2007-06-05 - CORALIE-007

CORALIE-007

Performances and new functionalities

Prepared by F. Pepe, La Silla 2007-06-05

The CORALIE-Upgrade Team: B.-O. Demory, M. Fleury, M. Jeanneret, C. Maire, D. Mégevand, F. Pepe, M. Pichard, D. Queloz, J.-D. Vuille, L. Weber

1.1 Preamble

In May and June 2007, the CORALIE instrument has been upgraded with the goal of increasing its optical efficiency. Besides instrumental changes, several new functionalities have been introduced. In the following we shall describe them to the CORALIE-007 User.

1.2 Opto-mechanical changes

Coralie suffered from a low optical efficiency. We identified two main components to be responsible for this low efficiency: The optical fiber which feed the instrument and the cross-disperser prism-grism assembly. Following changes have been made to the instrument:

- Exchange of the scrambler fibers (voie directe) with a new set of fibers without scrambler.

- The interfaces on the telescope and on the spectrograph side have been kept identical to the previous configuration, such that this operation does not impact the standard operational mode. However, the functionality to swap between the fibers with and without scrambler has been removed. In practise, the fiber-swap functionality and mechanisms on both the fiber-head (bonnette) and the spectrograph side have been completely removed. A spare fiber (with scrambler) is installed next to the standard fiber, but is NOT connected.

- The oscillation blade at the entrance of the spectrograph (lame oscillatoire), which was (never) used to do spectral flat fields, as well as its SW function, have been completely removed.

- The cross disperser has been replaced by a pure prism. The consequences are that the deviation angle is now of 90 degrees, and that therefore the cryostat containing the CCD is placed vertically on the spectrograph.

- Thanks to the cross-disperser change, the focal filter in front of the CCD within the cryostat, could be removed.

- The location of the flat-field LED has been changed.

- The red-cutting filter in the optical beam has be replaced by a higher-through put and shraper-edge filter.

- The shutter was replaced by a new Uniblitz bi-stable shutter.

- The exposure meter design was changed. A spherical filter with a central hole is now placed in front of the fiber exit and collects the beam outside the scientific F/15 beam onto a fiber leading to the photomultiplier tube. This has several consequences: a) The exposure meter is much more efficient than before. b) The exposure meter is located before the shutter and records therefore the flux entering the spectrograph at ANY moment. c) The flux in the reference fiber (ciel) has been separated and is therefore not recorded anymore by the exposure meter.

1.3 Performances

1.3.1 Spectrograph

1.3.1.1 Spectral format

Due to the change of the cross-disperser, the spectral format has completely changed. The orders are not equally spaced anymore, instead the blue orders are more separated than the red orders. In addition the focal filter on the CCD has been removed. Altogether results in a spectrum ranging 2 to 3 orders longer on the blue side while keeing the red boundary unchanged. No contamination is present anymore on the blue side by the second order of the grism. Ghosts (grism) and rings around the emission lines (parasitics reflections by the focal filter) have disappeared.

In the following image, a raw LOCO frame is shown, with the blue orders on top:

The spectral format of the reference fiber (ciel) is slightly different. Given the smaller inter-order distance in the red, we had to put a filter on the reference fiber, which cuts the light above 550 nm. About 1/3 of the orders and ¼ of the CCD are therefore “cut”:

1.3.1.2 Spectral resolution

Due to the chosen fiber diameter and optics, their effective image size on the CCD has been reduced by about 10%. The effective resolution of CORALIE increased from R = 48’000 to about R = 55’000 to 60’000.

1.3.1.3 Bias, dark and background

The BIAS value remained unchanged at about 1’020 ADU. The DARK has been measured to 1.75 e-/hr/pixel. The background in the inter-order decreased significantly due to the cleaner “prism” spectrum. We shall however not indicate any value here because it depends on the type of spectrum (OBTH, OBJO, etc.) and the SNR.

1.3.1.4 Efficiency

The efficiency of the spectrograph has increased by a factor of higher than 4 in average. The gain in optical through-put is shown in the following plot:

1.3.2 Spectro-photometric response

The expected counts per extracted pixel at 550 nm is computed using the following formula:

I = 16226*10^(mag/-2,5) *t [e^-]

The SNR, for the two read-out modes, is:

SNR_slow = I/sqrt(I + 100)

and

SNR_fast = I/sqrt(I + 25)

1.3.3 Guiding errors

When the star is decentered with respect to the fiber hole (see e.g. on the guiding camera), a systematic RV error might be introduced. Magnitude and sign of this error depend on the de-centering direction and magnitude, and on the seeing conditions.

The maximum error we have recorded is about 10 m/s when de-centered by 4 pixels. It is therefore recommended to ensure that the star is always well centered on the fiber, if necessary by re-centering the fiber manually with the “manuel” button in the GUIF panel.

1.3.4 Exposure meter

1.3.4.1 General

The design of the exposure meter has been changed. Following new properties apply:

- The exposure meter is located BEFORE the spectrograph shutter. The shutter measures therefore the flux entering the fiber continuously. This feature can e.g. be used to do a through-focus with the GUIF panel.

- The exposure meter records now only the flux in the object fiber. In particular in the case of faint objects, the count rate indicated by the exposure meter is the one due to the star and is not contaminated anymore by the simultaneous thorium (OBTH) on the reference fiber (ciel).

- The dark counts are now below 10.

1.3.4.2 Response

The exposure meter shows the the following response:

DARK: 3 cps

LOCO exposures: 5’100’000 cps (non-linear regime)

THA2 exposures: 62’000 cps

Cross-talk on fiber B: 1/2000

Following values shall be taken as reference for average seeing conditions (1-1.5 arcsec) and almost clear sky:

Mv=4.7: 550’000 cps (seeing ~ 1.5 arcsec)

Following formula applies: I [cps] = 42’000’000 cps * 2.514^-mag. We expect however up to 20% higher counts at perfect sky conditions.

For a 14^th magnitude star we expect therefore at least 120 cps. From cross-talk measurements we deduce that on a 15-minutes exposure the sim. thorium counts will be lower than 5 cps.

1.3.4.3 Exposure meter calibration

The curve below shows the relation between the totally achieved exposure-meter count rate and the signal-to-noise ratio (SNR) per extracted pixel achieved on the stellar spectrum. Blue, green and red points correspond approximately to the bands B, V, R in the spectrum.

1.3.4.4 Linearity measurements

The exposure meter is quite linear up to 3 Mcps, which corresponds to about a stellar magnitude m_v=3. Below this magnitude the PMT works still but delivers non-linear results.

1.3.4.5 FITS Headers

Following keywords concerning the exposure meter can be found in the fits-header of the raw images and all sub-products:

HIERARCH ESO CORA PM COUNT = 54283768 / s_pm_count / Nb de coups total PM

HIERARCH ESO CORA PM SAMPLING = 1 / s_pm_sampl / sampling

HIERARCH ESO CORA PM FLUX CTMEAN = 2163343 / s_pm_mean / CDG flux PM

HIERARCH ESO CORA PM FLUX CTMAX = 2.77978375E+06 / s_pm_ctmax / Comptage Max

HIERARCH ESO CORA PM FLUX CTMIN = 1.50290488E+06 / s_pm_ctmin / Comptage Min

HIERARCH ESO CORA PM FLUX RMS = 68842.99218750 / s_pm_rms / RMS signal

HIERARCH ESO CORA PM FLUX TMMEAN = 0.52673799 / s_pm_cg / CDG flux PM

HIERARCH ESO CORA PM FLUX SLOPE = +0.0134 / s_pm_slope / Flux PM pente

COUNT represents the total counts accumulated during the exposure.

SAMPLING is the time sampling unit, usually 1 second.

FLUX CTMEAN is the average counts rate expressed in counts per second [cps]

FLUX CTMAX is the maximum count rate recorded during the exposure.

FLUX CTMIN is the minimum count rate recorded during the exposure.

FLUX RMS is the standard deviation of the count rates during the exposure

FLUX TMMEAN represents the flux-weighted exposure-time center. This value is used to determine the effective time of the exposure (e.g. in case of clouds), which is in turn used to compute the barycentric correction of the Earth motion.

FLUX SLOPE is a first order trend of the count rate during the exposure expressed in relative flux change between the start and the end of the exposure.

1.4 New software functionalities

1.4.1 UIF and exposure-meter panel

The exposure meter has been separated from the UIF panel (left side on screen shot below). Instead, a dedicated exposure meter panel is now available below the logbook. The panel shows the counts of the photomultiplier at any time, i. e. also during shutter-close time. The exposure-open time is indicated by a green strip. The bold red line indicates, for the presently observed object, the expected counts in case of ideal conditions, while the plain red line corresponds to the 50% level.

In addition to these modifications, the hole screen has been re-arranged. In particular we have to mention the introduction of a clock to the top right indicating the UT, and the increase of the exposure time indication panel.

1.4.2 Wait on guiding and LED centering

In order to improve the centering of the star on the fiber, we have introduced two new features is the edp (see screenshot below):

a) “Attente Guidage”: This option allows you to ask the exposure to wait on the guiding. By default, no waiting time is asked (“pas d’attente”). Alternatively you can ask to wait for a fixed predefined time (“predefini selon config”), which is set by default to 30 sec. You can also choose the option (“manuel”) to make the exposure wait on your mouse click. In this case, a panel is shown on the GUIF screen, where you will have the possibility to abort the exposure or start it immediately (see screenshot below). The “attente guidage” option is particularly important in case of short expsosures < 120 s. The former, automatic option is preferred for a continuous and automatic use, e.g. when the User leaves for the mid-night lunch. The second is recommended for a perfectly controlled start.

b) “Centrage LED”: If selected (default), this option will ask the observing software to perform an illumination of the fiber entrance to detect the precise position of the fiber each time the telescope points on a new target. A small window showing the centering image will appear in the “afficheur” of the guiding screen (see screenshot below). The new fiber-center values are computed and send to the guiding automatically. If necessary, the can be verified in the “guidage” terminal. We recommend to select this option for the high-precision radial-velocity mode, although an time overhead of several seconds must be taken into account. Nevertheless, we recommend the User to always check the correct centering of the star on the fiber.

1.4.3 Autofocus

The autofocus option is a functionality (left, top on the edp), which already existed. At each exposure, the guiding procedure computes the best focus (M2 position) as a function of the telescope and external temperatures. This technique works very well and is very reliable. A small offset exists however between the best focus as seen by the camera, and the one corresponding to the smallest image size and the fiber entrance (highest flux in the spectrograpoh). There is therefore the possibility to set a M2-position offset with respect to the computed position. This value is expressed in millimetres of M2Z, and can be set in the “DEFOCES” field to the right to of the edp. The best defocus value has been tested to be +0.05.

The Users has the possibility to try a through focus by pressing “±focus” in the GUIF and measuring the flux on the exposure meter. Each “focus” sends a step of ±0.050 mm to M2. The User has to remember by how much he moved M2 when finding the best focus and correct the DEFOCUS value by the corresponding amount, e.g. if you pressed twice on the +focus button, then DEFOCUS_NEW=DEFOCU + 0.1 .

1.4.4 New calibration set

Given the higher flux, the calibration exposure times had to be changed. When clicking the “insert calib” button following calibrations are inserted:

2 BIAS fast

1 BIAS slow

1 LOCO fast 50 sec

1 LOCC fast 50 sec

3 FFO fast 50 sec

3 FFC fast 50 sec

3 THA2 fast 100 sec

In order to prevent the standard User to change the calibration values, all the parameters are fixed and cannot be changed. If, for an unknown reason, the calibration parameters (e.g. exposure time) have to be changed, this can be done in the “edp” using the button “Fenetres” and editing the default values in “Parametres de configuration”.

1.4.5 DRS – The data-reduction software

The DRS remained unchanged in its functionalities. The only difference will be that now 3 additional orders are available. At present, they are not yet extracted by the DRS, but the DRS will be soon updated to do so.