Two examples of the 4-component noise model fits to measured noise PSDs. The underlying noise estimation for these fits is described in Planck Collaboration XII (2016). It measures the sample auto-covariance of signal-subtracted and masked data to overcome effects of gaps and signal residuals. The blue curve is the measured PSD averaged over 30 pointing periods. The red curve is the 1 /f component, deconvolved with the bolometer transfer function. The black curve is the electronic white noise component and the grey curve is the photonic white noise component. The magenta curve is the measured a/b correlated component. The green curve is the sum of all noise model components. Vertical blue lines indicate the region where the fit was made and the vertical black lines indicate the 10 lowest spin harmonic frequencies.