Ask any diehard racing fan, "who won the Indianapolis 500 last year?" and you'll not only get the correct answer--Arie Luyendyk--but possibly a short list of favorite Indy victories over the past decade. Then ask, "Who came in second?" The answer will most likely be silence.
Nobody remembers the runner-up. And no sport more visually demonstrates the enormous void that actually separates first place from second than racing. A few inches over several hundred miles of competition measures the difference between fame and obscurity.
As part of their effort to place their team in the winner's circle, race-car designers apply computer-aided engineering methods today more than ever. The era is over when the mere use of CAE tools separated the leaders from the pack. The difference between the best and the rest lies in how the computer tools are used, not if.
Software plays an essential part of any competitive race-car development program, be it NASCAR or CART. In most classes of racing, the cars are so closely matched that the power and precision offered by CAD, CAM, and analysis programs can uncover a mere 0.1% edge over the competition.
Aerodynamic advantage. Take the work of Bryan Holzinger, for example. As a design engineer for Team Rahal (Hilliard, OH), Holzinger spent a great deal of his time this past winter using Parametric Technology Corp.'s (Waltham, MA) Pro/ENGINEER (Pro/E) and Pro/MECHANICA to prepare for wind-tunnel testing of this season's car.
Team Rahal spends several days a month in the tunnel all season optimizing the car for different courses. But 1998 presented engineers with some added challenges. Ever interested in the slightest advantage, the team began using a new wind tunnel at The University of Glasgow in Scotland. However, a new tunnel means new mounting systems, new balance attach points, and a host of other new components. All must work the first time out to avoid wasting the $300-$600 per hour the tunnel costs to rent.
"We needed to come up with the mounting structure to attach our car to the balance in the tunnel," says Holzinger. "For that to happen, we needed a general layout of the tunnel in Pro/E and then we needed to model all the parts of the car that interface with the tunnel."
Team Rahal uses a 40%-scale, carbon-fiber model of the car, supplied by chassis manufacturer Reynard. During testing, sensors gather downforce, drag, and center of pressure data as the model goes through a series of ride heights. A test run might take 10 minutes or so, after which engineers converge on the model to tweak wing angles, add or remove vortex generators, vary the wheel tracks, or completely alter the car's setup--changing wings and all--from a speedway configuration to a road course.
One feature that differentiates the model car from the real thing is the way the wheels are mounted. Instead of being fastened to the car, the wheels are attached to large A-arms that pass through the tunnel walls. This system permits the wheels to quickly be moved out of the way to allow for rapid changes to the car. The A-arms lie outside the rolling ground plane over which the car is mounted; they don't impact the aerodynamic measurements. Championship Auto Racing Teams (CART) rules closely regulate tire sizes, so the drag of the wheels (which can be substantial) is not measured. What's important is the wheels' affect on the airflow over the chassis.
Holzinger and other engineers worked with professors at Glasgow to obtain measurements that allowed them to create solid models of essential parts of the wind tunnel in Pro/ENGINEER. They also used Pro/E to model more than 250 parts of the car. "Pro/E gives us the ability to create the parts in a solid designer, then create a working assembly of our model car," explains Holzinger.
By placing the CAD model in the wind-tunnel layout, engineers could confidently design mounting blocks and the A-arms. After approval by the university, Pro/MANUFACTURE created tool paths used to CNC mill the various parts from aluminum, or create aluminum molds for composite parts. "This process probably saves us 30-40% working time," Holzinger says. "It also gave us confidence that when we got to Glasgow everything would work right off the bat."
Winging it. Team Rahal also applied Pro/E to the development of a special rear wing optimized exclusively for speedways. While Reynard supplies its own speedway wing--for a mere $16,000--engineers felt they could improve upon its performance for a fraction of the cost.
The first step involved importing the wing endplate models from Reynard into Pro/E. Aerodynamicist David Watson then determined the positions and angles he wanted to test. From there, a 40%-scale, wind-tunnel model of the wing was made from solid aluminum using Pro/MANUFACTURE to create CNC tool paths for the team's Mazak milling center.
Wind-tunnel tests revealed the fruits of the labor: a 4% gain in downforce, about 40 lbs on the actual car. "This doesn't sound like much, but cutting even a tenth of a second off a lap time makes us extremely happy," says Holzinger.
Engineer Brian Willis used Pro/-MECHANICA to optimize the wing assembly. The final models were passed to Pro/MANUFACTURE to machine molds and mold inserts for the final layup. Engineers worked with sponsor Textron Specialty Materials (Lowell, MA) to make use of a graphite/boron composite skin, called Hy-Bor(reg), which made the wing thinner, lighter, stronger, and stiffer than the stock item.
Holzinger reckons the use of Pro/E not only saved time, but made for a higher quality final product. "It has all these tools that work well together," he says. "You can go from design to stress analysis straight to manufacture. It's a nice package."
Pit-stop improvement. When victories are measured in fractions of a second, and a pit stop might take 10 secs, it's easy to see how shorter pit times can win races. It was this sort of easy analysis that prompted engineers at Tasman Motorsports--a Hilliard, OH-based CART team--to try cutting the time of an infrequent, but lengthy pit operation: changing a suspension spring.
"In the existing damper system, spring perches were fed through a C-clamp," says engineer Luke Stonis. Mechanics had to lower the perch, remove the C-clamp spring with a special tool, pull the top spring perch off, and then remove the spring. "Not only was it time consuming, but the mechanic was left holding several parts; not an ideal situation in the heat of battle," he explains.
Stonis began drawing an improved shock quick release with the help of EDS Unigraphics version 12. The new design consists of a collar that threads onto the eye of the shock absorber. A 1/4-inch hole lets mechanics tighten or loosen the collar with a punch and hammer. Simple, and effective. Now mechanics only have to handle one part, and the process is easy to do while wearing the requisite heat-resistant gloves.
Bottom line? The new quick release cuts the 30- to 40-sec spring-change process by about 50%. "This is huge," says Stonis. "On an oval, where the lap times are 30 seconds, this is half a lap difference." The assembly also weighs about a 1/4-lb less than the stock mount, cutting a welcome lb from the entire car.
Stonis found Unigraphics to be a powerful tool for rapidly producing blueprints for the shop. "Once I have a part modeled, I can get the print out in about an hour," he says.
The program also helped him easily address several fit and optimization issues. As a retrofit part, the collar had to interface with existing components. "UG let me test fit parts and clearances before cutting metal," Stonis explains.
To take out as much weight as possible, he began with a set of design requirements that were fixed, while optimizing other aspects of the assembly's size. "UG really helped us with modifying the car in an efficient manner," he says.
Shifting gears. When John Wolfe attends an Indy Racing League (IRL) event, he feels a part of every car on the track. As design and development manager for EMCO Gears (Chicago, IL)--the company his grandfather founded--he headed the engineering of the EGB16, the six-speed gearbox used in every IRL car.
Officials chose to run a spec gearbox, they say, in an attempt to lower the cost of racing. And they chose EMCO Gears to design the box based partly on the company's past experience making gears and shafts for CART cars. Even so, this was EMCO's first complete transmission design. As a result, several factors added significantly to the challenge:
- Time. "We were told at the end of June that we'd need a gearbox for testing by September," says Wolfe.
- Cost. A CART gearbox might run $140,000, and a custom rear cover for a particular chassis and engine might cost $20,000 more. A complete EGB16 costs just $30,000.
- Durability. IRL wanted the box to last three seasons (bearings excluded). Says Wolfe: "The idea was that you could run last year's box this year and still be competitive."
- Flexibility. Teams can use two engines--Oldsmobile Aurora or Nissan Infiniti--and three chassis--Dalara, G-Force, or Riley & Scott. The gearbox has to interface with all of them.
To design the transmission's gears, Wolfe leveraged Universal Technical Systems (UTS, Rockford, IL) TK Solver, a program he introduced to the company in 1993. In fact, Wolfe partly credits the UTS program with helping the company bring gear design in-house. Whereas EMCO had originally made gears designed by clients, the firm steadily developed its own design expertise to the point where it could take on a complete gearbox, like the EGB16.
TK Solver is a mathematical modeling program which is said to make child's play out of solving complex equations. "You can type loads of equations into it, and it will split up all the variables so that you can solve for any of them," says Wolfe.
He makes considerable use of a UTS add-on package developed specifically for gear design. For a gearbox like the EGB16, Wolfe would compute gear life using standard American Gear Manufacturers Association (AGMA) equations built into the program. "Input horsepower, bending strength, bearing strength, compressive stress, the material, and other factors, and the program outputs gear life," he explains. "It lets you know what you're going to get before cutting metal."
Taking the durability goal quite seriously, EMCO engineers produced a ring-and-pinion gearset for the EGB16 that has, in some cases, run more than 3,800 miles on the same gears and bearings. IRL rules fix these gear ratios. Nonetheless, previous Indy cars usually had to change the gearset for nearly every race.
The robustness had one downside: possible safety concerns. For 1998, EMCO applied a series of modifications to the original design, primarily to make the solid gearbox less of a potential hazard during accidents. Some of the changes include:
- A more crushable transmission case. This alone cut nearly 40 lbs from the unit.
- A hollow section in the transfer shaft that makes it lighter and gives it the ability to collapse in an accident.
- Aluminum spacers between several gears that provide a crush zone during an impact.
Engineers also plan to adapt an idea from motorcycle transmissions where the shift dogs are milled like pockets into the gears, instead of protruding from their faces. The resulting webbing would not only make the gears lighter, but stronger--and the gearbox three inches shorter.
The greatest advantage TK solver might offer Wolfe, however, is the ability to quickly run through a series of designs that would have involved lengthy hand calculations before. A single scenario worked out by hand might take 2.5 hours, versus 10 minutes with TK Solver. "Before UTS, I didn't do too many 'what ifs'," he says.
Getting a scoop on scoops. Horsepower and drag are the yin and yang of racing. When the maximum propulsive force equals the resistive forces (wind drag plus some rolling resistance) you've reached maximum speed. On a short road course, a car's top speed plays only a small role in its overall lap time. But on a long, fast course, such as Road America at Elkhart Lake, WI, even a small increase on the top end can pay significant dividends.
The quest for the last mph led engineers at Neuman/Haas Racing to design an air scoop that applies the ramming effect of a car going 200-plus mph to help compress the intake charge. Neuman/Haas uses turbocharged Ford-Cosworth XD engines, and like all CART cars boost pressure is limited by a pop-off valve to 40 inches of mercury.
Inclusion of a pop-off valve would seem to preclude any added benefit from a ram-air intake system. But the intake performs its magic another way. "If we can reduce the work that the turbocharger has to do," explains Peter Gibbons, the team's chief engineer, "then we don't heat the air as much, the charge is cooler and denser, and we get more horsepower."
The flip side is that the scoop increases drag as well as disrupts flow to the rear wing. "You've got to gain enough from the pressure recovery to offset the losses," says Gibbons.
Engineer Rob Lloyd developed a solid model of the scoop using CADKEY (Baystate Technologies, Marlborough, MA) running on a Pentium Pro 200 with Windows NT. He began with a standard NACA wing profile, tweaked it, and then revolved it to form the jet-engine-nacelle-like shape. The scoop's intake measures 3.5 inches in diameter, a bit larger than the engine intake it feeds. Final optimization was done with the help of wind-tunnel work and a 45% scale model.
To create the molds used to manufacture the scoop, Lloyd output his final creation to an STL file, then contracted an outside rapid-prototyping shop to make a Laminated Object Manufacture (LOM) model from paper. "We needed a pattern that we could make carbon-fiber tools from," he explains. "The paper has a thermal expansion rate close to the carbon."
The scoop's real test is still to come on a high-speed course this season. In a race such as the Michigan 500, the inlet scoop is worth about one mph on the straights. At the end of the day this 0.5% or so increase over a portion of the course could be worth half a lap or more--the difference between fame and fifth.