Optical Fibre scrambler/Brasseur de fibre optique

[Olivier GARDE]

Trouvé à la BU de l'OHP sur le rapport annuel 2010 du Nordic Optical Telescope par Thierry G. , un prototype de brassage de fibre en LEGO !!!
http://www.not.iac.es/news/reports/annu ... nRep10.pdf
à voir page 27

Found in the library of the OHP, in the annual report 2010 of the Nordic Optical Telescope by Thierry G., a prototype fiber scrambler in LEGO !!!
http://www.not.iac.es/news/reports/annu ... nRep10.pdf
see page 27

Thank you Thierry for this information !!!

[Christian Buil]

Ca c'est assez extraordinaire et j'adore. Je pensais être dans l'original, mais là... Bonne pioche !

Dans la photo jointe le brouilleur testé jusqu'alors pour ceux qui se demandent de quoi cela peut-il
servir (on agite en permanence la fibre pour casser des effets de tavelures et de non homogénéité
dans le sortie de fibre). La photo montre le dispositif au T60 du Pic du Midi où on était juste avant
le stage spectro.

Mais dans mon cas, cela ne s'avère pas suffisant. Pas assez d'amplitude de mouvement. Je teste en ce
moment une nouvelle version avec la fibre qui navigue accroché à des ventilateurs oscillant à 15-20 Euros.
C'est le mouvement de va et vient qui secoue ma fibre. Alors coté délire, hein, qui c'est qui gagne (et
là j'ose même pas poster une photo) ??!!!!

Christian

[Christian Buil]

Une manière de fortune pour secouer plus sérieusement la fibre. Il faut cela et la bonne nouvelle : c'est efficace.
J'ai retiré les pales des ventilleurs pour travailler dans le silence. Un spectre de Vega avant et après, avec une réduction très
forte du bruit structurel dans le continuum. Quelques vues aussi de l'instrallation avec le spectro sous la table,
bien protégé.

Pas sur que le concepteur du mécanisme du ventillateur ai songer qu'il aller servir à détecter des exoplanètes. J'aime
bien ce type de détournement de matériel, c'est assez vertigineux

My new fiber shaker : large amplitude, impressive and the most important, efficient... see the spectrum of Vega before scrambling et scrambling with oscillation fan mecanism

Also some view of my actual VHIRES-MO setup.

Christian

[Robin Leadbeater]

Wow! that is a lot of noise without the scrambler. Do you see this problem with the eShel too or only with the the VHIRES?

Robin

[Christian Buil]

Yes, this problem can exist with eShel (but with less amplitude because spectral resolution) and in all high resolution fiber fed spectrograph.
The amplitude is very dependant of the fiber, telescope aperture, .... Check your relative spatial noise on a flat-field spectrum (for a first quick diagnostic) with and wihout scrambling (shake by hand for test).

(spatial noise = spectral noise more precisely = here along the wavelength axis)

Christian

[Miguel Rodriguez]

As I could read, this is the same case as in the CARMENES spectrograph, under construction at Calar Alto Observatory, in Spain. It will incorporate an scrambler system based on sound waves, wich is necessary for the instrument to achieve detection of exoplanet signatures in the infrared spectrum of red dwarf stars.

[Christian Buil]

I have used a similar solution during my test: a 50 rotation per second small motor attached to the fiber with a deported center of mass on the rotational axis (a very efficient shaker !). Can solve near field non-uniformity (to be confirmed), by not efficient for my speckles structure (wavelength dependant). The actual only simple solution for the moment is large amplitude scrambler.

(Miguel, you have a reference about the CARMENES sound waves system ?)

Christian

[Martin Dubs]

From my own experience I can confirm the usefulness of the fiber shaking. I noticed it when recording flats, the S/N increased from 250 to 350, obtained by measuring the noise resulting from the division of two flat spectra (dividing out fringes and blaze function not related to the fiber). For star spectra the difference was not noticeable at a S/N of about 100 (weaker targets and poor grating efficiency. The moving of the fiber was manual, a slow motion (about 1 Hz).
In case I would need it on a regular basis, I would build a double pendulum attached to a rotating excenter, which gives a chaotic motion, favorable for suppression of mode noise.

Another thing besides fiber noise is the effect of fiber illumination on radial velocity measurements. A good description of the effect is given here: http://arxiv.org/pdf/1001.0794
While the fiber scrambles well the near field, the far field (as described by angle of incidence) is much less scrambled by the fiber. Changes in f-ratio of the illumination (e.g. caused by defocusing the telescope on the fiber) have a measurable effect on the line position. The cause is coma aberration of the spectrograph objective. In my setup this resulted in systematic errors of up to 1km/sec! with a not very good fiber, for a Canon 85mm f1.8 lens, the same as used in the Eshel . I found out about it by comparing the FeAr calibration with a terrestrial H2O line calibration around H-alpha. In order to reduce this effect it is highly recommended for precise radial velocity mesurements to first pay attention how the signal and calibration light is fed into the fiber and second use an optical fiber scrambler, which changes the optical near field and far field.
A description of a fiber scrambler as used with ELODIE is given by Baranne: http://adsabs.harvard.edu/abs/1999ASPC..185....1B

It is important to note that this second effect (far field illumination of the fiber) cannot be removed by shaking the fiber! Of course the introduction of additional optics introduces additional light loss, which decreases S/N. It has to be tested experimentally if the added precision outweighs the loss in S/N.

A challenge for achieving a 10m/sec precision!

Martin

[Christian Buil]

10m/s is a real challenge, I agree Martin!

Concerning the far field, for VHIRES, my hope is the high optical quality of the system (FSQ85ED refractor, excellent WFE of the grating). It is virtually free of optical aberration (the image of the fiber is very sharp for all orders and all wavelength - diffracted limited). Installation is almost symmetrical and I think the effects of field angle must have small effects in RV...

(the most critical point is a possible vigneting at the 90 deg. injection prism level, into the refractor optics, see photo. The size of this prism is a bit limited. I have limited distance between fiber output and prism surface for reduce this risk, but... The injection angle of the prism may also be critical with respect to the optical axis of the refractor. Actually the FSQ85ED pupil seem appear uniformly illuminated).

Systematic error in RV measure is not a problem. The problem is the stability of this error...

Christian