Spring To Action

Spring to action

TDC

Kevin Cameron

MOTORCYCLE ENGINES, HAPPILY, ARE in a constant state of evolution. Take the lowly valve spring, for example.

Early camshaft designers asked metal to do things it was incapable of, as “square” lift profiles impacted against tappets at steel-shattering rates of acceleration. When basic math was applied to lift curves, it was realized that valvetrain parts need some time to “load up,” and the more parts there are in the valvetrain, the longer this loading up process takes. Thus, high-rpm engines must use the fewest valvetrain parts-just a finger or bucket tappet between cam lobe and valve stem. Most engines use springs to make the valvetrain follow the cam profile.

A helical valve spring is just a long torsion bar coiled up. When you deflect it, you are using its own round shape as a lever with which to twist the wire. Unfortunately for reliable operation, the wire does more than twist; the coils must move with the valve. Since this is an intermittent motion, the coils are no sooner urged into motion by the cam than they are required to stop and reverse direction at the top of the lift profile, then stop again as the valve closes.

As observed by Newton and all the rest of us, objects in motion tend to remain so, while objects at rest resist our efforts to set them moving. Therefore as the cam begins to lift the valve, the coils at first lag behind, owing to their inertia. Then they catch up, and even rebound ahead of the valve’s motion somewhat. No sooner has the valve reached top speed than the profile asks it to slow again, but the coils of the spring show their inertia-this time by continuing at high speed until they pile up somewhat against the valve end, bounce back, and oscillate at a natural frequency.

You can see this happen with that grand old toy, the Slinky. Fasten one end to a fixed object (a friend, or a doorknob) and then move the other end back and forth, as though the hand holding it were the valve-spring retainer. You will see waves of compression and extension run back and forth from one end of the Slinky to the other as you do so, more strongly the more suddenly you move your hand. In a valve spring, these waves are not too harmful unless their frequency

happens to be in step with, or a low multiple of, the valvetrain opening and closing frequency. If they’re in step, the violence of the bouncing in the spring builds up rapidly, subjecting the spring wire to not one, but several back-and-forth oscillations per valve event. This, called spring surge, quickly eats up the spring’s fatigue life.

In the early 1960s, very “clean” steels became available-those remarkably free from the interior defects that often initiate cracking. Such materials allowed valve springs to operate at much higher stress levels without early failure. Before that time, most designers simply accepted the valve spring as the Achilles’ heel of the engine, and tried to design around it, keeping maximum wire stress low. With helical springs, the standard scheme was to make the spring’s natural frequency at least five or six times the crank speed-if the crank made 100 revs per second, the spring frequency should be 550 or more cycles per second. Even so, very vigorous lift profiles could still provoke surge, just as your hand can do with the Slinky. The only lift curves that such springs could reliably follow had long duration and moderate lift, qualities that contributed to the “camminess” of race engines in the pre-1960s period.

Mercedes-Benz in the early 1950s, followed some years later by Ducati, decided to do away with springs completely, electing to open and close the valves with paired cams. This desmod-

romic system (Mercedes called it the “Zed Drive”) required high precision to make the paired cams truly complementary, without looseness or pinch.

Yet the greater ease of changing profiles with normal cam-and-springs won the day as soon as better spring materials arrived.

Bypassing helical springs was also the aim of the hairpin springs used on Ferrari GP engines of the 1950s, on the Manx Norton, and so on. These resembled the spring on the familiar scissors-grip exerciser, and so had no 'f bouncing coils to surge. Poor materials still caused rapid sag, so designers had to avoid rapid and high lifts that accelerated this decay.

Honda’s approach to the problem in the late ’50s was to note that the natural frequencies of small springs are higher than those of bigger ones. This fact, more than any desire for greater flow area, prompted Honda to adopt multiple small valves and springs per cylinder; this pushed the surge point beyond peak revs. Later, engineer Peter Durr and others noticed that the higher properties of the new spring materials would permit springs to operate at far higher stress levels; you could, in effect, get more spring out of less metal. In practical terms, this has meant springs with very few coils— typically three and a half. Their extremely high natural frequencies have allowed the use of more vigorous cam profiles, those with desirably short duration and high lift, and so have made possible the wider powerbands of modern engines.

Now the standard valve-closing scheme in F-l auto-racing engines is the gas spring, or air closer. Each bucket tappet is sealed to its bore by a piston ring, and the space beneath it is pressurized from a central system. A check valve admits gas to make up for leakage, but prevents outflow. Valves bouncing? Just jack up the system’s pressure. Air closers may have “civilian” applications as well; a primary source of friction in production engines at low rpm is the valve drive. By using an air-assisted spring system, a very light spring could provide lowfriction running at lower rpm; bouncefree operation at maximum revs would be provided by adding gas pressure as the engine revved up.