When an idea is brand new, every variation of it is rapidly tried, but eventually a consensus emerges as to which best combines low cost, performance, and ease of manufacture. This is why engine cylinders remain round rather than oval or square; a simple, inexpensive boring and honing process produces usable bores.
Before World War I, motorcycle engines proliferated—parallel twins, fours, flat-twins, V-twins, radials. Valves jutted in all directions. But in Britain through the 1920s, the dominant form became the vertical overhead-valve single, its height shoehorned between the pavement and the flat bottom of a gravity-feed fuel tank. Simplicity. As this occurred, writers of the time commented that, like the bicycle, the motorcycle had assumed its final form.
The tank above required that the engine’s intake port emerge from the cylinder head at right angles to the bore axis. When in 1937 Edward Turner added a second cylinder to create the British parallel twin, the right-angle intake port remained because the fuel tank above the engine defined the space available for it. Because this design was near universal, it was assumed by many to be correct.
But in the course of trying to win races, engineers and tuners found a problem. During valve overlap (the period at the end of the exhaust stroke when the intake valve has started to lift but the exhaust has not yet closed) air and fuel droplets from the intake were partly going straight out the exhaust. When in tests the intake port was raised from the horizontal (given a slight “downdraft” angle) of 10 to 15 degrees, this loss was reduced. Better yet, because a downdraft port reduced the severity of the angle through which intake flow had to turn down at the valve, there was also a potential for increased intake flow. AJS was a leader in adopting this, but it gradually became general.
Another trend was to store some intake flow energy in the cylinder by offsetting the intake port (usually to the right). The flow bias created by a 30-degree port offset made the charge rotate, or “swirl,” as it entered (we’ve all played with swirl while filling a bucket from a hose), and this rotary motion continued through compression, creating turbulence that greatly speeded up combustion. The faster combustion takes place, the shorter the time during which hot gas is confined by the piston against the cooler cylinder head, rapidly losing heat and pressure. Fast combustion saves energy.
Naturally, because orthodox design placed the carburetor under the fuel tank, adoption of intake downdraft and port offset compelled constructors to notch the fuel tank on the right to clear the carburetor in its slightly higher position. Such notches are features of tanks on Velocette and Norton racebikes.
What is good for the intake can also be so for the exhaust. When the late Kenny Augustine tackled the heat problem of the Harley XR-750 dirt-track engine, he knew that increased exhaust flow, by getting the hot gas out of the engine sooner, would push less heat into the cylinder head. It is usual for the intake side to receive most of the airflow attention because, while there is only 14.7 psi to push the intake (atmospheric pressure), residual exhaust pressure is more like 100 psi. That may work on small-block Chevys (which surely consume 95 percent of the airflow work performed in this world), but it doesn't on hot-running air-cooled engines like the XR. When a water-cooled engine is tuned to higher power, its thermostat just opens a bit more to compensate. But when an air-cooled engine makes more power, the only way it can rid itself of the increased heat is to get hotter. And the hotter it became, the lower the maximum compression ratio it could tolerate before detonating. And that was why the XR's exhaust ports had to be raised and further refined.
In the 1960s, Englishman Keith Duckworth learned to ignore the majority opinion in favor of his own test results. While the intake downdraft angle of motorcycle engines remained limited by the inability of carbs to work when tipped, Duckworth was finding flow gains with steeper downdraft intakes, made practicable by fuel injection that was unaffected by installation angle.
When he switched to four valves, offset ports and intake swirl no longer worked, so combustion, lacking the turbulence they generate, became slow. He tried to use piston-to-head squish areas (forming jets as the piston squeezed mixture out from between), but it was ineffective.
Then he had the idea of storing intake energy by causing charge rotation around a different axis. An offset intake port made the charge swirl around the cylinder axis, but the right intake downdraft angle could make the flow cross the cylinder head, down the far cylinder wall, back across the piston top, and up again in what he called "barrel motion." By playing with intake-port diameter (which controls velocity) and intake-downdraft angle, he could allocate how much intake energy was used to fill the cylinder and how much was used to generate his barrel motion (now called "tumble"). With tumble, Duckworth's engines burned faster, needing only 27 degrees BTDC ignition timing—a big jump down from the previous 40 or more. The resulting Cosworth DFV F1 engine won 155 GPs.
At Ducati, Massimo Bordi combined Duckworth's combustion system with desmodromic valves to produce the long series of "Ottovalvole" V-twin superbikes that continues to this day. Japanese makers at first adopted and increased Duckworth's intake downdraft, apparently without understanding its purpose. When they did understand it, today's remarkable sportbike engines resulted. Because downdraft had come to outrank conventional fuel-tank shape, what had once been the front of a fuel tank now became an airbox containing steeply downdraft intake ports and throttle bodies—creating a new orthodoxy. For a time, fuel tanks swelled into mushroom shapes, but then after 2002 tanks began to grow a "foot," which projected rearward and downward, allowing much fuel to be carried under the rider's seat, also allowing the rider to move forward. To make room, swingarm bracing moved under the arm, and shock linkage had to do likewise—one change, leading to another and another, in cascade.