I’m picking up my new 808 clincher front wheel today from B&L. I will mount up a broken-in ProRace clincher, size 23.
I have a 404 with a Tufo S3 on it. No sealant is in that tire, and it has about 250 miles of racing on it.
The tires are almost the same actual width, and I’ll put 120 pounds in each one. The road out there is pretty smooth.
I’m going out to Fiesta Island this coming Saturday morning, if it’s not windy. I’ll do two loops with each wheel, and then repeat. Two loops is 5 miles. I’ll hold about 200 watts – what a good AG’r would do in an IM.
I’ll download the data and report back. I’ll re-read a Zipp article I have somewhere to give the appropriate amount of handicapping between the 404 and 808.
Will your tires weigh the same? I believe they should to even out any affects of rotational weight on your test. Perhaps you should add enough sealant to one of the tires to make them even in weight.
Hey Ashburn, have you considered doing this test at a velodrome?
Nah. I’m after the low-tech version. We’re looking for big differences here. I’ll wait for the data, but if the results end up pretty close, I have little interest in divining things down to the exact watt. People on the board want to know if these tires are “close enough” or are they “really, really different.” I would guess that I want to see at least a 5-8 watt difference in order to form any final judgments.
5 watts and I wouldn’t buy new tires and certainly not new wheels. 20 watts – now we’re talking about making wholesale changes. 20 watts in an IM bike is 10-12-15 minutes (depending on how fast you go).
I’m picking up my new 808 clincher front wheel today from B&L. I will mount up a broken-in ProRace clincher, size 23.
I have a 404 with a Tufo S3 on it. No sealant is in that tire, and it has about 250 miles of racing on it.
The tires are almost the same actual width, and I’ll put 120 pounds in each one. The road out there is pretty smooth.
I’m going out to Fiesta Island this coming Saturday morning, if it’s not windy. I’ll do two loops with each wheel, and then repeat. Two loops is 5 miles. I’ll hold about 200 watts – what a good AG’r would do in an IM.
I’ll download the data and report back. I’ll re-read a Zipp article I have somewhere to give the appropriate amount of handicapping between the 404 and 808.
Any other protocol suggestions?
My suggestion: give it up. While I appreciate your efforts, I really think you won’t be able to distinguish any effect of the differences in the two wheels given all the noise inherent in your trials (holding wattage and position constant, differing wind/temp.humidity conditions, etc.). Do you know that 120psi is the optimal pressure for both tires to minimize rolling resistance? Did you read Andy Coggan’s attempt to discern the differences in aerodynamics between two frames (http://www.biketechreview.com/aerodynamics/aero_frame.htm)? He did many more trials, and eventually determined that the differences appeared to be within error tolerances.
Having said that, you might want to check out http://www.analyticcycling.com/ForcesTires_Page.html, which allows you to see the differences between tires of different types. Perhaps you will see a significant difference (as seen between “Premium Clinchers” and “Premium Tubulars - Road Glue”)!
I think that this is exactly what we are after. I will try the roll down test that was outlined in my earlier post and report back the data. This should give us a few real world data points. I would offer to drive down and help but SD is too far for me to go on Saturday.
I believe they should to even out any affects of rotational weight on your test.
This isn’t a troll is it?
Rotational weight will have no measurable effect at the precision levels we are using in this test.
No, not trying to troll, but I also believe you have too many uncontrained variables to convincingly attribute any difference in runs to rolling resistance (similar to what Ken stated). I was simply trying to convince you to eliminate one of the variables. 5 unmeasured variables that have “no measurable effect” can cumulatively result in something that does have a measurable effect.
I really think you won’t be able to distinguish any effect of the differences in the two wheels given all the noise inherent in your trials (holding wattage and position constant, differing wind/temp.humidity conditions, etc.). Do you know that 120psi is the optimal pressure for both tires to minimize rolling resistance? Did you read Andy Coggan’s attempt to discern the differences in aerodynamics between two frames (http://www.biketechreview.com/aerodynamics/aero_frame.htm)? He did many more trials, and eventually determined that the differences appeared to be within error tolerances.
If the differerences fall into the narrow range Andy experienced, yes I will give it up by reporting “no conclusion.” I’ll work with some folks off-line to nail down the confidence interval properly. Again, if we can’t tell a difference at an appropriate confidence level, I’ll say so. Just because two bike frames are very close doesn’t mean two other components are close as well. Let’s let the data do the talking. If we treat messy data as messy, we can still get meaningful results. “Messy” does not always mean “inconclusive.”
BTW – there is no need to hold watts constant. I’ll stay in a range, but we will normalize the data for wattage variations across the runs (a routine and robust procedure). The main thing I need to control is to hope for no wind or traffic – or at least light wind that doesn’t change throughout the runs. I’m pretty good at staying still on the bike, certainly well within the big wattage numbers we are looking for.
I’m choosing 120 because that’s what virtually all tire makers that actually test such things (which is rather few of them, I think) say to use out on real-world roads. But again, we’re looking for big, chunky differences here. We want to know if we can reject the null hypothesis. Secondarily, we want to know how big the differences are, if any. We’re not trying to nail the actual rolling resistance numbers, only the differences.
The null hypothesis is this: “Tufos and Michelins have the same rolling resistance.” Can we, with confidence, reject that statement? That is the question, and I’ll try to find an answer.
BTW – if any San Diegans want to join me, I’m planning on being there at 6 am on Saturday morning. Sunday is the backup if Saturday is raining or windy. The weather forecast looks iffy.
Y’all are missing the point of this test. The point is to determine if the Tufo’s are so significantly different that all those other variables mentioned are insignificant in comparison. If the test yields results that are within a margin of error then that answers the question.
Andrew’s test of aerodynamics is relevant because the conclusion what that the frames were similar enough as to not care. Can the same thing be said about Tufo’s. To some degree we don’t know yet. Some evidence suggests there is a bigger difference there.
Personally, I don’t care which tire is better. What I care is if Tufo’s are significantly slower. Slow enough to justify changing tires. If the results are close then I’ll just continue on my way and worry about other things.
Can you explain how you will normalize for wattage differences? I think it is hard to hold wattage within 3%, whiich at 200W is within the bounds of what you are trying to measure (a 20W difference) but still very significant. Are you swapping rear wheels, and thus powertap hubs? Make sure you look at elapsed time difference (ideally from a flying start) rather than average speed to eliminate any differences in wheel diameter.
I actually think it’s worse than that. I mentioned this on another thread: fairly small changes in body position can change the weight distribution over the wheels, which would really skew the results. I think a fair test really needs to swap both wheels out.
I actually think it’s worse than that. I mentioned this on another thread: fairly small changes in body position can change the weight distribution over the wheels, which would really skew the results. I think a fair test really needs to swap both wheels out.
-jens
Since that means swapping out the powertap, is an SRM is the only way to do this test?
Not sure if you are aware, but two trips around Fiesta are never the same, unless you ride the same exact line you may be off by meters. There is a way to ride that course that I learned a year ago from a crusty old masters racer, and cut 5+ secs off my lap split, no kidding. He showed me which lines were the fastest, which I never thought about.
Not sure if you are aware, but two trips around Fiesta are never the same, unless you ride the same exact line you may be off by meters. There is a way to ride that course that I learned a year ago from a crusty old masters racer, and cut 5+ secs off my lap split, no kidding. He showed me which lines were the fastest, which I never thought about.
I won’t be using lap times – I will use the powertap distance data, and do a roll-out measurement on each wheel+tire setup. We can then clean up any differences in circumference after the fact. Again, we are not trying to determine the exact rolling resistance figure, just the difference.
Rick M. and Jens make good points, and I need to look into that further.
I actually think it’s worse than that. I mentioned this on another thread: fairly small changes in body position can change the weight distribution over the wheels, which would really skew the results. I think a fair test really needs to swap both wheels out.
-jens
Since that means swapping out the powertap, is an SRM is the only way to do this test?
So, you know, after all this work, you will still be forced to ask: “what % of the measurable result was due to the type of glue used on the tubular?”
This is only one of the MANY variables you aren’t controlling for. Not trying to troll here, but this is about as problematic a test protocol as one could draw up.
I raise the issue of the glue, because this is one of the documentedly important variables in tubular performance, and you don’t even mention it. A lot of the controversy regarding Tufo rolling resistance centers around their tape, which appears to be very high resistance - are you using it? If not, what are you using? What is the “dry” viscosity of the glue you are using? What tube are you using in the clincher? Someone already mentioned the difficulty of maintaining position discipline in a test like this… it is nearly impossible to maintain this discipline in a rolling test where you are looking down at a power meter readout… unless you choreograph your “looks” at the meter, you will eat your margin of error/variability up simply by extra helmet drag from looking down…
Blah, blah, blah, etc…
This might be fun to do - but you aren’t going to come up with useful data; I doubt that you would be able to discern a difference of < /= .1, which is what you will be looking for. Too much “noise.”
I hate it when I try to do a quote on this BB and my response gets lopped of the bottom. Am I the only one this happens to?
Rick:
I think the way you are proposing to test is still OK. It’s just that the differences will be pretty small, possibly approaching the margin of error in testing. RR increases with weight; the lighter the load on a wheel, the smaller the differences will be between good and bad tires.
If you do run the test, I’d be sure to be on the aero bars all the time. Not only will it put more weight on the front, it will also minimize position shifting that might affect aerodynamics and front-rear balance.
If there is wind at significant yaw, all bets are off, since the 404s and 909s perform quite differently in those circumstances.