Mr. Honda’s Big Idea

[Hugh Janus]

Honda made its racing debut at the 1959 Isle of Man TT in the 125cc class. (Honda/)

At the 1954 Isle of Man TT races, Soichiro Honda was stunned to learn that European bikes were making three times more power per liter than the best Japanese machines.

In his wartime business of manufacturing aero-engine piston rings, he had learned the value of research. When his first rings flunked quality standards, he sought understanding from university metallurgists and achieved success with an improved process. R&D is the key to moving forward.

Now Mr. Honda saw German Rupert Hollaus win the 125 TT on a high-revving NSU single. Four NSU twins swept the 250 TT, with winner Werner Haas averaging 90.867 mph over 113 miles. The 350 TT was won by Rod Coleman on a British AJS single at 91.54 mph.

The three major components of horsepower are 1) engine displacement, 2) stroke-averaged net combustion pressure, and 3) rpm. The well-traveled path to power was the English way: a low-revving but deep-breathing single-cylinder engine of great refinement. The new way, so powerfully demonstrated by NSU, was to make power through rpm.

Related: Honda CB750 - CLASSICS REMEMBERED

Mr. Honda explored both the traditional and the new. Because his company was producing small commuter bikes, the lightweight TT classes were especially interesting. He wanted to know more, so he bought an NSU and an Italian Mondial, both of which were designed to make their power through higher rpm, and shipped them back to Japan.

Now comes the really interesting part. Honda decided to develop his own machines to compete in a future TT. For the 1959 Isle of Man TT’s 125 class he built tiny twins with double overhead camshafts. Were they copies of NSUs or Mondials? Not in the least; Honda’s researchers equipped the new twins with four valves per cylinder. As a kind of insurance they also built some two-valve versions of the same engine.

Honda engineers Tadashi Kume and Kimio Shinmura knew the literature, and it was discouraging. At the time the majority opinion was that two valves per cylinder delivered the best flow, as seen in bikes like the AJS 350. Four-valve designs had been successful in 1912, but Harry Weslake had begun scientific flow testing in 1922, finding that in terms of specific airflow (cubic feet of airflow per minute, per square inch of valve-head area), a single intake valve flowed best. Four-valve engines fell out of fashion despite Rudge’s brief TT success with them in the early 1930s.

The high flow from a single intake valve arose mainly from its ability to attach that flow to the inside surface of a hemi combustion chamber, converting flow velocity back into pressure in the cylinder more efficiently. Decades of success by the highly refined British singles from Norton, Velo, and AJS underlined two-valve superiority. Modern airflow specialists will recall that for years, the instruction book supplied with Superflow’s flow benches identified the two-valve Norton twins as having the highest intake-specific flow in their experience.

No Room for Error

Yet there was a problem. As MV Agusta’s four-cylinder two-valve 500 engine neared 10,000 rpm, riders were told they had a safety margin of just 300 revs before destructive valve float could be expected (precipitating collisions between valves and pistons). Prevention required one eye on the tach and one on the road ahead. Why not just use stiffer valve springs? Add enough spring and your reward is not only cam and tappet galling but valve-stem stretch as well.

MV’s narrow 300-rpm safety margin was a no-go for Honda, whose first ambition was to win Japan’s Mount Asama races. These events were run on a rough cinder surface, and the constant wheel hopping as the rear tire broke and regained traction made engine speeds jump. You could hear it there, and you can still hear it in sound recordings of racebikes on the Isle of Man as the rear tire periodically loses grip over the rough surface. How tall are those spikes, measured in rpm? More than 300?

When you run into tough problems, you look at every “fact” again, in detail. Compare one big intake valve to two smaller ones of equal total flow area, and an important scale effect jumps out: The total weight of the two smaller valves works out to be considerably less than that of the single larger valve!

The Problem of Scale

Every dimension of the single, larger valve is bigger. Its head diameter is bigger, so the head must also be thicker to support combustion pressure over its larger area. The bigger valve needs a longer valve stem, to make room for its larger single port to turn down, and that stem has to be sufficiently thicker to withstand the force of the stronger valve springs. It’s like comparing two cubes, one whose sides are 1 inch long, and another whose sides are 2 inches long. What is the ratio of their volumes? For the 1-inch cube, it’s 1 cubic inch: 1 x 1 x 1 = 1. But for the 2-inch cube, it is 2 x 2 x 2 = 8 cubic inches.

Instead of scaling directly with head diameter, valve weight scales nearly as the cube of head diameter. The lighter the valve, the easier it is to make it follow the cam contour without valve float.

The Solution Appears

So here was a solution—provided adequate flow could be achieved. Choosing to build a four-valve-per-cylinder engine reduced valve weight enough that Honda could safely operate the valve train at engine speeds roughly 40 percent higher.

Getting adequate flow turned out to be a non-problem. All Honda had to do was make the cylinder’s bore slightly greater than the piston’s stroke. This gave enough room for four valves big enough to supply all the airflow they needed—even though four-valve’s specific airflow is less than that of a well-designed two-valve engine.

As race time approached, Honda’s two-valve 125 was making just over 15 hp, but the four-valve was giving nearly 18. The four-valve concept made it possible to both reach higher revs and provide the necessary wide-overrev safety margin.

Honda didn’t win at the first try. In that 1959 125 TT the highest-placing Honda was Naomi Taniguchi in sixth, with other four-valve Hondas in seventh and eighth. Ahead of them were two MVs, a pair of two-stroke MZs, and a Ducati ridden by some 19-year-old named Mike Hailwood. While outright victory eluded them, the Honda riders were awarded a manufacturer’s team prize.

Constantly refining its concept, Honda won its first world championships in 1961, Tom Phillis taking the 125 title and Hailwood the 250. By the end of 1967 Honda had won a total of 16 FIM world roadracing championships with four-valve four-stroke engines, and had made its name known worldwide.

Mike Hailwood on the 250 Honda six-cylinder at the 1967 Isle of Man. (Cycle World Archives/)

Honda’s R&D engineers had not respected accepted opinion, but had rather focused on ways to solve their problem: How to achieve winning power through higher rpm. Their success is still evident. Today, the great majority of auto and motorcycle engines are built with four valves per cylinder.

Source