A force transducer does not simply deviate by a fixed percentage. At the same force it gives a slightly different value when the load is decreasing than when it is increasing, it does not repeat a measurement perfectly, and it responds to the way it is mounted. On a calibration certificate those properties appear as hysteresis, repeatability, reproducibility, creep and zero return. In this article you can read what they mean, how they are determined on the force calibration bench and what you do with them in practice. How the measurement itself is carried out is described on the page force calibration in tension and compression.
Hysteresis: increasing and decreasing loading
Hysteresis (called reversibility in ISO 376) is the difference between the indication at increasing and at decreasing load, at the same force. A measuring body of steel or aluminium does not spring back perfectly: a small part of the deformation lags behind the force. In mechanical force gauges, friction in the mechanism adds to this. The result is that the indication during decreasing loading is usually slightly higher than during increasing loading.
On the force calibration bench, hysteresis becomes visible because each series is increased in steps to the maximum force and decreased again in the same steps. The difference per measuring point, expressed as a percentage of the force or of the range, is the hysteresis. For your application the question is: do you measure only at increasing force, as in a tensile test or a proof load test, or also at decreasing force, as when lowering a load? In the second case the hysteresis counts fully in your measurement error.
Repeatability: the same measurement again
Repeatability is the spread between consecutive measurements at the same force, in the same mounting position and under the same conditions. On the bench, two or more series are measured directly after each other for this purpose, without removing the instrument. Poor repeatability points to noise in the electronics, play in the mounting, a damaged measuring body or an unstable power supply. It is the lower limit of what the instrument can do: deviations can be corrected, spread cannot.
Reproducibility: mounted differently
Reproducibility goes a step further: the spread between series in which the instrument has been removed in between and mounted in a different position, for example rotated 120 degrees about the force axis. This reveals the effects of a force introduction that is not perfectly centric, of the mounting parts and of the assembly. ISO 376 therefore requires measurements in several positions. In practice, good reproducibility means that the instrument is insensitive to small differences in mounting, something that is directly relevant for load shackles and load pins.
Creep and zero return
Creep is the change in the indication over time while the force remains constant. The measuring body and the adhesive of the strain gauges continue to yield slightly for a while. In applications where a load is held for a long time, such as a proof load test or the pre-tensioning of anchors, creep is important. The zero return is the indication after the force has been completely removed. A transducer that does not return to zero has been overloaded or is mechanically damaged.
How do these values end up on the certificate?
In a regular force calibration we measure, per measuring point, the deviation at increasing and decreasing loading and, with repeated series, the repeatability. In an ISO 376 calibration this is extended with the reproducibility over several positions, the interpolation error, the zero error and, where relevant, the creep, and the results are tested against the class requirements. All contributions are included in the measurement uncertainty, calculated according to EURAMET cg-4. That uncertainty is therefore not an arbitrary error margin but the sum of what the instrument and the measurement itself do.
Worked example
A 50 kN dynamometer indicates 25.03 kN at 25 kN during increasing loading and 25.09 kN at the same point during decreasing loading. The deviation at increasing loading is then +0.12 %, the hysteresis 0.06 kN or 0.24 % of the indicated force. If you repeat the series and find 25.02 and 25.04 kN at increasing loading, the repeatability is 0.02 kN (0.08 %). If you only use the instrument to lift a load and read it at increasing force, the deviation of +0.12 % is your correction and the hysteresis hardly plays a role. If you also read while the load is being lowered, you have to reckon with a band of over 0.2 %. In this way the application determines which value on the certificate matters to you.
How to keep hysteresis and spread small
- Preload. Load the instrument several times to the maximum force before a series of measurements. The measuring body settles and the first measurement is then representative.
- Warm up. Give the electronics time to reach temperature; many indicators need minutes for a stable zero.
- Load centrically. Use the prescribed mounting parts, make sure the force passes through the axis of the transducer and avoid side forces and bending moments.
- Same direction. Measure as far as possible in the same direction in which the instrument was calibrated. Hysteresis only counts if you read in both directions.
- Let it settle. Let a swinging or jolting load come to rest first; a dynamic peak is not a measured value. For crane scales, practical tips are given in using a crane scale safely and accurately.
- Zero check. After use, check that the indication returns to zero. If not, have the instrument inspected.
In summary
Hysteresis, repeatability, reproducibility, creep and zero return together describe how stable a force measuring instrument is. They are determined on the force calibration bench by loading up and down in steps, repeating series and mounting the instrument in different positions, and they determine the measurement uncertainty on your certificate. Would you like to know how your force transducer or dynamometer is performing? See force calibration in tension and compression.