[Peter.M]

You are assuming the images are shot noise dominated. This is what I am saying, if we assume that we expose the 5nm image to be just shot noise dominated, to get the 3nm image to also be shot noise dominated will take longer. We are both making the same point Clive. I understand that the 3nm will collect more signal in that time but it will require more exposure (than the 5nm) to be shot noise dominated.

Obviously every variable you mention that could be changed is kept constant (camera telescope ect ect) and I am assuming the transmission of each filter is similar. You work out 5nm image is shot noise dominates at 30mins changing to a 3nm filter The image will have less noise, because it is no longer shot noise dominated. The 3nm filter will benefit from longer exposure. The 5nm image will not.