spectral method..numercial instability

DJ

Hiall, I am using spectral method to resolve N-S equation in 2D.. velocity profiles seem ok.. but isobaric curves are really strange and looks bad!! i guess this is due to numerical instabilty.. i have been using collocaion method with stagered grids.. so that collocation points for momentum equation and for the continuity equations are different.. any suggestion???? please!!!

How can we get he boundary conditions for vorticity( in stream funxction- vorticity formulation of N-s Equation) for the classical problem of lid driven cavity flow??

.. In the stream function, velocity formulation of the N-S equation.. I have a forth order partial differential equation for stream function.... is it ok??? .... Thanx in advance, DJ