I'm a new user, in the U.S. midwest. I have found that the simulation results are significantly different (ie. 13%) between Lat/Long generated NASA-SSE (synthetic) and TMY3, which i believe uses NREL data. Two questions: is this kind of variance to be expected between simulations using these different data sets? And is there a generally accepted viewpoint of which of these sets of data are considered to be more accurate?