After a day in the shop that yielded nothing except scrap, dulled tools
and renewed respect for the intractability of materials, my thoughts turned
to the Wright brothers. One hundred years ago, these two Ohio bicycle
mechanics and their helper Charlie Taylor scratch-built the world's first aircraft engine,
using little more than hand tools, a drill press and a lathe. The
water-cooled, four-cycle, four-cylinder 200 CID engine required just two
months to complete, during a period when the brothers were busy constructing
the Flyer and inventing the science of propeller design.
How did they do it? What habits of mind did they bring to the task? And what sort of technical support could be found for construction of gasoline engine in 1902?
Wilbur (born in 1867) and Orville (born in 1871) completed four years of high school, but without receiving diplomas. Orville missed the commencement and Wilbur elected not to take some of the proscribed courses. But the family put a high value on learning; their father, Milton, was a writer and bishop of the United Brethren Church of Christ and their mother had, remarkably enough, a degree in mathematics.
The brothers opened their first bicycle repair shop in Dayton in 1892 and, three years later, branched out to manufacture bicycles made to customer order. During this period, American factories were turning out about a million bicycles a year, using the techniques of mass production that had been pioneered by firearms and sewing machine manufacturers.1
Handmade bicycles, while more expensive than the factory products, exhibited
more rigorous engineering and higher standards of fit and finish. The Wright
brothers would apply the same patient craftsmanship to the construction of
airplanes and engines.
The wind tunnel with precision balances made by Charign They began their research into the flight in 1896 by reading everything they could about the subject. By 1900 the brothers had moved from experimental kites to tests of a man-carrying glider at Kitty Hawk, North Carolina. The site had been chosen for its steady winds and soft sand. The brothers returned with a new glider in 1901 and with another machine in 1902. That glider, developed with the aide of a Orville's wind tunnel demonstrated that roll and pitch control problems had been solved. Pitch control was ragged, but that could be addressed at a later time. The next step was to add propulsion.
The engine was only a component of the larger effort, which was to build a flying machine. They calculated that 8 bhp with an all-up weight of 180 lb would be sufficient to put the machine into the air. Back in Dayton at the end of the year, Wilbur went to work on propeller design. Orville mailed out inquires to automobile and engine makers describing what they needed. No suitable engine was available, or, more precisely, available at a price the brothers could afford. The bicycle shop made a profit of between $2000 and $3000 a year, which did not permit the brothers to commission a one-off engine.2
The only option was to build their own.
Wilbur and Orville were not experts in the new technology of gasoline engines, which were just then appearing on automobiles, although steamers were more popular. But, with that almost uncanny economy of effort that characterized the whole enterprise, the brothers had built themselves an IC engine, powered by illuminating gas, the year before. It was a massive single-cylinder machine displacing 200 CID. Even with a compression release, starting it must not have been fun. The engine developed 3 hp at 447 rpm.
Performance data was determined by tests on a homemade Prony brake. Whatever the brothers built, they tested.
This engine powered the lathe and drill press through a system of overhead shafts and flat belts. It was drove the wind-tunnel fan and, tuned carefully on the brake, was used in experiments to determine propeller efficiency.
Building an engine gave them experience in critical areas, such as ignition systems, crankshaft construction, and selection of valve materials. It was practice, as were the more than 2000 unpowered glider flights they made before December, 1903. It is also interesting to note that the aircraft engine induction system was an adaptation of the system used on the illuminating gas engine. They did not atomize gasoline with a carburetor; instead the fuel was dribbled into the intake manifold and vaporized by engine heat.
And, as with their aircraft, the brothers began with a mathematical description of the problem, however inadequate. An index card preserved in the Wright papers gives the formula for torque, horsepower and rpm.
Contemporary automotive practice seems to have been the model for the water-cooled four-cylinder inline layout and influenced many of the details of the design.
The bore and stroke were both 4" for a total displacement of 200 cubic inches. Why the brothers selected such a short stroke, when most engines of the period were "under square" is unknown. Perhaps it had to do with the maximum center offset of their lathe. The compression ratio was about 4.5:1 and suitable for the 60 octane gasoline then available.
No throttle was fitted. The engine appears to have been designed to run at constant power throughout the whole flight envelope. Ignition could be manually retarded as an aid to starting, but there is disagreement about whether timing could be adjusted in flight. At any rate, the control over engine speed provided by retard would not have been dramatic.
The originality of the Wright design was in the extensive use of aluminum to save weigh and in the ways in which the design was adapted to the limitations of their tooling. Some fabrication work may have been farmed out to local machine shops, but the bulk of it was done in-house on a lathe and drill press. Someday in that great shop in the sky, I would like to ask Wilbur and Orville why they did not purchase critical, hard-to-make parts, such as the crankshaft and connecting rods. Perhaps, caught up in the rhythm of the work, the idea of going to outside suppliers, working around someone else's givens, would have been a kind of violation.
Even so, Engine No. 1 did not approach the state-of-the-art of 1902. Some features, such as splash lubrication, vacuum-actuated intake valves and make-and- break ignition were obsolescent. Nor was the performance goal of 8 hp at 180 lb anything to write home about. But conservatism has appeal for aircraft engine designers.3 As Orville wrote in a letter of April 12, 1911,
"We look upon reliability in running as much more important than lightness of weight in aeroplane motors."
The Dec. 26, 1948, issue of Collier's Weekly published an interview with Charlie Taylor, whose job it had been to turn the brothers' ideas into metal. At the time of the interview, Charlie was 83 years old:
"If they had any idea in June of 1901 that someday they'd be making a gasoline internal combustion engine for an airplane and would need some first-rate machine-work for it, they sure didn't say anything about it to me. But when they returned from the South in 1902, they said they were through with gliders and were going to try a gasoline engine. They figured on four cylinders and estimated a bore and stroke at four inches.
"While the boys were handy with tools, they
had never done much machine work and anyway they were busy on
the air frame. It was up to me. My only experience with a gasoline engine
was an attempt to repair one in an automobile in 1901. We didn't make any drawings. One
of us would sketch out the part we were talking about on a piece of scratch
paper and I'd spike the sketch over my bench."
The lack of engineering drawings is supported by an entry in Orville's 1904 diary, "Took old engine apart to get measurements for making new engine."4 However casual the approach to drawings, the design was far too complicated to pulled off on an ad-hoc, piece-by-piece basis. One of the three men, probably Orville, must have been able to visualize the interrelationships between components early on, before much metal had been cut.
One researcher suggests that the aluminum crankcase casting, with its array of bosses, webbing and provisions for mounting other components, may have been visualized by using the wooden patterns as "three-dimensional drawings upon which design details could be worked out."5
The brothers moved cautiously into the unknowns of engine design, minimizing risk at every opportunity. Each newly machined part was weighed and two preliminary power output tests were made during the course of construction. "The first thing we did," Charlie recalled, "was to construct a sort of skeleton model, a test cylinder of about a four-inch bore…we hooked up the test cylinder to the shop power, smeared it with oil with a paint brush, and watched it run for short periods. It looked good, so we decided to go ahead with a four-cylinder model…"6
The second test was made six weeks into the project, this time using all four cylinders and the crankshaft. The parts were mounted in an open frame, oiled by hand, and powered up just long enough to get a torque reading from the Prony brake. 7 Satisfied with the outcome, the brothers proceeded with work on crankcase/water jacket.
The accompanying drawing illustrates a later version of the engine with pressure lubrication, spark plugs and pressurized fuel delivery. But, except as noted, other features were carried over from No. 1.
Cylinder Barrels. Cylinders were cast iron, with extremely thin walls to reduce weight. Securing the barrels to the aluminum crankcase and making watertight seals at the upper and lower ends of the barrels demanded precise machining on an expensive one-off casting. While some work might have been farmed out to other shops, Charlie Taylor said that he was responsible for the fit of the barrels.
As shown in the drawing below, the barrels had two sealing flanges, one at the valve end and one near the bottom of the bore. The valve ends of the barrels were threaded to mate with bosses on crankcase/jacket casting. When the barrels were screwed into place both flanges had to form watertight seals.
Barrel threads extended above the aluminum casting. The iron valve bodies made up on these threads to function as lock nuts and load the sealing surfaces by putting the water jacket in compression. This arrangement distributes clamping loads evenly around the largest possible area.
The IDs of the cylinder barrel bosses and recesses in the valve body castings formed the combustion chambers.
Valves. As mentioned earlier, the intake valves were "automatic" in that they opened in response to cylinder vacuum. Soft springs, aided by compression pressure, were responsible for sealing. A conventional cam and rocker arm arrangement operated the exhaust valves. The rockers were made of steel strips, riveted together. The two-piece valves consisted of cast-iron valve heads threaded on tool-steel stems.
In terms of design, the valves were the worst feature: automatic intake valves functioned rev limiters and the absence of any provision for cooling cost power. Upon startup, the tuned version of the engine developed 16 hp; a minute or so later the valve bodies glowed red and output dropped to 12 hp.
Pistons. Cast-iron pistons each carried three compression rings, but no oil ring. Pistons were individually fitted to their bores, since parts interchangeability was of no concern.
Connecting Rods. Reflecting the attitude of "dance with the one you brung," the rods were three-piece units, with shanks made of what appears to have been seamless-steel bicycle tubing. The ends were bronze castings and threaded to the shanks.
The most complete analysis of Wright engine technology appears in the paper entitled "An Evaluation of the 1910 Wright Vertical Four Aircraft Engine," and available on the Internet. 8 Cu-Sn-Pb rod end castings are similar to several commercial alloys of the period, including the popular Naval M bronze. The report notes that a sharp radius at the threaded joint and the almost certain presence of voids in the casting would have concentrated stresses. At least one Wright engine experienced catastrophic rod failure in service. Orville switched to conventional I-beam connecting rods on his six-cylinder design of 1911.
Crankshaft. Charlie Taylor made the five-main bearing, single-plane crankshaft from a blank of tool steel. That particular crank was borrowed from the Smithsonian by the Aero Club of America for a display in 1916 and never returned. But its composition was almost certainly identical or very similar to the 1911 crank analyzed by Pratt & Whitney in 1966, which contained 1.5% Cr, 2.0% Ni, 0.3% Mn and exhibited a Rockwell hardness of C-24.9
Only a small minority of tool steels contain nickel, used conjunction with chromium to improve toughness and wear resistance. The trace amount of molybdenum is a foundry convenience, although it does contribute to strength and toughness. What one finds unusual is the degree to which the alloy was annealed. Rockwell C-24 is about as soft as alloy steel gets. By way of comparison, the hardest tool steels rarely exceed C-70, file hardness averages about C-65 and good quality knives test at C-50 or less.
A soft crankshaft might contribute to more rapid wear, but that was hardly a consideration for an engine whose cylinder heads glowed red. Nor is the relationship between journal hardness and wear linear, especially when alloy steel journals turn against babbit bearings. The main advantage of using dead soft steel was that is permitted fabrication to proceed with hand tools and conventional lathe cutting tools. The material also provided insurance against catastrophic failure. Five main bearings, one between each throw, located the crank laterally. But the long, slender shaft had little torsional resistance. Soft steel permitted the crankshaft to flex without cracking.
Modern designers take the same approach for the heavily loaded crankshafts of large compressors. Typically these crankshafts have a Rockwell hardness of around C-27. After long service, compressor cranks can develop enough permanent twist to dislocate the drive keyway, but they rarely fail.
Charlie Taylor carved the crankshaft from a blank about 2 ˝ ' long by 4" thick. He began by outlining the shape with hundreds of small-diameter holes on the drill press. Much hacksaw and cold chisel work followed. Finally, the workpiece was moved to the lathe where the mains were turned between centers. Sizing the crankpins required a four-jaw chuck and a tailstock offset 2" on each side of center.
The flywheel, small for the period, weighed about 25 lb with the mass concentrated at the rim. It was shrunk-fit to the crankshaft.
Except for the cylinder tops and the upper half of the intake manifold, the engine was encased in an aluminum envelope. This complex, thin-walled casting comprised the crankcase, water jacket and lower part of the intake manifold. It also incorporated four hollow lugs for the engine mounts.
High winds overturned the Flyer after its fourth powered flight in the afternoon of December 14, 1903, damaging the airframe and cracking the aluminum casting. As displayed by the Smithsonian, the 1903 Flyer is fitted with a later crankcase that incorporates full-pressure lubrication.
However, fragments of the original casting survive, and Dr. Frank W. Gayle, a metallurgist employed by the U.S. Bureau of Standards, was able to make analysis of the material. His findings are interesting.10
The alloy is a commercial grade of the period, mixed with 8% copper for ease of pouring and containing 1.0% iron and 0.4% silicon as impurities. Wall thickness of the samples examined ranges between 4 and 5 mm.
The micro-structure exhibits precipitation hardening, resulting from exposure to heat after casting. The effect is to increase the strength and toughness of the alloy. Precipitation hardened aluminum was first described by the German researcher Alfred Wilm in 1909. It immediately went into production under the trade name "duralumin." The heat-treated Al-Cu-Mg alloy increased its strength over time at room temperature. First used in zeppelin airframes, duralumin has since become the model for all high-strength aircraft and aerospace alloys.
Of course, precipitation hardening of the Alcoa alloy by the Dayton Foundry in 1903 was an accident, probably resulting from heating the molds to cope with the complexity of the casting. Samples of the material Dr. Gayle had made up and cast in heated molds exhibited a similar microstructure.
Another intriguing aspect of the casting, mentioned in Dr. Gayle's paper, is the difficulty experienced in replicating it for a reenactment of the first flight at next year's centennial. The skills of modern foundrymen do not appear to be up to the task. Howard DuFour, Charlie Taylor's biographer, pointed out in a telephone conversation that the industrial age peaked around 1919. Many of the skills associated with that age have been lost. We forget things.
The case of the Iceman suggests how much we have forgotten. The body of this man who lived 5300 years ago was discovered in the Italian Alps in 1991. His equipment included an elaborate assortment of tools, weapons and medical supplies. One researcher remarked:
… his equipment contained 18 different types of wood. This means that for every tool that he used -- the axe handle, the knife handle, the arrow shafts -- he sought out the most suitable material. And this is knowledge we've forgotten today.11
Sources1 Peter L. Jabob, Visions of a Flying Machine, Smithsonian Institution Press, 1990, p. 8.
4 Harvey H. Lippincott, "Propulsion Systems of the Wright Brothers," p. 82, reprinted in The Wright Flyer: An Engineering Perspective, edited by Howard S. Wolko, National Air and Space Museum, Smithsonian Institution Press, Washington, D.C. 1987. 5 Ibid., p.81-2.
6 Collier's, p. 30.
9 Lippencott, opt.cit., p. 92 and 95.
10 Frank W. Gayle and Martha Goodway, Science, Nov. 11, 1994, Vol 266, pp. 1015-1017.