CFD Debugging - Negative densities

I have written an inviscid 2D flow solver for supersonic flow over a wedge. I eventually want to get to viscid flow. However, I cannot seem to find the source of error where the density, pressures, and temperatures go negative. The mathematical schemes are the MUSCL and AUSM + UP if you're interested. Again, please let me know if you find anything.

Thanks,

Thomas