Hello, For professional use it is crucial need to get a very accurate time of the exposure (5 ms resolution or better). I am using STT-3200ME and MaximDL 6.08. Camera is connected via USB to my laptop. My first approach was of course to use "Measure Shutter Latency" tool of MaximDL but I had a lot of problems with it: discontinuity in 100th of second cells lighting while exposure, probably meaning that the display sequence of my screen was not fast enough to capture this information. That made the interpretation of the shutter latency results quite difficult but at least I observed that this shutter latency is fluctuating a lot. And doing an average of it is not acceptable for my application. Then we replaced "Measure Shutter Latency" tool by an array of LEDs commanded by a Rapberry running a real-time OS. On the other side, the camera was commanded by a script. Both PC controlling the camera by a script and Raspberry controlling the array of LEDs were fed by a common very accurate NTP time. The first results we are getting seem to confirm the important fluctuation of shutter latency. Too much to pretend to a 5 ms accuracy. I don't know where this fluctuation is coming from. Even though I guess I am not the only guy for which time-stamping of the pictures is crucial, I dug it quite deep but yet failed to find a way to get accurately this time information. Then I am now thinking to control the shutter more "directly" by script or C, and then time-stamp myself directly the resulting FITS header with the correct value. But is there a way to do that? Could you please help me? Is there anything left to attempt? for example on the camera control side using SBIG Universal Driver/Library, for example to activate very accurately the shutter in very accurate reproducible timing? At the very end, I could imagine to buy an external shutter commanded by TTL signal... but I am not sure there is enough room for that before the focal plane. Thanks for reading... and helping!