Size of ACIS-2C pixels

In the preceding subsection we assumed that both the FAM relative spatial motion in Y and Z, and the size of the CCD pixel are exactly as predicted. Using the above subpixel test data it is possible to cross-check these assumptions. To do this we divided the stated total motion of the FAM (440 $\mu$m) into dummy pixels of size differing from the fabrication goal of 24 $\mu$m. In the ideal case there should be a linear correlation between the position of a beam inside an ACIS pixel and that of trial FAM pixels.

Values of the linear coefficient for different sizes of FAM `pixels' are shown in Fig. 6.30. It is clearly seen that only a narrow region of 24.60 +/- .025 $\mu$m pixel size exists where the data converge and give good statistics.

The value of 24.6 $\mu$m is 2.5% bigger than the nominal value of 24 $\mu$m for ACIS pixels, and measured with a high precision - 250 Å - by this method. This deviation is considerably larger than allowed by the measurements of the total CCD active area for 1024 pixels. The probable explanation for this difference is that the FAM scale itself is compromised and the true displacement is really 390 $\mu$m when readings show 400 $\mu$m.
 

Figure 6.30:  Plot ofcorrelation coefficient vs. assumed size of ACIS CCD pixel