I think one thing I did not understand about the single fit is that it tries to use all standards (although some of the standard's weights can be equal to 0 in the results), which can be detrimental to the fitting. Thanks a lot for the precisions, things are clearer now!

--