carpediemracing 

There's a picture of a guy looking at an "email" from Specialized. Apparently he's a Cannondale employee.

Okay, it's funny, right? But it's also true, something I learned yesterday.

Specialized held a job fair in Bethel CT (location of Cannondale headquarters) shortly after that Canadian company bought Cannondale. Bethel CT is a few thousand miles away from Specialized, and other than a couple good bagel shops or delis (and a higher-end-than-average LBS), there's very little in that town.

Anyway, I thought the ad was funny, but it looks fake. I just wanted to put it out there that it's a real email.

fwiw I'm both a Specialized and Cannondale fan. Actually, I like Specialized in some ways a lot more than Cannondale, but right now my primary bike's a Cannondale.

cdr

seppomadness

Lets be honest they both suck. Its like comparing turds in the dunny. sh^t vs poo still = crap.

waterrockets 

C'dale engineers are the ones who told me the following:


Yeah, no straightforward way. Like using a PTap and an SRM on the same bike.

Why use math and the scientific method when you've got intuition?

asgelle

So now you have all the combined losses through the drivetrain. How do isolate the effect of frame deflection?

ElJamoquio

No kidding.

That said, even if the engineers aren't the best, they still make one of the best reasonable cost aluminum frames out there. And their manufacturing facility is in Bedford. Combine the two and I think I should still look for one...

Where is Specialized's engineering facility?

ElJamoquio

Easy, do the same tests on two frames; even swap the same drivetrain.

waterrockets 

Yep, if the only variable is the frame, you'll see the difference. The human's power output will be variable, but if you can't detect a difference through that noise, you really have no case for frame efficiency.

Psimet2001 

- Translation: "Good question. We have no clue. We sit behind our computer all day and read Bikeforums. BF says it's important and our Marketing department agrees. So we did a computer simulation and it said, 'yes.'"

Tell these guys to go ask anyone in their machine shop (might require an international flight) how rigidity affects power application.

Psimet "I like blanket statements and examples that are orders of magnitude different than what we're talking about" 2001

garysol1 

Morgan Hills California I do believe.

patentcad

It is hard to overstate the complete idiocy of this post.

Well done.

damocles1

So sayeth the Cannondale fanboy...

ElJamoquio

Hm? Google maps gets confused, guesses that it's just north of LA.

garysol1 

Is someone looking for a job? They do have a few job fairs listed on there website....

Campag4life

I am an engineer. You can interpret that he didn't know the answer but he does.
These are the reasons he didn't answer:
1. Doesn't want public record of engineering analysis to be used for competitive gain.
2. Doesn't want to be bothered...easier to blow you off.
3. You lump a given frame design into one a single frame. There is much more to it. There is delta in frame stiffness throughout the size range XS -XL to compensate for BB stiffness. Heavy riders statistically load BB's more. Large frames are inherently more flexible then small frames due to engineering dynamics without compensation in section modulus.

In summary, he will not provide Cannondale's accepted internal standard for BB stiffness as it relates to power transmission. Cannondale likely has as good of engineering as any top frame maker today. All their frames are modeled on CAD including FEA. No question they have std loading algorithims for each frame size to give them the companion stiffness they need. There is diminishing return for frame stiffness and that is weight. They know what they are doing but just don't care to share it with the public.

lgraba

An experiment to measure the effect of bottom-bracket deflection on drivetrain losses only works if you are measuring the total energy that the rider is putting into the frame. If the instrument only measures torque in the plane of the chainring (and I don't see how it could do anything else), then it is not measuring out-of-plane torque, which is what is deflecting the bottom-bracket from side-to-side. In order to do this, you would really need to measure the forces (and their directions) that the rider is exerting on the pedals through the pedal stroke, which is not so easy.

This experiment can only work if you can accurately measure all force/torque inputs and outputs.

Chris_F

That would only show you the efficiency difference between those two frames. One couldn't conclude that BB stiffness will always show the same result. For example, let's say you have 3 frames. One frame has a BB stiffness of "5", and two other frames (which are different from eachother) have a BB stiffness of "3". You measure the efficiency of the "5" at 99%, and one of the "3's" at 95%. Could one conclude that the effect of BB stiffness is equal to (99%-95%)/(5-3) or 1% for each point of BB stiffness? Perhaps. But in actuality probably not. So you measure the final frame and the efficiency is 80%. Now what do you conclude? The conclusion is that there are other factors involved how a frame effects efficiency beyond BB stiffness. How do you segregate these other factors from BB stiffness?

It's a non-trivial engineering exercise. Modeling the BB as a traditional spring/mass/damper (traditional dynamic systems approach), the stiffness is the "spring" measurement, but the efficiency is related to the other components of this as well. How do you segregate these equally importent components (hint: the mass in this scenario is not equivalent to the mass of the frame)? So although two frames may have identical spring components (BB stiffness) some may have more damping than others and thus not "give back" quite as much energy when flexed.

Also, one would expect the power transmission to vary depending on the rider. A rider who doesn't apply as much torque or load to the system will be impacted less by a flexy BB than a rider who does apply a lot of torque. A high cadence, low torque rider would be impacted less than a low cadence, high torque rider (of equal power). How would you expect an engineer to answer such a question then? For what particular rider does the answer apply?

Ask the engineers from the other bike companies and see if they come back with better answers than the C'dale guy...

waterrockets 

All we care about is weather the f#*&ing frame stiffness is causing a power loss. I don't care if it's BB deflection, down-tube twist, chainstay bending, or paint color. An SRM will tell you how much power is going into the frame, and a PowerTap will tell you how much is coming out with drivetrain losses.

Change frames, note that there is no difference in power, then live with marketing ignoring the facts for the next 20 years.

Phantoj

Won't disprove the "impedance mismatch" theory of power loss in a noodly frame, will it?


If you think your flexible flyer is more comfortable, I call that a marketing lie.


Furthermore:

I have known some excellent engineers that were incorrect about things. I have known two good engineers to hold opposing views on engineering questions - one or both of them must have been wrong.

Finally, I highly doubt that the "best engineers in the world" are designing bikes.

patentcad

He slammed Specialized too. Why wouldn't I like great racing bicycles?

ElJamoquio

I was told there wasn't going to be any math on this exam.

ElJamoquio

Incredibly competent engineers developed the concept of Control Volumes for a reason.

Fat Boy

Well, then, physician heal thy self. If you know how to do the engineering work, then do it, big boy.

Sometimes the best answer to give someone from an engineering point of view is, "It's not that easy". Of course, you don't want to hear that. You want an answer that is 10 words or less and without ambiguity like you get in the magazines.

OK, I'll give you want you want.

1. The stiffer the frame, the better.
2. Over a certain (rather low) nominal stiffness, the powertrain losses are negligible.

I don't know which one you want, but pick one and get over it.

lgraba

No, the SRM will tell you how much power is going into the frame in the plane of the chainring, it will not tell you how much energy is going into the frame perpendicular to that plane; the SRM does not measure that. Because the pedals are offset from the bottom bracket, they will exert an out-of-plane torque on it. If you are not measuring power input that does nothing to propel the bike forward, then you are not measuring an important source of work, the work to continually deflect the bottom bracket.

Look, I'm not against you doing your experiment, but if you are going to come to a conclusion about what the experiment means, you have to expect some scrutiny of your experimental approach, because a flawed approach can lead to erroneous conclusions. This is just part of the scientific method.

roadwarrior

First off, it's "whether"...all my engineering buds know the difference.

Frankly this whole thread is pretty funny.

I sincerely doubt there is a single poster on this thread that's good enough to be worrying about such things. Which after laughing is probably why the engineer answered the way he did.

Triguy

I don't believe I've ever seen frame stiffness numbers correlate to a faster frame. Though I think the first step in doing so would be to have two frames made with the exact same dimensions but different deflection numbers. After this, you'd have to make them weigh the same, probably put the same size water bottle on each frame, fill the lighter frames bottle accodringly. Build each up the same, everything.

Using any kind of powermeter test them on an uphill climb of x amount of minutes at a set power. Do 10 runs per bike. Compare resultant data. Once we prove that frame deflection does indeed lead to a certain amount of power loss, that it would be easier to quantify with frames already on the market.

I think that installing a PT and an SRM on a bike is also a good way however, there are a lot of variables. Chain changing, geometry, aerodynamics. These all effect your data. PT and SRM respond to changes in data differently based on their measurement techniques.