Context

Just wanted to show an other simple method in order to calibrate low resolution spectra with the Neon lamp available with LISA and LHIRES3.

Using all the lines present in Neon lampe from Shelyak

Follow these steps to get the best from your Neon lampe:

1. First of all, take a short (typically 0.3 s) exposure spectrum of the lamp in order to get red Neon lines;
2. Take a long exposure (typically 15 s, be aware with remanance!) to get blue lines;
3. Merged 2d spectrum with the background unified (red edge of long exposure was set to 0).
   For Spcaudace (≥2.12) users, spc_lampmerge command does it all in one step for you:
      spc_lampmerge spectre_lamp_long_exposed.fit spectre_lamp_short_exposed.fit

You get the best of the two worlds!

Standard wavelength calibration on wide CCD:

Then, the lines profiles with identified main lines is made for you:


(Figure updated on 2012/01/29)

I recommand you to select widespreaded lines in order not to favour the red edge or blue edge.

• The following lines seem working fine for a deg 3 polynom:
  4065.65, 4348.06, 5400.56, 5852.49, 6402.25, 7032.41 and 7245.17
  Which lead to a rms of 0.6 A where the dispersion is here 2.57 A/pixel, so an accuracy of 1/5 pixel.
• With a deg 4 polynome and the following lines:
  4065.65, 4348.06, 4710.06, 5400.56, 5852.49, 6402.25, 7032.41, 7245.17, 6143.06, 6678.28 and 6929.47,
  the following dispersion law was found: Lambda=3917.50781533+2.43491203042*(x-1)+0.000230671871345*(x-1)^2+-1.61973682859e-07*(x-1)^3+4.63971559225e-11*(x-1)^4 with RMS=0.293443870835,
  So a linear law described by: 3917.50781533+2.56720166243*(x-1).

Method's limits:

Even if RMS of internal Neon is impressive, star's calibration with the associated Neon lamp calibration law isn't fine at all almost in the blue part, assumed that geometrics corrections are the same. Here are some tests:

• Calibration number 1: neon_regulus_loi-1.fit, regulus_loi-1.fit
  
  • Lines used (Å): 4065.65 4348.06 5400.56 5852.49 6400.25 7032.41 7254.17
  • Calibration law computed:  λ=3873.64043675+2.45329879185*(x-1)+0.000161324351193*(x-1)^2+-9.14163293013e-08*(x-1)^3+2.3751959068e-11*(x-1)^4
  • RMS=0.300161931476
  • Linear law: λ=3873.64043675+2.56470286273*(x-1)
  • Balmer's lines shifting:
    

    Line	Hα	Hβ	Hγ	Hδ
    Δλ (Å)	+0.3	+6	+15	+20

• Calibration number 2: neon_regulus_loi-2.fit, regulus_loi-2.fit
  
  • Lines used (Å): 4200.67 4348.06 4710 5400.56 5852.49 6400.25 7032.41 7254.17
  • Calibration law computed: 
    λ=3859.60617185+2.5432888818*(x-1)+-3.34322566205e-05*(x-1)^2+8.0436440161e-08*(x-1)^3+-2.92753132789e-11*(x-1)^4
  • RMS=1.40995880447
  • Linear law: λ=3859.60617185+2.57360695356*(x-1)
  • Balmer's lines shifting:
    

    Line	Hα	Hβ	Hγ	Hδ
    Δλ (Å)	+0.6	+5.5	+12	+15

Conclusions:

Without purchasing a ThAr lamp, Neon lampe available on LHIRES3 and LISA spectrographs seems to have enought lines in blue side too for a suitable wavelength calibration in low resolution spectrography.

However, even if internal Neon lamp calibration RMS looks fine, star's calibration with the associated Neon lamp calibration law isn't good in blue side.