We haven't found that the values for the saturation in the headers are very good. We like to examine the data and determine fairly conservative limits for each chip.
Here's how we've had to revised the headers. If you're looking at a raw image, don't forget to subtract something reasonable for the overscan (bias) level.
Aug 00-Feb 01
hedit obj*.fits[1] SATURATE 39000 add+ ver- show+ hedit obj*.fits[2] SATURATE 23000 add+ ver- show+ hedit obj*.fits[3] SATURATE 37000 add+ ver- show+ hedit obj*.fits[4] SATURATE 24000 add+ ver- show+ hedit obj*.fits[5] SATURATE 27000 add+ ver- show+ hedit obj*.fits[6] SATURATE 27000 add+ ver- show+ hedit obj*.fits[7] SATURATE 29000 add+ ver- show+ hedit obj*.fits[8] SATURATE 24000 add+ ver- show+
hedit obj*.fits[1] SATURATE 48000 add+ ver- show+ hedit obj*.fits[2] SATURATE 35000 add+ ver- show+ hedit obj*.fits[3] SATURATE 25000 add+ ver- show+ hedit obj*.fits[4] SATURATE 28000 add+ ver- show+ hedit obj*.fits[5] SATURATE 30000 add+ ver- show+ hedit obj*.fits[6] SATURATE 32000 add+ ver- show+ hedit obj*.fits[7] SATURATE 48000 add+ ver- show+ hedit obj*.fits[8] SATURATE 28000 add+ ver- show+