The toluene trick usually does it for me.
If the material really seems to love a volatile you will have to do several cycles of dissolving/evaporating for it to work. I'm assuming you are running the NMR in CDCl3, so after the runs with toluene do a few runs with regular chloroform. I've also heard that dioxane is able to azeotrope acetic acid more efficiently than toluene, so you might give that a shot as well.
Is your product an oil by chance? If so, after doing the above runs, drop a stir bar in your RBF before you leave it on high-vac overnight.
It also might be good to double check that your high-vac is pulling as well as it should.
If these tricks don't work, double check that your CDCl3 isn't contaminated.