Crank-Cam Connection

Crank-cam connection

TDC

Kevin Cameron

A FOUR-STROKE ENGINE’S CAMSHAFT is driven at one-half the crankshaft speed. This has to be so because a complete engine cycle requires two turns of the crank, or four piston strokes, while intake and exhaust valves open and close once during that time-half as often. This seems like a nice, cozy arrangement and, conceptually, it is.

But real materials are flexible rather than rigid. When cylinders fire, the crank twists like a torsion bar. We know that cranks twist because, unless well-designed and/or made of fatigue-resistant materials, they break from torsional oscillations. In car and truck engines, these oscillations are absorbed by torsional dampers: heavy metal rings on the crank, coupled to it only through a frictional medium such as silicone oil or rubber. Motorcycle cranks survive without dampers because (a) they are short and stiff, and (b) some damping already exists in the system-such as in the clutch hub.

Crank speed is also unsmooth as a whole because power comes in bangs. When a cylinder fires, the crank is accelerated, while the rest of the time, the load causes it to decelerate. Torsional vibration is superimposed upon this speed variation.

Next we gear, chain or belt this unsmooth rotation to the camshaft. In car engines, this is almost always done irrationally; the cam is driven from the higher-vibration, non-flywheel end of the crank. But not always. John Judd built his F-l V-Eight with the cam drive next to the flywheel, thereby minimizing torque fluctuations in his cam drive. The Rolls-Royce Griffon V-12 aircraft engine also used this sensible scheme. Other designers have put both power take-off and cam drive at crank center.

Now we add trouble to trouble. A camshaft is slender, so, like the crank, it is therefore a torsion spring. This springy cam drives an intermittent load: the opening and closing of valves. In low-speed designs this is okay because spring loads are light and the forces required to accelerate valves upward are likewise low. For high rpm, we need higher spring pressures. Valve acceleration forces increase, compounded by a need for higher valve lifts in limited duration. All this makes resistance to cam rotation lumpier yet.

Meanwhile, the crank is receiving bigger, more frequent bangs from cylinder firings. The crank has been designed with its major torsional vibratory modes up at some high and “unattainable” speed. Now, racers want to run the crank up in that very region. As a result, it vibrates torsionally even more than before. So we have violent motions upstairs in the cam department, and vigorous torsionals going on downstairs in the crank department.

Your mission, if you are a cam drive, is to somehow connect these two sets of rattlings and bangings.

The cam designer creates a profile carefully calculated to deliver accelerating and decelerating forces to the valves without impact, so the valvetrain will remain obediently in contact with the cam at all times, without float or bounce. His calculations are based upon the clearly false notion that the camshaft is rotating smoothly and steadily, like a giant flywheel. It is not. Instead it is varying above and below its own average speed as it is battered by crank vibration coming up through the cam drive. It is varying also because its load is lumpy and its drive has some flexibility in it. The cam is twisting, too, because it has its own flexibility. At high rpm, therefore, the cam designer’s careful work goes out the window, because all this violence is superimposed upon the smooth mathematics of the cam profile. The result is that valves are tossed unpredictably. Instead of doing what the profile tells them, they are doing what the cam’s unsmooth rotation tells them.

When valve action ceases to follow the profile, power and reliability depart. For this reason, engine developers often have to accept a lower-performance cam that keeps the valves on the profile, over a potentially higher-performance cam that goes nuts at peak revs. Yet on a test rig, when the hot cam is driven by an electric motor, the valves are happy and well-behaved at that same rpm, so we aren’t talking about classical valve float here-we are talking about valves being tossed by irregular cam rotation.

What do we do? Well, we can try to force the valves into line-Ducati has its desmodromic system. The Renault F-l car engine uses pressure-filled cylinders instead of springs to close its valves; if the valves rattle, just screw down the pressure regulator some more. Rigid drives may break parts; Honda has changed the cam-drive arrangements on its racing RC30s several times, no doubt for cause. We can drive the cams from a relatively quiet part of the crank (next to a flywheel, if any, or next to the clutch). We can use a cam-drive system with damping built into it; chains can run against springbacked and rubber-coated Weller blades. Rubber-toothed belts are also self-damping. We can try to stabilize cam rotation by putting flywheels on the cams-also tried by Honda and others. We can try to calm down the crank with a torsional damper. We can beef up the parts to increase their stiffness, putting their torsional frequencies beyond our peak rpm (we hope).

What we really want is a compliant cam-drive, something able to yield slightly to accommodate the crank’s antics, combined with damping to discourage the kinds of bouncy motions that toss valves. That suggests that rigid drives-gears and chains-aren’t the answer any more. As Superbike engines reach for 15,000 rpm and above, careful initial design must be combined with damped, flexible elements to eat up the torsional vibration that is created.

Could that be why Rob Muzzy was talking about toothed belts at Texas World Speedway?