Understanding fatigue-rated transducers

Original article date: November 1997

The term fatigue, when applied to a load cell or torque sensor, is one of the least understood terms in the transducer industry, says Paul Armstrong of Amber Instruments.

Typically, sensors that are manufactured for general purpose applications have been “de-rated” to one-half the normal output and then called “fatigue rated”. Since most applications do not fully demand the rigours of a true fatigue, this de-rating technique has been allowed to exist without much challenge.

To the measurement community, the term fatigue is understood to mean:

  • Designed to a stress level obtained from the S/N diagram (Fig 2) for the particular transducer material
  • A rated life of 100 million, at minimum, fully reversed cycles
  • Manufacturing technique and design considerations conducive to the total energy absorbed by the transducer during fatigue operation
  • Resistance to off-axis loading.

In static or general purpose applications, the output at full-scale (as expressed in mV/V) is traditionally the primary design consideration. Since the output of a strain gauge is proportional to the amount of strain it sees, a design stress corresponding to the strain producing the required output is obtained from the traditional stress-strain curve for a given sensor material. Fig 1 shows such a typical curve for an alloy steel.

In a fatigue application, life is or primary importance, that is the number of fully reversed cycles to failure. In this case, design stresses are obtained from a different diagram, conventionally referred to as the S/N curve (Stress / Number of loading cycles), as typified by Fig 2. The S/N diagram allows the design engineer to enter in the required life of the transducer to obtain the maximum allowable stress, irrespective of whether this stress will produce a certain output or not.

As the S/N curve becomes nearly horizontal, the ordinate stress is called the fatigue limit of the material. This is the stress level below which the material can endure an infinite number of completely reversed stress cycles without failure.

In most load cell applications, the load cell will be subjected to more forces and/or torques than just the primary measurement axis. This gives rise to a state of combined stresses. An empirically defined equation is used to estimate the values of the combined stresses in a particular transducer application. It is these combined stresses that should correspond to the required life in the S/N diagram.

While the design stress diagrams clearly distinguish between a static (general purpose) and a fatigue-rated transducer, it is not the only factor. Other considerations include optimum grain orientation, geometrically streamlined designs and avoiding stress raisers. From a manufacturing point of view, this means that EDM machining, welding and other similar techniques are usually avoided since they can induce microscopic cracks.

Because the S/N diagram is less well understood than the more common stress-strain diagram, transducer manufacturers have come to define fatigue life with various claims. Figures that range from 100 million to 1 billion and beyond have been stated. Because the curve is asymptotic after 100 million, any transducer that is designed for 100 million will generally last to whatever higher number a manufacturer elects to pick! When the S/N curve for a given material is developed, the limit of practical testing is typically 100 million, with the asymptotic portion being empirically derived. Therefore, use of 100 million fully reversed cycles for a fatigue life claim is arguably a more responsible figure.

  • Amber Instruments
  • Paul Armstrong
  • Tel: 01246 260250

November 1997