TEST OF AN ECHELLE SPECTROGRAPH

The page present my very first tests of a cross-dispersed spectrograph (April 2008). This moderate resolution échelle spectrograph was designed to allow stellar spectroscopy with resolution (R) from 5000 to 12000, depending on the brigthness of star and the scientific goals (the resolution is modified by changing collimator and/or camera lens).

The grating is an echelle model of 79 grooves/millimeter with a blaze angle of 63°. The spectrograph is equipped by a cross-dispersion prism (F2 glass) to separate spectral orders, so many differents are recorded in a single CCD exposure. The camera use standard photographic lens (nominaly a Canon 85 mm f:1.8 objective lens). The camera can be a CCD camera or a digital SLR (for education purpose - direct spectacular color spectra can be recorded).

The telescope and the spectrograph are linked by a 10-meters long fiber optics (50 microns diameter core). A special feature is used for center and focus the star image at the entrance of the fiber (the prototype is from F. Cochard - Sheliak instrument).

 For the first light I use a CN-212 Takahashi telescope (D=0,2 m). I exploit the direct Newton beam of the telescope (F/4 focus).

Use of a low cost digital camera  (Canon EOS 40D)

True color echellogram of Vega (Alpha Lyrae) taken with a Canon 40D camera (the red response is extended, see here) and a Canon 85 mm f:1.8 lens. The measured resolution power is R=8 000.
The H-alpha line is the large absorption feature at the top. The H-beta line is near the center.

Echellogram of Delta Scorpius (it is a Be star). Note the evident emission of Balmer lines. Canon 40D, 9 x 300 s exposure with the 0,2 m telescope. 

The calibrated spectrum of Vega (Canon 40D camera). I use orders from 33 to 38 only. I use a neon lamp for spectral
calibration of this small part of the total acquired spectrum (a thorium lamp is pending...).
The processing pipeline is fully automated (note the very good transition between successive orders). This spectral profile is the composit of orders 33 to 38.

Calibrated spectrum Delta Sco (Canon 40D camera).

Calibrated spectrum Chi Oph (another Be star). Modified Canon 40D - 9 x 300 s. The signal to noise ratio is fundamentally limited by the fixed pattern noise
of the camera (CFA structure, and only one pixel sensitive for 4 available in the red - the processing quality can probably be improved in the future). These documents show that it is possible
to obtain very educational echelle spectra with a relatively modest camera.

Use of a black and white CCD camera : Audine model equipped with a Kodak KAF-0402ME
(also a low cost camera but with a modest sized detector - only a part of the echellogram can be captured in one once)
100 mm focal length collimator and 85 mm focal length camera configuration for a 8000 resolution.

Spica (Alpha Vir), spectral type B1III (0,2 m telescope - 7 x 300 s exposure).

Sirius (Alpha CMa), spectral type A1V (0,2 m telescope, 15 x 60 s).

Procyon (Alpha CMi), spectral type F5IV-V (0,2 m telescope, 4 x 180 s).

Delta Sco (Be Star) - (0,2 m telescope, 5 x 300 s exposure).

48 Lib (Be star) (0,2 m telescope, 4 x 300 s exposure).

Chi Oph (Be star) (0,2 m telescope, 4 x 300 exposure).

Use of a black and white CCD camera : Audine model equipped with a Kodak KAF-1602E
100 mm focal length collimator and 85 mm focal length camera configuration for a 8000 resolution.

Echellogram of the daylight (image reduced by a factor 0.75).

Use of a black and white CCD camera : Audine model equipped with a Kodak KAF-0402ME
100 mm focal length collimator and 50 mm focal length camera configuration for a 6000 resolution.

34 Boo (spectral type M3III) - 0,2 m telescope - 10 x 300 s exposure.

Arcturus (spectral type K1.5III) - 0,2 m telescope - 12 x 60 s exposure.

Vega (spectral type A0V) - 0,2 m telescope - 12 x 60 s exposure.

Delta Sco (spectral type B0.3IV) - 0,2 m telescope - 5 x 300 s exposure.

Beta Lyr - Sheliak (spectral type B7Ve) - 0,2 m telescope - 7 x 300 s exposure.

P Cyg - 34 Cyg (spectral type B2pe)  - 0,2 m telescope - 7 x 300 s exposure.

Nova Cyg 2008 #2 (V2491 Cyg) the April 13.06, 2008 - 0,2 m telescope - 14 x 300 s exposure. See details here.

And some corresponding spectral profile...
(merge of order 33 and 34, i.e. only 1/10 of the total available spectrum is show here !)

 
Typical acquisition session. Right, the guiding section: The cross mark the position of nova Cygni 2008 #2 (at V=8.5) and indicate
precisely the entrance of the optical fiber. Left, raw CCD image of the nova spectrum (5 minutes exposure).

Nominal configuration of the echelle spectrograph : The CCD is a KAF-1600 type (1560 x 1024 pixels, 9 microns)
100 mm focal length collimator and 85 mm focal length objective for a R=8000 resolution power.

Full frame of a KAF-1600 type CCD (scale reduced by a factor 2).

Routine spectrum of Delta Sco taken the April 29.01, 2008 UT with a 0.2 meter telescope (5x300 s exposure).
The order 31 to 38 are used. The spectral calibration is done by using lines of neon, xenon and mercury gazes.
The CCD camera is an Audine KAF-1602E. The continuum is normalised to unity. The H2O telluric lines are removed (Visual Spec processing - see below).

Detail of the Delta Sco spectrum (1/3). Note the presence of He I 5876 A emission. The interstellar Na doublet lines are well resolved.

Detail of the Delta Sco spectrum (2/3). Note the presence of Si II 6347 A and Si II 6371 A emission. C2 emission is also visible near the red edge of Halpha line.

Detail of the Delta Sco spectrum (3/3). The He I 6678 and He I 7065 shows a double peak emission.

Aspect of the Delta Sco spectrum before and after H2O telluric lines removal procedure (automatic division by a synthetic spectrum of vater wapor under VisualSpec).
Discrete features near 590 nm are noticed after H2O cleaning (my observatory is sea level so, H2O lines are very presents).

And some objects taken with the same setup...

Some step of the pipeline processing...

The fundamental images. To, the target spectrum. Middle, the neon lamp spectrum for spectral calibration (only the red part of
the spectrum is here covered, a major actual problem - a thorium lamp is ideal and necessary for a full spectral calibration).
Botton, a tungsten continuum spectrum for the flat-field (halogen lamp).

From the individual order profiles to the reconnected spectrum.

Back