At the end of the 1967 season, Honda and Suzuki withdrew their factory teams from Grand Prix racing, and Yamaha did the same at the end of 1968. In hope of entries from smaller European manufacturers, the rules makers in blue blazers limited the number of transmission speeds to six and for 125cc and 250cc classes cut cylinder count to just two.

The 250cc grids were thereafter populated by over-the-counter Yamaha and Bultaco production racers, OSSA’s welded monocoque factory single, a couple of factory Benelli four-stroke fours, and the MZ two-strokes from East Germany. In 125cc, there was more variety. By 1970, the pattern was clear: The Yamaha production racers originally created for the AMA series took the top seven places in the 250cc championship, but “liberated” Japanese factory bikes and European projects ruled in 125cc.

The big change was the six-speed gearbox because it forced design away from the narrow power of the three-transfer orthodoxy of 1961 to 1968 in favor of something with wider range. Yamaha’s 1968 SAE paper described its efforts to give race engines a torque range of at least 3,000 revs (good luck!). Part of that was lower crankcase compression ratios, starting with 1.5:1. Then, in 1972, the 250cc production racer was given a 1.35 crankcase compression. MZ had followed a similar course, dropping crankcase compression from 1.6 to 1.5 for the 1966 season.

Why? Each time the crankcase volume was made larger, there was more mixture in it for the exhaust pipe to pull through the cylinder, scavenging it more completely. What had changed was that as the exhaust pipe took over more of the job of pumping the value of having more volume in the crankcase became greater than that of crankcase pumping. This would be a long-running trend, with crankcase compression ratios falling as low as 1.15:1 in the reed-valve era.

There was another advantage in letting the pipe do more of the job of pumping: Unlike the crankcase, the timing of its effect could be adjusted by properly dimensioning the pipe. Pipes gained efficiency when it was realized that there was no point in beginning the “megaphone” portion of the pipe right at the cylinder outlet. Just as had happened previously in four-strokes, there were those who believed it was best to begin expanding (and therefore cooling) the hot exhaust gas as early as possible. Bultaco in particular held to this belief, placing a lot of pipe volume quite close to the cylinder.

Physics disagreed. As the exhaust port begins to open, high-pressure gas exits at sonic speed, meaning that as long as the port is occupied by the resulting shock wave, no “message” from downstream can reach the cylinder. This means there is no point in beginning the megaphone portion of the pipe until its return wave will no longer find the exhaust port blocked by sonic choking. Pipes designed in this way had a cylindrical or slightly tapering head pipe beginning at the cylinder, followed by the megaphone section. Not expanding the exhaust gas early had the effect of retaining more of its energy for the useful purposes of scavenging and port blocking/supercharge. An obvious measure of how hard the designer is trying to use the energy in the exhaust blowdown process is the pipe’s area ratio, which is the ratio between the area of the head-pipe cross section (where it emerges from the cylinder) to the area of the cross section of the pipe’s center. In early designs, a common figure was 5:1.

California kart racer Erv Kanemoto was asked to “see what he could do” with a pair of cylinders from Yamaha’s 1965–’66 production roadracer, the TD1-B.

At first, Yamaha production racers had just the two transfer ports of the original Adolf Schnürle scavenge. But diesel engineers worldwide had been busy for decades improving the scavenging of two-stroke diesel engines in a variety of ingenious ways. None of them relied on just two or three transfer ports. Rather, they used all the cylinder-wall area available to them.

California kart racer Erv Kanemoto was asked to “see what he could do” with a pair of cylinders from Yamaha’s 1965–’66 production roadracer, the TD1-B. As the concept of using all available cylinder wall for ports was already well-established in karting, Kanemoto knew very well what to do: Add another pair of transfer ports. Because there was little thickness of material in the TD1-B cylinders, the best he could do was to cut slots in the cylinder wall, fed through holes that he made through the piston skirts. As the piston descended, raising the pressure of mixture in the crankcase, it would pass through the holes in the piston, up through the slots, and emerge into the cylinder. This modification boosted power and the word got around. Soon Kanemoto was asked to provide samples of this hardware to the Yamaha importer.

The next spring, Kanemoto’s ports had been incorporated—just as he had originally cut them—into the new TD1-C of 1967–’68. And a year after that, the revolutionary TD2 had something even better: two pairs of transfer ports per cylinder, fed directly from the crankcase; no more restrictive little slots. To make this possible, Yamaha had changed both the cylinder castings and the crankcase. Claimed output for the three 250s had been 35 hp for TD1-B, 38 for TD1-C, and a substantial increase to 44 hp for the new TD2. Evidently there was value in increased transfer-port area, and perhaps also in the improved directional control that four ports per cylinder could provide.

When you get a cluster of hard-working racers working on a problem, the variety and imaginativeness of their approaches to problems soon come up with something that works.

But why did this change come from kart racing rather than from degreed engineers armed with costly instrumentation? You could ask the same about why today’s 8,000-hp top-fuel drag engines came from hot-rodders rather than from automotive research centers. When you get a cluster of hard-working racers working on a problem, the variety and imaginativeness of their approaches to problems soon come up with something that works. Let’s try it, even if it shouldn’t work!

As progressive as the TD2’s cylinders were, they and many designs yet to come retained a feature of the past: Their main or “A” transfer pair were large—25mm wide at the cylinder wall—while the new “B” pair, added behind them and still inspired by Kanemoto’s little slot ports, were less than half as wide. This was even more true of Yamaha’s 350cc twin, the TR2. The roofs of all four TD2 transfers were flat, causing the emerging flows to stick to the piston crown. The “A” pair were angled away from the exhaust at a little over 35 degrees, but the “B” pair were aimed back only slightly.

Gordon Jennings, author of the much-loved Two-Stroke Tuner's Handbook (copies are going for $500 on Amazon), regarded it as important that all transfer flows should be aimed back at the non-exhaust cylinder wall in this way.

Just for the sake of confusion, however, Yamaha did not choose to redesign the crankcase of the 350 as it had that of the 250, so the 350s that scared the daylights out of the AMA roadracing establishment in 1968 by finishing second and third in the Daytona 200 were made with little-afterthought “B” transfers fed in TD1-C fashion through holes in the pistons. Kel Carruthers was able to improve on that in 1971 by cutting those “B” transfers all the way down to the crankcase, ending up with almost no area for sealing at the crankcase-to-cylinder gasket level. To make the best of that, he used copper base gaskets.

Although the main “A” transfers on the 350 had nice flat roofs, those of the piston-fed “B” transfers were aimed upward at about 45 degrees. The same would be true of the TZ750, about to arrive in 1974.

Visiting my permanent collection, I pulled out a Kawasaki H1 cylinder (that was the three-cylinder 500, arriving like the TD2 in 1969). It has fairly small “A” transfers with slightly up-angled roofs, aimed away from the exhaust port at 38 degrees, but then its quite tiny “B” pair have roofs sloping up at 30 or so degrees and aimed at the back cylinder wall.

Three years later, in the three-cylinder 750cc H2, Kawasaki did things completely differently. The “A” and “B” transfer pairs are now of equal sizes, all have quite flat roofs (no upflow angle), and are aimed away from the exhaust port at less than 30 degrees. And the “B” pair are aimed right at each other, not at the back (non-exhaust) cylinder wall.

Meanwhile Suzuki in its 500cc twin and 750cc triple resolutely soldiered on with just two quite large “A” transfers—no “B” transfers at all!

This leaves us with the question, “Where is the pattern here?” The diversity of solutions from the three two-stroke makers suggests there isn’t one. So where do we go from here?