My new home built spectroscope
Posted: Wed Apr 09, 2014 3:48 pm
François Teyssier suggested that I post some details about my new home made spectrograph.
For my first instrument I wanted to cover the whole visible range from 400 to 750 nm with the best resolution and throughput I could get. My design started with the slit. I purchased a 23 um photometric slit for the ALPY spectrograph. My telescope is a 311 mm f9.8 cassegrain so for good slit efficiency I decided on an f6.6 reducer which gives a slit width of 2.3 arcsec. The seeing at my location is often better than this. I wanted to use an 85 mm f1.8 Nikon lens that I had purchased on Ebay for the camera and my detector is a QSI583 that I have been using for imaging. I also wanted to use a mirror collimator in order to reduce chromatic aberrations. After playing with the numbers this led to the choice of a 25 mm square 600 g/mm grating and a 1 inch dia. 101.6 mm focal length spherical mirror for the collimator, both from Edmund Optics. At f6.6 the spherical aberration of this small collimator is well below its theoretical diffraction limit.
I don’t have very much room between the arms of my telescopes fork mounting so I needed a very compact spectrograph. My final arrangement uses a perforated folding mirror just below the slit and a 38 degree collimator/camera angle. The pupil of the telescope images approximately at the position of the folding mirror and the shadow from the cassegrain secondary is larger than the hole so no light is lost. With this design I am restricted to about 1.5 mm total slit width and it is critical that it is exactly on the optical axis of the telescope.
Modern enhanced Al coatings are actually pretty robust so for the folding mirror I purchased a plane mirror from Edmund Optics, mounted it face down with pitch on another glass plate, and drilled it at the required angle with a small diamond tool. It came out perfectly with no damage to the coating. I machined and anodized all the parts myself and am very pleased with the outcome. It is very rigid.
The estimated final design resolution came out at 1550 or 3.55 Å at 550 nm. At the detector I have 1.05 Å/pixel so I am a bit oversampled at 1x2 binning. For faint objects 2x2 binning is just slightly under sampled. Actual bench tests measuring line widths of the Relco calibration source came out significantly better than this at 2.8 Å FWHM over much of the range. Until recently I was a bit confused by this since ISIS was reporting R = 1400, more in line with the original designed goal. I have just recently realized that ISIS is using all the lines in my calibration line list to evaluate resolution and several of these lines were badly saturated. After removing those lines from the list ISIS is now reporting an R of 1860! The Nikon lens must indeed be very good! The resolution drops off to FWHM’s of 5.5 at 415 nm and 4.0 at 735 nm. Overall I am very pleased. Here is a spectrum of NGC6210 to demonstrate.
Tim
For my first instrument I wanted to cover the whole visible range from 400 to 750 nm with the best resolution and throughput I could get. My design started with the slit. I purchased a 23 um photometric slit for the ALPY spectrograph. My telescope is a 311 mm f9.8 cassegrain so for good slit efficiency I decided on an f6.6 reducer which gives a slit width of 2.3 arcsec. The seeing at my location is often better than this. I wanted to use an 85 mm f1.8 Nikon lens that I had purchased on Ebay for the camera and my detector is a QSI583 that I have been using for imaging. I also wanted to use a mirror collimator in order to reduce chromatic aberrations. After playing with the numbers this led to the choice of a 25 mm square 600 g/mm grating and a 1 inch dia. 101.6 mm focal length spherical mirror for the collimator, both from Edmund Optics. At f6.6 the spherical aberration of this small collimator is well below its theoretical diffraction limit.
I don’t have very much room between the arms of my telescopes fork mounting so I needed a very compact spectrograph. My final arrangement uses a perforated folding mirror just below the slit and a 38 degree collimator/camera angle. The pupil of the telescope images approximately at the position of the folding mirror and the shadow from the cassegrain secondary is larger than the hole so no light is lost. With this design I am restricted to about 1.5 mm total slit width and it is critical that it is exactly on the optical axis of the telescope.
Modern enhanced Al coatings are actually pretty robust so for the folding mirror I purchased a plane mirror from Edmund Optics, mounted it face down with pitch on another glass plate, and drilled it at the required angle with a small diamond tool. It came out perfectly with no damage to the coating. I machined and anodized all the parts myself and am very pleased with the outcome. It is very rigid.
The estimated final design resolution came out at 1550 or 3.55 Å at 550 nm. At the detector I have 1.05 Å/pixel so I am a bit oversampled at 1x2 binning. For faint objects 2x2 binning is just slightly under sampled. Actual bench tests measuring line widths of the Relco calibration source came out significantly better than this at 2.8 Å FWHM over much of the range. Until recently I was a bit confused by this since ISIS was reporting R = 1400, more in line with the original designed goal. I have just recently realized that ISIS is using all the lines in my calibration line list to evaluate resolution and several of these lines were badly saturated. After removing those lines from the list ISIS is now reporting an R of 1860! The Nikon lens must indeed be very good! The resolution drops off to FWHM’s of 5.5 at 415 nm and 4.0 at 735 nm. Overall I am very pleased. Here is a spectrum of NGC6210 to demonstrate.
Tim
