Some Examples of Failures in Bicycle Engineering
that have Caused Accidents

home          ARTICLES


A forensic engineer investigates accidents to determine the causes of failure. In most cases the cause, whether it is easy or difficult to discover, is not a failure of engineering design, but of conditions that exceed the capabilities that have been designed into the bicycle or the facility, or of the operator. Bicycles are not expected to survive undamaged when being ridden against curbs, curves have a maximum safe speed, human beings cannot make decisions instantly. Engineering knowledge enables us to provide reasonable explanations for many accidents.

However, there are other accidents that are caused by engineering error. As the eminent Professor Henry Petroski has wisely observed, "I believe that the concept of failure ... is central to understanding engineering, for engineering design has as its first and foremost objective the obviation of failure. ... To understand what engineering is and what engineers do is to understand how failures can happen and how they can contribute more than successes to advance technology." Every advance in engineering brings with it new failure patterns that must be understood before we can guard against them. I have investigated quite a few accidents in which the cause stemmed from the failure of new ideas whose failure modes were not properly understood by the initial designers. Several of the articles in this section discuss the analyses that explain such failures.

Front Brake Design Failures

Inadvertent strong application of the front brake can cause the bicycle and rider to pitchover and the rider to be thrown over the handlebars. In some cases this is not rider error but is caused by poor design of the bicycle brake that erratically changes its grip so the rider has no warning of how strong the brake application will be. The design may be entirely adequate when properly installed and maintained, but becomes subject to jamming under the conditions of normal use and typical maintenance.

In one case, a brake that had been designed to be especially powerful for the expensive mountain-bike market caused a rider to flip when on a city street. The cause was not the supposedly powerful linkage, which was not exceptionally powerful, but the mechanism designed to make it easy to adjust the brake. Jamming of the CODA Brake

In another case, the cause was also in the adjustment mechanism, but this was a case of designing what appeared to be a simple and cheap way to adjust the brake, but without seeing far enough ahead to the state of maintenance that would enable the design to jam the brake. The Neat Design that Jammed the Brake

In a third case, this was not so much a failure of the design engineers, but of later installing a very powerful brake onto a fork whose engineering error was insufficient, by itself, to cause normally powerful brakes to jam. Either piece by itself, the brake or the fork, would be acceptable, but the combination was very dangerous. Self-Energizing Brake and Misaligned Fork Cause Pitchover

Insufficient Steering Stability

It is commonly agreed that bicycles require steering stability, but there are two different definitions of steering stability. The ill-informed definition is that a stable bicycle continues on its path regardless of the wobbles of its rider and the slope of the road and doesn't fight with its rider. The well-informed definition is that a stable bicycle steers according to the slightest lean of its rider and tells him, by feed-back force felt at the handlebars, which way it wants to go, and goes that way.

A bicycle of the first type appears easier for the beginner to ride, particularly easy to steer because the steering is so light, but because it doesn't know which way it wants to go it wanders all over the place and doesn't respond to body lean. Since leaning and steering must be coordinated for the bicycle to stay upright, bicycles of this type tend to cause falls and to develop steering oscillation at speed. A heavy woman (weight makes a difference) on such a low-quality bicycle, descending only the few feet of a road bridge over a canal, got going fast enough to flip her front wheel sideways and fall to the roadway. I investigated that bicycle, and found it difficult to ride no-hands; it was not very stable.

A bicycle of the second type appears to the beginner to be unstable because it steers to his every wobble and its steering appears to require much force. However, it is this steering to every body wobble that keeps it upright, and the steering forces tell the rider which way to steer it. Such bicycles do exactly what you tell them to do and don't wander about or develop steering oscillation at speed. You can ride them confidently in a racing pack only inches from other riders. They are stable.

A cyclist on a bicycle of the first type, the DaHon folding bicycle, turned his head to look at his riding companion, and found himself on the ground. I investigated the bicycle. Stability of the Da Hon Bicycle, and of Bicycles Generally

Estimation of Cyclists' Speeds

Front Fork Damage Is Not A Reliable Estimate of Speed

It is often desirable to estimate the speed of a cyclist under certain conditions, particularly when an accident has occurred. It has been thought that, in cases where the bicycle hit a solid object straight on, the amount of bending of the front fork could be use to estimate the speed of the collision. This was attempted in the case of Johnson vs Derby Cycle Co., considerable money was spent in the tests and very accurate results were claimed. However, the formula which was claimed to have 5% accuracy had average error of 25%, and even under the best of transformations (made by someone other than the experimenter) the average error was no better than 6% . For the final formula which was claimed to have no more than 3% error on the basis of laboratory tests, the error when compared against the same data, was 84%. On average, the calculated speeds were only 16% of the speeds measured in the laboratory. Estimating Speeds by Fork Deformation Upon Collision

Proper Estimation Requires A Computer Program

To properly estimate the speeds of cyclists under particular conditions one must know not only the external conditions (gradient, wind, turn radius, brake application, for each section of road, and type of bicycle) but also the amount of power the cyclist is developing and his posture and clothing. The amount of power is most difficult to determine, although one can assume reasonable maxima when estimating the maximum probable speed. However, on descents where almost all the power comes from gravity, it is possible to make reasonable estimates. In fact, the first task for which American electronic computers were built was to perform this type of calculation for the trajectories of artillery shells. I have written such a program for the speeds of cyclists riding over varying terrain, using the accepted functions for the resistances encountered by cyclists (wind, slope, tires, etc.). This program calculates speeds at each point along the course that are in reasonable agreement with measurements. It is too complicated to post here, and, in any case, I regard it as proprietary.

Mysteriously Bent Back Front Fork

It is commonly accepted that a bent-back front fork has been caused by a collision with a solid object and not by application of the front brake. After all, the front brake should be powerful enough to create pitchover, throwing the cyclist over the handlebars, if too forcefully applied, and in those cases where this has occurred the front forks have not been bent. If the front fork is strong enough to carry the loads of the pitchover movement, then it can't be bent by the front brake, or by any similar restriction of the rotation of the front wheel. This is not quite true. There is a well-documented case in which a bicycle, carrying a mother with her infant child on the top tube in front of her, suddenly pitched over in the middle of an intersection and the bicycle ended up with bent back front fork, although all the witnesses stated that there was no other traffic about and that the road was smooth. Mathematical analysis shows that any sudden stopping of the rotation of the front wheel, far more sudden than the cyclist can apply with a hand brake, will develop sufficient force to bend back the front fork. Damage to Front Fork of Bicycle During Deceleration

"Safe" Brakes That Burn Up

Rim brakes can meet practically any cycling need. Coaster brakes, though also accepted by the US CPSC regulation, have both low stopping power (low deceleration) and burn up on long descents. This is an account of comparative testing of rim brakes and a coaster brake on a long descent.
"Safe" Brakes That Burn Up

Rear Brakes (Coaster Brakes) have greatly reduced stopping power on descents.

Bicycles with only rear-wheel braking are lawful vehicles. However, their braking is considerably less than that of bicycles with two-wheel brakes, and is greatly reduced as the descending grade gets steeper or the surface gets more slippery. The file you get through this hyperlink is a Mathcad worksheet that, when used with the Mathcad program, calculates the maximum possible deceleration rate for different shapes of coaster-braked bicycles on a range of slopes and with a variable coefficient of friction.

How The Bicycle Wheel Carries Its Load

For a hundred years after its invention, the physics by which the tension-spoked wheel carries its load was unexplained. Here is the explanation:
 Held Up By Downward Pull

A Fresh Look at Steering Geometry and Bicycle Handling

Chris Kvale and John Corbett, both custom frame builders, wrote an article on this subject that was never, so far as I know, published. The article discusses the relationship between bicycle front-end geometry and bicycle handling, indicating the considerable contribution of trail distance to ease of steering and stability. There are two tables relating head angle, rake distance, and trail distance.
BicycleEng\Kvale Geometry.pdf

To return to: Articles

To return to: John Forester's Home Page

Bicycle Engineering page last changed: 04-Feb-14