Comparison between the solar RVs reduced using the old HARPS-N (blue) and the new HARPS-N DRS (orange). Top panel: weighted rms for each chunk of RV data of 50 days, which corresponds to two solar rotations and therefore include activity signal. Orange and blue bars correspond to the weighted rms measured on the new RVs and on the old RVs, respectively. The number for each chunk corresponds to the improvement in weighted rms brought by new DRS (in percentage). Second panel: solar RVs as derived by the old and new HARPS-N DRS. The blue and orange lines correspond to a smooth version of the calcium activity index variation linearly fitted to the RVs, to remove the strong long-term correlation between RVs and the log(), and thus to mitigate the contribution from the solar magnetic cycle (see text). The colour scale for the new RVs corresponds to the S/N of the measurements. We note that we lost a significant amount of data around BJD = 2458100 (December 2017) due to a damaged fibre (see text). Third panel: difference between the new and old solar RVs, for which a long-term trend is clearly visible. We highlighted with red and green vertical lines the significant discrepancies in those RV differences. The red lines correspond to problems in the old DRS RVs, while for the green vertical line, it is not clear if the correct RVs are the ones derived with the old or new DRS. Bottom panel: generalised Lomb-Scargle periodogram of the residuals after removing the long-term trend observed in the RVs. The small inset shows a zoom around the periods 10–30 days to highlight stellar rotation.