Thinking about an Echelle spetrograph

[andrew s]

Well with the Lhires III optimised and ready to go when the skies clear and night falls I need something else to do the 90% of the time that is left. So I am designing an Echelle spectrograph optimised for my telescope and seeing conditions.

Thoughts so far are:

Givens

Scope D 300mm, F5.4, f 1600mm

Seeing 3-4 arc sec

CCD Kodak 8300

Echelle

R = 18,000 to 20,000

Spectra range 400 – 800nm

Fixed circular aperture

Cross disperser – prism, reflection or transmission grating – still to decide)

Scope mounted

Carbon fibre housing and space frame for temperature and mechanical stability.

If anyone has and views or comments I would be happy to hear them.

I have read Daniel Schroeder’s Astronomical Optics and Design Considerations for Astronomical Echelle Spectrographs but would welcome any other references.

Regards Andrew

[Merlin66]

The guys over on vds

Spektroskopie :: Index

Are at the cutting edge of amateur echelle design - they already have a few designs very well documented - well worth a visit.

Fibre optic link seems to be the prefered method

[Merlin66]

BTW what software have you considered for the analysis of the spectral matrix??

[FraserClarke]

Hi Andrew,

A nice challenge

I'd consider the fibre-fed option for stability. Doing high-res spectrographs that move around is tough (though not impossible of course). It depends how much stability you want. What is you goal with this spectrograph??

The CCD (5.4um pixels?) should be a good match to the slit size, though a bit challenging for the camera perhaps? At f=1600mm, your slit is 23 microns across (for 3"). So getting that down onto a pair of CCD pixels (to Nyquist sample the spectrum) you need a collimator/camera magnification of 23/10.8=2.1. You've got f/5.4 input beam, so you will need a f/camera = ( 5.4 / 2.1) = f/2.5. That probably goes up to f/2 by the time you include the dispersion of the gratings. So it might be easiest to bin the CCD 2x2 and relax the constraints on the camera.

Quite often you want to make the slit smaller than the seeing. This makes the instrument throughput lower (you throw light away at the slit) -- but makes you much less sensitive to varitations in seeing and/or guiding which would otherwise affect your wavelength stability. Again, depends what you want out of the instrument; max throughput, or max stability?

If you're not aware of it already, the Newport grating catalog is very useful. There is a very extensive list of echelle gratings that they can replicate *relatively* cheaply (i.e. ~£500 for a 50x50mm grating, off the top of my head). Gives you a lot more freedom in the design parameters than the stock catalogs from Edmund etc;

Newport Corporation | Richardson Gratings - Products: T3

[andrew s]

Thanks all for the input – I will follow up all ideas and comments. On a few of your points. I don’t have the engineering skills or kit to do a good job on a fibre fed system and I am put off by the light loss. I have been looking at rigidly coupling the spectroscope to the telescope mount (1” square cross section steel tube – home made) with a moving secondary lens to focus. As it has no moving parts it should not be any worse that the Lhires III.

I am looking to be able to do with the echelle what I can do with the Lhires III (Be stars etc.) but capture a number of lines and features simultaneously. I would also like to push to a lower magnitude limit if possible.

Other than that I like astronomy as a hobby as I can read up about the astrophysics (I have a PhD in Physics – long, long time ago), do it in the field and play with the kit. I am now trying to keep my scope up and running and build new kit on the side, otherwise it’s always in bits when the stars come out to play!

No problem binning 2x2 if that is best or I could use a different camera if needs be but the size is attractive for capturing the whole spectral range.

On software that is an issue! I will probably start but cropping out areas of interest and using Iris / Vspec (I have a TrAr lamp for calibration or will have soon as it’s a work in progress) other than that I will write a simple reduction process in Visual Basic.

Thanks again, Andrew