[ ESO ] La Silla


La Silla Science Operations: NTT


Commissioning report and
notes for the observer

First light: M8, or the Lagoon Nebula in Sagitarius. This is a false color V (blue), R (green), I (red) composite, 60s per filter. (full size). (pic Vanzi and the commissioning team).

In April 2002, the good old TEK CCD of the Red arm of EMMI has been replaced by a 2x1 mosaic of new generation MIT 2kx4k chips.

This page lists the characteristics of EMMI with that new chip almost in real time as we get them. This information will then drift into the main pages and into the EMMI manual. In the meantime, THIS is the reference.

This document wll also be eventually converted into a formal commissioning and -hopefully- acceptance report.

A quick reference guide is available for the inpatient and/or experienced EMMI users



This page lists the info on the new CCD almost in real time as we get it. Be sure to re-check it every time you deal with EMMI, until the new manuals are on-line.

  • May 16: the overscans were adjusted to take the gap geometry properly into account
  • May 15: very high wind. We did not open.
  • May 14: on-sky commissioning, night 2. Thick fog and high wind.Weather improved shortly: acquisition templates were commissioned, and many throughputs were measured.
  • May 13: on-sky commissioning, night 1. It is raining.
  • May 7-12: off-line commissioning
  • April 2002: TEK removed, MIT installed

Table of Content





Purpose and scope

This document lists the characteristics of the new EMMI Red detector, as measured during the commissioning period of 2002-May. It is a "live" document, wich is often updated as more information becomes available. It will eventually be formalized into a Commissionning Report, and the information will be passed to the other web pages and instrument manual.

The commissioning team was constitued by D.Sluse, C.Foellmi, P.LeSaux, M.Pizarro, A.Sanchez, M.Castillo, J.Miranda, D.Castex, K.Aubel, L.Vanzi from the NTT Team, P.Sinclaire and A.Pizarro from Detector Team, B.Lopez, J.C.Guzman, J.Ibsen for Software, choregraphy by O.Hainaut.


The following documents are used for the commissioning or referenced in this document.

CCD, bias and dark

Hardware and Geometry

Figure: CCD orientation in the various modes (also in PS with a better resolution)


The new EMMI Red detector is a 2x1 mosaic of 2kx4k MIT/LL, thinned CCDs, known as CCID20-14-5-3, nicknamed "Michele" and "Zeus". The new detector system is basically a copy of the FORS2 upgrade. The 15micron pixel (0.1665", cf below) match well the resolution of the EMMI-Red optics (0.21", [6]). The mosaic is controlled by FIERA which, together with the ATM connection between the CCD controller and the instrument workstation, ensure a very small overhead.

The two chips are mounted side by side. Each chip uses two independent amplifiers. The resulting frame therefore contains 4 independent sub images (which are stored as extensions in the FITS frame, cf FITS section below). The geometry and name of the subframes are listed in the table below. Refer to the Astrometry section for the accurate relative positioning of the chips.

Read-Out Modes

The mosaic is offered exclusively in the following modes

Table : value of the bias, gain (e/ADU) and RON (e).

Dark current

Dark current has been measured on a composite 3h exposure taken in "Slow" mode. Average dark current is 0.3e-/h/pix, with some peaks at 1.5e-/h/pix, and few hot pixels above 5e-/h/pix (ref: com. 2002-05-18: CcdDark).


The 4 sub-CCDs have been tested for linearity in the 6 read-out modes. The linearity is better than 0.5% over the full dynamical range (tests by Detector Team, confirmed by routine tests by NTT Team). More information will be available on-line on the detector page, as soon as the HTML interface is configured for the new set-up.

Read-Out Times

The table below lists

Table: Read-out times and total software overhead (s)

Mode1x1 2x2
Slow 43.9 / 48 11.7 / 18
Normal 20.1 / 24 5.7 / 18
Fast 9.9 / 18 3.4 / 18

Quantum Efficiency

The figure below displays the quantum efficiency of a MIT/LL ("Michelle") similar to those installed in EMMI-Red, together with that of the former Tek CCD (from [5]).
Figure: CCD Quantum efficiencies (ps)



Most of the information is summarized graphically in the figure below.

(ref: comm. 2002-05-15 astrometryGap) These values correspond to a rotation of 0.22 mrad. As this corresponds to a physical rotation of the CCD, it cannot be adjusted.

Table: Position of the center of rotation (and optical axis) and flexures

Figure: Field of View and orientation.
This image corresponds to a full 4x4k flat field. The grey area is the illuminated area.

Dome Flatfields

The following information was obtained using the 1x1 binning mode, with "Normal" read-out mode, slit width = 1". The exposure time given below are computed for a given expected peak level (EXPLEVEL) of 20 000 ADU. Halogen lamps were used. The Plot col. links postscript plots of cuts through these FF (H and V for horizontal and vertical cut respectively, taken at ~the center of the image), showing that the they are extremely good and flat over the available field of view.

Table: Imaging Dome FF conditions


Photometry and Throughputs

The following photometric solution was obtained on one single standard field with 4 stars, taken in poor seeing conditions; it obviously needs to be refined (ref: Comm.2002-05-11 stdPhot).

Table: Photometric solution. The left column refers to the standard magnitudes, the right-hand one to the instrumental magnitudes (obtained from the electron counts). The extinctions are the standard La Silla ones, Z is the airmass.

Table: Count Rates, for a Mag=15 star, in electron/s, at airmass = 1

Low Resolution Spectroscopy (RILD)

For the orientation of the spectra, refer to the figure earlier on this page.

Wavelength ranges, pixel size, etc.

Table: Wavelength ranges and resolution for RILD Grisms.

* Measured at 587.6 nm except grism #6 at 706 nm

With the new geometry of the CCD, Blue is down (small Y), Red is up (large Y).

Note that because of the larger size of the new CCD, the wavelength coverage is much larger than what it used to be. As a consequence, many grism now present a second order contamination. Use the table below for order separating filter selection.


Table: Order separating filters. (full list of filters available elsewhere)


Dome Flat-Fields

The following information was obtained using the 1x1 binning mode, with "Normal" read-out mode, slit width = 1". The resulting exposure time is automatically computed by the REMD_spec_cal_Flats template, for a given expected peak level (EXPLEVEL) in ADU. The CCD saturates above 60 000 ADU. The plots of the FF are roughly calibrated in wavelentgh. A example of horizontal cut in one of the FF can be found here.

Expected Peak Level (ADU)
Exposure Time (s)
40 000
ps gif
40 000
ps gif
40 000
ps gif
40 000
ps gif
40 000
ps gif
40 000
ps gif
40 000
ps gif



Wavelength Calibrations

Exposure times

The table below lists the updated exposure times for the new CCD. Note that they are similar to those used with the previous chip.

Wavelength Atlas

To be used for line identification and order selection filters. The plot present the full extension of the CCD (i.e. the 4000 pixels)

Table: Wavelength Atlas

gif ps
gif ps
gif ps
gif ps
gif ps
gif ps
gif ps



The table below lists the throughputs of the Grisms, in electron/sec/nm for a 12th magnitude AO star at airmass 1. A more detailed version of this table is available as an ASCII file.

Table: Grism Throughputs, in electron/sec/nm, mag=12 at airm=1
Wavelength[nm] Gr1 Gr1 Gr2 Gr2 Gr3 Gr3 Gr4 Gr4 Gr5 Gr5 Gr6 Gr6 Gr7
360 1 0 - - - - 0
400 0 53 29 - - - 0
440 1 131 149 - - - 0
480 0 164 181 1 - - 0
520 40 180 178 1 - - 0
560 180 188 184 2 - - 27
600 247 196 187 59 - - 111
640 286 200 179 119 - - 192
680 300 181 151 147 - - 238
720 310 158 125 157 - - 251
760 278 130 104 157 - - 237
800 241 125 97 151 - - 216
840 209 125 82 143 - - 193
880 212 113 74 146 - - 176
920 241 104 - 133 - - 179
960 225 79 - 87 - - 125
1000 174 - - 42 - - 78

Medium Resolution Spectroscopy (REMD-spec)

For the orientation of the spectra, refer to the figure earlier on this page.

Wavelength ranges, pixel sizes, resolutions, etc

Wavelength Atlas

The HeAr spectra corresponding at the "standard central wavelentghts" (defined in our calibration plan) are listd below. Of course, for scientific observations, any central wavelength can be selected.

Table: REMD-spec Wavelength Atlas

Exposure Time (s)
He + Ar
Central Wavelentgh (NM)
15 + 300
0 + 300
1 + 300
6 + 15
6 + 10
0 + 3*
0 + 7*
0 + 5*
0 + 5*
10 + 150
1 + 60
9 + 2*
0 + 4
ps gif
12 + 60
8 + 2*
3 + 2

*Note: with Filter OG530#645


Dome Flat-Fields

The following information was obtained using the 1x1 binning mode, with "Normal" read-out mode, slit width = 1". The resulting exposure time is automatically computed by the REMD_spec_cal_Flats template, for a given expected peak level (EXPLEVEL) in ADU. The CCD saturates above 60 000 ADU.



Not available yet (still raining...).

High Resolution Spectroscopy (REMD-ech)

This mode is described in more details on the REMD-ech page.


For the orientation of the spectra, refer to the figure earlier on this page.

Dome Flat-Fields

The following information was obtained using the 1x1 binning mode, with "Normal" read-out mode. The resulting exposure time is automatically computed by the REMD_ech_cal_Flats template, for a given expected peak level (EXPLEVEL) in ADU. We provide in the table below an image of the flat, and the corresponding FITS file as well, to allow users to use it.


Wavelength Calibrations

The following table list the recommended exposure times for a 2x2 bin CCD and links the "finding charts" for the orders and reference lines. It also links the FITS files countaining the flatfield and ThAr lamps. Potential observers are suggested to download these frames to experiment with the effect of the inter-chip gap on their data reduction software. Feedback is most welcome. Note that the gzipped FITS frames are big.

Table: Echelle wavelength calibrations

Cross Dispersor Grism
Exp.Time [sec],
Th lamp alone, Ar is included
not available yet
not available yet




FITS file, hints and tips


Observing Templates

The current Instrument Package (EMMI-IP-20020212) is still valid. However, it must be noted that

Templates wer fully commissioned. A new internal IP was released. It is not made a public release, as the current official one is compatible with the new detector.

DIMD mode is not offered yet. MOS templates have been commissioned "on dome". XMos package is being modified.

Final word: "Dans le sirop de poire, les souris coulent". Don't ask.




QE as table in [2]: verified in the lab: [OK]
Chip geometry [2] distance to flange and planeity measured in lab: OK. Gap and rotation: OK. [OK]
Binning modes required: 1x1 and 2x2: all certified, except 1x1/625kps, which did present twice during the commissioning an abrupt drop of bias on 3 chanels (BCD). [OK]/[PENDING]
Read-out modes required: "Slow" at ~100kps and "Fast" at ~300kps Modes provided are better than specs (100, 225, 625kps) [OK]
RON<4e at 100kps 3.5e at 100kps, ~8 at 225, i.e. within spec. However, a small pick up noise is detected, and will be eliminated. [OK]
Windowing in one chip Considering the read-out times, hardware windowing requirement was dropped, and replaced by software windowing, pending FIERA sofware upgrade for software windowing. [OK]/[PENDING]
Dark current < 2e/h < 1e-/h/pix average, few pixels ~5e-/h/pix [OK]
Linearity deviation < 0.5% up to 120000e- <0.5% in all modes up to 65kadu [OK]
Persistency: saturation leaves < 20e remnant in 30min exp no remnance detected [OK]
Cosmetics: <5 bad columns, no major defect 2 bad cols, very few bad pixels, no defect at all [OK]
Observing templates updated tested OK, ISF updated, backward compatible with pre-upgrade instrument package. New internal instrument package released [OK]
Set-up templates updated tested OK. [OK]
Initial calibration plan: Imaging astrometry and photometry performed. In agreement with expectations [OK]
Initial calibration plan: Spectro Wavelength atlas and flat field in all modes acquired. Spectroscopic throughput in 5/7 grisms, 0/4 grating, 2/10 echelle combinations in agreement with expectation. Others pending [OK]/[PENDING]
xMOS package update data acquired and processed. Implementation pending, scheduled for May 24. [PENDING]


[Observing] [ESO La Silla]  [The SciOps Home Page]  [ESO]  [Index]  [Search]  [Help]  [News]