[knut]

Hi all!!
I'm having some troubles in sampling some data for time statistic. I'm running a LES simulation on an empty box with a velocity inlet, pressure outlet on the opposite face of the box and periodic BC for the side walls. I want to monitor the velocity fluctuations in the 3 directions. To do that I create 3 custom field functions (eg Xvelocity- meanXvelocity). When I restart the calculation after enabling the sampling of the 3 custom field functions for RMS values, it becomes unstable (the residuals increase of more then one order and their time history becomes pretty unsteady). Should the solver be uninfluenced by whether I sample or not one of the variables? I noticed this behaviour only when I sample custom field functions. Any idea?