I didn’t change anything in my setup. What are some of the variables that could change this? Wind, not following the exact same line, or different speeds at the turn around?
I’d like to add a sub-point to the “Don’t use your brakes”. Make sure your brake isn’t rubbing! I did some testing where I switched from my H3 to a disc to try to test the difference. The results with the disc were really bad, which was confusing (roughly from .23=>.27). I then noticed that when I spun the wheel freely in my bike stand that it was very slightly rubbing at one point. It was subtle enough that I didn’t notice when installing the wheel or while riding (perhaps I shouldn’t listen to music while testing). Another day of testing bites the dust 
I’d like to add a sub-point to the “Don’t use your brakes”. Make sure your brake isn’t rubbing! I did some testing where I switched from my H3 to a disc to try to test the difference. The results with the disc were really bad, which was confusing (roughly from .23=>.27). I then noticed that when I spun the wheel freely in my bike stand that it was very slightly rubbing at one point. It was subtle enough that I didn’t notice when installing the wheel or while riding (perhaps I shouldn’t listen to music while testing). Another day of testing bites the dust
OTOH, it shows that the method can pick up pretty small effects.
I’ve got a bit of a mystery that I’m curious to get some feedback about. I did some field testing today and got some results I don’t believe. I’m looking for potential holes in what I did as well as a question of plausibility. My goal was to try to identify whether my desoto liftfoil trisuit or my new cycling team skinsuit was faster. I was expecting that they would be very close and guessed that maybe the trisuit would be a touch faster. I warmed up by doing some testing of two different helmets with so so results, it was hard to distinguish them, but I felt ok about my test setup. I had my race setup, except I was using my powertap rear wheel which has slow tires (it was on the trainer before). I was getting speed and power from the powertap, turning off the gps for speed on my 705 (I have had trouble in the past with my gsc10 sensor and quarq). I did three runs where each run went about two minutes from home to the start of my halfpipe, did two laps of the halfpipe and then went home. I didn’t turn off the garmin at home, just put down the bike, changed trisuit for skinsuit, and did the next lap. The results in aerolab look very clean. I was just looking for deltas in CdA, not absolutes, so I just left Crr as 0.005, although I did put in reasonable values for mass and rho.
-) Run 1 and 3: matching CdA of ~.217 in trisuit
-) Run 2: CdA of .192 in skinsuit
I stupidly didn’t do a 4th pass in the skinsuit, so I assumed that there must have been some error. After lunch, I decided to give it another shot. This time I got:
-) Run4: CdA of .193 in skinsuit
-) Run5: CdA of .215 in trisuit
The runs also look great in aerolab and match the first runs closely. So is it plausible that there is that big a difference between the skinsuit and my trisuit. Is there some other error that might manifest itself in such a way that my runs match each other so closely? For reference, the skinsuit is the one I’m wearing here. I didn’t think it was supposed to be exceptionally fast, it was just the one the cycling team I joined had in stock.
My first instinct is to ask whether you’ve adjusted your air density according to the time of day. Although they track each other fairly well over the course of the runs (skinsuit vs trisuit relative difference). The absolute CdA you’re quoting is probably a ways off, though.
As far as the delta between skinsuit and not…we’ll have to let some of the more experienced guys come in. 0.022-0.025 change in CdA is huge, though.
Just to be clear, runs 1, 2 and 3 were done over a total duration of 35 minutes, with 31 of those riding (just about 2 minutes per change of clothes).
Run 4 and 5 were done two+ hours later where the temperature had indeed risen by a few degrees, but they were also done within a 23 minute window (21 of those riding).
I’m sure the absolute CdA is inacruate which is why I mentioned that I was just looking for relative results. I also agree that >0.02 CdA change doesn’t seem right which is why I redid the test assuming a bug and then decided to ask here after my second test confirmed the results.
I realize I should also have put a picture of the trisuit. This is a slightly older picture of what that looks like. I certainly wouldn’t have thought it was a dog. The only thing that comes to mind is that the stitching is pretty raised on it. Otherwise it is nice and tight, etc.
FWIW, when I visited the wind tunnel last year we did some skinsuit comparisons. There were several riders and one brought a team suit that had raised stitching in a creative diagonal pattern that looks similar to your tri suit. It was a full skinsuit but the stitching was similar. It was clearly the highest drag skinsuit of all tested that day and added more than 200 grams of drag relative to faster suits and IIRC nearly 300 grams of drag relative to the Castelli bodypaint suit. The best guess of the testing folks was that those raised seams were in all the wrong places in terms of tripping the air flow and though they look kinda nice they may not be the best thing from an aero standpoint.
BTW, what are you assuming for Crr in your Chung testing? Are you basing it of roller testing or regression testing done in another session or just swagging it?
-Dave
The Crr is just a swag. My front wheel is a Jet9 with a GP4000s and latex. The rear is an American Classic 420 with a powertap and some cheap “strada-k” tire with butyl that I use on my trainer. I’ve got a Jet disc that I’ll use racing, but I was hoping the PT would give nice even results for the field testing.
If I remember my rules of thumb, its 50 grams of drag = 5 watts = .005m^2 CdA @ 30mph. So 200-300 grams would be .02 to .03m^2 CdA. So what you are saying is that it is not on the face of it absurd that there might really be that big a difference between my skinsuit and trisuit?
If my data is right I am sad since I really like my DeSoto stuff and would be sad that it is slow. I’ve certainly tested it in the pool and it is pretty fast there. I was just assuming it would be fast on the bike too
If my data is right I am sad since I really like my DeSoto stuff and would be sad that it is slow. I’ve certainly tested it in the pool and it is pretty fast there. I was just assuming it would be fast on the bike too
How high does the zipper go on the front of the skinsuit and the trisuit? Is the area around the neck equally tight?
I’ve never tested a sleeveless trisuit so I don’t really know but a difference of .02 m^2 is bigger than I would have thought.
The short answer is: absolutely. We’ve seen this happen a couple of times. To quote Paul Harder of Trek, “Skin is slow”. 
I know that the TriTalk podcast did some roll down aero testing of a few tri suits a while back and that of the 3 or 4 tested (not a huge sample or a controlled test, I know) the DeSoto LiftFoil tested the best. I always assumed it was a little faster than other tri suits because the stitching all runs parallel to the direction of the wind coming over the back and there is no pocket sitting perpendicular to the airflow on the low back to trip it up.
So, as far as tri suits go, I think you could do a lot worse than the LiftFoil.
If my data is right I am sad since I really like my DeSoto stuff and would be sad that it is slow. I’ve certainly tested it in the pool and it is pretty fast there. I was just assuming it would be fast on the bike too
How high does the zipper go on the front of the skinsuit and the trisuit? Is the area around the neck equally tight?
I’ve never tested a sleeveless trisuit so I don’t really know but a difference of .02 m^2 is bigger than I would have thought.
The zipper on the trisuit is not as high and the neck is probably not quite as tight. So I guess it could be a combo of the neck tightness/coverage, the seams and less coverage (skin is slow).
Slightly off topic, but I haven’t tried out a skinsuit under a wetsuit before. Any reason to think that won’t work well? I don’t have to run afterward; I just swim and bike.
Apologies if I sounded overly critical about your protocol–it looks solid.
Your testing makes me wonder about my own clothing selections! I have a baggy team one-piece, a tightly fitting Kiwami, and a new voler skinsuit that requires a second person to put on. Historically, I’ve just done TT’s in bibs + a long sleeve nike pro compression shirt.
What’s the pocket situation on your suits? I’m inclined to think that makes a pretty big difference, too.
Seems like it’s time for me to test and see where everything shakes out. Hopefully not as big a difference as you’ve seen. 
Unless my team suit is fast, then I’m okay
No problem about being critical. That is what I’m looking for! I was assuming I’d made some dumb mistake but couldn’t figure out what it might be. I was hoping someone might poke holes and figure out what I did wrong. On the other hand, if my testing is correct, I ought to be able to get a bunch of PRs since I used the trisuit in all my races last year including TTs.
Neither one of the suits I tested have pockets, so that isn’t the issue. As an aside, what does the second person do when putting on a skinsuit 
Second person helps get the skinsuit over the shoulders. I have to put both arms on at the same time, hike the sleeves up really high (so there’s enough fabric to get over my shoulders) and then readjust everything afterwards. I believe it was “Carl Spacker” who said that nothing motivates a diet like putting on a skinsuit…
Bummer, I was hoping you’d say, “well, there’s a couple big, handy pockets in the back of the Desoto”… crazy the difference!
Looks like you’ve got a good season ahead of you!
Geez, you really need a better handlebar setup. Something that allows you to angle the extensions.
Did you ever test the stacked hands versus high extensions positions?
Hey
If you want, you can send me your data, I have a code that corrects for change in temperature (if the powertap or your computer measures this), among other things.
And I dont know why theres not much actually done in skinsuits commercially but the gains can be huge! I dont know your skinsuit but maybe youre lucky. And congrats on your overal air resistance, I wish mine would be that low…
Hi, I’m keen on giving this a try. I live in a some what busy area, so I could use some tips for course selection. I have access at certain times to an empty parking lot, though there is a flag nearby and evidently it’s typically pretty windy there. I can also go a little further to a parking garage, which one would also expect to be empty. It’s not totally enclosed, but it has half height walls so the wind could be less there. I could probably put up cones to block it off without anyone complaining. Any other tips? Do you suppose a power tap and a joule 2.0 is sufficient for the data collection?
Thanks!
Hi, I’m keen on giving this a try. I live in a some what busy area, so I could use some tips for course selection. I have access at certain times to an empty parking lot, though there is a flag nearby and evidently it’s typically pretty windy there. I can also go a little further to a parking garage, which one would also expect to be empty. It’s not totally enclosed, but it has half height walls so the wind could be less there. I could probably put up cones to block it off without anyone complaining. Any other tips? Do you suppose a power tap and a joule 2.0 is sufficient for the data collection?
Thanks!
Alright, let’s get this going.
If there is a prevailing wind direction, I would align an out-and back course about 400-600m in total distance (out and back) with the wind. Then make sure you have a good handle on the air density at the time you test.
Once you have the data, we can have a look at it, ok? Have you read the protocols in the previous thread posts? Have you looked at jbanks’ website?
If so, then you’re ready to go.
G’luck!
I have been working a bit with the Aerolab function in Golden Cheetah, but I am finding it really hard to get good repeatable “loop” data. I do, however, have a really nice hill outside my front door. So I did some runs up a 600 m stretch with 20 m of elevation gain.
What I had hoped to do was to use the “Crr and CdA on constrained elevation gain” method give in on slide 43 of 112 in the R. Chung presentation/aero bible presentation. I coded everything in R and I think my data and estimated elevations are pretty good, but I notice that my graph doesn’t look like the one in the presentation. The big difference is that I calculate a zero-intercept, but the graphs on slide 43 are not constrained. Can you enlighten me?
Thanks
P.S. I can send R code and data files. I also implemented some optimization routines the find “best fitting” Crr and CdA values and I have fund that “best” is a relative term.