planes for you is responsive and it can corner well.
Nice to see an attempt at showing the principals involved here. I know that its going to be hard to give a perfect experiment when your not in a lab and can test for all sorts of variables. Though it at least shows that perhaps due to conservation of energy we shouldn't be so concerned that if there is some flex in your frame then you are just wasting all of your energy just to make a frame flex instead of making you go forward which is what your looking for. I am sure that some more more testing in a lab would help a little more but good one to get the ball rolling! Nice to also show you don't need to chuck out that steel/aluminium frame because it will flex.
What about the flex on the non drive side. Does it counter the flex?
The same thing happens, just mirrored. Therefore, it was a bit odd that he said the effective "clockwise" motion of the BB - on the non drive side it will be an effective counter clockwise motion of the BB that gives you the little kick when the frame returns into its normal position.
This is exactly what i thought. He's testing the "clockwise" side of the bike. (Maybe due to the setup being against the wall and inherently wanting to see the drivetrain during tesing).
On the other pedal the return kick would produce anti-clockwise energy?? Would this cancel out the clockwise energy produced in the previous half stroke [hence his IS Wasted energy] ?? Would this if tested similarly, force the drivetrain backwards and no wheel spin created, since it would just spin the freehunb? Or would it still "create" energy since no matter which direction clockwise or anti, still be energy and movement thta the bike has, and that is used for movement...?
(I ride a reasonable flexy cheap Allez so would love to know!)
The clockwise/anti-clockwise energy is at the level of the crank, but is applied in different points of time. All is transferred into "clockwise" (forward driving) energy at the level of the drivetrain. It's just a perspective thing. By the way, this doesn't imply the theory of flex-restored-energy is correct, just expanding the theory to the non drive side.
I would say that you see the same amount of flex on both sides. There may be a small variation as they are mechanicaly asymetrical, this may be insignificant, but pro teams with pedal stroke analysis may say different.
On a more flexible frame you would see a pendulum effect. As you are on the upstroke of the right pedal the frame flex is unloading to the right, while also being flexed to the right by the downstroke of the left pedal.
As the left pedal is on the upstroke the frame flex is unloading to the left, while also being flexed to the left by the downstroke of the right pedal.
Even the stiffer frame seemed to store energy though. A stiffer frame may still have same effect, perhaps to a lesser degree, just that the flex isn't so noticably focused around the bottom bracket and spread around the whole of the seat tube and down tube flexing to a lesser degree.
I once had a go at calculating the lost power due to an inch of sideways flex at the bottom bracket. It came out as about half a watt if I recall correctly.
Hmmm, anyone want to buy my old flexy frame? :D
There is a net loss in heat generated by flex, though small, its not insignificant, and also the stored energy released by unloading the frame comes at a different moment in time to the power stroke, so the effect will be (on a bike that flexes more) a more sluggish response, though a more compliant feel.
This experiment is fundamentally flawed and does not answer the question. The flaw: when you apply load to the pedal and the frame flexes, the pedal becomes immobilized between the block and force applied to the pedal. When you release the brake, the frame moves back to center while the pedal is still immobilized; however, as the frame moves while the pedal doesn't, the crank moves. This occurs because the crank axles moves upward with the frame while the pedal remains stationary; thus the crank turns. All you have done is pre-loaded energy into the system and forced it to the rear wheel. If you move the pedal to Bottom-Dead-Center and repeated the experiment, the results would differ because the rear wheel would not move.
Most likely frame stiffness contributes little to power gain or loss due the dynamic nature of pedaling. Frame stiffness is more likely to contribute to handling.
Thomas Johnson This is the only right opinion. My man!
Fabulous topic and video! I think Tom is spot-on correct. And a frame with a little flex is more comfortable too.
Simon has described two possible outcomes.
1. Is the experiment incorrect and stiffness does cost Watts. The solution would therefore be a stiffer bike reducing Watts lost.
2. Energy is transferred to potential in the frame, which in turn gets converted to kinetic energy in the drive train.
The problem with 2. is that no system is ideal. There will still be normal resistive losses in the drive train but added to that will be efficiency losses in the flex part of the experiment system. The solution would therefore be a stiffer bike reducing Watts lost.
Ultimately though, the Watt losses are tiny and insignificant. And most likely its only the riders in the top 1% of the sport can see benefits.
Amatures will affected more by other variables fluctuating (for example, weight or fuel choice) and probably shouldn't worry about minor stiffness improvements.
I have been thinking about this a lot. I agree with this to a point. Most of the energy goes back into your cranks of course, but you load it in a different direction than your pedal stroke is positioned when your legs ease off the power. So your crankset gets it's energy back in the wrong direction, not adding to your speed at all.. I think.
Does the wheel move when this test is done on the non-drive side (left)?
I'd argue it doesn't because if you put the force on the right side the flex in the frame would put some tension on the chain making the wheel move when released while putting the force on the other side would actually slacken the chain.
So.... what's the deal?
Christian Holmstedt I think the force is still carries in the spindle of thr crankset.
It might, but it would be in the opposite direction as the force applied on the right side.
That was my point.... the force is opposite.
GCN needs to do some quick science to check. :)
Christian Holmstedt Sorry for the confusion. I was trying to reply to rkan.
Johnny Cab I disagree in part. The experiment does exaggerate the amount of flex in the frame, but the bottom bracket does move, and that movement would provide force to the chain (or slack when applied to the pedal opposite the drive side).
Flexible frames obey Hooke's Law. When you apply a force to a frame (push down on a pedal) the frame will be displaced, as F=-kx.
When you release that force (at the bottom of the pedal stroke), the elastic potential energy, due to the displacement (U=1/2*k*x^2) is released. That energy just goes back into the drivetrain due to energy conservation laws (ignoring the very minor efficiency losses).
This is the first time I've seen it demonstrated empirically, and it's a very good demonstration, kudos to Tom Sturdy - never thought about it like that - great video!
My only minor dispute would be the focus on the vertical movement, as it's simply a component of the overall displacement. Assuming 100% mechanical efficiency, even the energy lost via the work done to displace the frame horizontally is recovered as the frame flexes back to its starting point, therefore, Work in = Work out so the net change (loss) is zero.
Loss of energy is not zero, there is little loss maybe in the frame but the tranfer of it through all the transmission components is not negligible, and above all there is the tire. Whatever it is the compound the gum suffer form hysteresis which probably wastes all the energy returned from the frame flex that managed to get there. There is also the problem of fatigue, when you repeatedly put the frame under load within the elastic range the material is weakened by a tiny bit due to imperfections. If a frame flexes too much that could shorten the life of it.
The Uprising of steelbikes has begone
Why is that? Steel has roughly the same stiffness per weight as aluminium.
im a 853 steel bike man , I love the feel and ride on the road
So the brake goes on, then the pedal moves down which puts tension on the chain (and flexes the frame as the chain can't move).
Of course the wheel will move when you release the brake, you've essentially just pulled the chain a bit. Yes, the frame 'absorbs' the tension and releases it but I don't think that says much about frame stiffness.
Love this vid., and some of the comments below are excellent. Perhaps a refinement of the set up to measure the heat loss/kinetic energy loss. That said, as a guy who recently ended his hard tail hold out status, my new full suspension mountain bike yielded me numerous p/r's on trails I've been measuring on Strava for years. Lot's of heat loss through making shock oil get warm yet the net result was quicker times on all sorts of trails, not just downhill segments. And lastly, fatigue is another component...even on the road, a comfortable (flexy?) bike may prove slower on short, high output situations, but end up being quicker during long days on a bumpy road in spite of the heat loss of bending and releasing the frame tension thousands of times.
Seams like the energy is stored in the chain.??
Yep, thats what i was thinking...
And in the spokes? They are put under pressure and thereby functioning in the same way as a spring in a wind-up clock, I would presume
Yes. I hadn’t seen your post before I posted. With the brake applied, turning the crank down 1 cm visibly pulls the chain ring rearward.
I personally think there’s something to this, after three quite high quality frames I moved to a mid priced modern steel frame (Ritchey Logic) and although I’m not some guy who is all old school the frame has a lovely springy and lively feel, despite being 1kg heavier and undoubtedly nowhere near as stiff as the carbon frames, over and over I am quicker on this bike, climbing, riding hard on the flat or to an extent sprinting.
Great video, no real errors in physics I can see, BUT... 2 key aspects are not mentioned. 1. One needs to calculate or measure the mechanical energy lost to heat as the frame is flexed. Your experiment assumes the energy is all stored in the frame material like a spring and all released. Any real material will have dissipation (think of it as internal friction). That energy is gone forever into heat in the environment. Maybe the loss is small, maybe not. WIll depend very much on the material (steel, carbon, Al, Ti). Unfortunately, this loss is VERY hard to measure! 2. One needs a DYNAMIC measurement in addition to a STATIC force measurement. The phasing of how the energy is returned will vary with cadence (frequency of spring). I would be that Al with the same static flex is going to absorb energy different than carbon composite flex while at 100 rpm!
In the end, I think this non-linearity is important. Stiffness is not everything. Peter Sagan needs a much stiffer frame for the same effect than I do :-)
Still a great video! Now Si needs to write a PhD thesis on it...
Two thoughts:
1.) The resonance frequency of an body depends on its stiffness to weight ratio and resonance phenomena are related to energy absorption. Thefore if the excitation frequency is in the range of the resonance of the frame, the energy "loss" could be higher.
2.) If the deformations of the frame are leaving the linear elastic region of the material, the osscilating stresses could lead to a hystersis loop. This could also mean energy absorbtion.
If it doesn't flex your bike will krak in two.pat from Belgium 🚲
There is also the matter that more sideways movement has an effect on the rider. If the frame flexes it will cause your body to move slightly to the side. This can be compensated, by pulling on the bars and/or using core muscles to balance. Either way, these muscles are using energy which could be going to your legs.
An exaggerated way to exemplify this is how you feel after a road ride compared to a mountain bike ride. For the former, just your legs are tired, for the latter, your whole body is tired from moving around on the bike and levering it into position.
I imagine the difference between modern road frames is pretty marginal though. Personally I always ride a steel, moderately flexible frame on road because if I used a super stiff carbon frame on the roads where I live, I would have no spine left after a few weeks.
Heres the answer for you.
So the frame flex shifts the position of the bottom bracket on the downstroke. But the pressure gets released on the upstroke, thereby putting the energy into reversed pedaling. Thats a clear loss right there. There is no gain or recycling of energy happening as explained in this video. Im surprised this wasnt discovered earlier, and that this idea reached so far. If anything, this experiment shows that a stiffer frame is indeed faster.
Sadly nobody will read my comment anymore... Thumbs up please.
You either dont have a clear understanding of whats going on here(the physics), or you didnt understand the video.. They literally diagram out the opposite of what you are saying.. The BB gets flexed(deflected downward), while the pedal stays in place.. When the system unloads the BB moves back up into its unloaded position while the pedal stays in place. This causes a FORWARD motion in the pedal stroke.. not a backwards one.. Thats is what is returning energy into the system.
Golf clubs are engineered for various amounts of flex in the shaft so that the quantity of energy storage and the timing of its release maximize clubhead speed at impact. A stiffer bike stores more energy for the same amount of flex (E=0.5Kx^2, K is the modulus and x is the deflection), but it also returns the energy faster. Perhaps if flex were engineered (in the crank arms?), the release would occur during the dead spots in the stroke, maximizing acceleration. It would have a negligible effect at steady state, though. In the meantime, a stiff frame feels better and accelerates faster, so all else being equal, is better.
I put my old aluminum TCR and new carbon TCR on the trainer and found huge differences on frame deformation. The softer frame killed me on my KOM event.
who cares, just ride!
I do! This is GCN Tech, it's supposed to talk about these topics. If you don't like it, that's fine. Go watch something else :)
i like these type of videos, but i do agree, just ride :)
gofundme.com/cycling-goals
Well, bike designers care. And people buying custom bikes care. And people buying bikes off the rack care. But you don’t have to care, because the caring has been done for you.
The energy that move the rear wheel came from stretched chain. The flexed Bottom bracket actually moved the big chain ring forward and stretched the chain in which pulled the rear cassette but wheel movement was stopped by the brake until it released. If you move the front chain ring to small one and rear cassette to the big sprocket then the energy transferred will be way less. In my opinion, flexible frame is not ideal for sprinting which require instant energy transfer.
I would assume a rider would be using energy to counteract the frame and keep the bike steady. Incorporating more stomach and back muscles this would increase rider temperature showing where energy is being converted.
Another factor could be aerodynamics tiny flexes in every revolution would cause the rider and bike to cut through the air less efficiently. As a straight line from A to B is always shorter a 2cm flex at the seat would be exaggerated at the riders head, meaning power is used moving more air out the way.
Really interesting content once again GCN, keep it coming guys.
Excellent demonstration of the frame flex returning stored power to the drivetrain. This definitely helps smoothen the power delivery. People have mentioned that there will be a loss of energy transfer, but a springy steel frame should have very little loss when returning the energy. Think of it as a steel spring and how it is pretty efficient. I would rather take a slightly flexy frame that works with my pedal strokes over a stiff one that bolts forward instantaneously. PS: i a not a sprinter or a racer.
So, I would like to add when the break was released, the resistance from the wheel from spinning is less than the resistive forces of the frame. However when at top output your application of power matches the resistive forces to moving forward. therefore when at a top sprint you will only lose power rather than getting it back. It is illustrated by the brake in the example in the video, where the force is only released back when the resistance drops.
One thing that was completely neglected is that when the frame yields under load gears and chain are no more aligned properly and this can be draining energy straight into friction and heat. Actually, this may be even more of an issue for belt drive systems where alignment is very important. Mine starts to make noise when I push harder and I have a suspicion that it may be because of the frame flex. Rear triangle tubes are quite thin on my bike.
In my opinion, the energy applied on a stiffer frame is transmitted "directly" to the drive train.
In a flexing frame, the energy used to flex the frame (let's discard the fraction that is actually transformed into heat at molecular level) is stored and will "spring" back, BUT the spring action will happen when there is no input from the rider (when the cranks are nearly vertical). At this point, given the bike in moving, the small amount of energy that the frame stored would not have the same effect (would not actually push the weight of the bike + rider) as it would "together" with the rest of the energy on a stiffer frame, all because of the way pedaling happens.
I might be wrong tho. In any case, a static test is leaving out many factors in play in a dynamic situation.
Simplistically, the frame etc. act as a spring. It stores energy as you flex it and releases the energy when un-loaded. As long as there are no losses during this cycle, it doesn't hurt. Less simplistically, there is likely some damping in the system which will result in small losses due to friction and such so you don't get back quite what you put into it. Would be interesting to expand this test to accurately measure the energy input and output. Good start and thought provoking!
Ralf Hoenhe had already brought up what I just posted about the energy being stored in the stretched chain in this demonstration (I didn't read all the comments before posting). To see if there's any merit to this idea, they could do the experiment with an old worn-out chain and see if there's any visible stretch in the chain. My guess is that they used a new chain that would only deform very slightly and not be visibly stretched. More energy would be stored in a new chain being deformed by a 1 cm pedal movement (because more force would be needed). An old compliant chain would be more easily deformed by the 1 cm pedal movement and wouldn't store as much energy.
Just to throw another confounding variable into the mix, recent research on the physiology of fatigue in endurance sports such as cycling are pointing to fatigue being largely a mental construct (i.e., as your muscles fatigue you need to recruit more muscle fibers to complete a given task and that it this extra mental effort to recruit additional muscle fibers that your brain interprets as fatigue).
This implies small differences in mechanical efficiency between frames could also have a multiplicative interaction with our perception of fatigue and therefore realized performance.
As a quick test, started swapping a power meter across three frames that I felt rode either lively or dead. Wheels and body position were the same. I rode to perceived exertion levels without power readings, then compared the average wattage. I tended to put out about 57 watts more (95% Confidence Interval: 39.0 watts, 76.4 watts) than the bikes that did not feel as lively. The power meter was crank based so any energy return from the frame would not have been included. I also compared average heart rate, which tracked the higher wattage suggesting higher effort despite perceived exertion being similar.
My interpretation, was that it was mentally "easier" to ride the bike that I felt pedaled more efficiently so my focus was better and my muscle recruitment was better. There of course may be energy return from the frame, but I think how this interacts with the mental component could be where the real big returns come from.
Anyway, to do this experiment properly, the rider would need to be blind to the bikes (e.g., built up with different tubing or layups but visually identical)
Tall guy who just got his first carbon frame after riding on an old steel frame.
I ride a 62 cm bike, I wonder if the length of the seat tube has any effect on the amount of flex. One of my more seasoned cyclist friends suggested to me that longer tubes flex more, which seems to make sense. 🤔 I know that my new carbon bike certainly feels faster, but I’m sure there are many factors at play other than just frame stiffness.
Very interesting vid.
The reason for the wheel turning was due to loading. The loading is possible because of flex in the frame and other bike parts. The flex in this demo isn't exact due to the brakes being engaged which meant force is put onto the break and the part of the frame it connects with as oppose to the hub to spoke to tyre.Flex rebound will be absorbed by your body and tyres, your body may be able to reutilise this energy. The real test is to have a flexible bike vs stiff bike and have the same wattage applied and see which goes further. the chain length between the sprocket and chain ring shortens on a flexible bike which means it will be harder to apply high wattages due to the give? Anyway the energy won't be resonating due to tyres.
Stiffer frames channels your power input directly into the drivetrain, with minimal elaatic energy being stored when you pedal.
Flexible frames channels a bit less energy directly into the drivetrain immediately, but gives it back with the elastic return later on.
So stiffer frames are essentially more responsive and crisper in feeling whereas flexible frames may feel like its blunting the power input slightly, but the frame can feel like its moving with you.
My old aluminum Raleigh was more flexible thanmy Cervelo, and I can feel that I am moving more relatively to the Cervelo's frame because its much stiffer, but it feels much crisper and more responsive, so if I put power down, I instantly feel the bike reacting and going faster.
Id say its a bit like the frame geometry argunent - slacker geometry gives more comfort and is more forgiving, but more aggressive frames are more twitchy, more responsive and less forgiving...
Back in 2015 when I was tossing up whether to get a new bespoke steel bike or just an off-the-rack Carbon Fibre bike. At the time I was reading claims from one or two steel bike builders about the benefit of steel frame flex/power transfer. Googled this matter but found nothing else on this issue. Consequently bought my plastic Giant Propel.
No regrets purchasing CF. I love the feel of the ride. One thing I am absolutely sure about is CF has much better power transfer. On steel, I literally had to fight to stay with on fast club rides after every traffic stop. (Hated it but it really made me much stronger). CF is so easy to take off with.
I am not an Engineer and so dont know the reason, but I can say whatever flexie power transfer a steel bike provides, it is not as efficient as simply having a stiffer CF bike. My guess is that you probably lose more power from the flexing process than what you get back.
heat is lost, also the dead spot in the pedal stroke may absorb the energy, with less or no chain tension i dont think energy will come back to help you go
Sean Kelly won paris nice 7 times and a zillion other races. He won most of his races on a Vitus bonded aluminum bike that was a flexy noodle. Damon Rinard, top cervelo engineer and now at Cannondale says that they don't design the supersix for ultimate stiffness but ideal stiffness.
That way I look at it, there are two related dynamics to consider. Stiffness/amount of flex and flex/ rate of spring to true.
It could be that this is a dynamic that can engineered and likely already is with companies that have the budgets for 5hat kind of research. Wanting a stiff bike that allows a combination of transference of power and flex to allow spring to occur and engineering a high speed of recovery springing back to True. For custom builds this could be beneficial for weight and average load could be measured and taken into consideration when calculating rate of stiffness and spring.
I think many people confuse actual squat with flex. Meaning, sure too much flex robs you of your power when applying force and the whole system is loaded, but lacking flex at all you Rob yourself of your excess energy that is transferred and returned in the form of spring. Interesting video highlighting that.
Well it's just like a guitar string being struck. If the bridge is not stiff, if the top is flexy, it will dampen the vibration of the string, so the sustain will be shorter and it will sound less bright/dull. For a bike it's kind of irrelevant because you don't expect your pedal stroke to be sent back at you. The effect that could be seen is conservation of momentum. It seems to me that most of it is lost through friction both with the tire rolling resistance and the freehub, unless you're riding a fixie. I'd be ready to sacrifice a little bit of efficiency for more compliance and comfort. However when a frame bounces from left to right when you're putting a lot of power you also lose time because the rear wheel is not travelling as much in a straight line. So while power is conserved it's less used to propel you forward. In any case I've won a few town sign sprints with my old steel bike during group rides, so the effect is probably negligible most of the time.