I want to compare the variance components between two instruments for measuring the mean fiber diameter (MFD) of wool. Wool samples were collected from 4 animals with different MFD. From each wool sample 5 subsamples were pulled. Each subsample was measured 3 times on one machine and then 3 times on the second machine. Samples were reloaded between each measurement such that the recorded measurement contained error associated with both inherent machine error (which is quite small) and loading error, which could affect which fibers are read and how they are presented. Please advise on the appropriate analysis