Campag4life

That's true Dean. I believe its faster. You may not. I believe for each pedal stroke more energy gets to the rear wheel with a torsionally stiffer bike.
In fact, if the industry didn't believe it, they would be less obsessed with it including the racer that works for Cervelo that just designed the new R-series who made a goal of making the frame and BB stiffer...when it was already pretty dam stiff. Frame stiffness has a cost for example, certainly in comfort as differential frame bending can not be completely divorced.

wphamilton

Drive train losses are proportional to power aren't they, in a given gear combination? Which you could relate to torque. Whereas frame losses would presumably be insignificant at lower power. So you could measure the drivetrain losses at low power and project those to the more extreme efforts.

Seattle Forrest

I don't believe a stiff bike is faster than a soft one within the range of modern road bikes. I do think an aero bike is faster, everything else being equal. If we all vote with our dollars, I've always cast mine based on how they ride.

​​​​​​I'd love to see the data, though.

Jack Tone 

I'm not an engineer, so I'm relatively stupid, but I always thought that energy can't be lost, only transferred to another form. My frames never get hot, or light up. I'm pretty sure that in a laterally soft frame, the energy is transferred in a larger arc of the crank as the frame unflexes. And as Nessism pointed out, Honda even found that when their road racing frames were too laterally stiff, the wheels wouldn't track well when leaned over almost to the sidewalls, though yes, that doesn't apply directly to bicycles.

Campag4life

Here's the problem, simply view the industry. Look at the boundary conditions of their $5K and up bikes. What is their priority? What gets discussed in bike reviews? I rode that bike. I could tell that every ounce of energy I put into the crank translated in propelling the bicycle forward commensurately. The bike is a rocket. That is what is stated.
Is this said about lower rung, lower carbon modulus frames with kiss your sister frame stiffness? No. The opposite is said. The frame lacks the response of a high level racing frame.

So whether skeptics here believe it...to me its plausible that a much stiffer bike transfers energy more efficiently. The bike 'doesn't push back on the rider' as has been conjectured when this is discussed in other threads which 'may' contribute more to rider fatigue...the hysteresis aka dynamic lag between input and output...if this energy is stored or pushes back on the rider. The industry does everything they can to make a laterally stiffer bike. Why in the world? Isn't too stiff bad when they can't get stiff enough? Not apparently.

Sorry, I explained why the motorcycle analogy of making a frame too laterally stiff doesn't apply. Different design boundary condition with a bicycle. Handling isn't the 'be all'. Power transfer and rider control is. Motorcycles have challenges on traction. Most riders can't burn the tire when exiting a corner or can corner at 1g with high slip angles and live to talk about it. Completely different design priority. Guess what? This is known. Why? Because of the sited GP motorcycle example. They made the frame too stiff for optimized traction. Makes sense to me. When is this 'ever' discussed with road bikes in the march to make them laterally stiffer in the last 10 years? Never. Only that stiffer is better.

Years ago when Specialized was making the best race bike on the planet...still argued they are in the Tarmac even tho aero is more in vogue and wins the day based upon particular riding venues especially, the SL2 Tarmac was considered a world class race bike. Of course Specialized wanted to make it better. This is about the time they were really learning about FEA and CAD application to bike frame design...and the SL2 Tarmac was even still an English threaded BB. So Specialized sought the advice from the pros. How can we make the SL2 better? The pros in chorus said, make it stiffer and more responsive. The bike is too soft. So Specialized did. They made the SL3 a fair amount stiffer. The SL3 Tarmac may be the most popular race bike of all time and in fact Simon Richardson who can own any bike on the planet and owns a bunch still has his SL3 Tarmac with Campy. So what was the feedback from the pros? Great job Specialized. You answered our prayers. But now you made the ride too dam stiff. What did Specialized do? They went back to the drawing board and softened the vertical flex of the SL4 but made it even more laterally stiff than the SL3. This of course engendered much praise for just how great the SL4 Tarmac is. Specialized was really on to something. Cake and eat it too. The bike out of the saddle accelerated like a rocket but wouldn't bust your balls like the SL3. SL5 and SL6? Even more vertical compliancy BUT even more lateral stiffness for even better power transfer.

Specialized who is probably the top or close R&D bike company on the planet, didn't have to make their top tier race bikes more laterally stiff with each generation. They could have made them softer with each generation or make them the same but they chose laterally stiffer because they believe this transfers energy more efficiently from cranks to back wheel.

Then you have BB to chainstay interface. Take the bike I just got for example. The new Cervelo R3. Stiffest 'its ever been'. Stiffest bike Cervelo 'has ever made' Design release: 2018. They believe today this is 'better'. Chain stays are massive. BB is a softball. Downtube is a log from a log cabin and is massive where it integrates into the head tube.

So Cervelo certainly believes stiffer is better for a race bike.

Skeptics can believe what they want all said and I agree the math or testing protocols to prove are complex. Coming from the world of R&D myself, I will tell you we often tried to computer model particular dynamics to save time on prototype testing and predict the future based upon changes to CAD but this many times fell short. Its not easy to do in many applications. But the engineers that design the bikes believe the design direction of the products they make and there has been a steady march to create laterally stiffer bikes for the last 10 years. In fact, if you want to look at another data point, look at the design evolution of aero bikes...what engineers have done to solve that riddle. When section modulus aka frame shape is the 'opposite' of what designers want for bike performance. Designers want wide and short frame members for lateral stiffness and vertical compliance and for aero dynamics tall and skinny frame members cut the air. Most that have ridden an aero bike can literally feel the lack of crispness to the acceleration compared to a TCR or Tarmac. But...designers have made huge strides to solve that riddle and now make most recent gen of aero frames more like conventional frames in performance, ergo more laterally stiff and vertically compliant.

Racing Dan

I think its a great idea replacing armchair debates with data. While not flawless given not only the frame would influence the results, its much better than nothing, and if not perfect still a great start.

Campag4life

Data is in. Test protocols will never be agreed so. Math and test strategies are an imperfect analog to reality. What is reality? How a bike performances on the road.

So what have top bike builders produced in the past fifteen years? Both soft and stiff carbon bicycles. What do ALL the top companies manufacture for their highest level racing frames? Stiffest frame in torsion they can produce. Why is that? Because the people that make them believe them to be faster. You many not believe it but I agree with them.

Certainly a stiffer frame isn't more comfortable...or generally not. Why wouldn't the big mfr's make a softer racing frame to promote more comfort and less rider fatigue because racing is a grueling test of endurance? Because stiffer is faster.

Racing Dan

Im not sure what data you refer to. I believe this is a reasonable test "protocol" that should be able to show any significant losses from the crank to the wheel in real world bikes, if any such losses exist.

asgelle

It's been done repeatedly and there's plenty of data. Losses amount to ~1-2% between pedal/crank and hub. While this includes losses to frame flex, from testing chains off the bike, we know that more than anything this test measures power lost to chain bending.

Racing Dan

OK? Please point me to a test similar to what is proposed.

asgelle

Just about every power meter review on dcrainmaker does it. Here's one. https://www.dcrainmaker.com/2018/04/...th-review.html

Racing Dan

? I have a hard time seeing how this is similar to what is proposed by OP ?

asgelle

It's a test "that should be able to show any significant losses from the crank to the wheel in real world bikes, if any such losses exist."

Racing Dan

But the test was supposed to compare losses in different frame materials. How is DC relevant for this. At best is shows differences between power meters in the same frame. That is hardly relevant.

asgelle

wphamilton

I think the experiment is feasible. Although the flexed frame may increase drive train losses due to altered orientation of the ring with respect to the cassette gears, which would have to be separated from the losses due to the frame itself. I ALSO speculate, without any hard data, that there are losses from extra tire deformation when flex causes the wheels to orient slightly off-line. This additional loss would not be captured by the experiment.

Alternatively, insulate the frame and measure the temperature of the metal.

zacster

Hey all. I just proposed a simple, real world, easily performed test using equipment that is readily available. There was no argument of stiffer is faster. The test is easy, put a Powertap wheel and a pair of power meter pedals on a bike, measure the power outputs of both, repeat with the same meters on a different bike and compare the ratios. Keeping the courses and the rider the same would even things out. Out of necessity the drivetrains will be similar since they need to be compatible with each other due to the swapped wheel. Repeat the test a number of times over different days to take rider fatigue out of the equation. Compare similar sections of the ride, hills vs. flats, sprints vs. tempo. It doesn't have to be perfect and we aren't going to change the industry by doing this. Maybe we'll get results that show something, maybe not.

wphamilton

The major (real) objection is that you'll be measuring drive train losses and not frame flex losses, because the latter is much smaller and will be lost in variable "noise".

You are going to need to stay in the exact same gears as well, because the 2% drive train loss on derailleur bikes is best case. it can range up to 4-10% depending on the gear and how well maintained the chain is. If your experiment measures losses of 2% plus or minus an error, the frame flex may be inside that "error", and therefore not measured. It needs some valid way to subtract out the drive train.

Jack Tone 

I think the key words here are "response" and "feel". No doubt a modern Specialized Tarmac is more responsive than an old Vitus. But power transfer? I'd wager that if you put a the same sprinter on both bikes, they would get to the finish line at the same time, but the Vitus would "feel" terrible.

Dean V

Absolutely.
When the fat tube Cannondales came out which were hugely stiffer than the Vitus and also steel bikes of the period did anyone seem to notice an actual performance gain?

Trakhak

Unsolicited testimonial from one of Italy's all-time greatest sprinters! What more could anyone want?

zacster

And if we keep drive train the same? Of course even identical components won't be the same after installation. We aren't measuring drive train loss anyway, but the change in drive train loss between bikes. That may still be well above any measurable frame loss.

Mark Manner

Hi,I don't want to get into an argument with anyone (there is a lot of fairly harsh back and forth here). But I would like to share my limited data set experience using powertap hubs and powertap P1 pedals on several different bikes. Note that the PT hubs are in different wheelsets (Zipp 404 tubular, Enve SES 4.5AR, ENVE G23, Reynolds Aluminum (a couple of those), Zipp 303 clincher). Two different sets of P1 pedals, and on a few tests recently some Garmin V3 pedals. I always calibrate my hubs and pedals before riding. I haven't been doing this to test frame materials, but I have done many rides with one or the other of those wheelsets, sometimes in steel frames (Tommasini and Gunnar), carbon fiber (Colnago and Wilier), aluminum (Bianchi), and titanium (Litespeed, Moots, Mosaic, Habanero). The cranksets have varied, including shimano Ultegra and DA, Campy Record, Sugino, Ritchey, FSA, maybe others. I have noticed on all the rides on any combination of these components over the past few years that the power meter pedals measure 1-2% more power than the hubs (they also do when put on a Tacx NEO trainer). Since the differences are pretty small, 1-2%, if there is any difference in any of the frame materials, hubs, cranks, chains, whatever, it is small enough that I don't see any pattern no matter what combination I have used. Given what I have seen, it may be that the resolution of consumer power meters may not be sufficient to show which frame is more efficient from a power transfer point of view. Please note that I didn't collect any of this info to prove anything, in particular anything related to frame materials. I just did it to see what the power differences were between the hub units and the pedals also know that I didn't use the same PT hub/cassette/chain/crankset/ chainline, etc. etc. However, I did (and do) consistently see the 1-2% more power at pedal than at hub that is predicted (and should be expected), and it is always in that same range no matter the combination. I have done it enough that I think I would have noticed a difference between frame materials if there was one that was large enough when compared to the drive train losses that I seem to be measuring. I offer this purely for what it is worth. My experience may be different than others.
Best,
Mark

wphamilton

I'm not sure if you're seeing it as much of a distinction, but you are measuring the drive train loss on each bike and then comparing, and not the delta between bikes. I do see that as kind of important, since you actually have two different sets of measurements each with it's own error margin. If you were actually measuring the delta, that would be more accurate.

I would doubt that any differences in the chainring and cogs would change your results much as long as they're the same size, and the chain lines are equivalent. Use the same chain on each bike perhaps. In my mind you have more of an issue with the small differences that you'll measure, and also with the precision of the meters. I think - maybe - you could measure the drive train loss at low powers (where the frame isn't flexing), and use that same percentage for higher power when the frame does flex. Just subtract it out. But that needs to be verified somehow.

Elvo

This is the fastest bike with the least "power loss":
https://feltbicycles.com/pages/pursuit-for-gold