Hi...
Unfortunately, the variations you've reported are too much to say that your compound is consistent with data observed in literature.
Changing the MHz from 500 to 400 should not effect the center of the peaks. Your signal to noise and resolution should improve, but the center of the peaks should not change, therefore they should be at the same ppm. Is it possible that you have a metal contaminant? Shimming should not change the ppm of your peaks, bad shimming can cause them to be mishapen, so you likely have a sample related problem. Another good way to check this is to see if the peaks for a standard sample are also off. If that is the case you might need to ask the NMR technician for help.