Who would imagine that such a robust-looking structure as an engine crankcase could flex? Often, there would have been no problem had the engine in question not been thrust into duty for which it was not designed—racing.

Remember—imbalance forces rise in proportion to rpm, squared. A stocker that revs to 8500, when modified for racing runs to 10, 250, increasing bearing loads and con-rod stress by 45%. That’s why Rob Muzzy—“Mr. Superbike” in the 1980s—said of the highly-modified Z1-based Kawasaki Superbikes he built, “Everything starts to go bad in a big way at eleven (thousand).”


That crankcase flexed so much in racing use that the base gasket, separating it from the cylinder block, was destroyed by the relative motion between the two. Oil leaked out. Not having the R&D clout of a factory behind him, Muzzy made solid copper base gaskets that controlled the problem. Git ‘er done with what’s at hand.

Honda discovered the same effect with their earliest 250 in-line racing four. They solved the problem definitively by casting the cylinder block in one piece with the upper crankcase. Thus stiffened, there was no further trouble. Same with their early Hurricane 600—they cast upper crankcase and cylinder in one piece.

But wait. Don’t they balance these engines? Wouldn’t balancing have cured such problems?

Oho! That question takes a bit of answering. The piston, piston pin, rings, and the small-end of the connecting-rod are referred-to as the reciprocating mass (parts that move back and forth, rather than rotating).
The big-end of the rod and the crankpin are rotating mass. These items can be 100% balanced by either drilling holes in the crank or by adding a heavy metal (lead, tungsten, depleted uranium) to the crank.

If we are offended by our engine’s up-and-down shaking (caused by the movement of the reciprocating mass) we can start to add counterweight mass at 180 degrees to the crankpin. This at first seems to work, so we add even more counterweight mass.

Now we notice something disturbing. The benefit of the counterweight does reduce the up-and-down shaking somewhat, but when the added counterweight is at the 90-degree or 270-degree positions, its mass is tending to shake the engine forward-and-back.

If we keep on adding counterweight mass until it balances 100% of the reciprocating mass (piston, &c), we have created a forward-and-back vibration that is just as bad and the up-and-down shaking had been before we started adding counterweight. All we’ve done is move the shaking from vertical to horizontal.

The terrible truth is that a rotating counterweight cannot perfectly balance the reciprocating mass of the piston.

Engineering is very practical. For many years engines were balanced at 50% of the reciprocating mass. This reduced vertical shaking by half, but created a new forward-and-back shaking of the same magnitude. This is “the sweet spot” because it cuts main bearing loads in half. Go past that 50% of recip mass and you start to increase main bearing loads in the fore-and-aft direction.

Riders of classic British parallel twins didn’t mind the vertical shaking too much when the engines were small-bore 500s. The small pistons were fairly light, and so their starting and stopping did not yank the massive engine around too much. Remember that in these engines both pistons moved up and down together (Yeah, I wondered about that too, but this allowed smooth running on a single carb).

Riders wanted more power, so displacement (and piston mass) were increased. 500s became 650s, 650s became 750s, with pistons and their shaking force growing heavier at each increase. Many riders wanted vertical shaking force reduced.

An associated problem was that of frame resonance. The cradle tubes that supported those classic British twins had a fair amount of vertical springiness. A mass (the engine) moving up and down on a spring (those frame cradle tubes) have a natural frequency. When engine rpm came into step with that frequency, the amplitude of the vertical shaking increased alarmingly. As the engine accelerated past that rpm, the vertical resonance decreased.

Chassis are stiffer front-to-back, so it was possible to balance these engines at more than 50% of their recip mass, as the forward-and-back shaking was less able to excite frame resonance, or to be perceived by rider and passenger.

So they just kept on adding higher percentages of the recip mass as counterweight, eventually reaching as high as 85%, moving almost all the vertical piston shaking into the fore-and-aft direction where it was less felt.

Muzzy’s Z1-based Superbike’s crankshaft had good self-cancellation of piston shaking force: when piston #1 was at TDC, piston #2 was at BDC, and vice-versa for #3 and #4.

Now consider that four-cylinder crankshaft as two 180-degree twins, placed end-to-end. Each crankshaft has a rocking motion – when #1 is at TDC, it is yanking the left end of its half-crank upward while piston #2 is at BDC, trying to yank the other end downward. Combine this side-to-side rocking with that of the other half of the crank and you get a quite large force, trying the flex the crankcase at its middle.

This is what created the crankcase flex that rubbed the cylinder’s base gasket to death at high rpm, a result never imagined by the original designers.

When experiments in racing showed how much lighter and stronger chassis could be if made of aluminum rather than steel, a new problem appeared. Steel has what is called a “fatigue limit,” a level of stress below which cracking from cyclic strain does not occur. Most aluminum alloys lack such a fatigue limit. As a result, aluminum chassis for production bikes were impossible unless their engine vibration was considerably reduced.

It was at this point that balance shafts became necessary. Honda’s 1100 cc four-cylinder ‘Super Blackbird’ of 1997 was given secondary balance shafts. My little import car has them as well, located down in the crankcase, driven by chains. At first, traditionalists made a fuss over balance shafts, imagining they “ate power,” and asserting that real men ignored vibration.

Chassis, carburetion, and other parts are, by this reckoning, un-manly. Kenny Roberts’s early three-cylinder 500 GP bike lacked a balancer to cancel its crank’s rocking motion. It therefore broke its seat subframe and shook its carbs such that at certain rpm, the mainjets were flowing air-fuel foam rather than solid fuel. Why yes, it did screw up carburetion. On a shelf at Bud Aksland’s shop in California (Bud is the wisest of wise old birds, knowing many things he doesn’t even know he knows) are unfamiliar-looking devices – carburetors custom-made by the team in hope of delivering consistent mixture despite metal-eating vibes.

The same happened to Albert Roder’s 1953 attempt to design a four-cylinder 500 cc GP bike for NSU. Like so many other designers in history he gave it separate cylinders to allow field replacement.

In WW I (the British call it “The Kaiser War” to this day) many in-line aircraft engines were built the same way – with separate cylinders. They did not crack their crankcases in half, however, because with direct-drive propellers limited by tip speed to 1250 rpm, they couldn’t muster the crankcase-flexing antics of Muzzy’s Superbike engines.

But during WW II Chrysler prototyped an inverted V-16 aircraft engine that packed a lot of power into a small frontal area (A reworked P-47 with this engine allegedly achieved over 500 mph). Alas, Materiel Command insisted it be given separate cylinders “for ease of service” (after all, the giant wartime air-cooled aircraft radials had separate cylinders).

But Chrysler’s IV-2220 turned 3400 rpm for take-off, and its crankcase would not stop cracking. Despite its promise, the lack of the stiffening effect of cast cylinder blocks (as on the classic Rolls-Royce ‘Merlin’ V-12) allowed flexure that fatigued and cracked the case. The program was canceled and plant and machine tools reassigned.

Why didn’t those big radials suffer crankcase cracking? One of the attractions of the radial design was its very short crankcase – inherently stiffer than the long cases of in-lines. Push an idea too far, though, and nature gets tired of you. Pratt & Whitney produced its four-row, 28-cylinder R-4360. Its forged aluminum crankcase, cantilevered far forward from its mounts, did crack from vibration. The company added seven aluminum struts, extending forward from the engine mounts to the first crankcase section join. Better, but not a solution. So they lengthened the struts to extend forward from the mounts to the second crankcase section junction.

Wright Aeronautical, having got its 14-cylinder R-2600 radial through type test and into production without secondary balancers, hoped to do the same with the 18 cylinder R-3350, which would power the B-29 bomber.

Oops! Boeing had built its own instrumented test rig, so soon after receiving prototype engines it was determined that balancers must be developed ASAP. Typical symptoms were fretted propeller mounting cones, loosening of the giant prop retaining nut, breakage of fuel lines and carburetor linkage, broken magneto drives, &c.

It took months to design, develop, and type test the eight eccentric balancers, gear-driven in front and rear cam cases at twice crankshaft speed. The big fear was that the B -29’s 16 foot, seven inch props would suffer vibratory blade fatigue, allowing pieces to fly off--possibly through the fuselage (this has occurred with other aircraft).

For the Roberts Team’s rocking KR500, along came Yoichi Oguma, recently retired from Honda, offering to redesign the triple with balancers. They worked. So did the carburetion. Parts stopped breaking off. Just as the bike was showing real promise, MotoGP replaced the previous 500 cc racing formula.

Nothing could help NSU’s 1953 500-four. With its separate cylinders and heads, it was like a 4-armed tuning-fork, all its parts in frantic vibration. Riders were unable to ride it for long because the bars “seemed to have swelled to 75 mm” (3 inches). Stormy seas battered the insides of its float-bowls, causing irregular running. The four was quietly put away (it still exists) and the company became dominant in 125 (a single) and 250 (a twin) GP classes.

Yamaha’s TZ750A, released for 1974, had a pair of twin-cylinder blocks and heads, but with nothing connecting right and left blocks. After a couple of racing seasons, some riders found their crankcases cracking from an edge of the gearbox sump plate opening. Same problem that forced Muzzy to make copper base gaskets—crankcase flex.

There was even talk a few years back of one MotoGP constructor having a crankcase so lightly built that chassis forces passing through it interfered with gear-changing.

Everything flexes.