An evaluation of the performance of the OFDA2000 instrument operating in OFDA 100 mode

Abstract

The OFDA2000 was designed both as a portable instrument for use on-farm, and as a laboratory instrument not only for the measurement of mean fibre diameter, diameter distribution, and fibre curvature, but also the diameter-length profiles of suitably-prepared full-length fibre specimens. Descriptions of its use for on-farm work have been published in appropriate forums. This trial was carried out to assess the performance of OFDA2000 instruments operating in OFDA 100 mode, in order to confirm that the instrument used in this manner would comply with the requirements of IWTO-47.

Data obtained in a trial of 4 OFDA 2000 instruments and 4 OFDA 100 instruments in two laboratories indicates that the OFDA2000 in 100 mode gave equivalent accuracy and precision to the OFDA100 for mean fibre diameter, for both scoured wool and tops.

On standard deviation of diameter, equivalence was also demonstrated for scoured wool using IWTO-0, but the high sensitivity of the trail allowed the results on tops to show a very small difference that was marginally statistically significant. However, the magnitude of the difference (0.02 µm), and the fact that most of the responsibility for the statistical failure of this test could be attributed to the results on the coarsest top, suggests that to all practical effect, the two models of instrument are equivalent for this parameter also.

There was a very small and consistent bias of less than 1 º/mm between the two systems on curvature. Whilst this is almost certainly of no practical importance, the precision of this difference suggests there may be an anomaly in the calibration for this parameter in the two pieces of software, and this should be investigated by the manufacturer.

The OFDA2000 in 100 mode gave higher counts than the standard instrument. However, whilst additional raw data is always notionally of benefit, there was no compelling evidence that the small percentage increase produced any statistically significant increases in precision, except perhaps on tops, where the high precision of the trial allowed a small reduction in confidence limits to be detected.

Whilst IWTO-47 does not specifically mention OFDA 100 as the model referred to, it is recommended to include a minor text change to the scope of the standard to the effect that the OFDA 2000 when specifically used in 100 mode satisfies the requirement of the standard. 

Citation

"An evaluation of the performance of the OFDA2000 instrument operating in OFDA 100 mode", B. P. Baxter, Report RWG03, Technology & Standards Committee, Raw Wool Group, IWTO Barcelona Meeting, May, 2002.