2 Observations and data reduction

Table 1 gives a log of the spectroscopic observations of HD 826: in Col. (1) the PN and central star designations of NGC 40 are given; in Col. (2) the spectral type and V magnitude of HD 826; in Col. (3) the telescopes and spectrographs used; in Col. (4) the observed spectral range, the adopted spectral resolution and resolving power; in Col. (5) the continuum signal-to-noise ratios for each night given in Col. (6); in Col. (6) the dates of observation; in Col. (7) the number of consecutive spectra acquired each night.

We used the 1.52 m telescope at the Observatoire de Haute-Provence (OHP, France) equipped with the Aurélie spectrograph (see Gillet et al. 1994). The detector was a double linear array Thomson TH7832 of 2048 pixels (Gillet et al. 1994). We used a 300 l/mm grating, leading to a 2.8-pixel resolving power of 5000 ($\approx$1 Å spectral resolution at 5500 Å). The spectral range was centered on 5625 Å and covered 5250-6000 Å. The entrance aperture of Aurélie is circular with a diametre of 3 $^{\prime\prime}$. We also used the 1.6 m Ritchey-Chrétien, Boller & Chivens telescope at the Observatoire du mont Mégantic (OMM, Canada) combined with the Perkin-Elmer (model 31523) spectrograph at the f/8 focus. The detector was a THX CCD with $1024 \times 1024$ pixels (before 1996 July 29), or a Loral CCD with $2048 \times 2048$ pixels (starting on 1996 September 26). We used a 600 l/mm grating as dispersive element, leading to a 2.2-pixel resolving power of 2000 ($\approx$2.8 Å spectral resolution at 5700 Å). The spectral range, centered on 5630 Å, covered 5300-5960 Å. The width of the slit was 2.5 $^{\prime\prime}$.

Figure 1: HD 826 typical normalized spectrum indicating the most obvious emission or absorption features (1998 January 23)

Tests at higher resolutions (0.5 Å and less) were conducted at the OHP (23 spectra taken in January and March 96). At high spectral resolutions, we were forced to significantly increase the exposure times in order to reveal the subpeaks (hence losing time resolution). However, despite longer exposures times (about one hour and more), we only achieved poor S/N ratios precluding any detailed statement concerning the appearance of subpeaks observed at different resolutions. The spectra were reduced in the way described in Paper I with the MIDAS package (OHP data) and the IRAF package (OMM data).