When fitting the Spectrum, the error on the peak position is only accounting for the statistical uncertainty.
One also observes an undererstimation of the error when looking at the Vcal Callibration Fit.
This might me due to not include the systematical uncertainties coming from the PH->Vcal Conversion.