ISIS Echelle - calibration

[Andrew Smith]

hi, I have been trying to validate my use if ISIS for processing my echelle spectra. I attach a screen shot which shows order 33. The problem I have is that the calibration solution is very good but measuring the peak at 6929.467A show an error of about - 2A error compared to the oc_33 listing of -0.0027.

Am I doing something wrong?

The calibration is from a Relco lamp rather than ThAr.

Thanks for any advice.

Regards Andrew

[Andrew Smith]

I think I have solved my problem. I have adjusted the gamma setting to make the "Camera Test" circles match the spacing of the lines I picked on order 34. Now the RMS fit values are even better and the error in the line in profile 33 gone. The only underlying concern that the value is lower than that which I measured.

Regards Andrew

[Andrew Smith]

As a physicist I have followed the rule - "if all else fails read the instructions" and have now adjusted the alpha angle as advised to do in the eShel instructions. Sorry for being such an idiot!

Regards Andrew

[Martin Dubs]

Hello Andrew,

I follow with interest your progress in echelle calibration. To get an agreement of the spectrum with calibration lines, it is in my opinion more convenient to adjust alpha and maybe the focal length instead of gamma. The latter has only a small influence as dispersion varies with cos(gamma), almost constant for small gamma. In addition you have to check that you assign the correct order, otherwise you cannot fit more than a single order. Notice that for the central wavelength the following relation holds: order * lambda(order) = const.
A while ago a recorded a spectrum of a starter lamp (ST111) and calibrated it with an FeAr spectrum. This may be helpful in identifying lines. Notice that different gas mixtures are used in starter lamps, they are not intended as spectral calibration lamps. For calibration, select only single lines, well separated from nearby lines. Most useful for adjusting the echelle calibration is check.fit. Try fitting a few orders first, extending the range when successful. Notice that the starter lamp has few lines in some orders, it is definitely an advantage to use a FeAr HCL or similar (I got mine on eBay, with a homemade power supply. FeNe is more readily available with the disadvantage that the neon lines are very intense).

Regards, Martin

[Andrew Smith]

Thanks for your continued interest and support Martin. I think I now have got the calibration correct for orders 32 to 52. I used Richard Walker's "Glow Starter Relco SC480 - Atlas of Emission Lines" to guide me. From not being able to reliably identify any orders I now see the patterns as old friends! The visualisation of the spectral strips in his atlas were of particular value along with the list of wavelengths. Indeed he sent me the line list in excel format so I did not make any errors in transcribing them.

I now have at least 6 lines in each order and RMS residuals better than 0.02A (mostly much better) and checking a sample of lines shows no errors such as I posted initially in this thread. I also have a nice monotonic set of dispersion across the orders from 198A/pix to 121A/pix as I go from order 32 to 52.

I do have NeFe HCL and home built power supply. I will use this to validate the calibration, I just need to construct a housing for it. I will also see if I can get a ArFe HCL for a good price.

Once I have completed the NeFe validation I think I will be good to go on the stars. I still have to build a stable platform/enclosure for the echelle but I will get some stellar spectra first with the prototype so I can practice doing the instrumental response.

Thanks again for your help and advice and to Christian for ISIS and his help.

Regards Andrew

[Christian Buil]

Andrew, your success on this task (not so easy) is a very good new (ISIS eShel tab is more universal than expected)!
Fine use of your spectrograph and ISIS,

Christian