LISA : de l'influence de la turbulence

[Francois Teyssier]

Bonjour à tous les utilisateurs du LISA.

Nuit très étrange en Normandie, un courant d'air froid descendant plein sud entre deux dépressions. Atmosphère instable à une échelle de temps très courte. Très turbulent

Spectre de alpha leo. Les images sont issues d'une séance d'acquisition de 16 poses, d'un seconde chacune, sur une durée totale d'environ 1 minute.

Voici le résultat sur 5 images prises parmi les 16.

alpleo.png (31.63 KiB) Viewed 15503 times

Quelques spectres très bons tel celui du haut, le reste dégradé à très dégradé.

Sur l'apparition de la "queue de poisson", j'ai également noté que la distance de la CCD est un facteur essentiel.

a+

François

[Martin Dubs]

Hello Francois,

maybe it is my poor french but I do not quite see what the fishtails have to do with the turbulence of the air, which in my humble opinion should widen the whole spectrum (movement of star image alonfg slit. My explanation would rather be a defocusing of the spectrograph (between slit and CCD), maybe by a shaking of the spectrograph or attached CCD in the wind? Or you have defocused the spectrograph permanently and move a defocused star image perpendicular to the slit. By selecting only the center portion of the defocused star image (in a moment of little turbulence) you would increase the appparent f-ratio and reduce as a consequence the fish tails. With turbulence you would average over the full aperture of the telescope and work at design f-ratio. What is the normal appearance with little turbulence and good focusing? I hope you understand what I mean.

Regards, Martin

[Francois Teyssier]

Hi Martin,
I understand what you mean, but there's the whole spectrum

almost well focused on all the spectral range

All the images are taken in only 1 minute duration.

The fish tail aspect appears randomly
There's no significant variation on the mean intensity of the spectrum from an image to another.

Best regards

François

[Christian Buil]

François, cela fait plusieurs fois que tu souligne que la qualité du spectre dépend de la "distance" du CCD.
De quelle distance parle tu au juste ?

Peut être l'épaisseur de la cale support de la caméra ?

Mais dans tous les cas, cela ne devrait par avoir d'mpact. Si cette cale change d'épalsseur on compense
en avance ou reculant l'objectif (visa son filletage) de manière à avoir un spectre net (son iùage arrive
toujours de l'infini à l'entrée de l'objectif).

La distance caméra/objectif reste alors la même (lorsque le spectre est net) et cela
n'affecte pas la qualté globale du spectre (queue de poisson).

Christian

[Robin Leadbeater]

Hi Martin,

Seeing is wavelength dependent (something which is well known to planetary imagers for example where IR images are often sharper despite the theoretical negative effect of wavelength on resolution.) How rapidly it deteriorated into the UV in this case is a surprise to me though.

I did a bit of Googling and came accross this interesting observational anecdote
http://www.astropix.com/wp/2012/02/06/s ... avelength/

Cheers
Robin

[Martin Dubs]

Hi Robin,

it is always a pleasure to argue with knowledgeable specroscopists. But as a physicist I disagree with you. Seeing as I understand it is caused by temperature and pressure fluctuations of the atmosphere. The changes in refractive index are somewhat wavelength dependent (4% larger at 340 nm than at 760 nm) but the difference should hardly be noticeable. This is why adaptive optics works in the IR with a visible laser guide star. So I remain with my opinion that the observed fishtails are created after the slit.
To your observation that seeing is better in the IR in planetary images. In my opinion this is probably more related to the higher contrast in IR images (depending on the object), especially for planetary clouds absorbing in the red to IR (Jupiter, Saturn e.g.).

Best regards,

Martin

[Robin Leadbeater]

Hi Martin,

As a fellow physicist I find no fault in your argument, though of course evidence always trumps theory The evidence from the planetary imagers though, I agree is pretty subjective (using an IR channel seems to be most effective in Mars and Lunar imaging). In this case though we should easily be able to distinguish which side of the slit the blurring occurred by looking at binned spectra. If the blurring was pre slit we should see a drop in the flux in this region but the Balmer lines should remain sharp. For blurring after the slit, the flux should be unchanged but the lines should be less well resolved. Francois - can you compare good and bad binned spectra?

In general from what I have seen, UV fish tails in LISA spectra (assumed to be chromatic aberration) seem to be quite common but the severity appears to vary significantly between users.

Cheers
Robin

[Andrew Smith]

Hi Robin, if the blurring was due to CA I would have expected the lines in the UV to have been more washed out than they appear to be in François’s spectra. I.e. given the width of the fan I would have expected the spectra to be almost completely washed out in this region.

I have a LISA on order, due any day now, I will have a look at the UV response and see if I can get any useful data to help explain the fan.



Andrew (a fellow Physicist and beginning astro spectroscpist)

[Francois Teyssier]

I come back on this subject after a discussion with Bernard Tregon (La Rochelle 2012)

Here's the 24 spectra (1 second exposure, every 4 seconds). The sequence duration is only 1,5 minutes.
One can see the strong defocalisation which appears randomly

According to Bernard Tregon, this could result of a phenomenon of 'focus/defocus' : the wavefront of the turbulence should be (at observation time) of the dimension of the pupil ("pupille" in french, i'm not sure of the translation)
So in that case, turbulence acts as an external lens, whose focal lenght would change with very short time scale.

More, on the following picture, one can see that "fish tail" appears when the spectrum is out focus.

Waiting for your comments
Best regards

François

Bernard Tregon (ENS Ulm) presentation :

http://astro-proam.com/larochelle2012/p ... Tregon.pdf

[Christian Buil]

I have a possible geometric interpretation linked to the slit width and seeing (tilt-tilp + possible piston or natural defocussed star because presence of focal reduced and its associated optical chromatism). Look the attached drawing.

At f/10 (for example !) the LISA aberrations are significantly reduced (an also, the aberration associated to focal reducer - an important point).
About the F/D impact of the spectrum aspect see the page: http://www.astrosurf.com/buil/small_tel ... pectro.htm
and also this ARAS topic (fish effect): viewtopic.php?f=8&t=255

For confirm, try to find a relation between the spectrum intensity (at a given wavelength) and the fish effect.

Christian Buil