Has anyone had experience with calculating the enthalpy of formation from DFT? I have had a few calculations with the planewave-pseudopotential method. What puzzles me is that in one case the agreement with experimental values is excellent, while another case was about 30% more stable. Surveying the literature, it seems like a hit-and-miss affair. Some calculations show good agreement to 5%-10%, but there are the occasional reports of more than 30% off. Just to note that the experiments are done at 700K (for the calorimeter), and the DFT calculations done at 0 K. Made sure the DFT energies are converged to 1 meV.