cavitation...I'm in a trouble! SOS

Student

I'm trying to model a cavitating flow with the first versione of fluent 6. I've a pressure inlet of about 90 bar and an outlet of 2 bar with a meatus of 0.5mm (it's an homogenizing valve). I've tried with an unst simulation bringing the pressure outlet from 85 bar to 2 bar with a udf. The time step is 0.02 s and for 6s. The underelaxx: 0.3 for momentum, 0.1 for pressure, 0.001 for vap mass,0.1 for volume fract, 0.8 for all turbolence parameter.The bubble number=9e6. Segregated solver, k-e realizable,second order upwind for all the eq, Standard pressure coupling. If I use incompressible gas model for the vapour the solution seems to converge but the pressure still remains negative (-240 bar!!!!) and the vapour is in "high" pressure zones. If I use the compressible model the solution diverges at about 40 bar.