RChung

Did you measure a difference in CdA for these runs compared with two years ago?

GhostRider62

I was on an M5 CHR two years ago. A bit less than 0.150 bent vs 0.320 m^2 on my upright with bolt on aero bars, a big top tube bag, and two water bottles. The upright is a Cervelo S3 with Farsports Ventoux wheels (56 mm with 16F and 21R carbon spokes). I also had a normal road helmet on the upright vs more aero helmet on the recumbent. Previous test was with Flo60 wheels with NTN low friction bearings. I would expect such high quality bearings to be 1-2 watts or an order of magnitude less than what I am seeing here. I know I am going to learn something. I just do not know what.

Thank you very much for the confirmation that I need to do more work. THAT makes this thread valuable for me. It could be the road or I have a problem with the bike or of course, I made an error. But, I can a quick test to confirm a suspicion that I had some linear factor slowing me down. Before I do more like looking onto bearings, I will do better testing Crr before this weekend's 400K. I just do not recall seeing real Crr values for rough roads and your opinion that 0.0066 is too high is very helpful. Thanks!!!!

GhostRider62

Thanks Burnt. I know you are a top time trialer and appreciate your views. I am just a schmuck randonneur who wants to maximize what little power my old carcass can make. I had a goal of breaking 1 hour for 40km this year but so far testing shows that this is not remotely possible, sadly. Rolling resistance or friction somewhere is one problem. My position also is not very good. I did spend 2 years optimizing my recumbent position getting it down from 0.200 to a little below 0.150. Small things matter. I found that fun. I have given myself a year to get myself more efficient on an upright bike. I know I will make mistakes along the way. I just want to thank you for accepting that a Fred like me wants to get as efficient as I can get. In the end, it might simply be that I can get a couple hours sleep on an endurance race or just that I had fun with the experiments. Seriously, thanks

martianone

is the size [area] of contact patch the same between tests ?

RChung

OK, as you already mentioned it's good to get a pretty wide range in speeds and power during your out and back runs. The averages needn't be different but some parts of the runs ought to be slow and other parts should be pretty fast. You're trying to pry apart two coefficients, one of which is linear with speed and the other with the cube of speed, so you want a reasonably wide range of speeds or else it'll be hard to figure out which is which. In addition, with a wide range in speeds it'll be easier to tell if wind screws up your tests. Keep me clued in on your testing.

Koyote

I've got nothing to add on the crr itself, but will offer this: for many riders, the main reason for going tubeless is to reduce (but not eliminate) the possibility of punctures. Be sure to consider that impact along with the counteracting effect of potentially higher rolling resistance.

burnthesheep

You’re not a schmuck. Your contribution has value. It interests me to get to “real” crr as knowing that can make my Notio aero testing more consistent and perhaps able to find smaller CdA differences.

I found an out and back route and distance now to get good laps, but nailing the crr is next. I take a scale with me to weigh all of it day-of and that has helped me.

I let out a swear word when too many vehicles suddenly came past and ruined a lap this week. But, let autolap do its thing and go again.

dedhed

I can tell none if you guys put down miles of HMA in your careers.

"When asphalt surfaces are newly laid, the microtexture of the aggregates is masked by a film of bitumen binder that tends to lower skid resistance during the early life of a pavement, typically the first two year dependent on traffic volume. However, as it is subjected to traffic polishing, the skid resistance of a newly laid pavement starts increasing as a consequence of the scouring of the bitumen binder film, which gradually reveals the microtexture. Once fully exposed, the microtexture is progressively polished by the traffic tending to cause a decrease in skid resistance and leading to pavements becoming more slippery again in the long term."

​​​​​​https://www.sciencedirect.com/scienc...43164817316101

70sSanO

I’m not that familiar with the subject, but I did a quick Google search and on NYVelocity found this...

"Crr is a number assigned to a tire at a certain psi that allows you to compute horizontal drag force as a percentage of down (‘normal’) force. This means a better tire will create less drag, and a heavier rider’s tires will have bigger contact patches and create more drag. The formula is simply: F = Crr * Nf"

It may just be your extra weight; especially since Bicycle Rolling Resistance has a standard GP5000 requiring about 20% more effort than a TR S. BRR is not real world, but it is a more controlled environment and its test results are probably better from a comparative perspective.

If Tire A has 20% less rolling resistance than Tire B in a controlled environment, I doubt Tire B will have 20% less rolling resistance than Tire A in a relatively uncontrolled real world test.

John

Edit Added: Plus if you consider the surface changes dedhed brought up, I don’t think you can make an accurate comparison.

PeteHski

I'm not suggesting Contis are superior to other tyre choices, merely stating that I don't think they have an unusually high rolling resistance. They are clearly top tier in that respect.

livedarklions

I've given up on this correction. The other one is using "cannot be underestimated" for "must not be underestimated".

livedarklions

Honest question, though, from a person (myself) who admits to knowing absolutely nothing about the subject. Doesn't the relative size of the effects make real world comparisons impossible when you can't control wind speed and direction?

I am going to be fine if you tell me that's a stupid question, I don't know enough to know whether or not it is.

GhostRider62

There is an upcoming USA Cycling 10 mile TT on that very road that I muse about entering, so, my testing wasn't just for fun.

It is windy all week. So, any testing using the Chung method is probably out.

Looking more closely at the Silca blog, I am guessing the surface I rode on was maybe like what they consider course asphalt. It appears they got Crr of around 0.0058 if my eyes are seeing it correctly. Of interest, it also shows the improvement in aged asphalt compared to the initial laydown.

AeroCoach shows how too much sealant is also a problem. My sealant injector also can suck out the sealant. I am going to check. I think I have 45 ml of sealant in these 25 mm tires.

I also do low speed rolldowns on one particular hill. Rolling 10 mph until I stop. I have done hundreds of such tests. The hill and runout used to rough chip seal on the first part and then smooth pavement on the other half. It has since been repaved with nice asphalt. The tubeless GP5000 run out nearly as well as the latex tubed GP5000. As well as 28 mm GP4000 with latex tubes and better than 28 mm Schwalbe Pro One tires, which are also about equivalent to Herse 650x38mm with latex tubes. On my little hill, the only tires better were Conti Supersonic with latex tubes, Vittoria Speed tubeless, and the aforementioned GP5000 with latex tubes. The Vittorias actually made it to the stop sign, I had to brake. I do three runs on any test. Why? I have once found bad tire casing out of the box and two other times, I had damaged the casing riding gravel. It showed up in my rolldown test. Since the hill is near my house, it is an easy quick and dirty test.

https://silca.cc/blogs/silca/part-4b...-and-impedance

https://www.aero-coach.co.uk/tubeles...ing-resistance

livedarklions

OK, so I'm basically nerdy about how people discuss things even when I have no clue about the subject matter. I think what happened on this thread is a bunch of people were sucked into the red herring of an easily answerable question--can you ever know for sure why your results are different from a test done 3 years ago?--that a lot of people missed that the real questions you were asking were quite reasonable.
burnthesheep is right, even in General Cycling we shouldn't just jump in to sarcastically answer the easy question and just get in the way of people trying to answer the hard ones.

This being General Cycling, though, I reserve the right to ask naïve questions about specialized methods and issues.

GhostRider62

I understand that BF does not allow linking to other forums, but I found something useful on a forum decided to time trialers and triathloners. Someone did get really high Crr and the gentleman who does the bicycle rolling resistance website linked by another poster in this thread, made the following comment.

A possibility that I had not considered is simply that the break point in "impedance" is now at a lower pressure. The difference in speed isn't trivial, it is around 1 mph or perhaps a touch more. It is noticeable. I also know my power meter isn't lying even though I switched from a powertap to the SRAM AXS crank PM. I know this because of measured time to climb hills that I have climbed for years. I know what power/weight gets me what time. The PM if anything is reading a little low by 2%.

To be clear, I doubt there is any problem with the tires. I might be doing something wrong or have simply underestimated the effect of worn pavement. At this point, I do not know.

GhostRider62

I am sorry.

Where did I miss the easy answer?

Unless it was mixed in with all the nasty vitriolic comments. Since all bikes have tires, questions about bike tires seems to be a general question.

seypat

30 extra lbs is going to make everything except for descending seem a little slow.

RChung

So, you're mostly right: If Tire A < B in a BRR test, very rarely do you find B < A in a road test. That's partly why Ghostrider62's findings are so anomalous, and we're trying to nail down the reasons. Generally, Crr on real roads is higher than on a smooth roller, but the relative rankings almost always stay the same, so you can buy a tire that's near the top of the BRR lists and almost always be confident that you're getting a low Crr tire.

That said, your quote from NYVelocity is almost right, but you've misinterpreted it. Extra mass is accounted for in the N term, not the Crr term, so when we do the calculations we get a Crr that's independent of mass. That is, we agree that the extra mass results in higher drag; however, we account for it when we calculate Crr.

The place where NYVelocity got it wrong is kinda subtle: the extra drag doesn't come from the size of the contact patch. Thinking that larger contact patch causes the drag is what lead most people to over-inflate their tires, since higher inflation reduces contact patch. We now know that it's not the compression of the tire where energy is lost, it's the *rebound* from the compression. Low rolling resistance tires rebound better, so they lose less energy. That's why latex tubes reduce rolling resistance *even though* they don't affect contact patch size. In some but not all cases, tubeless tires will rebound better than tubed tires: they appear to almost always rebound better than tires with butyl tubes, and sometimes do and sometimes don't rebound better than tires with latex tubes. Over-inflation limits the compression, but compression isn't the reason for the loss. Nonetheless, when you're riding on an ultra-smooth surface, like a wooden track on a velodrome, you don't need much "suspension compliance" from the tires so you can reduce overall rolling resistance by super-high inflation pressures. You can't get away with that on a real road because without enough compression you're kind of lifting the bike+rider a tiny bit on every bump.

70sSanO

Thanks for the response. I think there is merit in looking at rolling resistance within a range of tires that fit a particular purpose.

I’ve been riding Rubino Pro III’s for a while and went to BRR and found both the newest models of the Rubino Pro’s were slower and no more puncture resistant. Even if the testing is not reflective of a real road surface, it was the determining factor to search out the discontinued III’s.

John

livedarklions

Post #7 is an example.

Sorry if I'm being unclear--the question of whether you will ever know why exactly the 3 year old test is different from this one has an easy answer--obviously, no. I think many of the nasty vitriolic comments were basically variations of "no duh", glib pointless snark. I also don't think it was really the question you were asking.

I think your questions about what factors and how best to measure them are really good, tough questions that are way over my head to answer. My reading of your OP is "here's why I think I'm getting a different answer, how can I best control for these" which is a really good solid question. I get that you don't need me to validate that so ignore it all you want, but it was not my intent to imply that you were asking an easy question, just that people were interpreting the question that way so they could give their glib answer.

RChung

So, your question is exactly why for so long so few people even tried to do drag estimation. Direction has no effect but wind can definitely screw up the calculations. However, wind affects the aero drag component, not the tire rolling drag component, so modern analytical methods that are used to estimate the rolling and aero drag can distinguish between them. That's why I suggested up thread that you vary your speed a fair amount during testing: Crr will be unaffected by wind but CdA will be, so getting a lot of speed variation helps to identify the difference. Ideally, you either want to measure the wind or do the tests on a dead calm day but even if you can't measure the wind we can do the next best thing and see that the tests were screwed up, and then discount those results. That's why there's a new generation of windspeed sensors for people who really want to do drag estimation (like @burnthesheep): they want to measure the wind so they don't have to wait for a dead calm day. (I have reservations about the current generation of wind sensors and especially how the current generation of analytical software use their data, but that's a separate issue).

RChung

Actually, if he still has the original raw data and his test protocol is/was robust enough, we can often figure out why the 3-year-old test is different. It's a little bit easier if there were more and shorter laps rather than a 7-mile long out-and-back (the extra laps give us more bites at the apple so we can find the worm) but I've done comparisons for tests done over time and been able to figure out why things were different. The key is having enough contrasts in the data and recording enough about the conditions of the test so you can go backward and see what changed: it turns out that different kinds of changes leave different kinds of fingerprints on the data. It's sort of like CSI: The Aero Test. You can do forensic investigations to figure out the source of the change.

livedarklions

Thank you!

I think Dunning and Kruger may have just met for coffee in my head. It's wonderful to get such patient and clear explanations from a real expert.

WhyFi

Only so many of us on the forum, like you and I, can earn our keep with our warmth and charm; guys like RChung need to contribute in other ways.

GhostRider62

i did 8 runs today. Approx. +/- 5% variation.