The main issue is their claims are based on copy/pasting data from other people’s aero research using completely different test setups and presenting it as an equal comparison.
Also, claims that other aero bikes are on the order of 50% higher watt expenditure at a certain speed when the aero game has been chasing tenths of a percentage for a long time.
The frame isn’t even smooth or speed-dimpled carbon, it looks like it was welded by Orange. Pure Road Warrior bodge job.
Pure BS from top to bottom, but people are being polite.
do dimples matter on a bike frame shape? my non expert understanding was they helped on shapes with lots of wetted area exposed to free air, so maybe on fork legs or headtubes as well as the rim sections we’re used to seeing, but not parts of the frame in already turbulent air
so the thing is, laminar flow is good. turbulent flow is worse. fully separated flow is worst.
but laminar flow is more susceptible to separating than turbulent flow. so you add in some surface discontinuities to trip the flow from laminar to turbulent before the flow separates.
On a round & spinning object like a golfball, you don’t have any idea what the orientation is going to be/which way the air is going to be coming from, so you just cover the entire surface uniformly with weird little turbulators.
This strategy may also make sense on things like rims. It doesn’t really make any sense at all on forks or frames - there are better (easier, simpler) ways to design turbulators if you know which way the air is coming from, that have the added advantage of maintaining good laminar flow as long as possible (not that there’s much of that to begin with on bikes…).
But everyone “knows” that golfball dimples reduce drag so sticking them all over your “aero” bike is good marketing
From what I remember about the GB team track bike, the rider’s legs actually caused a lot of turbulence. That was why they put the fork legs where they did.
I thought the Zipp dimples were to help maintain aerodynamics in a cross wind. But really, we should just go back to drillium and call it “speed holes.”
The Performance Process podcast from the Escape Collective goes so deep in aero you will probably be cured of it. Also a fair bit on the Marginal Gains Josh Portner one. I cant listen to that one without imagining that one of his co-hosts is Peter Griffin from Family Guy, his voice is so comical.
Actually Im confused, is his usb converter in that pipe or is that just a switch?
Edit: watched again, its a converter and switch.
Jinkies, the wife of one of my bikepacking/cx buddies just dropped off some of his gear. He died of a malignant brain tumour a year or 2 ago. 3 klite lamps. 2 of them 3rd gen, never used. So sad.
Sort of interesting for $30. Wonder if it can resolve a rando bike?
The idea is, from a 20 second video with any phone I can generate a 3D model and then perform a full-on CFD test to determine, with seemingly quite high accuracy, CdA (aerodynamic drag). .
I bet folks at GRM would be interested in affordable CFD to tune amateur race car aero also. wonder how he’s turning a walk-around video into a functional open foam model
appears that software bro has not validated his results with wind tunnel testing. Notably he’s not simulating rotating wheels at all. low-speed fully-turbulent flow is notoriously tricky to simulate with any accuracy and the results will be highly dependent on the quality of the model, size of mesh elements, and boundary conditions. Probably OK for checking the relative effect of certain parameters but I would take the absolute numbers with a large grain of salt.
he hasn’t invented anything new here, just glued a bunch of open-source stuff together and provided an easy way for you to pay him to click “go” on the scrips. But he’s making a lot of bold claims with seemingly zero expertise in FEA/CFD or understanding any of its foibles.
EDIT: ok to be fair he has actually put some work into generating a watertight solid from a series of images shot with a cell phone, but the quality of the solid he’s generating raises even more questions about the validity of the results
i like that he both doesn’t know what a significant digit is (the leading zero doesn’t count) and also doesn’t get that this comment is making fun of him.
This is a place where a machine learning model could be useful. CFD is really expensive but a statistical model could be trained on CFD results for a bunch of slightly different bike positions. So long as someone doesn’t try to use it on a fully-faired recumbent or a racing trike, you’d probably be OK and it would be faster/cheaper than running the full model every time. Finding/creating a good training dataset is probably the big challenge for this.
