Wind tunnels fall into two general categories. Some are housed at institutes of higher education, and examples are the tunnels at University of Washington, Wichita State and, most famous among those who follow cycling, the tunnel at Texas A&M.
Then there are the tunnels developed by private industry, and airplane and car makers are among their most obvious customers. When I visited the San Diego based low-speed wind tunnel at Allied Aerospace this past weekend I spoke at some length with the person who heads up aerodynamic testing at a high-volume aircraft manufacturer. He was observing the cyclists in the tunnel on Saturday, but on Monday his own company's airplane control surfaces were slated to be tested. Allied Aerospace is not attached to a university. Its business is running wind tunnels.
A "low speed" tunnel is one that can produce a wind of up to around 250mph. Even in modern aerospace this has utility, as airplane makers must test control surfaces during low-speed maneuvers, such as take-offs, landings and stall speeds. High speed wind tunnels can produce winds of up to mach-3.5 and above — Allied Aerospace also owns such a wind tunnel adjacent to LAX airport.
This tunnel facility is considered quite advanced, although the tunnel itself was owned and built by Convair in 1943, at the same time it was pushing out B-24 bombers at a rate of one every hour (Convair's massive San Diego factories built their large bombers, PBY sea planes, and a hefty assortment of historic war birds throughout its history).
Big differences are evident between this tunnel and the tunnel at Texas A&M. The most obvious initial realization — though of no real consequence — is that the wind travels counterclockwise in Texas Station's tunnel, and clockwise in San Diego. The prop that generates the wind is maybe 50 feet behind the object or subject to be tested at A&M (the photo below right shows the A&M prop from the vantage point of the test subject). The prop at the San Diego tunnel is on the other side of the oval at Allied Aerospace (the wind in a wind tunnel goes through a donut-shaped tube of varying diameter, passing through a propeller once per revolution so that the wind is reaccelerated).
Having the prop on the opposite side of the "donut" from the testing balances causes the San-Diego tunnel to be amazingly quiet. When you're in the control room you can't tell when the tunnel has been turned on. But when you look through the window at the subject and see the ruffles in his jersey as he's pedaling his bike, you know a test is being conducted.
The prop that accelerates the wind is much larger at Allied Aerospace than at Texas A&M. I don't know that this means anything, it's just an observation. The propeller at A&M (above) came off a B-29 bomber. The prop at the San Diego tunnel (left) is constructed of a variety of very large wooden blades that were built decades ago specifically for the purpose. These six prop blades were each longer than I am tall, and were reminiscent of beautifully-constructed balsa "guns" on which one might've surfed big waves back in the 60s.
The meat of this tunnel, however — the thing that sets it apart from other tunnels, as far as cyclists are concerned — is the "turret" (my word) on which the bike and rider are suspended. This is a metal disc perhaps eight feet across underneath which is a set of balances (precise scales that measure drag) and a couple of motors. The turret sits about 18 inches off the ground. The point of this tunnel (for cyclists) is not only to measure the drag of an object, but to measure the drag in as close to a real-life situation as possible. Therefore, almost nothing is visible above the turret, and only the bike and rider are subject to the wind, save a couple of very thin support blades attached to the bike's rear axle. One nice element to this tunnel, as one can see in the picture above, is a readout of all salient data on the floor of the tunnel in front of the rider, so he can see what the control room sees. This turret can be turned on its axis to facilitate testing in a "yaw," that is, in a sidewind.
There are a pair of motors that independently turn the front and rear wheels. The tunnel's wind speed is 25 miles per hour for these tests, slower but more real-world (for age-groupers) than the industry-standard 30mph used in most bicycle tests. The front wheel is therefore spun at 25 miles per hour by its motor, and the rear wheel has a load generator that is reminiscent of that made by Computrainer. The load generator is not from Computrainer, however, but was built on-site from scratch. It is incredibly smooth, and is calibrated to an SRM. In other words, it is a very accurate resistance unit, and the resistance is modulated by a person in the control room. The rider can also modulate his power output by changing to a higher gear and thus increasing the velocity required to turn the resistance unit.
The concept behind the bike testing fixtures and balances here at Allied Aerospace is to have the subject ride at a given (race pace, let us say) power output, for a reasonable period of time, and sample his drag throughout the exercise. Sampling takes place 20 times per second. This protocol then generates an average for all the drag samples over a time increment — say 30 seconds or 10 minutes — and produces one drag number emblematic of a performance in a given rider position at that power output. Then changes can be made to one's position followed by a retest.
I liked everything I saw at this tunnel. There is one small problem in the protocol. Since the rider can modulate his speed by changing gears, it is possible that he might be riding in "position-one" with his rear wheel spinning at 24mph, and "position-two" at 27mph. While the front wheel, and the wind to which the rider is subjected, are both travelling at 25mph, position-two would generate a slightly higher drag number even if the rider position generated precisely the same drag number. When I asked the tunnel's scientists about this they acknowledged the problem, and said they were already working on a solution that would be apparent in the next round of testing. In the meantime, it seems to me there is an easy enough fix: just make sure the rider's back wheel velocity is very close to 25mph, and that it is basically the same for both tests to which he is subjected.
Otherwise, Allied Aerospace appears to have taken about as much noise out of the system as anyone, and they've hit the ground running as a site for bicycle testing. That is to be expected, as Allied Aerospace has as a core business owning and running wind tunnels. They own several around the country.
One element of their attention to detail can be seen by looking at their mannequin for testing helmets. One can see the split between the top and bottom of the mannequin's head, in which is placed the balance (shown at the bottom of the photo). The strain gauges in the balance record the helmet's drag. I don't know if I like the way this mannequin is built — whether it can be positioned to mimic a rider's aspect into the wind — but it shows what the engineers at this tunnel are capable of doing.
How would the economics of such a facility work? One can imagine a tunnel costing $X per day, and maybe that's $4000 or $9000. You're running a camp, let's say, and you process a dozen riders who receive a half-hour of tunnel time each, at a rate of $500 or $1500 per head (strictly for the tunnel time). Assuming the revenue exceeds the day rate of the tunnel, you've got a working business model. That's precisely what Roch Frey and Paul Huddle did this past weekend with their Multisport School of Champions.
The utility of this tunnel and its protocol are apparent. Instead of taking a drag "snapshot" of a rider on his bike, a rider can pedal at race effort and his drag over a period of time can be averaged to produce a single number that tells a 5- or 10-minute "story."
How might this work? I can envision a bike fit expert positioning an athlete in two distinct positions, one in which comfort is maximized, the other in which the comfort position was optimized for aerodynamics. Let's assume that both positions are usable and ridable, but that one is just a bit more aggressive than the other. The subject's bike is fixtured in the tunnel (taking 10 minutes) and the subject performs a 5-minute test at race effort in position A. Five minutes of changover to position B, another 5-minute test, and the subject can see how his drag and pulse change while pedaling at a fixed power output. He'd then know how much a position maximized for comfort "costs" him in drag, if anything. Then drag is translated to time, and voila, the rider has decisions to make based on valid and precise testing.
Ironically, I think the secret to whether this sort of testing will work lies more with the expertise of the bike fitter than with the testing facility. To be sure, this tunnel has the killer app — the ability to test a rider over a period of time during a real-time, real-world effort. The trick is to make sure the positions being tested are the best, most valid, most optimized (for comfort, power and aerodynamics) rider positions available to the subject. Otherwise, a subject would waste his time for half a day in an expensive, precise wind tunnel, not finding out anything worth knowing.
If this is the camp of the future, those who're hosting it will need to be first-rate bike fitters. Then they'll need to have a first-rate tunnel. Do the Multisport School of Champions have both? I haven't spent enough time with them to know. But based on my Saturday with them, they're on the right track, and further along it than most.
