Pour une version en franšais de ce texte, cliquer ici.

This page reports the status of a Littrow type spectrographs, the first results obtained with different configuration designed and describes a processing procedure for spectra calibration.

Tables of Contents

Calculation of the throughput
First results
Calculation of the signal-to-noise ratio
Partial conclusion
Preprocessing (manual method)
Preprocessing (automatic method)
Importation of 1-D spectral image
Spectral calibration
Radial velocity correction
Normalization process
Radiometric calibration
Remove the telluric lines
Flux Calibration
Remove the continuum
Importance of the slit in the case of surface objects 
Acquisition and processing procedures 


The object of this page is the description of the Littrow type spectrographs. This type is relatively simple to build, which constitutes a good means of demystifying spectrography, then to show all the interest of it. By associating a spectrograph to a CCD camera the amateurs have in hand the possibility of obtaining scientifically exploitable spectral data (see for example the project stars Be on this site). Even with a telescope of  100-mm diameter only it is perfectly possible to usefully study a great number of stars. The force of the amateurs, like very often, comes from their numbers and their capacity to observe objects over a long period of time, which is not always possible in the professional world. Thus a collaboration amateur/professionnel takes all its interest, in spite of the relative smallness of the instruments of amateurs. But, the experiment shows that the valid spectral data acquisition requires more than ever a certain rigor and a good methodology on behalf of the observer, but there is nothing insurmountable !


Several prototypes of Littrow spectrograph were built (Robert Delmas, Christian Buil, Thierry Maciaszek). The spectral resolution go from 1000 to 10000 following the model. One of these spectrograph is used since 1999 on the programme of monitoring of Be-stars. The most recent model is also more the resolvant (R=10000, click here for details).

The principal parameter to be defined when designing spectrographe is the spectral resolution. This one must be selected according to the type of observations and stars concerned, because there is not universal spectrograph which allows to do everything. For example, which says a high resolution says observations which this confines with relatively brilliant stars to compensate for the strong dilution of the light in a very spread out spectrum.

A low resolution version (R=dl/l= 1000 approximately) gives access to the spectrum of objects of weak glares. The applicability is overcast then the spectrophotometric study of variable stars, the evolution of the novae or certain supernovas, the analysis of the spectrum of the most brilliant planetary satellites , the study of comets...

The high spectral resolution is equivalent to increasing a particular area of the spectrum so for example analyzing the temporal evolution of the profile of a line (case of the survay of the Be stars with the observation of the line H-alpha). It is this type of high resolution spectrograph which makes it possible to obtain results most immediately usable within the framework of a joint program amateurs/professionnels. But the low resolution should not be neglected for as much because it is it only with the telescopes of amateurs who give access objects such as weak active variables, novae, comets...

Practically (and at first approximation), the distinction between a low resolution and high resolution spectrograph is given by the level of the number of grooves per millimeter of the diffractive grating used and by the focal distance of the objective/collimator section. This last parameter determines also the size and the mass of the spectrograph and thus, its capacity of carrying on a type of given telescope. To fix the orders of idea, a prototype version of the low resolution spectrograph  includes a photographic objective of 35 mm focal-length (Nikon in F/2) and a grating of 600 lines/mm (origin Edmund Scientific). The version used on the star Be program (magnitude limit of 8 with a 8-inch telescope and R=7000) includes an objective of 180 mm focal distance (Nikon in F/2.8) and a gratings of 1200 grooves/mm. The Littrow spectrograph in R=10000 uses an achromatic doublet lens of 300 mm focal-length and a grating of 1200 traits/mm (limiting magnitude of 6 with a telescope 200 mm in diameter). There is however a higher limit for the value of the focal length, imposed by the dimension of the grating (and its price) and also by the mass. Beyond some weight the spectrograph must be deposited with the foot of the telescope and it is necessary to bring the light through optical fiber.

The figure hereafter shows the spectrograph with R=7000 (dispersion of the spectrum of 0,38 A/pixel - assembly out of wooden) in place with the Newton hearth of a telescope CN-212 Takahashi.

The optical configuration adopted for the spectrograph is of the Littrow type because it optimizes the size and the cost (the same lens objective is used as collimator and camber objective and this saves room...). The Littrow mode is characterized by nearly equal incident and diffractive angle on the grating (click here for more information about spectrograph setup).

With small telescope spectrograph can be used with a large input slit (width of a few millimeters). The traditional function of the slit in a spectrograph is here fulfilled by the star itself, which is viewed as a point at the focal plane of the telescope. This configuration, known as slitless spectrograph, avoids having to very precisely position the image of star at the center, which saves plenty of time when pointing and makes it possible to use the spectrograph with moderate stable equatorial mountings. In addition with a wide slit the photometric efficiency is maximalized (with a fine slit of a few tens of micron part of incidental flux is blocked at the entry of the spectrograph, and this is the same in a case of a spectrograph with optical fibre). In the opposite, the use of a wide slit requires additional work during the calibration stage and limit the accuracy of this calibration.( reminder, the spectral calibration consists in associating a wavelength to a pixel number).

Diagram in 3D showing the various components.

The open box of the high resolution version (R=7000 version) on a Takahashi CN-212 telescope. The photographic objective of the spectrograph is a 180-mm Nikon model. The grating is an Edmond Scientific of 1200 lines/mm.


Another prototype (R=5500 version) on a Takahashi FSQ-106 refractror (notice the presence of a comparator for precise focusing). The photographic objective is here a 135-mm focal length Olympus model.
Click here for R=10000 setup.

Resolution test of a moderate resolution version of the Littrow spectrograph (dispersion of 0,515 A/pixel - R=5500) on a small portion of the spectrum of the star Zeta Cepheus (spectral type K1) in the region of the Sodium doublet (D1-D2). The wavelength of these lines is 5889.97 and 5895.94 A. Doubled, drawn aside 6 A approximately, is perfectly separate. Small line is also visible inside, which indicates the good resolving power of the spectrograph.

Goto index page                                                           Next page