The Bishop's Boys (34 page)

Working with Wilbur’s 1900 and 1901 data, Chanute struggled to explain the discrepancy between actual experience and the performance calculations. After analyzing the results of a single glide of August 8, 1901, he admitted: “I have tried to figure out this glide with the Lilienthal coefficients, but they do not fit at all.”
¹⁰

Still, Chanute found it difficult to accept the possibility of a fundamental error in data that had always been thought trustworthy. Perhaps he simply did not understand the mathematical procedures involved. He sent the Wrights a copy of the chapter on Lilienthal’s air-pressure tables reprinted in Major Hermann W. L. Moedebeck’s
Taschenbuch zum praktischen Gebrauch für Flugtechniker und Luftshiffer
, the standard engineering handbook to which all serious aeronautical experimenters referred, and asked that the brothers double-check his calculations.

Wilbur knew perfectly well that Chanute understood the “mode of calculation.” The problem was not with the equation, but with the numbers that were plugged into it. To grasp the difficulty, we must return, as Wilbur and Orville did, to the basic formula devised to calculate lift:

L = k X S X V
²
X C,
_L

Recall that in this equation:

L = lift measured in pounds

k = a constant coefficient for air pressure

S = total surface area of the machine in square feet

V = total velocity of the machine (headwind + forward speed)

C
_L
= a coefficient for lift that varies with the shape of the airfoil surface and the angle of attack

The surface area and velocity were known quantities. The coefficients for air pressure and lift had been inherited from their predecessors.

The Wrights, like Chanute, had followed Lilienthal’s practice, using John Smeaton’s figure of .005 as the coefficient for air pressure under standard atmospheric conditions. But that figure had been experimentally derived, and not everyone agreed that it was accurate. During the course of his extensive research, Chanute had identified up to fifty separate coefficients ranging from .0027 to Smeaton’s very high .005. As Wilbur pointed out in 1908, the variation was not difficult to understand in view of the problems faced in trying to measure minute differences in pressure with primitive instruments.
¹¹

It occurred to him that both the low lift and the reduced drag of the 1900–01 machines could be explained by assuming that Smeaton’s coefficient for air pressure was too high. On September 26, only a week after his speech, Wilbur called Chanute’s attention to the fact that “Prof. Langley and also the Weather Bureau officials found that the correct coefficient of pressure was only about .0032 instead of Smeaton’s .005.”
¹²

The longer Wilbur thought about it, the more certain he became. “While I have not personally tested the point,” he told Chanute a week later, “I am firmly convinced that it [.005] is too high.”
¹³

In fact, Wilbur had already done some testing. Referring to his meticulous notes on the performance of the 1901 glider, he calculated a lower, and much more accurate, value for the coefficient of air pressure. Both he and Chanute had developed tables of reduced data for six particularly well-documented flights made between July 3 and August 17, 1901. Will used the data on wind speed and aircraft weight to calculate a coefficient of lift for each of those flights, then proceeded to work back to a calculation of the coefficient of air pressure. For the six glides in question, those coefficients ranged from .0030 to .0034. The average was .0032. He could see “no good reason for using a coefficient greater than .0033 instead of .005.”
¹⁴

Wilbur had identified and corrected one source of error in the Lilienthal calculations, but there might be others. Next, he turned his attention to the problem of lift and drag coefficients. Unlike the pressure figure, the lift and drag coefficients varied with each airfoil shape at every angle of attack.

A great many nineteenth-century engineers and physicists had struggled to chart the tiny fluctuations in lift and drag dancing back and forth over a wing set at various angles to the wind. In planning their first machines, the Wrights had referred to Lilienthal’s coefficient tables for his airfoil (arc of a circle, camber of
1
/
12
). That information was now in question.

Wilbur and Orville had a low tolerance for guesswork. If they were to continue, there was no choice but to devise a means of checking the Lilienthal coefficients and, if necessary, correcting them. The temptation to replace uncertainty with hard facts and figures was irresistible.

It was a problem made to order: their genius for visualizing mechanical solutions to theoretical problems would be fully exercised in conceptualizing the experiment and designing the necessary apparatus. Their personal qualities—an insistence on absolute precision and an unflagging confidence in the results of their own research—were equally essential. Nothing less than absolute confidence would do, for the Wrights would be staking their lives on the numbers they produced.

The first task was to prove that a problem really existed. To accomplish that, they designed a mechanical analogue of the Lilienthal tables. According to Lilienthal, a cambered wing with a surface area of 1 square foot set at a 5-degree angle of attack should produce just enough lift to equal the pressure on a flat plate measuring .66 of a square foot set at a 90-degree angle to the air flow.

The brothers recreated that situation, mounting a bicycle wheel, free to turn, on its side with a cambered Lilienthal airfoil of the correct dimension fixed to the front of the rim at a 5-degree angle and a flat plate positioned vertically one quarter of the way around the wheel from the airfoil. If Lilienthal was correct, the wheel would remain stationary when placed in a strong wind. If he was wrong, either the lift of the airfoil would overpower the pressure on the flat plate and turn the wheel in one direction, or the pressure on the plate would be higher than the lift of the airfoil, turning it in the other.

When the Wrights discovered that they could not get satisfactory results from their testing apparatus in a natural wind, they mounted the wheel horizontally over the handlebars of a bicycle and rode madly up and down the street. The pressure on the plate was much higher than the lift of the airfoil.

The experiment proved the existence of an error somewhere in Lilienthal’s work. Was it the result of his use of Smeaton’s coefficient, as the Wrights suspected, or did additional errors lurk in Lilienthal’s measurement of the lift and drag coefficients for his airfoil? The whole business gave them pause. Could the great German gliding pioneer, a man for whom they had an extraordinary respect, have made such a mistake? Why had no one noticed it before them?

The Wrights constructed their first wind tunnel to double-check the results of the bicycle-wheel experiment. It was a simple device, a square trough with a fan at one end driven by the overhead line shaft. Confirmation of the error was quick in coming.

Nothing was left but to recheck all of the figures in the Lilienthal table. They would fashion a small model of his wing, test it at every angle of attack, and calculate the coefficients for themselves. And they would go a giant step further. Lilienthal and most other experimenters had tested only single airfoil designs; Wilbur and Orville would study a wide range of shapes and sizes, searching for the most efficient lifting surface.

The original small “channel” was inadequate for the task. They constructed a new wind tunnel, a wooden box six feet long and sixteen inches square on the inside. To the uninitiated eye, it looked like nothing more than a long packing crate resting on two sawhorses. In fact, it was a delicate instrument that would permit the Wrights to unlock the secrets of a wing.

The purpose of the tunnel was to move a smooth, steady stream of air through the box at a constant speed of 27 miles per hour. A sheet metal hood at one end partially shielded a two-bladed fan that was driven through the gearbox of an abandoned grinder at a speed of 4,000 rpm. It would not do for the wind to careen off the walls of the tunnel, creating eddies, swirls, and cross-currents, so the Wrights placed a “straighter”—a cross-hatch of thin wooden strips covered with wire mesh—just in front of the fan.
¹⁵

Disappointed and puzzled by the poor performance of their machine, the Wrights returned to Dayton in the fall of 1901 and set to work on a testing device. Initially, it consisted of an airfoil test rig mounted on the handlebars of a bicycle.

The 1901 wind tunnel offered a far more precise method of gauging the performance of airfoils.

A section of the wooden top of the tunnel was replaced by a pane of glass. This window into the heart of the tunnel looked down onto a spindly metal balance bolted to the tunnel floor: carefully crafted out of bicycle spokes and hacksaw blades, it was designed to balance lift against drag, just as in the case of the far more primitive bicycle-wheel experiment. This time the Wrights would be able to make precise measurements of what was occurring as the wind flowed around a small airfoil mounted on the balance.

The balances were not much to look at—they were delicate things, small enough to fit in a shoe box. The slightest jar would dislodge the many pins on which the various parts rested, reducing the device to an assortment of bits and pieces on the tunnel floor. Reassembly was exasperating, something akin to building a house of cards; one slip and the entire edifice collapsed.
¹⁶

Perhaps because they were so small, fragile, and totally devoid of the aura that surrounded the gliders, it is easy to underestimate the importance of the balances. In fact, they were as critical to the ultimate success of the Wright brothers as were the gliders.

George Spratt had first suggested the notion of a testing machine that would balance lift against drag. He had played with the idea back in Pennsylvania, devising a testing instrument that operated in the open air. Spratt had not been successful, but he had described the episode to the Wrights at Kitty Hawk.

Small wonder that Spratt had failed. The difficulty of designing such a device required Wilbur and Orville Wright to draw on the full measure of their ability to visualize a complex physical problem in mechanical terms. How could they suspend a small test surface in their tunnel so that it would be free to move in response to changing conditions? How could they measure that movement? How would they translate those measurements into coefficients for lift and drag?

In all, the Wrights built three balances for their wind tunnel. The first, inspired by Spratt’s notion of obtaining all the required information with one simple test, was designed to measure both the lift of a test surface and the ratio of lift to drag. It was a failure.

The Wrights then split those functions and developed two subsequent balances, one to measure lift and the other to measure the lift to drag ratio. In simpler terms, the first balance provided the raw materials for calculating the lift coefficient (C
_L
); the second enabled them to calculate an accurate drag coefficient (C
_D
).

The tunnel and balances were complete and in use by November
22
. The next month was a time of hectic activity, as the brothers moved rapidly through a series of tests with perhaps 150 small model airfoils. Milton once remarked that Wilbur “systematized everything.” Nowhere was that more apparent than in the wind-tunnel experiments.

Procedure was everything. No one but the operator was allowed to stand near the tunnel during a run, nor could any change be made to the upstairs room at the bike shop where the experiments were conducted. They discovered that the tunnel actually set up a circulation of air around the entire room, so that even moving a piece of furniture had an impact on the readings.

The test procedure was quite simple. The Wrights had manufactured all of their test surfaces out of 20-gauge sheet steel, cut to size with tin shears and hammered into shape. Each was about six inches square and
1
/
32
of an inch thick. Two metal prongs were soldered to each surface so that it could be clipped in place on the balances.