Bug in all interpolate2grid routines
In the surprisingly not-so uncommon case, that 2 subsequent raw data points of a meassurement are already aligned to a frequency grid and are temporally seperated by a gap of size 2*freq, a BAD-flagged linearly interpolated (-or whatever interpolation method was selected-) value gets inserted to the harmonized timeseries in between the 2 values. However, the expected behaviour would be a NaN value to get inserted.