New ArNe lamp for LISA

[Olivier GARDE]

je viens de remplacer la lampe Neon d'origine du spectro LISA par la lampe ArNe.
Voici le spectre obtenue avec cette nouvelle lampe et les valeurs des raies.
Un grand plus pour l'étalonnage des spectre surtout dans la partie bleue du spectre qui présente de nombreuses raies de l'Argon.
Certaines raies sont en fait le mélange de 2 raies très proche qui du fait de la faible résolution du LISA n'ent font qu'une.

I have replaced the original Neon lamp of the spectro LISA by a new one (Argon Neon lamp).
Here the spectrum obtained with this new lamp and values ​​of the emission lines.
A great upgrade for this spectrograph to have more lines in the blue part of this spectrum (to produce more accurate spectra).
Some lines are actually mixing together (because they are very close each others for this low resolution spectrograph)

[David Boyd]

I also have a new argon-neon lamp for my LISA. I have found that using the following 16 lines, I get a good fit with rms ~0.1.
7383.98
7272.94
7067.22
6965.43
6506.53
6402.25
6143.06
5944.83
5852.49
5495.87
5187.75
4965.12
4879.90
4764.87
4545.08
3948.98
I have created a .lst file with these lines which I have provided to ISIS in the File Mode of Spectral calibration (General tab) but I am not clear which is the best procedure to follow in carrying through the full spectral analysis from this point. Clearly the Calibration assistant no longer applies and I am not sure whether I should be using the Response assistant on the Calibration tab (which seems to automatically switch ISIS back into Predefined dispersion equation mode) or the Response command on the Profile tab to correct for instrument response.

Advice from Christian on the best procedure to follow would be much appreciated.

Thanks,
David

[Robin Leadbeater]

Hi David,

Is the file mode wavelength calibration working ok with your calibration file with ISIS finding the lines automatically given the position of one identified line ?( I also had to generate one of these for my ALPY 200 )

I have not used the assistants so normally calculate the instrument response by first generating a wavelength calibrated (but not response corrected) profile of the reference using the General, Calibration and Go pages. I then display this on the Profile page and using the "response" and "continuum" tabs to generate the instrument response. The file name of this is then entered in the General page and the full reduction of the target run from there using General, Calibration, Go

Cheers
Robin

[Robin Leadbeater]

The procedure I use does not allow a wavelength calibration combining both lamp and Balmer lines though. I cannot see an easy way to do this without using the assistant. It would be nice if the user could set this up in ISIS for any given dispersion and combination of lines.

Robin

[Christian Buil]

David,

The spectral assistant does not work with the file mode (the response assistant yes of course). Some info here:

http://www.astrosurf.com/buil/isis/tuto ... ion_us.htm

The ISIS algorithm is very simple. First, ISIS find approximative X coordinates position of the lines (in pixels) by using dispersion law coefficient (a) from the file and constant (b) from the Calibration tab : X = a (wavelength) + b

It is a first guest. Next ISIS search in the range {X-dx, X + dx} the exact X position for each wavelength (max pixel intensity pixel and next, a gaussian fit of the line position).

Finally, ISIS fit a polynomial function lambda = A X'^2 + B X' + C (for example), with X' the exact (fractional ) X position of the lines.

I suggest order 2 or 3 for the spectral law fit (max).

The principal problem concern the initial linear fit. No real problem if the LISA spectrograph (or Lhires of course) is concerned, the base dispersion is correctly linear. The situation is more complex with Aply600 because the dispersive element is a grism. For the Alpy600, the predefined mode use is a 2nd order polynomial function for line search (internally coded).

Your line list seem correct and a 0.5 A RMS is normal (note the low contrast of some lines). I recommand acquisition in bin 1x1 for the best result.

Here a typical line list (for Alpy 600 is line mode) :

3
2.4548
3948.98
4158.59
4199.45
4545.05
4657.90
4764.87
4965.12
5400.56
5495.87
5852.49
5944.83
6143.06
6266.49
6334.43
6506.53
6678.28
7067.22

Christian

[David Boyd]

Thank you Robin and Christian for your replies and for the web reference. I had looked for information on the .lst file format but had not found that reference.

I take 60 sec exposures of the lamp spectrum binned 1x1. So far the rms values from wavelength calibration using file mode with the new lamp are ~0.1.

The procedure Robin described is the one I have used successfully except that I used the Filter rather than Continuum functions to smooth the initial instrument response. Which one is better to use?

I'm not clear why you would want to use both lamp and Balmer lines once you have lamp lines covering the full spectrum range?

I have noticed that ISIS now calculates a considerably lower value for the spectral resolution (<900 rather than ~1100) using the file mode method of calibration with the supplied list of lines than previously using neon+Balmer lines. However the FWHM of the lines in the spectrum look the same as before. What is the reason for this?

David

[Robin Leadbeater]

Hi David,

Christian reports that With the ALPY, using Balmer lines with the NeAr lamp gives a slightly better fit at the UV end beyond the last reliable NeAr line. See bottom of this page.
http://www.astrosurf.com/buil/isis/guid ... uto_en.htm
I have not tried this myself though.
The improvemernt might not be significant with the LISA though if it is more linear.

I like using the continuum function to smooth the instrument response as it has a feature to crop out bad sections eg telluric bands or where the lines have not exactly cancelled (double click on either side of each bad area) It fits a spline function rather than using a gaussian filter and can even chose the fit automatically if you are particularly lazy

This has prompted me to try the response assistant which sounds like it could be even easier. I do find the process of generating a good instrument response manually very satisfying though.

Robin

Robin