Failure-Prone Crank Designs

See other failures under 000.html.

``Sweet Wings'' brand bicycle cranks were some of the lightest and stiffest when released, but reports on the web indicate that an unusually high percentage of the cranks failed. A common failure was reported in the spline area, and riders often reported such failures after little use.

The failure was that the spline got loose and noisy. ``Loose and noisy'' failures are far better than many other crank failures: you do not get injured or stranded. However, the cranks cost several times most other cranks of the time and several failures were reported with less than a year of modest use.

Some crank failures are likely because the spline attachment is not preloaded. That is, the parts are not drawn up tightly as with a tapered spline, pinch clamp, etc., used on most other cranks; nor does the spline use a tight interference fit. Because the attachment is not preloaded, cyclic loads cause the joint to move, both in bending and in torsion. Motion causes noise and wear, wear leads to failure.

To visualize the torsion failure, imagine a spline where the spline ``teeth'' are smaller than the slots they fit in. If you stand on both pedals with your right foot forward, the loose spline will shift one way; if you then rotate the cranks and stand on both pedals with your left foot forward, the loose spline will shift the other way. In the Sweet Wings crank, the spline is not obviously loose. However, the spline loading is high, so even a loose interference (press) fit may still allow miscroscopic motions. When such motions occur, they wear the joint. As wear progresses, the joint gets looser and looser. With sufficient wear, the motion becomes noticible without measuring tools.

(Note that this failure occurs also in Shimano Octalink(tm) splines, which are axially tapered but not circumferentially tapered.)

Such motions also tend to loosen the fixing bolt. The bolt is attached to both parts of the spline. When the two parts of the spline move relative to each other, three things happen: the bolt threads move relative to the threads in the crank; the bolt twists; and the bolt head moves relative to its support. The first causes wear, the second fatigue. When the third occurs, the bolt tends to loosen because the tensioned screw threads act like a ramp to apply a biasing force. Thus, when the spline parts rotate in the bolt unscrewing direction, the threads are more likely to turn; when the spline parts rotate in the bolt tightening direction, the threads have a higher force, and the bolt head is more likely to slip relative to its support. The slippage may be microscopic but can ultimately cuase the bolt to unscrew and loosen the joint, exagerating the other problems.

To vision bending failure, again imagine a loose spline, this time with the inner spline smaller in diameter than the outer spline. When you stand on the pedals right-foot-forward, your weight pushes down on each pedal, causing the center of the spline to rise. If you unweight the pedals and rotate the cranks they are are left-foot-forward, the center, which was up, is now down. When you stand again, it again makes the center rise. In the SweetWings crank, the spline is comparatively tight and the motions may initially be microscopic. Nonetheless, when such motions occur, they wear the joint; as wear progresses, the joint gets looser and looser. Rocking probably also helps loosen the fixing bolt. Rocking the joint also tends to rock the bearings, likely shortening bearing life.

It should be noted that microscopic motions occur in many common joints. However, if the initial forces in the joint are very high, typical joint loads have a hard time overcoming the high initial forces. Thus, the joint will only move under very high loads. As long as the joint moves only rarely, it will survive.

As an aside, note that the left bearing is mounted outboard and carries no axial load. The right bearing outer race is ``caputred'' axially between the right cup and the bottom bracket sleeve, which is pressed axially by the left cup. The right bearing inner race is ``captured'' axially between the two crank halves.

The following images of a non-damaged crank were graciously provided by the seller, who did not respond to queries how s/he wished to be credited.

SweetWings manual (PDF, 150KB)










The following summarizes comments from as of 2004/01. Note the sample here may overstate the rate of problems -- people with problems are more likely to post.

According to as of 2009/12:
[Most failures were] at the axle/arm weld or at the axle near the threaded hole.

Sweet Parts, manufacturer of Sweet Wings, is no longer in business. It may be possible to track down the original makers, though. From as of 2004/01:
John Kruse -- Exercise Physiologist

Mr. Kruse is the owner and operator of CycleSimulator. He holds a BA degree in Kinesiology with an emphasis in exercise science. He has also competed in the sport of cycling since 1988 and competed in the 1992 and 1993 Collegiate Track Nationals for California State University, Northridge (CSUN). As an undergraduate, he worked at the Encino Velodrome as a youth cycling instructor. After graduating from CSUN he worked as a fitness counselor and one year later went to work in the bicycle industry for Sweet Parts, the innovative crankset and stem manufacturer. Mr. Kruse is curently finishing work on his Masters degree in Special Education and works as a science teacher for special education students at South Gate Middle School in South Gate, CA. Most of all, Mr. Kruse enjoys the physiology of training athletes and started CycleSimulator in order to fulfill that passion.


From as of 2009/12/07:

These are Craig Edward's latest creation, the eebrake. Craig was originally known for his "Sweet Wings" cranks that came out during the early '90's [...]
There is a link to (as of 2009/12).

See also FAIL-007.html