Thanks Jeff, for the reply. Your program stabilizes around the same value on my machine, so indeed it is likely something in Snack. Too bad, I really like this toolbox. I'll try to contact the author, and perhaps in the meantime look for an other solution. All best, Gabriel