While fitting the two disk model, we realised that intrinsically the two functions describing the thin and the thick disk are the same, so both parts must be constrained heavily. When fitting they have to be given different starting values and fitting boundaries to prevent the fitting function from giving unphysical solutions and to be able to distinguish the results for the thin and the thick disk.
For the two disk model we expect to see two vertical regions, the inner part dominated by the thin disk and an outer part dominated the thick disk. This is surely the case for the thin disk as it is by definition the dominating disk, as the thick disk is expected to be fainter and more extended.
For fitting one will give preference to the smallest possible amount of fitting parameters. Of all the parameters of the two disk fit, the scaleheight of the thin disk is always the most accurate and stable parameter because of its high S/N ratio, while the thick disk component will be faint with few datapoints and a lower S/N.
A 1D Two Disk fit is often ill constrained as it uses four parameters to describe 15 datapoints, so the results can be erratic if there are irregularities in the inner part of the profile. One solution is to constrain the well defined thin disk, by constraining the value of 
by taking a small selection of datapoints which clearly belong to the thin disk and from these determine the slope for each cut. A mean or median of all profiles over the radius can then be used to determine a constant value for . Using this value as a constant for each cut of the 1D two disk fit shrinks the amount of fitting parameters from four to three, making it easier and faster to fit the profile.