125x125 banner
Stylish Motorcycle Riding Gears

Four Valves Per Cylinder, Part 2

Kevin Cameron has been writing about motorcycles for nearly 50 years, first for <em>Cycle magazine</em> and, since 1992, for <em>Cycle World</em>.

Kevin Cameron has been writing about motorcycles for nearly 50 years, first for <em>Cycle magazine</em> and, since 1992, for <em>Cycle World</em>. (Robert Martin/)

By 1914, auto racing engines had adopted the four-valve DOHC concept as optimal. The idea had been pioneered by three of Peugeot’s racing drivers and their designer/draftsman Ernest Henry, and another feature of their work came with Henry’s conviction that cylinder bore should be one-half of the stroke; i.e., a small bore with a very long stroke. This seems remarkable today; nearly all motorcycle engines take the opposite approach and have bores substantially bigger than their strokes.

The small bores of those early times were dictated by the fact that their iron or steel pistons conducted heat poorly. Since both materials have low thermal conductivity, the only way to prevent very high piston-crown temperature was to reduce piston diameter. and shorten the heat path from hot piston-crown center to cooler cylinder wall. Such a design made detonation and piston-crown sagging less likely. If you look at pistons from this era, whether for auto, motorcycle, or aircraft engines, you will often see piston-dome support struts extending downward to contact the wrist pin through a slot in the con-rod’s small end. Imagine piston temperature high enough to make steel slowly yield.

Related: Four-Valve Beginnings, Part 1

The need to use such small bores gave further reason to provide four valves per cylinder as a means of fitting adequate valve area into such limited cylinder-head area.

Since the early experiments made in the 1903–1905 period, it was accepted that to avoid destructive crankshaft torsional, or twisting, oscillations, racing engines should use no more than four cylinders in a line. The engines of the “giant racers” built around 1910 were limited to around 1,200 rpm; Peugeot’s big step forward roughly doubled that.

Technological change most often occurs when “progress” logjams against obstacles. We can understand how the Peugeot could overcome the giant Fiat: The Fiat had twice the displacement, but the Peugeot could turn twice the revs; that is, it could perform its power-producing cycle twice as often. Close-to-equal power in a smaller, lighter car added up to equal or better overall performance.

Important obstacles remained. Despite turning twice the big Fiat’s revs, the Peugeot was still limited by early bearing technology. And no one had scientifically investigated airflow through engine ports. Additionally, gasolines with poor anti-knock qualities limited compression ratio to around 5:1, while hot-running steel pistons constantly threatened to provoke detonation.

Fiat’s Two-Valve Revolution

Outstanding engineers, working at Fiat under Guido Fornaca, pushed Grand Prix car engine design forward in several ways, leaving behind the four-valves-per-cylinder concept so lately considered essential. The first part of their revolution was cylinder multiplication, dividing the engine’s displacement among six cylinders instead of the previous four. How was this possible, when the first inline-sixes of 1903 had suffered so badly from crankshaft torsional oscillation? The Fiat 404′s crank was less than 22 inches long, but featured substantial 40mm journals. The emerging trend was the proliferation of higher-revving straight-eight engines in auto racing, followed by such engines in production cars; I loved the sound of the straight-eight in my uncle’s 1948 Buick convertible.

The second big innovation was to recognize that as a cylinder is made smaller, its valve area does not decrease as fast as does its displacement. This is because displacement is determined by cylinder dimensions, cubed, while the area available for valves is cylinder dimension, squared. That favorable relationship made it possible to simplify the engine by reverting to two valves per cylinder. To ensure adequate flow area (scientific development of port flow still lay roughly five years in the future), the stems of each cylinder’s two valves were widely splayed at a valve-included angle of nearly 100 degrees, something that would continue to influence European engine design for years. If you imagine a cross-section of the combustion chamber as a pup tent, it’s clear that you can fit bigger valves into the sides of the tent than into its floor.

The use of an all-rolling-bearing crankshaft, which was able to survive on very little oil, allowed the Fiat 404 to reach double the rpm of Peugeot’s 1912/13 design. The Fiat racers were very successful, and their success made the two-valve design attractive.

As a concrete example, when engineer Piero Remor designed Gilera’s postwar 500cc inline-four GP motorcycle engine, he gave it Fiat’s wide valve-included angle. He did the same for MV in 1950. In addition, engines such as Edward Turner’s 1936 Triumph 500 “Speed Twin” and Harley-Davidson’s EL overhead-valve V-twin of that same year used 90-degree valve-included angles. Two valves at such a wide included angle had become the official “right answer.”

Two-Valve Aircraft Engines

In the fast-growing aviation industry, air-cooled radial engines overwhelmingly used two valves per cylinder. The reason was clear as early as 1916, when Professor Gibson at England’s Royal Aircraft Factory observed that the fewer holes you put into a cylinder head, the fewer problems you have with cracking and heat distortion. Major liquid-cooled aircraft engines of the 1930s and ‘40s such as the Rolls-Royce Merlin and the US Allison V-12 used four valves per cylinder, but the big air-cooled radials produced in much greater numbers by the US, Germany, and Japan during World War II were all two-valve designs. Even today, the makers of classic air-cooled big-twin motorcycle engines have had to adopt “topical” liquid-cooling to keep the valve seats distortion free.

When formal study of engine airflow appeared in the mid-1920s, it revealed that the orifice coefficient of a single intake valve in a hemi combustion chamber could be made quite a bit higher than that of paired intakes in a four-valve chamber. Small paired exhausts had been more reliable in the early days of poor valve materials, but in the 1920s the coming of stainless alloys rich in chromium and nickel made this irrelevant. At the time, both racing and production motorcycles produced moderate revs, and this gave a “scientific” basis for setting aside the four-valve concept. Although the Offenhauser engines used in US Champ Car racing continued the use of four-valve design, as late as the mid-1950s even Mercedes gave its M-196 straight-eight GP car engine two valves per cylinder.

So what about the 1930s’ TT successes of Rudge and Excelsior four-valve motorcycle engines (with all those rocker arms)? I suspect there was a short window of opportunity for four-valve engines to perform well just on the basis of their large valve area, but as the airflow capability of two-valve designs was rapidly elevated by the airflow work of Harry Weslake and others, their simplicity and good performance left the mechanical marvels behind. Right through the 1950s European auto and motorcycle engine design was dominated by two-valve engines.

Enter the Overhead Cam Layout

But not all was sweetness and light. As overhead cam valve drive (OHC) showed its superior ability to control valve motion as revs continued to rise, the OHV concept of operating overhead valves via rockers and pushrods was falling out of use. First to go were engines with the largest, heaviest valves, the 500s. The last win by a pushrod OHV in the Isle of Man TT races came in 1936, by a 250 New Imperial.

Another problem of two-valve engines was the inertia of their large single valves. This made opening and closing them easier if they were given very long valve-open timings and limited lift, resulting in long valve overlap (often over 80 degrees) and late intake closing (75 or more degrees after BDC—such timings given at a 0.020-inch checking clearance). Both aspects boosted peak power but also gave engines “megaphonitis,” the inability to accept much throttle below 70 percent of peak revs without burping and hiccuping. The late intake closing allowed back pumping of the slow-moving fresh charge as well, resulting in weak bottom- and mid-power.

A tremendous amount was learned with two-valve engines. But with the 500 fours from Gilera and MV running to 10,500 rpm and Mondial’s 125 single revving to 13,000 by the late 1950s, valve control was becoming a potential issue.

This is Part 2 of a three-part series. Stay tuned for the concluding segment.

View full post on Cycle World