Measured and Predicted OFAD

In Table 2, results of theoretical analysis of the OFAD for the "as built'' PF camera using Zemax tm are presented along with the values empirically determined by Guo et al.(1996) from plates taken in 1995. In both cases, the two ADC doublets were set in a "neutral'' position with their dispersive axes opposed in the north-south direction. UBVRI bands in the observations are approximated respectively by the conventional wavelengths of 3600A, 4400A, 5500A, 7000A and 9000A.

Table 2: Theoretical and Experimental OFAD Coefficients for the PF Camera

Source Band f.l. d3 d5 Comments
Guo U 11465.4 360.0 775000 Measured
Guo B 11467.3 360.2 695000 Measured
Guo V 11467.8 357.6 687000 Measured
Zemax U 11461.9 360.2 880000 Predicted
Zemax B 11464.2 357.4 840000 Predicted
Zemax V 11465.6 354.7 820000 Predicted
Zemax R 11466.2 352.4 810000 Predicted
Zemax I 11466.9 350.7 800000 Predicted
d3 and d5 are dimensionless
Images have been refocussed for each passband
All values are from 2mm BK7 filters

The photographic exposures were made through UG-5, GG385 and GG485 filters with thicknesses of 2.73mm, 1.96mm and 1.84mm respectively. The focal lengths in Table 2 have been corrected to the values they would have had if the filters had all been 2mm thick and made of BK7.

These two sets of models predict star positions with an rms difference between theoretical and empirical positions of less than 13µ over the field in all three colors. Nowhere does the predicted position of a star image differ from its measured location by more than 17µ (.31 arcsec). Still, the measured focal lengths of the system are slightly greater than the theoretical values. The rms difference between measured and predicted positions can be reduced to 4µ by adjusting the focal lengths of the system to be 1.8mm greater than predicted. As we will show, an adjustment of this order is what might be expected as a result of manufacturing tolerances.

Theory and experiment agree that there is a gradual increase in f with increasing wavelength while d3 and d5 decrease. Most of this variation of the OFAD is caused by secondary chromatism coming from color dependencies in the image distortion. Moving away from the optical axis, the blue images at first fall slightly closer to the center than those in the red so the focal lengths are lower in the blue. Farther out in the field, the shorter wavelength images begin to be displaced more because of larger values of d3 and d5, passing the longer wavelength images near the edge. The U images are as much as .36'' from the I images in parts of the field. This shift is the reason the broad band images often required by Argus are somewhat larger than the narrower band images generally used for CCD exposures. With Argus and the PF Camera, there is a small additional shift of image position with color resulting from primary chromatism caused by the incorrect filter thickness.