For access to all of our training, gear, and race coverage, plus exclusive training plans, FinisherPix photos, event discounts, and GPS apps, sign up for Outside+.
Pro triathlete Cameron Dye recently visited the San Diego Low Speed Wind Tunnel, located just outside the city’s airport, with his Pearl Izumi teammates to test fabrics and refine his position on the bike. The former collegiate swimmer has already proven he’s capable of riding away from the sport’s fastest cyclists—he used his biking prowess to win St. Anthony’s, the LA Triathlon and Rev3 Costa Rica—but he wanted to find a few more seconds of advantage. Here’s how he did it. Also, make sure to check out our photo gallery from the session.
Cameron Dye’s Kestrel 4000 rolled out of the San Diego Low Speed Wind Tunnel’s mechanic shop and the technicians started to prep their hyper-sensitive balance, a scale designed specifically for bikes capable of measuring vertical, horizontal and side-to-side forces. Bike companies typically bring their athletes to the wind tunnel, as time costs about $800 and hour, but apparel manufacturer Pearl Izumi recently decided to foot the bill for its athletes, partly because they wanted to measure the aerodynamic influences on various fabrics, textures and garments, since clothing plays a part in wind drag. In addition to Dye, Angela Naeth, Caitlin Snow and Jesse Thomas all had their positions analyzed.
The concept of a wind tunnel test is pretty simple: Blow air against something and measure how hard the air pushes. Application, however, is much more complicated. It requires advanced aeronautical facilities, the keen eye of cycling and fit experts and a willing subject. All this in pursuit of a super-position that can make the difference between Dye breaking the tape or being run down by those fleeter of feet. Armchair aerodynamicists can provide guidelines for a fast position, but, as Dye’s test demonstrates, changes that actually influence drag are often counter intuitive. The Pearl Izumi crew hoped to find a way to make Dye speedier, but there was no guarantee they would find a position that was faster than his original setup while still preserving comfort and biomechanical function. Despite this uncertainty, the tunnel is the best way to know for sure whether a change works or simply looks fast.
Metal struts stretch out from the platform in the middle of the test chamber to mount the bike. Extra long, quick release skewers go through these struts and both wheels to secure the bike directly to the balance. Although the rider doesn’t have to pedal for the system to measure drag, pushing against race-like resistance helps settle the rider into a realistic position, so the balance is equipped with a roller drum that creates resistance, and the front wheel spins to create road-like conditions.
Test engineers started by mounting the rear wheel while LSWT test manager Stephen Ryle used aluminum foil tape to seal the opening around the front roller drum. The opening is small, but everything in the tunnel influences passing air. While he set up the front wheel, another operator installed a cadence sensor on Dye’s chainstay that is wired into the data system. His churning legs impact wind resistance, so the test operators closely monitored his cadence to keep it consistent. The bike was dressed in its full race-day setup—Shimano Dura-Ace C75 wheels, XLab Torpedo front water bottle mount and Aero TT downtube bottle—so the test reflected race conditions as closely as possible. With both wheels connected to the balance and the cadence sensor in place, the bike was ready for its rider.
The building that houses the tunnel has the architectural flair of a Soviet barracks. Its exterior looks like a warehouse, and the structure is bland for a reason. The wind tunnel itself isn’t assembled inside the building; it is part of the building. Almost all pictures of the tunnel are taken on the testing platform or in front of the oversize hand-carved wood fan blades that circulate air. The entire chamber, however, is a continuous loop molded out of concrete that circles the building and connects those two landmarks. Every portion of the loop works symbiotically to create smooth airflow—engineers call it laminar flow—over the test subject. The enormous fan blades are in the widest section of the loop. They slowly drive air over an enormous cone extending from the center of the fan toward the first bend in the loop. It’s one of many features designed to stabilize the airflow. Since the tunnel is one loop, the same volume of air is passing through every region of the tunnel, regardless of size. Therefore, air rushes faster through the narrower sections of the tunnel than the wide portions. This first chamber in front of the fan gets progressively smaller, which forces the air to accelerate. At the end of this first room, it hits a wall of concave blades that guide the air through a nearly 90-degree bend. Then it enters a small open room before moving through another set of concave blades. After passing through the second set, the air comes to a honeycomb wall with chambers a couple feet in depth. Facing the other side of this permeable wall is the rider. These hexagonal tubes settle the air into the laminar state needed to produce accurate, repeatable drag readings. Between the honeycomb and the balance, the loop narrows further, again forcing it to accelerate. It passes over the rider at the precise speed entered by the test administrators, then immediately starts to slow behind the rider as the chamber widens rapidly. By the time the air reaches the third turn, it has slowed to a trickle. It passes through another set of concave blades and once again meets the wood fan blades. Everyone other than the subject exits the tunnel, which is closed off with heavy metal doors, while testing a position. The fan is killed between trials, and the frigid air continuous to circulate through the loop even with the fan stopped. Test operators and other hangers-on can enter the test area between runs.
The concrete-chilled air was too cold for Dye to comfortably warm up, so he wore a pull-over while Ryle explained the data projecting on the ground in front of him. The wind tunnel records drag, wind speed and cadence. Power is set to a specific value. Dye could see all this data, and all he had to do was maintain a steady cadence. When riding outside, the rider gets “proprioceptive feel of what’s working,” says Ted Barber, Pearl Izumi’s director of innovation and advanced development. “Because the power is constant in here, they don’t have that feedback so we have to give them the trace [of the drag they’re creating.]”
Tim DeBoom, also a member of the Pearl Izumi tri team, has been in the tunnel many times in his career and flew out from Colorado to offer guidance to Dye and the other young athletes, as well as test his own setup. He instructed Dye to pick a spot and stare at it for every test to fix his head position, which is one of the many variables that has to be controlled to accurately measure the drag differences between positions.
An aerodynamic riding position is only useful if Dye can sustain it—either for a full 40k bike leg or a short breakaway effort. Indeed, some athletes come to the wind tunnel and progressively drop their bars lower and lower to see improved stats, but impressive drag numbers are only irrelevant if the rider can’t translate them to the road. To ensure he doesn’t waste time testing unsustainable positions, he was fit by Retul bike fitter Mat Steinmetz in Boulder, Colo. before coming out to California. “We know what moves we can and can’t make [to Dye’s position] to best blend aero, power and speed,” says Steinmetz. “The position Cam is in right now, we don’t have a lot of movement to go lower so we’re looking at arm width, hand position, head position and small little back angle changes… We’re looking to find the most aerodynamic bike position while working in the constraints of his personal biomechanics.” After a few quick words, the tunnel emptied and Dye started to pedal.
DeBoom, Steinmetz and Barber all stared at Dye’s position with arms crossed until DeBoom spoke up to tell Ryle to increase the resistance. Dye was slouching. “Draw him into his race position,” instructs DeBoom. They upped the resistance to 280 watts, not quite Dye’s race intensity but enough to keep him honest. With the cadre of advisers satisfied, Ryle started the first of two baseline runs designed to establish the drag of his original position so they could understand the magnitude of changes they made throughout the test. Tilting the aerobars upward was the first thing DeBoom mentioned, but Dye’s aerobar extensions could not rotate independent of the basebar. Spinning the bars up meant the airfoil basebar became a giant wind blocker, “and that kills it,” said DeBoom, so they start talking about other ideas while the tunnel collects data.
A fixed video camera captured Dye’s pedal stroke and displayed it on a bubble-screen TV on the side of the console. A rider can shimmy and wiggle into many different positions on the same bike, so Barber traced the outline of Dye’s back and head on the screen with a dry erase marker. He drew black dots over both the front and rear axles to align his eyes with the bike in the same position every time. They used this trace to control his body position through the trials.
After the baseline run, Barber, DeBoom, Steinmetz and tunnel technicians shuffled into the testing chamber. Dye didn’t have much room to move downward. His back angle was already 18 degrees, which was near the lower limit Steinmetz prescribes for high-level triathletes. He was also right at the bottom of the bike’s range for aerobar position, but they wanted to test a slightly lower setup so Steinmetz removed Dye’s elbow pads to drop him an addition 5mm. His forearms rested straight on top of the basebar for this trial. Instead of spinning the fleshy part of his arm downward onto the bar, he let the bone press straight into the carbon bar to maintain his standard position in case the small drop saved substantial drag. Ryle fired the tunnel up for the second run. The data streamed in front of Dye, but there was nothing interesting to see. Lowering his elbows 5mm didn’t drop his resistance, and data shows that it actually produced a tiny bit more drag. Steinmetz put the pads back on. Next up was a higher position. They added a 15mm spacer to Dye’s steerer tube and lengthened his reach to the bars. Back in the tunnel, Dye was generating almost identical drag despite the additional lift. Steinmetz eyed him up and said, “I think he’s drawing forward.” DeBoom agrees. Lengthening the reach distance to the aerobars allows Dye to rotate his hips further forward and extend his back more naturally. He found a position that was just as fast but more comfortable than the one he started with. Counterintuitive information like this—that a position assumed to create more drag doesn’t—is the reason people travel across the country to come here. But despite this useful revelation, they still haven’t found a faster position.
With Dye’s bar height set, the team started looking at the rotational orientation of the aerobar and his forearms. The aerobar extensions on his Pro Missile bar can’t rotate upwards without also spinning the basebar, so Barber instructed Dye to ball his hands into fists and rest them on top of his S-bend extensions to angle his forearms upward. Lifting the hands high above the elbows allowed the forearms to block air from flowing between his upper arms and getting caught against his torso. Many riders including Tour de France road racers and forward-thinking triathlete TJ Tollakson have found this strategy reduces aero drag, and DeBoom held high hopes it would also work for Dye. Barber crouched down in Dye’s line of sight and used his own arms to simulate the position, telling Dye, “We’re going to try and split the air with your arms so it doesn’t hit your torso and your pelvis. Sometimes it works, sometimes it doesn’t.” Steinmetz reminded Dye to keep his hands high to really seal off his torso and pulled Dye’s elbow cups in tighter, to further help direct air around the side of Dye’s body rather than up the gut.
Ryle started the fan again and data quickly started flowing in. They found a way to make Dye faster.
Thinking in grams of drag, the unit measured by the balance, is conceptually foreign to most, so Barber used an equation to convert the grams saved to watts. This new hands-up position saved roughly 16 watts. To find even more free speed, the advisers coached Dye in another technique frequently used by time trialists: turtling. In almost all positions, even aggressive ones, the head pokes above the back and collides with otherwise smooth airflow, but if a rider tucks his head down between the shoulders, which some call it turtling, it can reduce drag. However, turtling can also can strain the back, hamstrings and shoulders. Dye practiced this position before the next run. Steinmetz told him, “Try to bring your shoulders to your ears. Shrug.” When the fan was up and running again, Dye’s head was still sticking up a little too high. Ryle got on the microphone and instructed him to drop his head. Barber liked what he saw and whispered to himself, “Oh come on, hold that position.” The numbers registering on the small screen in front of the technician raised the excitement level in the control room. Shrugging his shoulders turns a 16-watt savings into 28 watts—a real game-changing number. “Drastic,” DeBoom said. “He should be able to hold that for 40k, easy.”
Dye wasn’t so sure. “Narrow I feel is doable. Raising the hands without lifting the back, I don’t know,” he said. Pointing the hands up also lowers the elbows slightly if they hang off the back of the pads, as was the case for Dye. Simply lowering the bars didn’t yield any aerodynamic improvements but did threaten his biomechanical function; lifting his hands had a similar effect on his ability to pedal the bike but did reduce drag. He must acclimate to this position to reap benefits on race day.
Dye’s head protruded above his back in every trial so far except for the turtle tuck position. His Rudy Project Wingspan helmet has a short tail that doesn’t close the space between his head and back. Rudy is one of Dye’s biggest supporters, but the advisers were curious to see if another helmet is faster. Dye hesitantly tried another lid for a repeat of the “hands up” position.
The new helmet eliminated the opening behind Dye’s head and turned his back into a textbook aero profile, but when the data started showing up, everyone was surprised. Dye was substantially more aerodynamic wearing the Rudy Project Wingspan. Somehow, the position with an unsightly space behind his neck was faster. Barber summed it up: “The one thing about coming here, you can think that something is fast, but until you test it, you have no idea.”
After nailing Dye’s helmet selection, attention turned back to his hand position. The group decided the basebar could be rotated upward slightly without presenting a giant wind-blocking surface. Steinmetz spun the bar upward to elevate the tip of the aerobar extensions. The final segment of Dye’s S-bend aerobar extensions were angled slightly upward, but they didn’t position his hands as high as ski-tip style extensions. To simulate the other position option, Barber once again instructed the test subject to ball his hands into fists and rest them on top of the shifters, just like his original run before rotating his basebar. Coupled with the upturned bar, this setup positioned his hands several centimeters higher than his previous test runs, relative to his shoulders. Before closing the door and testing this position, DeBoom instructed Dye to switch positions part way through the trial. Dye was to keep his head in a comfortable, neutral position for the first portion, then he would be given the cue over the loud speaker to drop his head and shrug his shoulders.
The first data that came in—hands on top of the aerobar extensions with the entire bar rotated upward—showed slightly more drag than the same position with the bar flat. Then Dye dropped his shoulders. “That’s a hell of a number,” exclaimed Ryle. “If he can rid that—wow. It’s unbelievable when he gets his head position right.”
“It drops so drastically,” confirmed DeBoom. Despite the increase in drag created by rotating the bar upward, Dye’s total position improved, probably because of the taller hand position. This position saves 30 watts over his original, which is worth roughly 90 seconds in an Olympic distance tri. Although these calculations are estimates, Dye is the rare athlete who might actually see those results on the road. His race strategy is always the same: He swims fast then tries to bury the field on the bike, so he breaks the wind for himself from T1 to T2, making these savings more real for Dye than most. He never bides his time catching a subtle draft off his competition; he’s overcoming max wind resistance the entire ride. For the final trial with the bars in this position, Dye dropped his head and looked straight down at his front hub. His savings scaled back from 30 watts to 27.
Steinmetz’s fit data taken before this tunnel test confirmed Dye’s ability to hold a lower position than the one used for the last few trials, so the crew decided to apply his fastest tuck position with the most drop he could tolerate. They removed the 15mm spacer below the stem to drop his aerobars, lowering him to a position similar to his original in stack height. Counter-intuitively, dropping his bars increased his aero resistance a tiny amount. With his head in the neutral position, Dye was facing slightly more drag without the 15mm riser. Next, he tucked into his fasted position—shoulders shrugged and head lowered—and his drag once again plummeted, registering the fastest position of the day with a savings of 32.7 watts below his baseline position. Steinmetz, the bike fit expert, was quick to point out that despite the incredible savings Dye reaps by burying his head between his shoulders, that position probably isn’t sustainable for a complete race. “Now he can train [in this position] and use it in a breakaway, trying to rest or going really fast,” said Steinmetz. He will save this fastest position for key moments in the race when wind drag can influence the outcome even more than usual.
Although his test didn’t reveal a simple change that dramatically reduces his aero drag, Dye found a moderate but substantial improvement that didn’t hamper his biomechanics, and a super-position achieved by contorting his body without moving the aerobars that he was unable to sustain for a long period at the time of the tunnel test. To see the real benefits, Dye must to return home and start training in these two positions.
Dye needs to use a different aerobar to get his hands into the position they found to be fastest. Rotating his current bar upward to accommodate this position cost him a little bit of speed, so he swapped over to the Pro Missile EVO, a bar with more adjustability, after returning home to Boulder. It’s also taller than his current bar. Situating Dye in his optimized position using the Missile EVO requires a smaller frame size. After he returned to Boulder, Dye swapped his 57cm Kestrel 4000 for a 55cm and installed the other bar that allowed him to tilt his forearms upward.
After about six weeks riding his new position and occasionally testing his ultra-aero tuck, Dye found that the basic position was easy to adopt, but the tucked super-position wasn’t. “We carried over the changes we made to the cockpit and implemented these changes on a slightly smaller frame. I’ve been working on training my body to do better in the faster aero position, as far as my shoulders and my head go, just getting used to riding like that.” Growing accustomed to the upturned aerobar was simple, but the turtle tuck is still a strain.
“I had tight shoulders, tight neck. It’s still not comfy and it’s never something that I’ll be able to implement over the entire 40k because I’m not a cyclist and have to be able to use by shoulders to swim and run afterwards… As far as something I can do in shorter spurts or into a headwind or in a flat straightaway where it’s super advantageous, it’s gotten to the point where my body can handle it better,” said Dye. He can hold the shoulder shrug position that saves him roughly 30 watts for about 2-3 minutes at a time.
And his final impressions of the upgrades: “I’m more comfortable on the smaller frame. It’s not like I go out and feel the bike moving dramatically faster or anything,” but the knowledge that he has a fast position, is riding the best helmet for his body and can scrunch into a turbo-charged position for short bursts have prepared Dye to assault the Olympic-distance tri circuit this summer.