AAO Newsletter October 1996 - Page 5
Figure 2: Vector diagram showing a comparison between two measurements of the same plate: (a) two measurements in the same orientation; (b) one measurement made at 90° to the other.
blocks of standard length from the National Physics Laboratory. These tests showed the principal axes of the table to have RMS error in absolute accuracy of ~ 0.2µm in both axes; the standard error of short timescale repeatability was shown to be 0.05 to 0.lµm.
Repeatability and absolute accuracy over the area of a 14 inch plate
The intrinsic repeatability of the table can be easily measured by scanning the same plate twice, without removing the plate from the machine between scans. Figure 2a shows an array of vectors over the area of a survey-grade 14 inch Schmidt plate scan, representing the mean shift in position of all objects within 1 cm boxes. Figure 3a shows histograms of the x and y shift values. The standard errors of these distributions, divided by 2 for a single measure, are x = 0.07µm and y = 0.03µm. These are consistent with the values found previously using standard gauge blocks. By measuring the same plate twice, but this time at two different orientations (in this case 0° and 90°), we can check the absolute positional accuracy by rotating one measurement in software back on to the co-ordinate system of the other and examining the shift vectors as before. Such a
Figure 3: Histograms of x and y shifts from Figure 2.