Is there any consensus on the time interval to be used to calculate noise for S/N system suitability purposes?

Stabilise the system, Inject the blank, and use S/N from the part of maximum noise.

I would just make sure that it is something that is chromatographically relavent. Short term ASTM is more for electrical noise. I think something on the scale of a few seconds to a few tens of seconds.

The only official guidline I've seen is from a 1988 BP (Appendix III D p. A85), which states that S/N = 2H/h, where H is height of pk coresponding to pk of interest, and h is the "absolute value of the largest noise fluctuation from the baseline in a chromatogram obtained after injection of a blank solution and observed over a distance equal to 20 times the width at half height of the peak in the chromatogram...situated equally around the place where this pk would be found." Personally, I don't like it- What is "the baseline" - What is a "fluctuation"(Spike? Long-term wavyness?) What if you have a 1 min. wide pk. in a 7 min chromat? I think 20 times the width is way too much and will include long-term noise that's obviously different from your peak. I like to use 30-60 seconds or so, or at minimum nothing less than one peak width. Does anyone actually use this BP procedure? I don't have access to a newer BP.

Present edition of European Pharmacopoeia (page 64 of 2002 Ed.) use the same criterion of 1988 BP. With the Agilent Chemstation software I use the Point-To-Point method to obtain noise.

I had always thought that one would take the noise (pk-pk) at a width equivalent to that of the widest peak. The kicker of using this is where exactly to pick that width. of course you could always go to 3*SD, always gets me pretty close to pk-pk