The advection and diffusion of all variables in the 1D Model and The 3D-Model are calculated.

If problems in the main loop occur, don't close the ENVI-met window frustrated.
First look at the output and the status bars to see which of the variables have cause troubles. Normally one or more of the variables display show an unrealistic high or low value. Then close the program frustrated.

This is either a result of a bad wind field with too high divergence (see here) or it is a general numerical problem.
If it seems to be a general numerical model, decrease the time step of your model (see here).

Also, you should have a look at the PANIC DUMP ENVI-met has produced. It is a set of normal Output Files except that is was written outside the usual save interval. Surf through the file and try to discover the spot where the problem is caused. Normally the point of trouble is quite visible due to unrealistic high and/or low values: Consider to modify this area slightly.
Sometimes it helps to change the inflow direction of the wind a few degrees. This does not actually effect the model results much but might help in case of numerical problems.

Unrealistic jumps in the sources emission rate might cause numerical problems. Make sure that the emission values you use are realistic and show a smooth diurnal cycle.

If you area contains a huge amount of vegetation, problems with the internal CO2 system might occur. In this case, go to the Advanced Settings Panel and enable the option “Use small time steps for CO2 model”. This might solve the problem.