On January 15, 1919, the storage tank at the Purity Distilling Company exploded, sending a wave of molasses eight to 15 feet high through Boston’s North End, killing 21 and injuring 150. After the 2.3 million-gallon mess was cleaned up, an inquiry determined that the builder, Arthur Jell, had not tested the tank for leaks. The tank had been filled to capacity eight times since it was built, putting the walls under an intermittent, cyclical load.1 Jell, it appears, failed to perform a fatigue test on his storage tank.
Fatigue testing may be performed on raw materials such rubber, steel, plastics and composites, small components such as rivets or springs or objects as diverse as automobile frames and spinal column implants.
Fatigue: What is it?Fatigue is caused when a material-whether it’s an aircraft wing or a spring-is repeatedly loaded and unloaded (force is applied, and released). Over time, it becomes deformed and its performance falters. Once material becomes fatigued, it’s impossible to get it to return to its original condition. It’s estimated that fatigue is responsible for up to 90 percent of in-service part failures that occur in industry.2
Material fatigue passes through three stages: A crack develops; it spreads; the material breaks completely, or fractures. There are three ways to apply stress to material: axial (tension-compression), flexural (bending) or torsional (twisting).
Fatigue tests can be performed in several ways, depending on the exact nature of the material or the machine involved. In most cases, the item is repeatedly exposed to the same type of motions that it would be go through in regular operation until the item finally fails. If the item is used in an up and down motion, for instance, then it would be moved in an up and down motion until it fails or reaches the required cycle number.
After learning the results of the fatigue test, the next step is to determine how to lengthen the effective life of each component. You can extend the life of a machine substantially by lengthening the fatigue life of its shortest-lived component. While you still may not be able to make all of the components last the same length of time, you can still make things last much longer than they normally would. To do so, you have to understand what causes fatigue in the first place.3
Several different types of machines have been built to perform many kinds of fatigue tests4, from contact stresses to multi-level stresses. In the spring industry, we deal with axial and torsional stresses.
Fatigue testing is common in the spring manufacturing industry, where the performance of a spring can literally be a life-or-death issue (think of the hydraulic valves in airplane control). Spring life is also exceedingly important to manufacturers who want repeat business from customers who buy their springs.
Because springs do change over time and use, some spring manufacturers “exercise” or pre-set their springs before they deliver them to their customers. This actually increases spring life, and gives the spring more consistent operating characteristics.
Fatigue-Testing a SpringIf you hired a worker to fatigue-test springs with a manual spring tester, chances are, the repetitive motion (not to mention the boredom!) would send him to the company nurse before coffee break. With a motorized spring tester, you can set the machine once and it will happily test the same spring all day long and well into next week without stopping or complaining. Some testers will allow you to run as many as 10 million cycles on a part.
For many years, speed has been the primary criteria for fatigue testing. An 1,800-rpm motor, with an appropriate diameter output wheel connected to a linear arm can produce high cycle counts in a short time.
However, high-speed testing does not necessarily produce the same results as fatigue produced at as-sembly or operating speeds. It’s important to a spring manufacturer-indeed, any manufacturer--to know how long a part will last when a certain amount of force is applied to it. More complex tests must be performed, and more controls are needed.
Typical Fatigue Test on a SpringIn a typical test, the spring must be cycle tested 90,000 cycles at compressed lengths of 2.2 and 1.8 inches. At various points of the cycle test, the spring must maintain load ratings.
To test a spring, a machine operator places a spring on the lower platform of the spring tester and sets the machine parameters by choosing a target force or length. Next, the operator decides whether to test the spring in increments or to an absolute number. He then specifies whether he wants to test the spring Length to Fatigue or Fatigue to Force.
The operator then enters the desired force or length value and the number of cycles to run, and presses the start button. The tester’s upper platform will move downward (if testing for compression) or upward (if testing for extension) to the specified length, then return to its starting position. This cycle will repeat itself until the tester completes the number of tests specified; in this case, three rounds of 30,000 cycles will be performed.
The fatigue test can be programmed to stop if the spring length gets out of tolerance or the spring breaks.
Our spring for this fatigue test is 3 inches long. Its maximum compression-the point at which it be-comes a solid object and cannot be squeezed any further-is 1.4 inches. The tester will apply loads (force) of 45 and 68 lbs., plus or minus a margin of 10 percent, and compress the spring to lengths of 2.2 and 1.8 inches. As Chart 1 shows, the spring maintains a constant force at length ratio of 57.5 percent through 90,000 cycles and does not appear fatigued.
ConclusionFatigued parts or materials can easily shut down your operations, costing time and warranty money and leave the company open to possible lawsuits. It makes good economic sense to know the endurance limits of your parts, equipment and products. Fatigue testing can help you plan future capital expenditures and improve your quality. It may also help you establish a reputation for quality and help you fight off foreign competition.
Arthur Jell’s failure to fatigue-test the storage tank before filling it with molasses created the first class-action lawsuit in the United States and cost his employer, the United States Industrial Alcohol Company, more than $600,000 in damages, the equivalent of $6.6 million today9. In today’s litigious business climate, fatigue testing makes more sense than ever.
References “Boston Molasses Disaster”, Wikipedia, http://en.wikipedia.org/wiki/Boston_Molasses_Disaster
 “Experiment: Fatigue Testing”, ME 3701, Materials of Engineering Laboratory, Louisiana State University
 Paraphrased from a blog, www.sorbothane.com/blog/fatigue-testing
 Fatigue Testing And Analysis Of Results, Chapter III, Fatigue Testing Machines and Equipments, Waloddi Weibull, 1960, www.barringer1.com/wa_files/Weibull-1960Book-2.pdf
 The length of a spring in its natural, uncompressed state.
 Rate, or spring rate, is the change in force that the spring exerts, di-vided by the change in the deflection of the spring. This principal is also known as Hooke’s Law, which states that load is proportional to dis-placement. Rate may be expressed in pounds per inch or Newtons (N) per mm. In spring manufacturer’s terms, spring rate is the amount of weight required to deflect a spring one inch.
 Solid height is the maximum amount that a spring can be compressed. Our 3-inch spring can be compressed to a solid at 1.4 inches.
 Force at Length is the force exerted by a spring at a certain length. Length at Force is the length of a spring when subjected to a certain force.
 “Was Boston Once Literally Flooded with Molasses?”, Cecil Adams, www.thestraightdope.com , December 31, 2004