I went to a wind tunnel and learnt a few things. The first thing I learned is that a wind tunnel isn't necessarily a tunnel.
I had an image in my mind of a fully enclosed and pristine tube like you see in the car ads. Octagonal, probably. Maybe with smoke trails, and definitely with a big fan at one end. The reality was a bit different.
The GST wind tunnel at the Airbus Defence and Space facility on the outskirts of Friedrichshafen scores on the fan front. But basically it's just a room. There's a table in it, with a bunch of tools on, and a stepladder propped up in the corner. There's some feather flags filling some space, and a projector hanging from the ceiling. Pristine it ain't.
It's an old facility; it used to be the Dornier wind tunnel when Dornier were making planes, and there's still a goodly number of switches and filament bulbs in the control room. Now it mostly splits its time between bike parts and scale models of drones. But it's a good tunnel, says Jean-Paul Ballard of Swiss Side wheels, who's invited me to spend an hour or two looking at the process. He's got a bit of experience with aerodynamics, having worked for 14 years in Formula 1 as the lead engineer on a number of teams. Between them, the team behind Swiss Side have over 50 years of F1 experience, which obviously includes a whole stack of wind tunnel time.
"It doesn't look like much but it's incredibly consistent", he says of the GST tunnel. "One of the problems with an enclosed tunnel is the blockage effects. This is important when testing with a rider as results can be significantly influenced. Here the flow isn't constrained". And, for the record, the fan sucks. "We laugh when we see pictures of wind tunnels in adverts with the bike pointing the wrong way". It's obvious once you think about it: if you want a clean flow of air you don't want a massive fan chopping it up a few feet from your test rig.
So the fan pulls the air through, and the intake at the other end is as smooth and uniform as possible to create the best conditions for consistent results. And the extra space in the room means the natural flow expansion and turbulence created has somewhere to go without affecting the readings on the rig. And all that makes for lovely, repeatable numbers. Stepladder or no.
We spent a bit of time running a standard series of tests. There's plenty to take out of the few tests that we did (and some we didn't), so let's pull out some of the most interesting bits of information we garnered. Starting with this...
Well, up to a point.
That's right: all you choppers out there – and I count myself firmly among your number – need aero most of all. That's madness, right?
Wrong. This seemingly counter-intuitive bit of information is explained by a single graph from the basic tests we ran during our hour at the tunnel. Here it is:
So what's this showing? It's showing the drag of a range wheels at various angles of yaw. The blue line at the bottom is an Argon 18 time trial bike; ignore that slippery beast for now and concentrate on the four lines at the top. That's four different wheels aboard the same frameset, a Cervelo R5 (Not a P5 like the legend says). They're all Swiss Side wheels: The box section Heidi (light blue) and three depths of their toroidal Hadron rim. The 80mm Hadron 800+ is in red, The 62.5mm Hadron 625 in orange and the 48.5mm Hadron 485 in green. Got that?
Yaw is the angle of the wind relative to the bike. So if you're cycling directly into a head wind the yaw angle is zero, and if you're at a standstill with a direct crosswind, it's 90°. When you're moving, the yaw is a combination of the wind speed, the wind direction and your speed.
The first take-home point here is to look at the drag from the bike at a yaw of zero. Whatever the wheel fitted, the drag of the bike as it faces directly into the wind is more or less the same. We (well, I, anyway) mostly have a notion that those big, chunky rims make a nice shape that slices through the air straight on. In reality, that's not where the gains come from.
Once the yaw angle starts to creep up then you start to see a marked difference in the aero-optimised wheels compared to the box section. The Heidi wheel keeps increasing in drag over about 6° while the Hadron wheels begin to drop back down again until at 20° of yaw their lines are well underneath the standard rim.
Why? Because the air flow stays attached to the rim, rather than breaking off and causing turbulence that drags against the bike. It's at these higher yaw angles that the gains are most apparent. That's where you're making your time up.
So, imagine a rider travelling at the speed of light. The wind will always be straight on, a yaw angle of zero: however hard it's blowing, the wind speed and direction will have no discernible effect. As you slow up, the fact that you're moving more slowly means that a greater range of yaw angles are possible for any given wind speed. And if you slow right down until you stop, anything goes: the yaw angle is whatever the wind direction is, relative to the way you're facing.
Followed all that? The practical upshot is that the faster you go, the lower the range of relative wind angles you'll experience. Professional riders in normal conditions won't see anything over about 10°, whereas us sportive bashers and lower-category chuffers will see much higher yaw for the same wind, because we can't go as fast through it. And it's those higher yaw angles that see the biggest gains, up until about the 20° mark when the airflow detaches from the rim and you lose the aero advantage.
"Faster riders generate more drag", Jean-Paul adds, "because drag is proportional to the square of velocity. But faster riders are also on the course for less time, and experience a narrower range of yaw angles. Through our simulations, we see that slower riders actually save more absolute time. They're out on the road for longer and can therefore benefit from the bigger aero gains for longer."
So go out and buy some new aero wheels, ordinary rider. Tell them I said it was okay. Make sure you fit the right tyres, though. Because...
This was perhaps the most surprising bit of information from the whole session. It's not something that we tested on the day, because unfortunately there wasn't time to swap masses of tyres out and in, but Jean-Paul from Swiss Side told me they'd been testing plenty, and the results are interesting. More than interesting. It's all about what kind of airflow is passing over the wheel.
You'll often see trips on aerodynamic profiles; small raised or rough sections. Some bike wheels even have them. These trips cause turbulence, and the turbulent layer of air helps the flow to stay attached to the surface. The tread pattern of a tyre can act as a trip and help the wheel to remain aerodynamically efficient at higher yaw angles. It can do that. It doesn't always do that.
"The best tyre we've found is the Continental GP4000S", Jean-Paul told me, and that's what our test bikes were shod with. Is the tread pattern designed for that job? Probably not, he conceded, it's more than likely just a coincidence. But there are other tyre manufacturers that are definitely designing their sidewall profiles with aerodynamics in mind, whether they're talking about it or not.
And how much difference does it make?
"If you fit a GP4000S to one of our Hadron wheels the flow will stay attached to 18 or 20° of yaw," Jean-Paul said. "With another tyre, one with a completely smooth profile, the figure drops to 8-10°".
One look at our yaw graph above shows that you're losing almost all of the aero advantage of a deep section if that's the case. That's a major issue, and one that most people won't have considered at all in the whole aero equation.
So how will you know whether your tyre is a good one, or a bad one? Well, you won't. Except if you're running GP4000S tyres, in which case you're in luck. Or if you're running tyres that don't have any tread profile at all, which is probably bad. The Swiss Side guys have been running tests on a wide range of tyres: there's boxes of them knocking about at the wind tunnel. I'm not sure they have plans to release that information to the general public, at least for now.
So, you've got your deep section wheels and they look pretty pro, and your tyres with the (hopefully) effective trip profile. But those big old rims are heavy: maybe a couple of hundred grams over a climber's wheelset. Isn't that going to negate all that precious aero advantage?
Well, no. because...
What's more important, light weight or aero? That's simple, according to Jean-Paul. Aero trumps light almost every time.
Swiss Side's aero range has - up until now - used a non-structural carbon aero profile with an alloy rim and braking surface. It's simple enough to get the right profile with a non-structural fairing, and the alloy braking surface means more consistent performance from your brakes. The alloy bead section is easy to extrude, too, which means the wheels are cheaper. The penalty in terms of weight is about 100g over going to a full carbon construction, and you're adding that on to the extant weight penalty of just having a bigger rim made of more stuff. What kind of time penalty is that going to add up to?
Swiss Side feed aero and weight data into a model that crunches the numbers for different types of ride profile and length, and then spits out the likely speed and timing penalties based on a reference ride. One of their programs is a 120km rolling ride with 1200m of height gain. Their 'average' rider completes this parcours at exactly 30km/h (211.4W average power), which left me wishing I was average. But enough about my failings. What difference would one hundred grams (from an 8kg bike to an 8.1kg bike, with a 75kg rider) do to the ride time?
Well, it would increase it.
By three seconds.
Adding weight to a rider going that fast, over that terrain, makes precious little difference, really. 100g is 1.25% of the wheel weight; even at four times that, the penalty is just 17 seconds.
Changing that 1.25% weight penalty to an aero penalty - upping the overall drag of the bike by the same percentage – gives a 22-second penalty, and quadrupling the drag penalty pretty much does the same to the time lost: 87 seconds at 5%.
Now these still aren't big numbers: just under a minute and a half in four hours of riding. But the difference is certainly significant: aero gains are worth six times what weight gains are, and a fair conclusion from Swiss Side's stats would be that on rolling terrain it's worth going heavier and more aero.
But surely there's a tipping point? Of course, but it's a long way north of where you might expect.
This graph shows what a average rider might do on a 20km climb with an average gradient of 4%, knocking out 211W for not much under an hour. That's certainly achievable for many of us.
This is the tipping point, more or less, for this imaginary rider. If your ride has an average gradient of 4% or more then the weight gains will mean bigger time savings than the aero ones. So you're still permitted to take a drill to your levers for the local hillclimb. But unless it is a hillclimb – or a really, REALLY hilly ride – then you're better off optimising your bike to cut through the wind than you are shaving off the grams. If that's a surprise to you, well, it was to me as well. Obviously it doesn't mean that a 10kg beater will be as quick as a 6kg superbike. But any time you're spending stressing over the odd hundred grams here or there is basically wasted.
Disc brakes, in their current incarnations, aren't aero. I'm sure we all guessed that. The magnitude of the difference is fairly significant.
"We've measured a 16% increase in wheel drag between a disc-braked wheelset and a standard wheelset", Jean-Paul told us. "We performed a direct back to back test of the Zipp 303FC in standard version and disc brake version, for our own competitor comparison purposes. That 16% is a constant offset in the performance curve across the entire cross wind angle range."
So there's work to be done there, but how much can actually be done? The extra drag essentially comes from three sources. The rotor itself adds drag, and because disc wheels need more spokes to cope with braking forces there's more drag there too. On top of that, a disc hub contains more material and needs to be built to withstand torque across the hub body, as the braking forces are on one side only. Because of that, the hub body is generally bigger and that increases drag as well.
What can be done about those three sources of extra drag? In reality, probably not a huge amount, and disc systems will likely continue to be at a disadvantage in terms of aerodynamics. So if you want to be as slippery as possible, it's rim brakes for the time being. And hide them, if you can.
A big thanks to Swiss Side for their time at the wind tunnel.
Dave is a founding father of road.cc, having previously worked on Cycling Plus and What Mountain Bike magazines back in the day. He also writes about e-bikes for our sister publication ebiketips. He's won three mountain bike bog snorkelling World Championships, and races at the back of the third cats.