Curve Fit Problem

Typical. You are fitting the data directly, but you are plotting on a log scale. The log scale accentuates the small errors, so that they look large (although they are in fact very small). If you want something that looks like a fit on a log scale, fit the logs of the data.

Your data also shows a lot of scatter. Almost like you have mixed in two different data sets, the upper values and the lower values. Don't read too much into whatever values you get from minerr.
__________________
� � � � Tom Gutman