Using PowerTap test to decide a new frame?

I don't think you can. The range of variation in your sprint power is probably greater than the range of variation in frame stiffness. Add to that the fact that PT's (IIRC) aren't particularly precise at the shortest durations, and "meaningful & repeatable" moves out of reach.

If I were overly concerned with stiffness, I'd look at a) lab test numbers from Tour Magazin and b) how each frame felt when I sprinted (using same wheels, tire pressure, bar, stem, cranks.) Not sure you can find a shop that will let you do the complete version of b) though, so I'd settle for same wheels at same pressure.

FWIW, Time frames are noodles in the lab. In your price range I'd look at Canyon, Storck, C'dale System 6, Cervelo R3, and custom builds from Crumpton or Ruegamer.

Yeah, but assuming you could find a route that would give you basically the same conditions, you could pedal at 250 watts for say 30 minutes and then look to see exactly how many feet you went. Do the test multiple times for each bike. If one bike consistently travels further, then you know it's transferring more of your energy into forward movement.

donrhummy said:

Yeah, but assuming you could find a route that would give you basically the same conditions, you could pedal at 250 watts for say 30 minutes and then look to see exactly how many feet you went. Do the test multiple times for each bike. If one bike consistently travels further, then you know it's transferring more of your energy into forward movement.

Assuming no changes in wind, air density, passing traffic...... Try some outdoor CdA regression testing and you'll see how tricky it can be to remove environmental variables.

vladav said:

on a flexy frame

How do you know your frame is flexy?

beerco said:

How do you know your frame is flexy?

Well it could be a combination of components, but it's an '05 Trek 2100. alum triangle + carbon stays. A $1200 18.9lb road bike can't be all that rigid - can it?

Also, I notice on the trainer the front wheel rolls about an inch forward with a shift in weight.
Also out-of-the-saddle in a sprint, I can feel the frame snap back after each pedal stroke. My arms have to do an exaggerated pre-torque of the bars to counteract the pedaling forces, and then act as shock absorbers mid-pedal stroke because everything is springing back.

Dave

vladav said:

Well it could be a combination of components, but it's an '05 Trek 2100. alum triangle + carbon stays. A $1200 18.9lb road bike can't be all that rigid - can it?

heavy and stiff is very easy to do. Light and stiff is not. Your frame is probably quite stiff.

vladav said:

Also, I notice on the trainer the front wheel rolls about an inch forward with a shift in weight.
Also out-of-the-saddle in a sprint, I can feel the frame snap back after each pedal stroke. My arms have to do an exaggerated pre-torque of the bars to counteract the pedaling forces, and then act as shock absorbers mid-pedal stroke because everything is springing back.

Dave

For $1200, I doubt your bike has a carbon steer tube, but that is a source of flex that many can misinterpret as frame flex. "Feel" is not something that we can use to determine how flexy a frame is.

donrhummy said:

Yeah, but assuming you could find a route that would give you basically the same conditions, you could pedal at 250 watts for say 30 minutes and then look to see exactly how many feet you went. Do the test multiple times for each bike. If one bike consistently travels further, then you know it's transferring more of your energy into forward movement.

I like the theory behind your idea; however that would be a design for measuring a frame for the steady state forces of a TT.
I'm pretty sure a thrashing violent sprint would require a different aproach.

And as Dave R W pointed out there's too many variables over that time frame. (And possibly same for what I want to do too...)

Dave

Oh yea, and to actually answer your original question, power loss due to frame flex couldn't be more than a few watts (less than one I would expect). It's totally impractical to measure it at the rear wheel.

beerco said:

"Feel" is not something that we can use to determine how flexy a frame is.

Agreed - which gets at the heart of this thread 😎

beerco said:

Oh yea, and to actually answer your original question, power loss due to frame flex couldn't be more than a few watts (less than one I would expect). It's totally impractical to measure it at the rear wheel.

From where do you arrive at that number?😕

vladav said:

From where do you arrive at that number?😕

It's an educated guess based on my experience and a degree in engineering.

Remember that power is work over time. Work = force x distance. So, even though forces are high when sprinting, most of that force goes into turning the cranks so maybe your frame is flexing what, 5 to 10mm with every stroke absolute tops?

So you've got power lost due to frame flex = Force(high)xDistance(very very small)xTime (slow, even at 120rpm) = a low number. Tire flex probably accounts for a lot more power loss. There's also speculation that at least some of the power lost due to frame flex would be returned when the frame flexes back.

I believe if you search on biketechreview.com archives you'll find some discussions using actual numbers.

edit: p.s. I haven't actually run the numbers so I will grant that I might be wrong and it could be as many as like 5w but I would still guess that it's closer to less than one.

beerco said:

It's an educated guess based on my experience and a degree in engineering.

Remember that power is work over time. Work = force x distance. So, even though forces are high when sprinting, most of that force goes into turning the cranks so maybe your frame is flexing what, 5 to 10mm with every stroke absolute tops?

So you've got power lost due to frame flex = Force(high)xDistance(very very small)xTime (slow, even at 120rpm) = a low number. Tire flex probably accounts for a lot more power loss. There's also speculation that at least some of the power lost due to frame flex would be returned when the frame flexes back.

I believe if you search on biketechreview.com archives you'll find some discussions using actual numbers.

edit: p.s. I haven't actually run the numbers so I will grant that I might be wrong and it could be as many as like 5w but I would still guess that it's closer to less than one.

Fair enough - Thanks for the input!

beerco said:

So you've got power lost due to frame flex = Force(high)xDistance(very very small)xTime (slow, even at 120rpm) = a low number.

Before anyone yells at me too much, this should be:

Force(high)xDistance(very very small)/Time (slow or relatively large, even at 120rpm)

Sorry for the error.

peterpen said:

I'd look at a) lab test numbers from Tour Magazin...

After spending about 2 hours looking for it, I have to ask how I get a copy of that?

Found www.tour-magazin.de site in German > translated on babelfish > still clueless. Of course googled too...

Some places to check:

http://weightweenies.starbike.com/phpBB2/viewtopic.php?t=25692&highlight=tour+tests

and

http://fairwheelbikes.com/forum/viewtopic.php?t=2617

i reckon the only possibly useful test would be to have a powertap and srm on the bike.

then compare the relative difference in power between the two meters, on different frames.

if there were any measurable loss between the cranks and the hub due to flex, then maybe this would find it.

I've got a 15 year old steel frame set up identically to a new (claimed to be) uber stiff carbon frameset. Same post, crank, bars, stem, pedals, seat, wheels - use the same PT SL. Bike position is set identical on both bikes: saddle hight and setback, bar drop and reach. (give or take a few mm)

Doing sprints in training, I've never come closer than 80 (peak) watts on the steel bike to what I do on the carbon bike. Could be coincidence, random varience, losses in frame. Who knows. But the numbers are consistantly higher by about 80 to 100 watts on the carbon bike.

I'm an engineer too. 🙄

robkit said:

i reckon the only possibly useful test would be to have a powertap and srm on the bike.

Unless the difference is due to pedaling ergonomics, not inefficiency in the mechanism.

Presumably, a flexy frame would have a higher max pedal velocity, and therefore a lower max torque (aren't pedal velocity v. torque charts roughly linear?) and potentially a lower power per stroke.

HammerHead said:

I'v But the numbers are consistantly higher by about 80 to 100 watts on the carbon bike.

I'm an engineer too. 🙄

So do you really think that the frame is sucking up 80w? (seriously)

Is there something wrong with my gedanken experiment?