Top: Radial velocity, FWHM and CONTRAST of the CCF for spectral order 145 (spectrograph order 13) as a function of time, with no correction for blaze variation on top and correction for it in the bottom (see text for details). We clearly see that the blaze variation correction significantly reduces the saw tooth pattern induced by periodic detector warm-ups. The obvious jump seen in the corrected FWHM and CONTRAST is induced by the change of the detector cryostat. Bottom: Radial velocities averaged across all orders with and without the correction for blaze variation. In their difference (bottom panel), we see up to one m s⁻¹ jumps at each detector warm-ups shown as the red vertical dashed lines.