The Wings of Voyager

Its historic flight may never have happened without behind-the-scenes design work at Ohio State University

When the famed Voyager aircraft traveled around the world last year, it did so with the help of several Ohio State University (OSU) professors who used Harris super-minicomputers to modify the airplane's wing design. The modifications were deemed necessary after checkout flights proved Voyager flew well in clear air but suffered sever lift problems in rain. In one early test, a rain-soaked Voyager plunged into a wild dive before it pilot regained control.

After considerable testing, the professors added a series of small triangular metal pieces--called vortex generators--to the craft's forward wing, or canard. This allowed the plane to fly well in bad weather without losing the lift it ordinarily achieves.

By so doing, the professors kept Voyager on schedule and did not jeopardize its historic flight with a costly and time-consuming redesign of the entire canard. 

How did they do it? Dr. Michael Bragg, associate professor, Department of Aeronautical and Astronautical Engineering at Ohio State, explains that, upon some initial test flights, it became clear that Voyager's surface roughness actually changed in the rain. "With the Voyager, rain would bead up on the front wing and the plane would nose down," says Bragg. "There wasn't enough control for the pilot."

Wind-Tunnel Tests

These tests were conducted in Ohio State's Subsonic Wind Tunnel located at the Aeronautical and Astronautical Research Laboratory (AARL). The laboratory is situated about eight miles northwest of the main campus in Columbus, Ohio. The tunnel is of conventional design with a three-by-five foot test section, eight feet in length.

The model of the Voyager canard airfoil, built for the test by Roncz, has a chord (the straight line joining its leading and trailing edges) of 22 inches, identical to that of the full-scale aircraft. Roncz took special care to build the model using the same construction techniques used on the actual aircraft.

The model was constructed with a hot-wire technique in which electrically heated piano wire was employed to cut a foam contour of the wing. Next, the model was wrapped with fiberglass. It included a 25 percent chord simple flap, which acts as the elevator on the full-scale aircraft.

Bragg says initial test were done with a Harris H-100 super-minicomputer. These tests included airflow experiments, where bits of data were gathered and fed into the computer. As part of the tests, Bragg and his colleagues conducted a wake survey to determine airflow section drag as a result of rainfall.

Simulating Rain

The tests quickly confirmed what the earlier checkout flight had already uncovered. When simulating cruise conditions, the tape produced a loss in lift of about 40 percent, which would be a terribly dangerous condition for the Voyager crew. Once Bragg and his OSU colleague, Dr. Gerald M. Gregorek, professor of aeronautical and astronautical engineering, pinpointed the problem, they were determined to find a solution that would stabilize the Voyager’s lift during rainfall without increasing the craft’s drag, which would waste precious fuel.

The Vortex Alternative

Bragg notes that loss of lift on the original Voyager design occurred when air lost energy as it traveled over the curve of the airfoil or rough wing surface. "air separated or left the surface, which caused the loss in lift," he says.

Mini Tornadoes

The vortex generators, four-tenths of an inch high, came in three types and were tested at several locations on the canard. "In three days, we tested the generators and recovered most of the lift lost," says Bragg. It took another three weeks before the professors had perfected the design to their satisfaction, however.

On-line test data was collected by the laboratory’s in-house data-acquisition devices. The system is based on Harris H-series super-minicomputers—an H-100 and H-800, to be specific—and associated sensory equipment. Both computer systems feature a 24-bit word structure and a 48-bit-wide central system bus architecture.

In a paper presented at the Harris Users’ Exchange 1987 Annual Meeting and Symposium , held last summer in Illinois, Bragg, Gregorek and Rick Freuler, laboratory manager, spoke of the benefits of using Harris systems for their research.

The Real-Time Advantage

Although it sounds complicated to those not familiar with computer-based data acquisition, Gregorek say the concept behind the Harris systems is a simple one. "These machines allow us to get answers to problems as our research is being done," says Gregorek. "all of this used to take so long for us to do. Now, complex calculations are made in real time, so as our experiment proceeds, we get data back answering our questions immediately."

Freuler, the laboratory manager, say the Aeronautical and Astronautical Research Lab has used Harris systems since 1973 for on-line data reductions from wind-tunnel tests.

He also notes that the main advantage of the Harris systems lies in the interactive interface they provide for on-line data acquisition and reduction. For instance, the paper explains that a wind-tunnel run period of less than 30 seconds—the kind of testing used on the Voyager’s wing—requires that most of the data-acquisition devices and software respond quickly to an operator’s instructions and hardware interrupts.

The "Slow" Human Factor

According to Freuler, the wind-tunnel software package was highly modularized, and each module provided a logically complete function. "Each module was a separate program that was loaded into memory and executed under operator control, usually by push-button selection," says Freuler.

During his presentation to the Harris Users Exchange, Freuler said the separate programs were held together by a small amount of information in a common area in main memory. A larger database of information was sorted on a shared disk file.

By using an interrupt-oriented, load-only-when-needed structure, the resultant conservation of memory space allowed data-acquisition activities to be carried out concurrently with other computing tasks going on in the lab. "I like to break data-acquisition and reduction into a number of separate tasks, then form them into one functionally complete package on the H-800," explains Freuler.

Anatomy of a Test

Next, Freuler says he calibrated the system, to account for any electronic variations, using a program designed for this purpose.

Once the output of a sensor is determined, it must be converted to engineering units for the particular task at hand. With the Voyager experiment, Freuler says there were many pressure points that had to be measured. After a final check to see that the system was ready to accept data, the wind tunnel was turned on and measurements were taken.

"After collecting the data, we could perform any number of calculations," says Freuler. "With Voyager, for example, we analyzed 48 pressure measurements on the surface of the canard." One such analysis examined the amount of drag as a function of lift. "We wanted to maintain the lift without increasing drag," notes Freuler.

Although the Voyager has completed its epic flight around the world, the Harris systems will likely assist in still more tests, including the archiving of raw data.

Why are the Harris systems appropriate for the Voyager tests and other projects like it? Says Freuler, "The Harris H-800 series super-minicomputer is ideally suited for real-time environments. As a 48-bit machine, it is optimally suited for scientific and engineering applications."

Rick Maule, director of product marketing for Harris’ Computer Systems Division, says he was pleased to see the results of the OSU projects. He notes that the Harris H-series is "the only super-minicomputer family to combine real-time interrupts with the processing power of a super-minicomputer."

The Voyager work, says Maule, is a "classic application," because the experimenters were able to gather real-time information during the wind-tunnel tests, even as numerous operator interventions were going on. Meanwhile, other computing activity in the lab continued unabated.

The Final Analysis

After more wind-tunnel work and refinement of the vortex design, the drag penalty fell to 10 percent. For an airplane that needed to span the globe without refueling, however, even this factor seemed too high.

So Gregorek and Bragg returned to the wind tunnel again and took another look at the canard. In the original wing, the aft portion was left dotted with rough fiberglass, a design decision originally made to conserve weight. Yet after testing the wing’s efficiency with smooth forward and aft sections, the professors discovered they could reduce drag by 15 to 20 percent.

In this way, the small drag loss caused by the vortex generators was made up, with a net increase in efficiency of 5 to 10 percent. What’s more, the new design could fly in rain. It was the scientific equivalent of having your cake and eating it, too.

And what was the final result of all this work with the vortex generators? 

Voyager’s pilot, Dick Rutan, radioed his ground crew—after encountering a rainstorm during his memorable flight with co-pilot Jeana Yeager—and told them to inform Ohio State University the generators ". . . . worked as advertised."