I am currently having trouble getting a non newtonian simulation to converge to my satisfaction, i am modelling a steady state shear thinning fluid with a large range of instantaneous viscosities. Residuals and spot values oscillate and although the simulation doesnt diverge as such there is no stability. Which differencing scheme is best for such a situation and which handels the range of viscosities the best?

I would recommend MARS as the differencing scheme. It isn't obvious though that your problem is caused by the differencing scheme at all. Have you tried under-relaxing the viscosity?

cheers for that, after closser inspection and fiddling with the viscosity relaxation its still not helping although it is the pressure that is unstable, should this be effected by the viscosity?