with a small instrumentation, Digital SLR and some filters

The setup

The situation of the observatory, a country largely polluted by artificial light (Castanet-Tolosan, near the city of Toulouse -South of France). The naked eye limit magnitude is near 2. The telescope used for all the observations of this page is a William Optics Zenithstar 80 refractor (aperture of 80 mm at f/6).


The camera is a modifed Digital SLR Canon EOS 350D (the internal IR cutoff filter is removed, click here for details about the operation).

The 350D is remotely controled by an homemade low cost USB interface, easy to made (QuickRemote interface, description coming soon).

An USB 2.0 hub link the computer, the QuickRemote interface (command of the shutter) and the DSLR USB interface (for download the images).

Only one USB cable come to the computer. IRIS software is used for remote the DSLR.

Use of dedicated filters is a well know and efficient method for observe in urban condition. The key principle of a Light Pollution Removal filter (LPR) is illustrated by this images. Upper image : The spectrum of a high pressure sodium lamp. This lamp emit the characteristics yellow light of your city. The major contributor is the largely saturated sodium doublet lines near 5893 Angstroms (click here for a description of this spectrum). Lower spectrum : The same high pressure sodium lamp is observed through a LPR filter (here, the Deep Sky Filter of Lumicon International). The major part of the sodium emission disappear. In the same time, the important astrophysical windows near Halpha line (6563 A), [OIII] lines (4960 - 5007 A) and Hbeta line (4861 A) are preserved because the added filter is transparent for these wavelengths. The setup used for obtain these spectra is a Grism spectrograph, described below, and the detector is a modified Digital SLR Canon EOS 350D (the internal IR blocking filter is removed). For this two images an external IR-cut filter is added (Baader UV/IR rejection filter).


The LPR filter is inserted in the optical beam for this spectrum of a high pressure lamp, but the infrared rejection filter is removed. An intense emission is visible in the far IR (only detectable with a modified DSLR or a B&W CCD camera, of course). So, the use of an IR-cut filter is a good idea for a maximum rejection of pollution light.




The observation is confortable. Left, a typical result under urban condition - NGC7000 - stack of 13 x 120 seconds exposure @ ISO 400 with a Lumicon Deep Sky Filter and the Baader IR/UV filter (cumulative integration time of 26 minutes).


Measured relative transmission curve of the Lumicon Deep Sky and Astronomik UHC filters. The spectrograph used is LISA (click here for details about this spectrograph).


Measured relative transmission curve of two IR cutoff filter (Baader and Astronomik). LISA spectrograph + Canon 350D as a detector.


Measured transmission curve of the assembled Lumicon Deep Sky + Baader IR/UV set of filters and of the assembled Astronomik UHC + Astronomik IR set of filters. The position of some remarkable spectral lines is indicated : Halpha (dashed red line), [OIII] doublet (dashed green lines) and Hbeta (dashed magenta line). The Astronomik setup is slightly better if the goal is pure emission line imagery (narrow band pass in the green-bleue region and better transmittance in the vicinity of Halpha).


Crop of a 120 seconds exposure (M27 field). No filter user. Unprocessed image (no flat-field correction) but the standard white balance is applied (coefficient for the color layers are R=1.85, G=1, B=0.65). Notice the color of the background, dominated by the artificial light, and also the effect of the doublet refractor lens chromatism for the infrared photons (halo around bright stars).

Same exposure, but a Baader IR/UV cutoff filter is added. The severe infrared chromatism is largely eliminated.

Same exposure, but with a Lumicon pollution rejection added only. The chromatism is present but the colors are more natural. The limit magnitude is affected because the filter stop an appreciable number of photons in the green-yellow region. So, the emission lines of the nebulae are transmitted.

Same exposure, but the filtering consist to the addition of the Lumicon pollution rejection filter (Lumicon Deep Sky) and the Baader IR/UV rejection filter. The stars are more sharper.






Median level of the sky background on a RAW image (CFA) for a 120 s exposure @ ISO 400

No filter

153 ADU (or counts)

IR cutoff (Baader UV/IR)

108 ADU (or counts)

LPR (Lumicon Deep Sky)

37 ADU (or counts)

LPR (Lumicon) + IR cutoff (Baader)

22 ADU (or counts)

Comment: From my observatory, with the LPR+IR cutoff filters added, the sky background is 7 time more darker relative to the unfiltered case, i.e.a differernce of  two magnitudes (a very significant improvement!).

120 seconds exposure. Full frame. No filter used. Unprocessed image (notice the presence of dust and optical vigneting).

120 seconds exposure. Full frame. Filter : LPR + IR cutoff. Unprocessed image (notice the presence of dust and optical vigneting). The contrast of M27 is considerably improved.

120 seconds exposure. Full frame. Filter : LPR + IR cutoff. Processed image (bias, dark and flat-field correction, background equalization an white balance applied - see the roadmap here).




Standard method for acquire bias and dark (thermal signal) reference images - the entrance of the objective lens is obtured.

Standard method for acquire the flat-field (gain sensitivity of each pixels) - halogen lamps and white screen are used.



One exposure of 120 seconds (LPR + IR cutoff filters).

Stack of 8 x 120 seconds exposure (LPR + IR cutoff filters)



Messier 57. Add of 10 x 120 seconds exposure. LPR (Lumicon Deep Sky) + IR cutoff filter (Baader).


"The rainbow sky"
A very low cost solution for spectroscopy

The dispersive element for spectrography mode is simply inserted in the T-mount interface of the DSLR.

The disperser is composed of a Rainbow grating (200 grooves/mm) and a refractive prism (Edmund Optics). This two elements form a low cost Grism ensemble. For details about this type of spectrograph, click here.



Typical aspect of a star spectrum taken with the Grism setup and a modified Canon EOS 350D (click on the image for display this document at its original scale). Here, the Vega spectrum of type AOV. Up, image acquired without filter. Down, image taken with a Baader UV/IR rejection filter (notice the suppression of the glow around the zero order image, i.e. reduction of the refractor chromatism). Of course, the "visible" part of the spectrum is not affected by this filter. The chromatism of the used refractor is not corrected for wavelength shorter then 4500 Angstroms and wavelengh longer then 7000 Angstroms. It is the explanation of the widening of the spectrum (defocus of the image for this extreme wavelength). One shot of 1 second exposure time.


Same spectrum of Vega (with IR rejection filter), but enlarged along the spatial direction for a better visibility of spectral lines. For obtain this effect the AD motor is turned on during exposure. The unfocused Hg line, at the limit of UV wavelength, is detectable.


Example of drift spectra. Up, Antares, a "red" star. Down, Beta Lyrae, a "bleue" star (complex Be type object).


The field of Wolf-Rayet stars WR138 (V=8.1) and WR139 (V=8.3). This two stars are of type WN5. Stack of 6 x 180 seconds exposures. Emission lines of these very luminous stars are clearly visible (click here for spectra of some Wolf-Rayet stars and a catalogue).


Spectrum of the Wolf-Rayet star WR140 (V=6.9). Spectral type WC7 (notice the doppler enlargement of the emission lines). 7 x 180 seconds integration time. Some spectra of this star taken with different spectrographs are here.


Spectral image of the field of the M57 nebulae. Upper image,  part of the full frame (rescaled by a factor 0.5). The zero order image of the ring nebula is at right. Near the center, the emission spectrum of the ring, dominated by the red Halpha and the bleue-green [OIII] lines. Down image, crop at the original scale. Composite of 8 x 180 exposures. Click here for a sample of planetary nebula spectra.


Science is also possible ! Spectrum of the just discovered nova Aql 2005 (V1663 Aql, see IAU circular 8540 and IAU circular 8544) at V=10.8. Date : June 12.92, 2005. Observation condition are not optimal : stack of only 4 x 120 seconds and the IR cutoff filter is not used for darken the bright sky background. But this result is correct if we consider the observatory location (naked eye limit of 2.5 during this run!) and the small size of the instrument (80 mm diameter). The Halpha emission is detected and the very pronounced reddening is also evident (note the lack of bleue photons). The coordinates of the object are AD=19h05.2m, DEC=+514.2. See also the nova page.


A spectrum of V1663 Aquilae taken at the date of June 14.97, 2005 UT with the 80 refractor and a Canon 350D from the Castanet-Tolosan observatory. The integration time is more longer : addition of 24 x 180 second images. An IR cutoff filter is also added, so the wavelength up to 760 nm are blocked. Comparatively to the June 12.9 observation, the Ha emission at 656 nm is considerably more intense. The object enter in its nebular phase. The two faint bleue emission lines are Hb at 486 nm and probably FeII at 493 nm. More subtle details are also detectable. The very red aspect of this peculiar object is confirmed. Spectral dispersion of 7.3 A/pixel.


The Sun
with a spectrograph...


 The spectroheliograh mode of LHIRES3 high resolution spectrograph.
And about spectroheliography...