Combining several measurement sets

this is a generic question about simultaneous fit of several measurements.
Here is a simplified template problem:
Suppose I have one observable and two sets of unbinned measurements for it,
say 100 measurements with accuracy 10, and 200 measurements with accuracy 20,
so I essentially want to fit Gaussian of width 10 to the first data set, and Gaussian
of width 20 to the second dataset, keeping mean value for both fits 100% correlated.
How the corresponding RooFit model could look like?