The actual exposure time of your image may differ by some fraction of a second from the requested exposure time due to the fact that no shutter can open or close instantanously. In practice, our ``big" shutters (used for direct imaging with the T2KA and T2KB chips at the 4m and 0.9m) have a shutter correction time of about -0.12 sec, uniform across the field. Our brand new big shutter used at the Schmidt with S2KA seems to have a negligible shutter correction time. The ``medium" shutter (used for direct imaging with the T1KA chip has a correction time of roughly +0.026 sec at the center and +0.004 at the edge. Our ``small" shutters (used for direct imaging with the T5HA and TI chips) are said to have a nearly negligible shutter correction time. By contract, the shutters in our spectrographs are quite slow, with a measured correction time for GoldCam of -0.25 sec, for example.
If you are attempting absolute photometry or spectrophotometry, and plan short enough exposure times that you are worried about the shutter correction, there is a simple routine in IRAF called findshutcor which will help you evaluate the correction. Written by Rob Seaman, the routine determines the shutter correction from a series of flat-field exposures of varying lengths (1 sec - 20 sec, say). You must be careful not to have saturated on the longest exposures, of course, and the frames must have been processed for removing the overscan level, but not flat-fielded. In taking the data one must also carefully take into account the possibility that the flat-field lamps intensity may drift slightly with time; i.e., a good exposure sequence might be 1sec, 20 sec, 2 sec, 2 sec, 20 sec, 1 sec. The findshutcor routine currently lives in the nlocal package.