I use the ML intervalometer for astro-imaging. Typically, I take 60 sec subs in bulb mode and get an image every 63 seconds. The additional 3 seconds per image I attributed to the storage of the raws and perhaps some lazyness of the intervalometer.
But recently I astro-modded a 1100D (filter & mirror removal, cooling). Thereby I also replaced the shutter by an emulator and did some tests with short exposures. When I came to exposure times below one second I left the bulb mode and selected the exposure time directly. And suddenly, the time loss of 3 seconds was reduced to half a second per image!
Is this a know effect, and if so: where does it comes from?