The strain tensor, principal strains and precision of the estimates of these values are derived for a range of different layouts of three‐dimensional strain rosettes. These values are based on the Monte Carlo technique applied to experimental work which was carried out on transducers tested in different laboratories. The estimates of precision are determined theoretically and compared with results based on experimental findings. A new design of a miniature tri‐rectangular tetrahedral rosette was manufactured and tested. Results suggest that this transducer does not perform as well as the rectangular patterns.

The strain tensor and principal strains were derived for a range of different patterns of three‐dimensional strain transducers. The precisions of the estimates of these values were based on the Monte Carlo simulation applied to experimental work that was conducted on strain transducers tested within the same strain field. The estimates of precision were also determined theoretically and compared with results based on experimental findings. Results suggest that the tetrahedron transducers in experimental work do not perform as well as rectangular patterns and that estimates of principal strains, which were not coincident with the axes of the transducer, showed slight discrepancies.

This paper is a review of different methods used to reduce the effect of strain gradients in experimental models with 3D embedded strain gauged transducers. A detailed analysis of various methods used in relation to research carried out on a prosthetic hip implant is investigated. The methods to reduce these gradient effects detail the importance of the selection of the most suitable 3D transducer pattern and establishing the locations on the model where strain gradients are least problematic. It is also shown that the use of large-scale models helps minimise the effect a gradient may have on data. Another way of avoiding strain gradients is to use axisymmetric models and embed the gauges at the same radius in the model but at a different angular orientation from the central axis. The strain gradient across the 3D embedded strain gauge transducer can be reduced using the methods described here so that significant errors in the final tensor do not materialise.