uhm I was really wondering if you'd thought about the what the CdA results implied about the repeatability of the PM (not accuracy). Just a sense check you know eh? I'm not good at dis English stuff.

Damn you.

It appears that on occasion the results are sensitive enough that one can diagnose problems with the PM. Alex mentioned that his buddy had a problem with the cover screws in a track hub. I've noticed a difference in a PT Pro hub that varied a little with clamping force -- it turned out my cone needed adjustment.

Right. It's not the math so much as the constraints.

I knew I could do that for an SRM but I didn't know it was possible to put another magnet inside the PT. Where? How? I'm intrigued.

Yes. In fact, there was a second P3C run in this session (after the P2K run) that I did that didn't make any sense elevation-wise until I increased the Crr to .006 from .0056. As I related to Andy and Robert though, I had possibly "over-tightened" the skewer when swapping the wheel back into the P3C for this final run. Also, since it was getting later in the morning, I was starting to experience occasional automotive interference on the venue, and later after realizing how sensitive the runs had been to ambient wind conditions, I didn't think that run was worth presenting. The calculated zero yaw (again it was under dead calm conditions...except for the cars, of course) was on the order of .215 to .218 for that run. I was hoping it would be a good "confirmation" run as I've experienced at my other venue, but I really don't have much confidence in the validity of that particular run for the various reasons noted above.

This is my first experience with a PT SL hub (mine is an older Pro model) and I have to say I'm not very impressed by the lack of adjustability/locking on the bearing preload. When I first borrowed the wheel, I noticed that the NDS axle nut was loose. Expecting there to be some sort of locknut, I was surprised when I researched the bearing adjustment and found that the factory recommendation is to just torque each side to 12 N-m, and to use some loctite to prevent loosening. Even though they're cartridge bearings, you still need to be able to adjust the preload correctly, right?...especially if you want to account for skewer tension when it's installed.

On the question of repeatability...yeah, I think the PT hubs are a lot more "repeatable" than the quoted precision value. I also made sure to zero the torque before each run. I'm sure that helps.

---

http://bikeblather.blogspot.com/

The magnet is embedded in the large nut/spacer on the left end of the axle, and the reed switch rotates around it. It therefore seems to me that it would be a fairly simple proposition to just bore another hole in that spacer and insert another magnet. I haven't tried this myself, though, so don't know if the reed switch would be happy with such an arrangement.

Cars and aero field tests simply don't mix. Indeed, I suspect that some of the residual variability in our testing out in LA is the result of having two cyclists on the track at once (although we were still able to detect a difference between a LG Rocket and a Bell Meteor II helmet).

okay just keeping you guys honest eh? :-)

One thing I'd thought about was changing the wheel rollout value in the PT head. I can double it -- but then I'd have to make sure not to exceed 50 km/h. I haven't done this, though.

According to this article http://www.tritopics.com/...ycling&Itemid=53 there could well be different rolling resistances between the p3c versus the p2k based on differences in vertical compliance of the frames themselves ( In spite of testing with the same wheel, tires and PT.) This would suggest that some of the differences measured are not purely related to aerodynamics.

what are your thoughts on this?

---

-----------------------------------------------
www.true-motion.com Triathlete Casual Wear since 2007
(Twitter/FB)

True. But to be honest, in the past on my original venue I didn't seem to be bothered by an occasional "car on course". Of course, that could well be because of that course's longer lap length, greater elevation difference, and higher overall speeds. So, like the wind sensitivity issues that appear to be magnified on the "mini half-pipe" course, I was a bit surprised at how the car interference apparently affected the data quality in the second P3C run. Robert thought the first P3C run looked "less consistent" than the P2K run...the second P3C run is even less consistent.

It's a good thing I started so early and got in the first 2 runs before the neighbors woke up and started going places :-)

---

http://bikeblather.blogspot.com/

"Check this out though...so I took a look at AFM's chart. I looked at the percent difference between my tires and the Vittorias and adjusted the .0038 Crr. I then looked at the percent change from latex to butyl and adjusted it again. That gave me an assumed Crr of .0056. When I put that into the spreadsheet, the calculated CdA for the zero yaw condition worked out to be...ta da! .228 m^2, just like the Piru data. Pretty cool, huh? Shows pretty clearly to me the validity of Al's testing."

And the validity of Tom's method for converting roller to flat data.

Thoughts?

My initial thought after reading just the first sentence, i.e. "Rolling resistance is the amount of energy required to overcome the friction between the road and tire.", is that whoever wrote that needs to read the article I wrote for Slowman to become better informed :-)

http://www.slowtwitch.com/...ling_events_226.html


My second thought is that the main "suspension" of the bicycle is the tires themselves. As long as they aren't overinflated and defeating the purpose of being a pneumatic tire, the compliance they have is much higher than any other part of the structure. When you put springs in parallel, the combined spring constant is going to be closer to the spring constant of the weaker spring element. The vertical compliance in this case (with rigid frames) is nearly completely driven by the tires.

I'm not seeing how any differences in the vertical compliance of either of those two frames in that test can make a measurable difference in the total "resistance to forward" motion.

---

http://bikeblather.blogspot.com/

That's part of the problem that I have in trying to find a good place to do such testing: anything rural enough to have limited traffic is too rural to have any short loops. (The other problem is terrain: almost every possible loop is too hilly to avoid braking for corners, or at least deviating from a perfect aero position.)

"My second thought is that the main "suspension" of the bicycle is the tires themselves."

you have a great point there - the article was, after all using a softride as an example which is of course a different kettle of fish...

now I'll go read your article and edu-ma-cate myself...thanks.

---

-----------------------------------------------
www.true-motion.com Triathlete Casual Wear since 2007
(Twitter/FB)

Instead of loops...what about "out and backs" like I've been using? Just find a road between 2 hills in a rural location...

---

http://bikeblather.blogspot.com/

That would require driving my bike, etc., somewhere...in which case, I might as well stick to my "natural wind tunnel".

Well...depending on what all you want to test out, that might be an advantage in that you can carry more tools, equipment, etc. with you.

Just trying to help...

---

http://bikeblather.blogspot.com/

The only real advantage of Robert's method is that you don't need to know the exact elevation profile of a course. It would therefore be quite handy if I could do as you did, i.e., make the measurements within riding distance of my house. Unfortunately, I haven't been able to locate a suitable location to do so...which means that I might as well keep on using the same stretch of flat road that I have been using for the past few years.

Who are you and why are you posting under Andy Coggan's name?

Yep, with most cyclecomputers you get speed with accuracy of about 1 part in 400 (0.1km/hr resolution divided by 40 km/hr speed). This has always struck me as unfortunate, because the computer is just counting revolutions and measuring the time delay, and it's easy to measure time to a very high level of accuracy.

Now V = dx/dt, and of course dx is the tire circumference, which I think you can measure accurately to say +- 0.5 mm. So that gives an achievable accuracy of roughly 1 part in about 2000, about five times better resolution than the computers report.

I've tried the multiple magnet approach and never had very good luck.

Another approach I've tried that works quite well is to use a small digital voice recorder. Just cut off the external microphone and splice a reed switch in. The resulting sound file is just "tick, tick, tick", and can be run through a simple script that pulls off the time interval between ticks. The hassle then comes in synchronizing the sound file to the Powertap file. Basically, the sound file gives you speed at fixed intervals in space (wheel circumference), while the PT file gives you data at fixed intervals in time (1.26s). It's doable and I've played with it quite a bit, but I haven't yet tried it on a real field testing file.

BTW - I suspect that the ibike uses high resolution speed (for power calculations) but the data you can download from the device (and view) are of limited resolution.

-Eric

I think Al's cat might've hacked into Andy's account...damned cat! :-)

---

http://bikeblather.blogspot.com/