Cycle World reader “RedRadio” expressed interest in how the peak-rpm capability of motorcycle engines has risen so high. A big, slow-turning engine of 110ci/1,800cc giving peak power at 5,000 rpm, and a little 37-inch/600cc four revving to 15,000 arrive at similar horsepower by different routes. Understanding these two paths requires an appreciation of the factors limiting and permitting high-rpm operation.
First, consider why a builder would want to raise an engine’s rpm limits. The three variables from which we compute horsepower are:
- <b>Cylinder displacement</b>, which tells us what volume of fuel-air mixture our engine can, in theory, pump into itself.
- <b>Stroke-averaged net combustion pressure</b>, which in turn depends on how well the engine breathes, what its compression ratio is, and how efficiently it burns its fuel-air charge. A much bigger increase can be achieved here with supercharging or turbocharging, but neither has ever found much use in production bikes.
- <b>RPM</b>, which measures how often our engine can perform the power-producing cycle.
Displacement is often set by a racing class or commercial segment, such as 600 Supersport or 450 motocross. Techniques for maximizing No. 2 above are known to all manufacturers and tend toward a common level. That leaves us with No. 3, raising an engine’s redline, as the most effective tool for boosting horsepower.
Problems With Raising the Redline
The classic problems with increasing operating rpm are numerous. First (with four-strokes) is the difficulty of making the valve train accurately follow the contour ground into each cam lobe. When an engine reaches its limit in this respect and the valve train fails to follow the cams, it experiences “valve float.” Uncontrolled valve motion leads to destructive effects such as valve-to-piston impacts. It is typical in high-performance engines for valve-train parts to be made lighter at each model change, often requiring use of upgraded materials.
Early cam lobes had simple geometric shapes that imposed hammerlike accelerations on valve trains. It would take years before improved lobe shapes applied loads to valve trains in a more gradual and survivable way.
For a given valve-spring pressure, the lighter the valve train, the higher the rpm at which the system can accurately follow the cam lobe. The late English F1 designer Keith Duckworth enunciated his “square root of two” principle, which states that when you abandon two valves per cylinder in favor of four, you divide your problems by the square root of two, which is approximately 1.41. In a spooky confirmation of this, consider the 10,250-rpm peak of Rob Muzzy’s Z1 Kawasaki-based 1,000cc two-valve 152 hp Superbike engine of 1983 with the 14,600 revs presently used by Jonny Rea’s one-liter ZX10RR four-valve World Superbike engine. The arithmetic is 14,600/10,250 = 1.42. Both of these engines employed the very best metal valve springs of their time.
The more parts there are in a valve train, the heavier it is, and the greater the spring pressure required to make it follow cam forms. This is why a pushrod-and-rocker valve train typically requires twice the spring force needed by an overhead-cam design. A change we have seen quite recently has been the adoption of very light F1-style pivoted finger followers as a substitute for the older “inverted bucket” piston-type tappets in overhead cam engines.
Next is the physical strength or fatigue tolerance of the connecting rods that join the pistons to the crankshaft. Con-rods can break at high revs, and pistons, subject to violent acceleration/deceleration cycles, develop fatigue cracks and come apart. Crankshafts that are well-behaved and long-lived at stock rpm may develop vibrational modes at higher revs that fatigue and break them. Cam drives, whether gear, chain, shaft, or toothed belt, can interact with the crankshaft to produce unstable herky-jerky twisting motions that can toss valves and twist the drive into scrap.
Special attention is often required here while developing racing engines. At Harley-Davidson in the ‘90s I was shown a box of what had once been experimental cam drives for the VR1000 Superbike. Undamped vibratory motions had destroyed the lot.
Frictional Losses and Parts Failures
Also of importance is the rapid rise of friction loss with higher rpm. Inertia loads on bearings (from the reciprocating parts) increase as the square of rpm, compelling designers to reduce the weight of pistons and con-rods as much as possible. This motivates the use of titanium instead of steel in rods, and the shrinkage of pistons from their 1950s resemblance to small buckets to the present ultralight “ashtray” look. Modern pistons are little more than a round disc to hold the piston rings, connecting via an under-the-dome box structure to a very short wrist pin, with a short pair of “skirts” to stabilize the piston against rocking in the cylinder bore.
High-duty pistons today tend to be forged rather than cast, and are carefully shaped in accordance with dynamic stress analysis to avoid stress concentrations that lead to cracking. This gives modern pistons their flowing organic shapes and freedom from sharp edges and sudden changes of cross-section.
Damage to or failure of the reciprocating parts—the piston, its rings and wristpin, plus the small end of the con-rod—result not from combustion loads but from the rapid acceleration and deceleration of these parts, which in ultra-high-rpm engines can reach 10,000G. Such accelerations are directly proportional to piston stroke and to the square of rpm. Since the path of development is often to increase rpm, the obvious line of attack is to shorten the stroke, then rev the resulting engine up to the stress limit of its parts. Two approaches have been used.
- In its classic high-rpm four-stroke GP bikes of 1959-67, Honda kept the bore-to-stroke ratio close to 1.1 because larger pistons were much harder to cool. If, as an exercise, we begin with a 250 single and seek a shorter stroke through adopting more and smaller geometrically similar cylinders, we find that a 250 twin’s stroke is 79 percent of that of the single. If, like Honda, we move on to a 250 four, stroke becomes 63 percent of that of the single. For a six, the number drops to 55 percent. Only when we go to the extreme of building an eight-cylinder 250 do we arrive at 50 percent of the stroke of the original 68 x 68mm single. In 1967 there was speculation that Honda might follow up its 250 six-cylinder with a V-8, but the company withdrew from racing.
- Formula 1 development has historically used cylinder multiplication (V-12s, a V-16, and an H-16) but more recently has adopted much larger cylinder bores and shorter strokes—going to the 2006 extreme of a 2.4-liter V-8 of 98 x 39.7mm dimensions, a bore/stroke ratio of nearly 2.5. In that year F1 engines supposedly reached 20,000 rpm.
As happens so often in life, going to extremes can create at least as many problems as it solves. In the early 1950s, BRM couldn’t afford to effectively develop its complex supercharged V-16. The same was true of Moto Guzzi and its legendary V-8 500. The greater the number of cylinders, the greater the heat loss from their much larger combustion-chamber surface area. More cylinders can also increase the friction loss from the greater number of parts. And think how hard it is to achieve a high compression ratio in a 98mm cylinder with a very short 39.7mm stroke: With a modern ratio of 13:1, a uniform combustion space between piston and head at TDC would be roughly 1/8 of an inch thick. That cramped space slows flame propagation—ignition must be timed to occur more than 60 degrees before top center in such engines—greatly increasing the time duration of heat loss.
The engineer’s choice of bore and stroke is also tied to valve area. There are limits to how fast flow can move through valves and ports, so when an rpm increase is desired, bigger valves may be needed—often bigger than will fit into the existing cylinder bore. This requires that bore be increased and stroke shortened to make room.
Bore and Stroke Ratios
Motorcycle engines have tended to avoid such extremes, mainly because their very limited tire footprints can’t handle the radical torque curves of F1 design (although some illustrious F1 organizations have tried their hand at designing engines for racebikes). A clear symptom of this is MotoGP’s limiting cylinder count to four and bore and stroke to 81 x 48.5mm (yielding a much more moderate bore/stroke ratio of 1.67). As motorcycle manufacturers are smaller than automakers, they simply cannot muster the $300,000,000 annual budgets of the larger F1 teams. Such limits make sense.
Ducati’s 1299 Panigale V-twin did reach 116 x 60.8mm bore and stroke, setting a benchmark for extreme bore/stroke ratio in a production machine: 1.91.
Valve Springs and Desmodromics
At various times in history valve springs have been the limiting factor in rpm, owing to quality issues and fatigue. Quality greatly increased with adoption of “cleaner” vacuum-remelted wire materials (that is, containing fewer crack-nucleation sites such as oxide inclusions) after 1957, and by shot-peening the wire to reduce surface cracking by placing the surface in compression.
Valve springs in production bikes are stressed to a lower percentage of the wire’s ultimate tensile strength than in race engines. As wire quality improved, it became possible to operate the springs at progressively higher stress levels. Even so, by 2004 MotoGP teams were having to change valve springs after each day’s free practices just to avoid failure. This point had been reached earlier in F1; first Renault and then the other manufacturers abandoned metal in favor of pneumatic springs in the mid-1980s. Ducati avoided this issue altogether by its long tradition of closing its valves not by springs but by separate closing cams and levers—the desmodromic system.
In containing the violent motions of internal parts at high rpm, engine structures flex. Honda, in trying to develop its four-stroke V-4 oval-piston NR500 GP bike, had to increase stiffness with extra metal. Inline-four engines with separate cylinder blocks can flex so much that they roll their base gaskets into little balls as parts chafe against each other. This led to the practice of making the upper crankcase in one piece with the cylinder block. After initial dyno testing of a new design, engineers eagerly await teardown to see if there is fretting, or marks of mutual rubbing and generation of wear particles, between major parts, indicating unacceptable levels of flexure in operation. Designs that are satisfactory at a product’s initial level of operation may not be after updates increase power and rpm.
Crankshafts and camshafts, because they are subject to sudden cyclically repeated loads, wind up and unwind in torsion: this is torsional vibration. In general, the longer the shaft, the more vulnerable it is to such motions, and the more intensive the measures adopted to suppress such motions must be. In F1, according to a Honda technical paper, 10 or more vibration dampers on crankshaft and cams may be necessary to achieve this.
As engineers have tackled all these problems, solutions have emerged that are able to combine affordability with the greater power that increased rpm can deliver. What was radical stuff a decade ago is normal production tech today.