In assembling many a Yamaha two-stroke twin, I knew that one essential step was to install the "oil bucket" on the right-hand end of the gearbox output shaft. This was a disc about an inch in diameter with a collecting trough on one face, plus a short central spout that fitted into the hollow shaft retained by a flat-head screw. Its job was to collect some of the storm of oil droplets constantly flung around the primary gear case and send them into the hollow shaft. From there, drilled radial holes led oil to the free-spinning gears to prevent their seizing to the shaft.

The gears themselves were normally lubricated by filling the gearbox pretty much to the shaft centerline with your choice of lubricant. Along the way, someone with access to a dyno had discovered that some power could be saved by draining the standard fill of 1.6 liters after first practice and refilling with just a quart. That first run with the full amount got everything started with plenty of oil, and the 1-quart refill saved a bit of power because you can run faster in water up to your ankles than you can if it’s up to your knees. With just the big first and second gears dipping into the oil, plenty was being thrown around the gearbox thereafter to keep everything well-lubed.

In the early 1960s, Yamaha engineer Noriyuke Hata was tasked with developing a 125cc Grand Prix roadracer. Did he “motor” a gearbox—use an electric motor and torque meter to spin an engine or other apparatus—to measure its power loss? Or was he aware of wartime practice in the lubrication of propeller reduction gears on aircraft engines? In any case, he dropped the idea of lubricating the gearbox by filling it halfway and substituted a small gearbox oil pump. Oil was sent to the hollow centers of each gearbox shaft and by individual oil jets to each gear pair in the gearbox. All gears were far above the oil level. In this way, the considerable oil-churning loss of conventional gearboxes was largely avoided.

When Yamaha’s TZ750A Daytona bike appeared in 1974, its gearbox was lubricated in Hata’s style by a tiny pump and oil jets. When the TZ250H debuted in 1981 (Yamaha’s first 250cc production racer not based on the RD streetbike crankcase), it too lubricated its gears by pumped oil jets.

Having such fast-moving gears half submerged in oil produced tremendous churning loss.

Today’s four-stroke production bikes share the same oil between engine and gearbox so it is natural for their gears to remain high and dry above the oil level, jet-lubricated and with low churning loss.

When in 1976 I had the opportunity to work with a Kawasaki KR250 roadracer, I saw that not only did it lubricate its gearbox by the ages-old system of having the gear oil half-submerge the gears, but it also drowned the two 100mm (4-inch) phasing gears that connected its two one-ahead-of-the-other tandem crankshafts. At 12,000 rpm, those gears were meshing at 12,000 feet per minute. At such high mesh speeds, Dudley's Gear Handbook advises that best results come from applying air/oil mist to the opening side of the mesh. Having such fast-moving gears half submerged in oil produced tremendous churning loss.

More power went to the rear wheel and less into boiling the gear oil.

Prototype and first-year 1975 engines were built on magnesium crankcases painted black. In testing, the cases became hot enough to burn off that paint. That heat represented many horsepower, so to reduce the loss, recommended gearbox oil fill was cut from 1,100cc to 600. Why so little? Because this engine was designed only for racing; it had no kickstarter, so its case volume was small to begin with.

I built a little open-topped sheet-metal box to keep gear oil away from the high-speed crank gears, with only a 1mm drilled hole (0.040 inch) to let oil into the box. Sitting still, the housing would fill, but upon engine starting the gears would throw all the oil out, save for what trickled in through that hole. It worked: More power went to the rear wheel and less into boiling the gear oil. Rider Ron Pierce was able to qualify the bike at the front. I should have put more of that work into ensuring reliability; in the race, the water-pump shaft that I should have replaced gave up and all Ron’s good riding was for naught.

The Pratt & Whitney geared turbofan engines now going into commercial airline service employ a gear reduction between turbine and fan, allowing each to turn at its best speed. Losses in their “star gearing” are said to be just seven-tenths of one percent. Yet in lots of written sources, transmission loss in motorcycle dyno testing is estimated at 10–15 percent. Okay, power is passing through two meshes, one in the primary gears (connecting engine crankshaft to gearbox) and one in the gearbox (whichever of the usual six ratios is driving). The other five ratios windmill without load, generating some loss. Then there is the drive chain connecting gearbox to rear wheel. I usually hand wave around a 2 percent loss for each, so that’s 6 percent. On roller dynos, there is also loss in the drive from rear tire to roller.

More than one person I know has had his back beaten up by rubber flung off of tires overheated by dyno running. They strap powerful bikes down hard on those dynos to prevent tire slippage. The resulting rear tire flexing is surely at least as great as when running on the Daytona banking, so the loss there and heat generated have to be substantial. If we imagine a tire-to-roller loss of 7 percent on a 180-hp bike, that is 0.07 x 180 = 12.6 hp. Since each horsepower is 746 watts, that is the equivalent of a dozen red-hot kitchen toasters being dumped into that rear tire. Think maybe that would make it pretty hot pretty quick?