Depends on your frame, an aero bottle usually fills the gap and makes the bike more aero but without testing you couldn’t say 100%.
Thanks. Iam using a Giant Trinity and the aerobottle is wider than the downtube of the bike…so i dont know if that is more aero…
Giants own brand aero bottle costs 2 watts or so i’m told by someone i would consider to be an expert on the matter.
every little helps as they say
if you dont need a drink, get rid
No, not from this. There are just too many variables. You need more data, and I would suggest using Golden Cheetah aerolab (or other online tool). Using the tool over and over on different routes and different days will help you lots.
My first attempt at a TT. 12km course, a bit of rolling terrain but nothing major. Treated this like a giant over under. absolutely cooked at the end. Goal was to crack 40km avg speed and just missed out.
You’ve got the steed ( ) to do it in future, nice effort
Saw some recent aerolab analyses on here - I’ve been playing with this as well and find it pretty useful and consistent, usually within 0.01 CdA on our local 10 mile out and back flattish to moderately rolling courses that are open to light traffic. Typically see one clear discontinuity in the model from braking at the turnaround, but on days without crazy wind it can be pretty good over the rest of it, while on days with moderate wind, the model will drift off in one direction and converge on the other and sometimes still works it all out pretty well. I’ll usually try to pin the end with an offset to match the known braking offset.
One thought - I was surprised to see people using 100% drivetrain efficiency… If not on a hub power meter, seems like it should be a couple percent loss from pedals to road, even with a waxed chain. Adding some drivetrain loss will mean lower CdA to compensate.
I also run gp5000s and have used a bit over 0.004 crr based on rolling resistance/ pressure/my weight and not perfect roads.
I’ll post a couple examples of what I’ve seen on a hilly course (more variation) and flatter out and back course even on nights this spring with very strong crosswinds - like 10+ mph. Surprising how well it works considering!
Hilly with very strong cross winds:
Flat course with relatively mild wind -
For what it’s worth, this run came out at 26.9 mph on 290W at pedals, standing start, and the turnaround on a 2 lane road (e.g. near complete stop which leads to the spike from braking…) - recently just lowered my front end more and seems to be ~0.5mph faster on some test runs, so will be curious to see how the CdA changes.
ps - kind of funny story about power vs. CdA…
For years, I rode a stages L-only meter. Finally this season I got Assioma Duos - and am averaging a 44L/56R power split
So, I was really happy that reported power is now way up - but depressed that CdA is also way up when doing this kind of analysis Stronger than I thought, not as aero as I thought!
Lesson being - it may be one of the few applications where dual power meters really have some meaningful value beyond sheer curiosity about something that is not really that actionable.
Great post, @hvvelo - I think I see broadly the same consistency as you are reporting too, many thanks for sharing that
That’s really interesting - I have never thought of attaching an offset for braking (I just try to avoid braking full-stop!!) What kind of offset do you use? Is it a given value for each braking event? Or do you use judgement depending on how much you have to brake? I think I can see the discontinuity in the graphs.
You are completely right, but… drivetrain power losses with a good setup should be small (1-3%) whereas estimating Crr could be a similar size (2-4W) or gusts of wind or traffic effects are likely 2-3 times more error-prone I guess. I have used a Crr of 0.004, for example, but it could easily be 0.00375 or 0.00425. And from traffic, I have seen a change in effective CdA from 0.185 to 0.155! In these terms, drivetrain losses get lost in the rounding I think?
Or, to put it another way, at 40kph, I would estimate drivetrain losses at 3-10W, rolling resistance at 40W and aero drag at 160W - would you agree?
Brilliant post - its really got me thinking!!
Yea, you can see it pretty much exactly halfway through on the blue trace as it’s 5 miles out, 5 back. These are roads with one lane in each direction, so no roundabout - we have to slow down pretty significantly - not full stop, but close, and all in one shot as quickly as possible! Since the model has no expectation of brakes being used, you see that very clear discontinuity. So, I try to have the end point offset by the same amount. This is why the blue line is a bit higher at the end of each - I did that intentionally when tuning the CdA to fit everything else.
Absolutely - if you play with this you’ll see it’s not a huge factor - and it’s still a guess… I browsed some stuff on a few different sites and it seemed like 2% was a good guess for a very clean / waxed chain using power pedals/crank.
I just figure there are some things we can’t know/control, but for those we can, it never hurts to refine the guess as much as we can. We know the value is something > 0% loss, but probably < 5% for anyone taking reasonably good care of their gear. The chain lube testing sites are a pretty good resource to make these guesses slightly more educated.
From what I’ve read on forums like slowtwitch, one of the other best things you can also do is to have a handheld weather device to get the air density dialed in better - I have to use an airport that is about 30 miles away. That value varies a lot for us from week to week and has a meaningful impact on the entire calculation. I’d be curious to know if anyone on this thread has used such a device!
For CRR, I’ve never tried to extract it - you can theoretically solve for this with test runs at multiple speeds if you have a really good testing ground, but I don’t have anything I’d trust enough for that.
These courses will also usually have a few cars passing in both directions over the 10 miles so it’s certainly a source of random variation for us, but nothing like what I think the UK has where it sounds like people TT on the side of what we would call a freeway / interstate highway - I can imagine that would make these analyses almost pointless
I’m rather amazed how often I get within 0.01 of the same value if I am not making any other changes when I consider how windy/gusty it can be, +/- 15F temperature, plus random cars thrown in! I’ve not yet had as much success finding a good halfpipe testing ground, but I have my eye on one to try on the next rare no-wind day around here!
I have allready buyed a TT suite (Bioaracer Speedwear Concept TT) size 2/S fits really well and cant get the suit on by my self. Someone need to take the suit over my back, so i think thats good. hehe.
Now i need a helmet and i read thats really personaly and so difficult. I have allready a Kask Mistral and can send it back when it is not good. I cant test it outdoor, only indoor. So no aero data… I will take some pictures to look how it fits, maybe you will help me to look at it… The Mistral is good in my shoulders and neck, but the frontal is brutal…what is more important? The frontal or drag?
I saw also the Lazer Volante but i think it is not going a good fit for me… I saw the same helmet by MvdP and it has a big gap between the helmet and neck:
Helmet are very individual…without being able to test, it is impossible to say which one is “best” for you. A helmet that test fast for some can test poorly for others.
That said, there are several that tend to test well for most people…the Giro Aerohead is one, along with the older Giro Advantage 2. The POC Cerebellum also tends to do well.
The new Rudy Project helmet also has had promising data, according to Jim @ ERO, but the data set is still on the small side (at least what is publicly available)
If it were me and I was just getting started, and could not test, I’d probably lean towards the POC or the Rudy Project.
Don’t worry too much about the helmet / back issue….lots of people test fast with gaps back there. The old “eyeball wind tunnel” evaluation for aero in that regard doesn’t really hold up anymore. Head position is the critical element.
I dont have the ability for a windtunneltest and a virtual aero test will only test the frontal and not the drag. I cant test the helmets outdoof because of sending back to the webshop.
What is eyeball windtunnel?
I have also bought a Giant Rivet TT in small. Hopefully not to small for my head.
I read that the aerohead is fast for peolpe thay cant turtel. I think my head is really low in front of my body.
If this is a good position for you, and your head is comfy, you have a great starting point. The Giro Aerohead is going to be a good fit I think, because of the angle of your head. Personally, I struggle to see with my head that flat (relative to horizontal), so I have previously thought helmets like the Kask Mistral and Lazer Victor would be best, (but I seem to be learning to make the Aerohead work).
With most positions/helmets, the aim is to get your head in-line with your body when viewed from the front, and to have a nice smooth transition when viewed from the side. This is where the ability to “turtle” comes in. The “eyeball wind tunnel” is just using your eye and judgement to see if this is the case! The trouble is, its not very good (if it was, real wind tunnels wouldn’t exist!)
The best you can do is try a bunch of helmets, get a friend to take photos in front and from the side, and take a best-guess. Then record data outside and start the neverending journey of refining and getting aero-nerdy!
Greg, thank you. No problem to post this picture.
This picture was taken before a went to a bikefitter. So my sit is now a little difference. I have no a Kask Mistral and next week i get a Giant Rivet TT and i will also order a Aerohead. Someone will make pictures of my and i share it in this thread. Thanks for helping!
I know someone with a aerohead and he have problems with seeing. The helmet is really long in front, so he has to put up his head to see something…i think i get the same problem.
I feel allready an aeronerdy due the struggling of buying a good helmet…haha
It can be a bit of an issue for me w/ the Aerohead, but my bigger problem is I tend to bounce the front of the helmet off my hands when I try to get really low.
The Eyeball Windtunnel is just what we are doing….looking at positions and evaluating things based on how they look / best practices.
Next week i have a Mistral, Aerohead and Giant Rivet TT and i can only test it indoor, otherwise i cant send it back. I will put some pictures here
Position looks like a great starting point. I would maybe slowly start trying to stretch out some by pushing the bars/pads more forward. That will help flatten the back.
In reference to the suit, tighter isn’t better. Before the modern fabrics, it was common to undersize but now with ridges in the fabrics you want a normal fit with no or limited wrinkles. If it’s too tight, the ridges will stretch flat and not work as intended.
This is my old position before the bikefit. Iam sit now more stretched with longer extensions and saddle slightly more back. I will ask @GregElwell to remove the above picture
The suit fits well. Tight but not to tight. Difficult to say. I have only ridges on my upperhands and shoulders. It is a S from Bioracer and my other bibs and shirts from Bioracer is also S. The pipes long enough, just above my knees and the the long arms just above my wrists.
Ok @Power13 @GregElwell @Majoeric , you all convinced me and i like to be an aeronerdy , so i will try to test the helmets outdoor with Aerolab. I have a round with two braking corners and 6km (a little traffic). How much laps do i need to do per helmet?
Nearby is also a cycling track (without traffic) of 1.4km with no braking corners. Is that better and how muchs laps per helmet do i need to make for a good measurement.
Is it important to do al the tests with +/- the same power? So i need to choice a power output i can sustain for more test laps? So when my FTP is 285w, but i can sustain 250w for more tests, i can better choice 250w?
Cyclingtrack of 1.4km (i think a cyclingtrack is also saver, because of lack of hearing through the helmet)