FAMILIAR OBJECTS: Motorcycle Springs

Kevin Cameron's Top Dead Center banner
Technical Editor Kevin Cameron shares his wealth of motorcycle knowledge, experiences, insights, history, and much more.Cycle World

Think about all the springs on a motorcycle. First to come to mind are suspension springs and valve springs, but there are also springs in the back of the clutch basket (which smooth out the “combustion thumps” of the engine’s cylinders firing). There are usually 5 or 6 clutch springs or perhaps a single diaphragm-type clutch spring. In the gear selection mechanism, there’s a spring behind the shift-drum detent (to hold it in each of the 6 gear positions, plus a shallow detent for neutral. Other springs are required for the ratchet action of whatever connects the shift pedal to the drum itself, and there’s a thick wire stiffie wrapped around the shift shaft with two tangs to keep the shift pedal centered (you feel its resistance every time you shift).

Back in the bad old days of drum brakes there were the really strong shoe retraction springs (installing them was a treat) but in disc brakes the return function is performed by elastic deformation of the piston seal O-rings in the calipers.

My introduction to spring breakage as a result of millions of stress cycles came from race bike exhaust springs (which hold exhaust pipes into their slip-fit couplers) and from drum brake shoe retraction springs. On my 1965 Yamaha production racer TD1-B, each of these springs had a length of black rubber tubing over it. After some thought I realized that the purpose of the rubber tubing was to prevent side-to-side sympathetic vibrations excited by the engine. At, say, 150 or 300 vibrations per second, it wouldn't take long to accumulate enough fatigue cycles to activate a defect somewhere in the spring wire, open it into a crack, and fail the spring.

In 2000, when Honda first took their RC51 Superbike to Daytona (It would win two World Superbike championships), Harley's VR1000 engineer Steve Scheibe pointed and said to me, "Look there. They don't have any rubber sleeves on their exhaust springs. Do you think they know something we don't know?"

The missing information was supplied in early practice as the springs began to break regularly. Somewhere in Honda engineering, someone had missed a step! Could they be only human?

In 1973 my rider the late Cliff Carr finished 2nd at Laguna on our home-made 750 – riding low in the rear with one broken suspension spring. That was understandable for two reasons; first, those 3-cylinder Kawasaki Triples generated heavy vibration, and second, I'd bought those springs from a Girling distributor in 1966. Lots of stress cycles.

Another time, my 500 Triple began shifting erratically. Gearbox troubles are serious business because pulling the clutch doesn’t free the rear wheel from a gearbox lock-up. Upon pulling the clutch and primary cover, I saw that the little spring whose job it was to hold the shift claw against the pins in the shift drum had broken at one of its ends (which was a 180-degree bend). Using a little magnifier, I could see that one of the holes through which the spring ends were hooked had ugly sharp edges – that plus a zillion vibration cycles had just worn down the spring wire in one place until it opened up. After that I carefully smoothed and radiused any holes used for retaining springs. And a rubber sleeve for good measure.

Another aspect of that shift claw spring failure was that its hook ends were of the same gage wire as the spring itself. Yamaha, because they initially had failures of the hook ends of exhaust springs, decided to find a way to make the ends of heavier wire. They joined the two by cold-heading one end of the hook (cold-heading is the same process that produces nail heads) and then wrapping the end of the spring small enough that the head of the hook was trapped.

Another kind of problem caused the occasional loss of a Yamaha TZ wristpin clip (arguably a sort of spring). To make it easy to extract the clips, Yamaha over-bent one end of each clip so it looked like the letter ‘G’. You could then extract it from its groove with needle-nosed pliers. Now visualize what happens as the wristpin moves endwise in its bore. Ideally, with all of the wire clip in its groove, being hit by the wristpin would only drive it harder into its groove. But with the bent-over end, the wristpin hit the over-bent part first – and if it hit hard enough, it could sproing the clip out of its groove. After one of those failures we always cut off the bent-over part of the clip using a thin one-inch abrasive cut-off disc. No more lost clips and no more wristpins gouging up cylinders.

Really crucial springs are the valve springs of four-strokes. Springs in production engines are moderately stressed for long life, but the racier the application, the harder the engineer works the metal in the spring wire. If the engine does a million revolutions (that’s a bit over three hours at 5000-rpm), each spring undergoes 500,000 basic stress cycles. But that’s not all, because the sharpness of valve accelerations also excites end-to-end vibration in the springs (you can see such end-to-end vibrations by tweaking a stretched ‘Slinky’ toy). If you Google “valve springs in a running engine” you are sent to a YouTube video of a valve spring continuing to vibrate and wobble after its valve has closed. To prevent these extra oscillations from leaving angry customers walking instead of riding, various damping schemes are used to prevent them. Most basic is the use of dual springs wound in opposite directions and dimensioned so that inner and outer coils rub together. This engineered friction is like placing your thumb on a ringing bell. Another scheme is the use of a helical flat wire damper which fits tightly against the ID of a valve spring. And a third scheme is to wind the spring in “beehive” form, so that its natural frequency is not constant, but varies with lift.

Manufacture of valve spring wire begins with low-defects steel – ideally the vacuum-remelted kind pioneered long ago by S&W – provided with a very smooth surface (cracks love to form from nicks and scratches) and 100% shot-peened to place the surface layer in compression.

Why? Tension is necessary to produce cracking, so if the surface begins life in compression, it can withstand more bending to reach a given level of tension at its surface.

Valve springs are not a concern for the street rider, but back when MotoGP was beginning to give up metal springs in favor of gas springs (ever see a fatigue crack in nitrogen?), the evening ritual for every team still using metal springs was replacing valve springs for the next day. If the rider was under way for half the time of each day’s two practice sessions, and if average engine rpm was 12,000, that’s roughly half a million valve spring stress cycles. If we now jump to pro-class drag racing we find teams running their springs between one and four runs before replacement. If the engine runs at 10,000-rpm and tire prep plus a run total ten seconds running time, that’s maybe 2000 spring cycles. The difference in spring life is mainly caused by how high a stress level is required of it by design.

Some years ago I asked Claudio Domenicali (Now gen. mgr. of Ducati – Good going, man!) what valve acceleration Ducati closing levers could produce, he replied, "That depends on how long you want the levers to last."

His words reveal that those closing levers are like springs in that their life depends upon the stress they are asked to undergo. On a streetster, the levers may last indefinitely, but as you push the stress level at racing rpm, life dwindles.

The (usually) six springs in the back of clutch basket are there as a spring drive. When the engine fires, its torque compresses the six springs, somewhat smoothing out the torque going to the gearbox. When in 1994 Harley began racing their VR1000 Superbike, they were told by their clutch supplier that a spring drive would probably not be necessary, as Yoshimura didn’t use one in their race bikes at the time. When practice started, things began breaking. Could it be that two great big cylinders produce larger whacks of torque than four smaller ones? (In those days, twins got 1000-cc displacement and fours, which could rev much higher, got only 750). I walked past the Harley garage about then and saw that everyone inside was on one of the new-fangled cellphones, trying to get answers. So yes, that end-to-end circle of springs in the clutch is there for good reason.

Coil springs gradually lose free length as a result of time-at-stress, so in service manuals you will usually see a minimum free length specified for both valve springs and clutch springs.

In closing, it must also be said that threaded fasteners are springs. Head bolts are the conceptual equivalent of many really strong rubber bands, holding the head, cylinder, and crankcase against one another. Tightening head bolts or nuts-on-studs is really just elastically stretching those bolts or studs. Their elasticity thereafter provides the force that holds the parts together. I repeat; bolts and studs are just extremely stiff tension springs. In the case of the old Triumph 500-cc twins, the degree of connecting-rod cap bolt stretch resulting from tightening the nuts could be measured directly with a micrometer; recommended bolt stretch was .004 - .005-inch.

Basic stuff, springs.