La Silla Observatory High Accuracy Radial Velocity Planet Searcher
Contents
1. 8TZYH 18 S s seg o Bulsas Gl 21 S002 IUSIN ouatonyja Figure 9 1 Appendix A Description of archived HARPS data A 1 Data naming rules The raw frames are stored in FITS format by the DFS with the ESO VLT standard naming rules HARPS YYYY MM DDTHH MM SS SSS fits with YYYY MM DD and HH MM SS SSS being respec tively the date and time of the start of the observation Raw frames are written in extended fits format each CCD being on a different plane of the frame This is effective as of January 1st 2004 Pipeline products are stored in FITS format with the same generic names plus an additional suffix describing its format see next section for details and the specific fibre name A or B For example HARPS YYYY MM DDTHH MM SS SSS_E2DS_A fits is an E2DS format image of the fibre A product by the DRS derived from the HARPS YYYY MM DDTHH MM SS SSS fits raw frame Tables in ASCII format are also produced by the DRS The relevant log books of the DRS is named DRS2. eXtended exposure Time Calculator HARPS User Manual Issue 1 3 3P6 MAN ESO 90100 0005 Page of viii 49 49
3. ESO 90100 0005 5 3 3 Calibration templates Several calibration templates are available However all calibrations necessary for a proper data reduction with the online pipeline are performed by the ready to run calibration OB named RV Page of viii 49 Standard Calibration The details of this OB are described in section 3 1 2 Only if additional calibrations are deemed necessary one needs to use one or more of the following templates e ech cal bias for taking bias frames e ech cal dark for taking dark frames e ech cal thoAB for taking a wavelength calibration through both fibres e ech cal tun for taking order location frames through fibres A and B e ech cal tunAB for taking spectral flat field exposures simultaneously through fibres A and B e ech cal tunAI2for taking spectral flat field exposures through the iodine cell using fibre A e ech cal tunUSER for taking user defined tungsten exposures If the number of exposures is set different from one in the ech_cal_tunAI2 ech cal tun AB or ech cal thoAB templates the pipeline will wait for the last exposure sum all the exposures and then process the resulting frame A concise description of the observation template is given in the HARPS Template Guide 5 4 Overheads 5 4 1 Execution time overheads telescope preset incl dome rotation Fibre automatic redefinition and centering of object on the fibre start of guiding instrument configuration readout ti
4. The EGGS mode gains a factor of 1 75 in flux with respect to the base HARPS mode with a seeing of 0 8 arcsec see fig The best RV accuracy reachable with the EGGS mode is of 30m s due both to the different injection mechanism and the absence of the image scrambler The ghost contamination is higher in EGGS than in HAM In particular the reflected order which crosses the detector perpendicularly to the main dispersion direction has an intensity in EGGS ranging from 10 to 1 of the flux of the regular echelle orders the higher value being in the blue side The same reflected order in HAM has an intensity generally below 1 of the flux of the regular echelle orders The diffuse light at 590nm is about 2 5 for EGGS to be compared to 1 of HAM In table 9 1 the HAM and EGGS mode performances are compared The EGGS mode may prove particularly useful to RV programs studying faint objects where the RV accuracy is strongly limited by the photon noise Users willing to use this mode should declare it in the proposal as this requires a change in the instrument configuration HAM EGGS flux seeing 0 8 1 1 75 best RV accuracy lt 1m s 30m s diffuse light at 590nm 1 2 5 strongest ghost intensity lt 1 1 10 Table 9 1 Comparison of the HAM and EGGS performances 39 HARPS User Manual Issue 1 3 3P6 MAN ESO 90100 0005 Page of viii 49 40 v ujSuajosem 0004 0059 0009 00SS 0005 00S
5. Their quantum efficiencies are given in figures 2 5 and 2 6 respectively The read out mode of 50kpix s has shown to be more noisy than expected when measured at the telescope This mode has therefore been decommissioned and a new high gain low noise mode with a readout speed of 104kpix s has been made available The characteristics of the two modes available for scientific purposes are spelled out in the following table Property Jasmin red 78 Linda blue 77 Read out noise 104kpix s e7 3 0 2 8 Read out noise 416kpix s e 4 5 4 8 Gain 104kpix s ADU e 2 04 2 04 Gain 416kpix s ADU e 0 74 0 76 Bias 104kpix s ADU 198 196 Bias 416kpix s ADU 186 248 The read out time 4296x4096 pixels is of 87s and 23s for the 104kpix s and the 416kpix s modes respectively The 104kpix s read out mode or slow mode is particularly useful when observing faint objects 2 7 Exposure meter The spectrograph possesses an exposure meter which serves to measure the stellar flux and to accu ratly measure the mid time of the exposure flux weighted mean of the time The mean time of the exposure delivered by the exposure meter is not used yet to correct the RV value This exposure meter consists of two photomultipliers one for each of the two fibres entering the spectrograph from the HCFA which use the light picked up at the gap between the two sub gratings 10 Page of viii 49 HARPS User Manual Issue 1 3 3P6 MAN ESO 90
6. in the Night field the only one which is white pressing the Enter key validates the entered value 4 wait until all the sizes are calculated it takes a minute or two 5 check the check boxes for the files to be backed up 6 click the Go button to start copying the files may take up to 30 minutes 7 de select all the boxes 8 press the Compute size button to see the total size of the files 9 select the box Prepare DVD 10 press Go to start creating the image may take 15 minutes 11 de select the Prepare DVD box select the Make DVD one 12 press Go to start burning the DVD may take 45 minutes If the total data size is too big for one DVD the raw files will be automatically compressed by about 50 approximately 170 gzipped FITS files fit on one DVD If the volume is still too large the user has to split the night in two e g lt date gt 1 and lt date gt 2 directories To archive both directories the complete name of the directory has to be put in the Night field Both direc tories have to be archived separately The DAU will automatically search data raw lt dir name gt data reduced lt dir name gt etc The Compute size button computes the total size of the data saved in the BACKUP under the lt date gt directory once the data are copied in the warch disk 45 46 Page of viii 49 HARPS User Manual Issue 1 3 3P6 MAN ESO 90100 0005 Appendi
7. the ThAr2 lamp will be used as the reference The ON time of the ThArl lamp is minimized in order to keep it as an absolute reference along the years Each exposure is used to build a wavelength solution The instrumental drift with respect to the previous calibration frames is measured expressed in m s If accepted by the built in quality control the wavelength solution is stored in the local calibration data base and used for the subsequent reduction of the scientific exposures of the following night Template HARPS_ech_cal_thoAB The user may then repeat a sequence of flat fields with more than 5 exposures if a SNR higher than 300 is aimed at in later science exposures The RV Standard Calibration made of 5 exposures and reaches a Signal to Noise Ratio SNR of about SNR 400 at 450nm 500 at 550nm and SNR 900 at 650nm In case the RV Standard Calibration is not taken the DRS will use the youngest available cal ibration data This might introduce offsets and possibly have a negative effect on the achievable precision The pipeline performs quality checks on each frame In case one of the frames does not pass the quality check the youngest available calibration data will be used In this case is however advisable to contact the support astronomer on site in order to make sure the general health of the instrument is not compromised HARPS User Manual Issue 1 3 3P6 MAN ESO 90100 0005 Page of viii 49 15 3 1 3 Observing temp
8. whadrs YYYY MM DD It is stored with all the other logs in the msg directory It is automatically archived by the DAU on the DVD A 2 Data formats A 2 1 Raw frames The raw frame corresponds to a 4296 x 4096 integer 35 242 560 bytes matrix written on disk in extended FITS format see Fig A 1 each CCD being on a different plane of the frame This image includes a 4096 x 4096 sensitive zone plus 4 overscan zones of 50 pixels each The following generic descriptors are used by the DRS MJD OBS Modified Julian Day start float EXPTIME Total integration time s float DATE OBS Date and Time of observation string RA RA of the target float DEC DEC of the target float EQUINOX Equinox of observation float The DRS needs as well the following HIERARCH ESO descriptors 41 42 Page of viii 49 HARPS User Manual Issue 1 3 3P6 MAN ESO 90100 0005 DET1 READ SPEED CCD Readout mode speed port and gain string DET OUTi GAIN Conversion from electrons to ADUs of port i float DET OUTi RON Readout noise of port i e float DET WIN1 DIT1 Actual sub integration time s float DET WIN1 DKTM Dark current time s float DET DPR CATG Observation category string DET DPR TYPE Observation type string INS DET1 TMMEAN Normalized mean exposure time on fibre A float INS DET2 TMMEAN Normalized mean exposure time on fibre B float INS OPTI5 NAME Lamp name on fibre A string INS OPTI6 NAME Lamp name on fibre
9. 1 The output of the pipeline is archived together with the raw data in data reduced night For a description of the reduction performed by the pipeline in the different observing modes see sec tions 3 1 4 3 2 4 and 3 3 4 Pipeline execution times are given in section 5 4 1 For each frame that processes the pipeline performs basic quality checks When a problem is en countered an error message is displayed to warn the user Calibration frames are also processed by the pipeline If they pass the quality check then the cali bration database is updated Otherwise an error message will be displayed and the latest entry in the calibration database will be used The manual of the DRS can be found online http www 1s eso org lasilla sciops 3p6 harps manuals html 2 9 System efficiency The overall efficiency of the HARPS system over the complete wavelength range is given in table 2 4 In this table the atmosphere telescope transmission is standard atmospheric transmission plus alu minum reflectivity The slit efficiency indicates an average value corresponding to z 1 and is calculated from the HARPS 1 fibre together with the average La Silla seeing of 0 9 The instrument HARPS User Manual Issue 1 3 3P6 MAN ESO 90100 0005 Page of viii 49 11 efficiency includes the fibre link scrambler and spectrograph collimator in triple pass echelle cross disperser grism and camera Instrument and CCD efficiencies are measured in the l
10. 100 0005 Stellar magnitude total count rate error on mean time worst case PA cps of exposure emrE sec RV error m sec saturated 2 850 000 456 000 12 000 120 19 8 Table 2 3 Expected count rate of the exposure meter as a function of stellar magnitude and estimated errors associated with the error on the photometric mean time of the exposure Count rates are indicative depend heavily on atmospheric conditions and slightly on stellar spectral type Calculated of the echelle mosaic no light is lost due to this design The flux in both photomultipliers can be read at the instrument console It is also recorded in the FITS header cumulative average and center of gravity The expected count rates as a function of stellar magnitude and the estimated errors in RV are given in table 2 3 The number of dark counts per second fluctuates between 10 and 15 2 8 Data reduction software A sophisticated Data Reduction Software DRS is an integral part of the HARPS system It allows the complete reduction of all spectra obtained in all three observing modes in near real time in about 30 seconds Once the triggger software is started the start of the pipeline reduction is automatically trig gered as soon as a new raw data file appears on the data reduction workstation whaldrs in the data raw night directory where night is the name of the subdirectory named after the day on which the observing night started e g 2004 02 1
11. 5e 03 0 00415 B 5e 03 pS RASEI LOCKE 3 8 pen Lu W er D7 na 0 0012 38 K iert Lu A egn 437 28 0 0012 8 7 53e 03 0 0039 B 1e 03 ai ma un 8 7 7456508 0 00378 2 7er03 5 31 12 0 0012 B 6 6 66e 03 0 00365 G ert 425 11 0 0012 B S 6 92e 03 0 0035 2 5e 03 2 25 13 0 0012 B 4 6 72e 03 0 00333 2 4e 03 1 Sa un 32 em Lu Peres 19 31 0 0011 B 1 6 220 03 0 00298 B 3e 03 78 416 46 0 0011 8 6 23e 03 0 00262 2 3e 03 75 413 65 0 0011 2 9 5 82e 03 0 00268 2 1e 03 75 am kumm 2 8 Site 0 00254 ideia 405 13 0 0011 2 6 5 17e 03 0 0024 Lien 05 43 0 0011 2 5 4 62e 03 0 00225 Lie 67 02 76 0 0011 2 4 4 41e 03 0 0021 1 6e 03 e 00 13 kumm 2 2 er 0 00196 Liege 897 53 0 0011 2 1 B 02e 03 0 00181 L Le 03 54 894 97 0 0011 2 2 26e 03 0 00166 Siet 46 392 44 0 0011 L 2 120 03 0 00152 7 76 02 45 389 94 0 0011 17 Ser um Sien 387 47 0 0011 1 5 1 66e 03 0 00127 Se D 385 03 0 001 1 4 Liest 0 00115 4e 02 87 852 63 0 001 1 3 927 0 00104 B 4e 02 25 380 25 0 001 1 2 1 37e 03 0 000938 Be 02 85 Figure 2 3 Table showing the spectral format of HARPS and the expected number of electrons for a 1 minute exposure on a G2V star of magnitude 6 A seeing of 0 8 airmass 1 and new moon are the values of the selected parameters In the table the spectral bin is defined as one pixel This table is obtained by the HARPS Exposure Time Calculat
12. 90100 0005 5 Preparing The Observations 21 Dal Cini trOGUCHON ou te A Ee E E Bae Sede Sees ae Eee 21 5 2 Introducing Observing Block 21 BB PRP EE EE EE a WO ede ans 21 5 3 1 Acquisition templates e 22 5 3 2 Observing template e 22 5 3 3 Calibration templates e 23 BA Overheads Ls ee ee nem BE BA ee Bee DO R eee ee ee Ee 23 5 4 1 Execution time overheads aoaaa 23 5 4 2 Off line overheads e 24 5 4 3 Fast time series observations asteroseismology e 24 5 4 4 Iodine cell out of service since May 2004 24 5 5 The HARPS Exposure Time and Spectral Format Calculator 25 6 Observing with HARPS 27 6 1 Before th night ia LA A a Dee iere A e A E 27 62 During the nights ns A tt A AA os ee 27 6 2 1 Target acquisition guiding focusing e 27 6 2 2 Pointing restrictions Por e e ELE ee 28 6 3 Night calibrations cs ise AA AA RE E a he a A 29 6 4 R al time display seua passat ai naa a e DEE Be A a 29 6 5 Observing very faint Stars ee 29 6 6 Ateroseismology lt 2 2 044 mesma e ee VE REA A RA a dae ee 29 6 7 End of the night sussa gui ban EEE E EVA CE AR ao e E EA CE IA A a Va 31 7 Data products and archiving 33 TA Data products a A o Phe A e Mg 33 T2 Datacarchiving eso Manos tots A A a Oe eh ed e 0 34 7 2 1 La Silla and Garching archive 34 7 2 2 La Silla data archiving unt 34 7 2 3 Use of archived HARPS data 2000000022 ee 34 8 The Reductio
13. AN ESO 90100 0005 3 6m telescope fibre adapter science fibres 2 calibration fibres 2 Calibration unit HARPS spectrograph Figure 2 4 The main components of the HARPS system Table 5 1 illustrates the fiber illumination scheme as a function of the template used 2 5 Fibre adapter All optical fibres are connected to the HCFA which forms the interface to the telescope The HCFA provides several functions 1 Illumination of the object and the reference fibres each can be separately fed by the object the sky light from a calibration source or it can be dark 2 Correction of atmospheric dispersion by means of an ADC 3 Switching between HARPS and CES from end 2004 on fibres 4 Feeding of the fibre viewer technical CCD camera for guiding 5 Introduction of the Iodine cell into the object light path 6 Attenuating the reference light beam from the ThAr lamp via a neutral density wheel to an equivalent exposure time of 40s at zero density This can be done for exposure times from 40s to 5400s 2 6 Detector and read out electronics The detector is a mosaic of two 2kx4k EEV CCDs It is mounted in a ESO detector head and cooled to 148K by means of an ESO continuous flow cryostat CFC The detector is controlled by the standard ESO CCD controller FIERA The detector head is mounted on the optical bench while the CFC is fixed on the outer wall of t
14. B string OBS NAME OB Name string OBS START OB Start Date and time string TPL ID Template signature ID string OBJECT TYPE What is on fibre A and B string OBJECT SP Object spectral type string OBJECT RV Object expected RV string TEL TARG RADVEL Object expected RV string 1 the use of this keyword made by the DRS is wrong the keyword shall be later replaced by OBJECT TYPE not yet implemented by the DICB 2 shall replace some of the DPR TYPE current function not implemented yet in the DICB 3 not implemented yet but needed for optimum RV computation 4 this keyword shall be later replaced by OBJECT RV HARPS User Manual Issue 1 3 3P6 MAN ESO 90100 0005 Page of viii 49 x 4296 y 4096 x 1 y 4096 x 4296 y 1 RED A 4296 raws PRINTER RESO BLUE A gt PRINS 4096 columns b Ly 1 Figure A 1 Raw frame format 43 44 Page of viii 49 HARPS User Manual Issue 1 3 3P6 MAN ESO 90100 0005 Appendix B Use of the Data Archiving Unit This appendix describes the use of the Data Archiving Unit DAU This unit allows the VA to write his her data raw and reduced to DVD The disks have to be DVD R General In order to archive the data of one night the following has to be done 1 launch the Data Archiving Unit on the wharch machine as harps user the command is dau csh 2 put a writable DVD DVD R General in the burner 3 indicate the night date
15. European Southern Observatory Organisation Europ enne pour des Recherches Astronomiques dans H misph re Austral Europ ische Organisation fiir astronomische Forschung in der s dlichen Hemisphare 00 La Silla La Silla Observatory High Accuracy Radial Velocity Planet Searcher HARPS User Manual 3P6 MAN ESO 90100 0005 Issue 1 3 Date May 25th 2006 La Silla Observatory ESO x Chile ii Page of viii 49 February 2003 March 5 2003 March 10 2003 June 16 2003 July 8 2003 July 31 2003 Sept 17 2003 Nov 25 2003 May 25 2004 August 2005 May 2006 Change Record sections affected all all all all all all DRS related all all Observing HARPS characteristics amp EGGS Editor G Lo Curto ESO La Silla HARPS User Manual Issue 1 3 3P6 MAN ESO 90100 0005 Reason Remarks First version for comments first public issue for P72 CfP GRU DQU DRS end of Comm2 DQU DAU info DSo data format answer to comments DQU more information added and structure revised GLO full update GLO overheads and asteroseismology issues GLO include slow readout mode THAR3 lamp and EGGS GLO Contents 1 Introduction VELO SCOP A ere Pee See OS Bote RL aoe OS BaP AI 1 2 Additional information 1 3 Contact informati n scort a e e Alen 1 4 Acknowledgments ooo ee 2 HARPS Characteristics 2 1 Instrument OVer vie Wani ee e ede Oe Se AA a A a GB Be 2 2 O
16. HARPS with the simultaneous Thorium reference method has been demonstrated during the three commissioning phases to be below 1 m s The RV accuracy can be affected by several factors external to the instrument e photon noise e telescope focus e centering errors e Thorium calibration errors As an example during the second HARPS commissioning June 2003 a 7 hours series of short exposures on o Cen D was recorded yielding a RV rms of 0 52 m s Of these 0 45 m s are due to the stellar oscillation 0 17 m s to photon noise 0 08 m s to Thorium calibration errors drift tracking and the remaining 0 18 m s to centering errors telescope focus errors and any other error source not yet identified The following systematic study is from data obtained with the simultaneous Thorium reference method only 4 1 Photon noise For a G2 star a RV rmsphoton 1 m s due to photon noise only is reached with a S N ratio of about 100 per pixel at 550nm The photon noise introduced in the RV measurement scales approximately 19 20 Page of viii 49 HARPS User Manual Issue 1 3 3P6 MAN ESO 90100 0005 as 100 4 1 S N550nm rms m s photon xX 4 2 Telescope focus A defocus of the telescope of 30 encoder units introduces a RV offset of 1 m s In a typical focus sequence which can take from 5 to 10 minutes is reached a precision in the determination of the optimum focus of 5 10 encoder units 4 3 Centering errors A de c
17. PS is made available to the users EGGS Usage of this mode has to be declared in the phase 1 proposal preparation as it involves installation of the fibres in the HCFA In this case the CES fibre will be removed and the EGGS fibres installed instrument or mode changes during the night will be allowed depending on which 6 Page of viii 49 HARPS User Manual Issue 1 3 3P6 MAN ESO 90100 0005 SE Pb me E 527 73 0 001945 L 5e 04 0 0068 5 4e 03 1 2e 02 523 22 0 0014 4 4 1 476 04 0 00664 E 3e 03 Liest 515 75 0 0014 4 4 1 34e 04 0 00649 Wie L est 514 43 0 0014 4 4 1 39 04 0 00631 Beat 1 20 02 510 14 0 0014 4 3 1 4e 04 0 00613 Siet 1 2e 02 505 92 0 0014 4 3 1 41e 04 0 00595 5 1e 03 1 2e 02 501 75 0 0014 4 2 1 29e 04 0 00579 4 7e 03 L Le 02 97 70 0 0014 2 1 32e 04 0 00564 4 8e 03 L 1e 02 493 69 0 0013 41 1 31e 04 0 0055 Wie L 1e 02 89 74 0 0013 AI 1 31e 04 0 00536 4 7e 03 Liest 85 85 0 0013 4 1 09e 04 0 00521 je Iert 2 03 0 0013 4 1 25e 04 0 00505 M e 1e 02 475 26 0 0013 3 9 1 23e 04 0 00492 4 5e 03 1 1e 02 474 56 0 0013 4 1 21e 04 0 00489 W ert Liest 470 91 0 0013 4 1 14e 04 0 00486 Viet Liest 67 31 0 0013 8 1 14e 04 0 0045 4 Le 03 Lie 1163 77 0 0013 4 1 16e 04 0 0047 Wie 1 1e 02 60 29 0 0013 4 1 12e 04 0 00458 4 1e 03 Liest 056 85 0 0012 3 9 1 126 04 0 00446 Wiert Liest 653 47 0 0012 B 9 P Je 03 0 00435 B 6e 03 po 450 14 0 0012 B S 1 09e 04 0 00424 West jet men 0 0012 nz D 7
18. aboratory Wavelength 380 400 450 500 550 600 650 600 nm Table 2 4 Overall efficiency of the HARPS system See text Page of viii 49 HARPS User Manual Issue 1 3 3P6 MAN ESO 90100 0005 QE efficiency Hi a v e l e n g t h Figure 2 5 Quantum efficiency of Jasmin the red CCD Quantum efficiency i HERE 1100 wende nm Figure 2 6 Quantum efficiency of Linda the blue CCD Chapter 3 Observing modes HARPS as it is implemented now offers three observing modes 1 Simultaneous Thorium Reference observation 2 lodine self calibration observation 3 Classical fibre spectroscopy with and without sky 3 1 Simultaneous Thorium reference method The Simultaneous Thorium Reference mode is the base line observation mode to get the best short term accuracy in radial velocity determination from the instrument In this mode fibre B is fed by the Thorium lamp located in the calibration unit see figure 2 4 while fibre A is on the stellar target A variable neutral density ND filter is used to keep the Thorium spectrum at a flux level equivalent to a 40 seconds exposure with zero density Since the density to which the ND filter is set is computed by the instrument software from the exposure time as defined in the template the exposure time should not be modified from within BOB Broker for Observing Blocks Otherwise the flux level
19. alculation can be turned off by entering 99999 The RV must be provided with an accuracy better than 1 2 km s in order to reach the expected performance For more details about the pipeline Data Reduction Software DRS please refer to the DRS user manual For pipeline execution times see section 5 4 1 3 2 Iodine self calibration method out of service since May 2004 HARPS offers the possibility to use an Iodine cell as an alternative to the standard Simultaneous Thorium Reference method In this mode fibre A is on the target fibre B on DARK and the Iodine cell inserted in front of the fibre A entrance to superimpose an Iodine absorption spectrum on the stellar spectrum 3 2 1 Performance The Iodine cell used in HARPS absorbs about 40 of the continuum from the source Exposure times supplied by the ETC have to be scaled accordingly The precision of HARPS using the Iodine self calibration method is still under investigation 3 2 2 Calibration A calibration sequence similar to the Simultaneous Thorium Reference method is recommended before the beginning of the night Moreover the 5 tungsten exposures series should be repeated 16 Page of viii 49 HARPS User Manual Issue 1 3 3P6 MAN ESO 90100 0005 with and without the Iodine cell However a specific sequence of observations is additionally needed during the night to later extract the Iodine information This sequence includes the observation of a star of spectral type B with
20. alibration exposures Bias Tungsten Tho rium because they are needed for the pipeline to produce optimum results A calibration OB RV Standard Calibration is available at La Silla and ready for execution to take these exposures De tails are described in section 3 1 2 More calibration exposures are necessary during the night when using the Iodine self calibration method see section 3 2 2 6 2 During the night Observations are performed in the standard VLT way i e OBs are selected by the Visiting As tronomer VA with P2PP and fetched by the Telescope and Instrument Operator TIO into BOB 6 2 1 Target acquisition guiding focusing Target acquisition is done by the TIO The object is centered on the entrance of the science fibre and kept there by an automatic dynamic centering algorithm with an accuracy of better than 0 2 pixels 0 05 This guiding accuracy introduces radial velocity errors of the order of 20 cm s and is therefore negligible within the accuracy attainable with HARPS Users with crowded fields close binaries faint objects etc should prepare finding charts The guide camera can guide on stars of magnitude up to 17 In the fibre AB spectroscopy mode object sky the observer should verify that the sky fibre is not contaminated by light from other sky objects This should in the first place be done by checking on the Digital Sky Survey At the telescope it can be verified by e offsetting the telescope th
21. and without the Iodine cell whenever a template of the target star is produced The target star template is produced by observing it without the iodine cell with a high signal to noise ratio 3 2 3 Templates The necessary acquisition and observing templates are available e HARPS_ech_acq 12cel1 acquisition and setup for Iodine cell exposures e HARPS ech obs all for Iodine cell exposures For a detailed description of the templates see section 5 3 and the HARPS Template Guide 3 2 4 Pipeline data reduction In the Iodine self calibration method the pipeline does spectrum extraction and applies the wave length calibration If TARG_RV is not set to 99999 a radial velocity is computed using the CCF technique Conversely RV calculation can be turned off by entering 99999 Considering that 12 lines pollute the spectra the radial velocity should be considered as an approximate value For more details about the pipeline Data Reduction Software DRS please refer to the DRS user manual 3 3 Classical fibre spectroscopy Classical fibre spectroscopy can be done in two different ways depending on the target and the goal of the program 1 fibre A on target and DARK on fibre B objA observation 2 fibre A on target and fibre B on the sky obj AB observation objA observation should be preferred for objects much brighter than the sky moon background where a careful CCD background correction may be needed For this type of obse
22. anual Issue 1 3 3P6 MAN ESO 90100 0005 1 high SNR template spectrum of the science object without T cell 2 high SNR spectrum of a B star taken through and without the I gt cell during the same night as the template spectrum of the science object 3 Fourier transform spectrum of the Iodine cell For the HARPS Iodine cell this FTS is available from the HARPS web site http www 1s eso org lasilla sciops 3p6 harps The extraction of the 12 information itself is left to the observer The description of one method to model the 15 data is given in the paper The planet search program at the ESO Coud Echelle Spectrometer I Data modeling technique and radial velocity precision tests Endl M Kiirster M Els S 2000 A amp A 362 585 also available from the web http aa springer de bibs 0362002 2300585 small htm Chapter 9 HARPS high efficiency mode EGGS Since October 2006 the high efficiency mode of HARPS dubbed EGGS has been made available to the users This new mode uses a set of fibres with a projected aperture on the sky of 1 4 arcsec and a diameter of 100um while the standard HARPS fibres used for optimum radial velocity accuracy have an aperture of 1 arcsec and a diameter of 70um The light injection mechanism is via image injection for the High Accuracy Mode HAM and pupil injection for the high efficiency mode Moreover in order to minimize light losses the EGGS mode does not use the image scrambler
23. dware part of the system HARPS User Manual Issue 1 3 3P6 MAN ESO 90100 0005 Page of viii 49 5 Halogen lamp 3000 K Wavelength nm Diffused stray light 90 680 116 527 160 383 G3V star 5700 K Wavelength nm Diffused stray light 90 680 116 527 160 383 Table 2 1 Level of diffuse stray light and ghosts as a percent of the flux in the order Figure 2 2 Spectral format of HARPS Blue orders are down red are up is schematically illustrated in Fig 2 4 2 2 Operations overview The HARPS instrument is mounted on the Cassegrain focus of the 3 6m telescope in La Silla It shares the focus with other instruments EFOSC2 and CES TIMMI2 retired at the end of period 77 For this reason it is not always attached to the telescope Since the connection of the CES fibre into the HARPS adapter and the retirement of the TIMMI2 instrument end of P77 instru ment changes at the telescope are needed only when switching between EFOSC2 and HARPS and viceversa The instrument change is a lenghty operation that takes a full day and it is scheduled at the beginning of the period when time allocation for the telescope and the instruments is defined In contrast the switch between CES and HARPS requires only the movement of the HFCA fibre carriage and it takes less than one minute In one night it is possible to use both HARPs and CES with minimum overheads Since the beginning of the ESO period 78 a new mode of HAR
24. e in simultaneous thorium exposure mode is 15 seconds minimum exposure time to achieve a 15 cm sec instrument drift tracking For each CCD frame there is an overhead readout attachment of fits header etc of 32 seconds in the fast readout mode 416 kpx sec With 15 sec exposure time on sky 50 seconds cycles have been achieved The pipeline presently implemented is able to reduce this flood of data in nearly real time Reduction of one frame lasts about 24 seconds the pipeline no frames will be eft behind 5 4 4 Iodine cell out of service since May 2004 The lodine cell needs to be in a thermally stable state before it can be safely used This means that it has to be switched on at least two hours before the first exposure through the cell should be done This constraint holds both for science and for calibration exposures Target acquisition with the Iodine cell is done through the cell This makes that the position of the fibre hole image on the guide camera changes when the iodine cell is inserted in the light path In addition the focus position of both the auto guider and the telescope change and have to be re adjusted This takes about 5 10 min once during the night If more changes cell on cell off are required during the night the previously defined optimum values of the guide camera and telescope focus as well as the fiber position can be re set very quickly overhead 30 s For a proper reduction of the data taken with the I
25. e instrumental errors Third the spectrograph optics which is very similar to that of UVES is very efficient e Improvement of the online data reduction includes better corrections for instrumental effects and zero point definition wavelength calibration and it is substantially faster HARPS is a fibre fed cross dispersed echelle spectrograph located in the Coude floor of the 3 6m telescope For the sake of thermal and mechanical stability the spectrograph is enclosed in a vacuum vessel evacuated to a pressure lt 10 2mbar and maintained to a temperature of 17 C constant within 0 005 C RMS No moving parts are located inside the vacuum vessel VV The spectrograph itself has only one possible mechanical configuration All necessary moving parts are located in the Harps Cassegrain Fibre Adapter HCFA with the exception of the shutter which is located just outside of the vacuum vessel The optical design shown in figure 2 1 is similar to UVES at the VLT 4 Page of viii 49 HARPS User Manual Issue 1 3 3P6 MAN ESO 90100 0005 Figure 2 1 Optical layout of the spectrograph Its echelle grating is operated in quasi Littrow conditions off plane angle 0 721 deg blaze angle and the collimator in triple pass mode A white pupil configuration has been adopted with the cross disperser placed at the white pupil The dioptric camera images the cross dispersed spectrum on a detector mosaic of two CCDs Two fibres A and B feed th
26. e sky fibre entrance is exactly 114 west of the target fibre e watching the count rate of the exposure meter photometer B It is important that the telescope is well focused at all times It is recommended to have a through focus sequence performed using the guiding camera and the exposure meter two to three times per 27 28 Page of viii 49 HARPS User Manual Issue 1 3 3P6 MAN ESO 90100 0005 3 6m polnting limits DEC dag Figure 6 1 The sky area accessible for HARPS night or whenever the image quality observed on the guiding camera deteriorates significantly or whenever there is a significant temperature change few C A defocusing of 30 encoder units introduces an RV error of 1m s The accuracy of the telescope focus determination is within 10 encoder units 6 2 2 Pointing restrictions The usual pointing limit restrictions of the 3 6m telescope apply see Fig 6 1 The telescope dome shall be closed when any of the following weather conditions occur e Wind speed gt 20 m sec on the 3 6m monitor e Humidity gt 90 on the 3 6m monitor e Temperature within 2 of dew point on the MeteoMonitor e Dew on the dome the TIO will check the dome in person when there is reason to believe that condensation may occur The telescope shall not be pointed into the wind when the wind speed is more than 14 m sec 3 6m monitor Note Weather conditions at the 3 6m telescope may be significantly different fro
27. e spectrograph one object fibre and one reference fibre science fibres The spectra of the light from both fibres are formed by the spectrograph side by side on the detector Although all care has been taken to avoid stray light and ghosts both are present at some level most noticeably in the blue part of the spectrum table 2 1 Ghosts seems to be due to third order reflections in the grism The instrument is coupled to the telescope through an adapter the HARPS Cassegrain Fiber Adapter HCFA Two calibration fibers transmit the light from the calibration unit located in the Coude floor to the HCFA and inject it in the two science fibers for calibration The characteristic optical data are given in table 2 2 The spectral format is shown in Fig 2 2 Since the beginning of Period 78 a new mode is offered for HARPS In the new mode the photon collecting efficiency is increased by a factor up to two dependent on the seeing at the expense of Radial Velocity accuracy The new mode could be useful for faint objects for which a radial velocity accuracy of no better than 30m s is required The characteristics of the new mode and its operation are described in section 9 In the following is presented a brief description of the HARPS components the fibre adapter on the telescope the calibration unit and the fibre links connecting these components A sophisticated online data reduction pipeline is also part of the system section 8 The har
28. em NGAS 7 2 2 La Silla data archiving unit For convenient archiving of raw observation data and pipeline products a dedicated Data Archiving Unit DAU is available at La Silla It allows the observers to write the results of their observing run on DVDs choosing the data product they want raw data reduced data or both log files Appendix B contains the instructions on how to do this ESO will make available the necessary blank DVDs for normal observing runs e g planet search classical fibre spectroscopy to store the raw and reduced data in visitor and service mode However due to the enormous amount of raw data produced during a typical asteroseismology run gt 20Gb night the backup media will be a DLT tape if raw and reduced data are requested In case only the pipeline products are requested they can fit in a single DVD and the standard procedure Appendix B will be used to back up the data The reason for this is the impossibility to write all raw data to DVDs within a reasonable amount of time during the observing run with the resources manpower and hardware available at La Silla The raw data and all the pipeline products will anyway be stored in the Garching Science Archive using NGAS 7 2 3 Use of archived HARPS data HARPS data can be requested from the Garching Science Archive Data taken by observers in Visitor or Service Mode are subject to the usual proprietary period of 1 year According to the Agreeme
29. entering of 0 5 introduces a RV offset of 3 m s The present guiding and the dynamic centering systems introduce a RV error at most of 0 2 m s rms 4 4 Thorium calibration errors An error of about 0 5 m s is given to the Thorium calibration drift tracking and zero point The zero point error is by far the dominant source Chapter 5 Preparing The Observations 5 1 Introduction HARPS uses the standard ESO way of observing i e pre prepared Observing Blocks This chapter describes the philosophy behind this concept the available tools and the HARPS specific input In order to reach the full performance of HARPS with respect to the determination of accurate radial velocities the following items should be noted 1 to achieve an accurate solar system barycentric Radial velocity correction of 1 m s the target coordinates must be known to within 6 including proper motion 2 the RV of a star needs to be known to within 1 2 km s to give the pipeline a reasonable starting point for the RV computation 5 2 Introducing Observing Blocks An Observing Block OB is a logical unit specifying the telescope instrument and detector parameters and the actions needed to obtain a single observation It is the smallest schedulable entity which means that the execution of an OB is normally not interrupted as soon as the target has been acquired and centered on the fibre An OB is executed only once when identical observation se
30. equired during the night In order to produce the correct calibration sequence the available observing block RV Standard Calibration should be executed without changes before the beginning of the night It includes e 1 bias exposure The CCD bias is very stable only one bias is therefore needed by the pipeline Template HARPS_ech_cal_bias e 2 Tungsten lamp exposures where respectively fibre A and fibre B are successively fed by the Tungsten lamp These exposures are used for order location which is done automatically by the pipeline The processed products are stored in the calibration database if they pass the quality control of the pipeline and used for the subsequent reduction of the scientific exposures of the following night Template HARPS_ech_cal_tun A sequence of 5 Tungsten lamp exposures defined by NREP 5 where both fibres are simul taneously illuminated This sequence is used by the data reduction pipeline for producing a spectral master flat field which will be stored in the local calibration data base if it passes the quality control by the pipeline and used for the subsequent reduction of the scientific exposures of the following night Template HARPS_ech_cal_tunAB e 2 for reasons of redundancy Thorium exposures in which both fibres are simultaneously fed by light from the Thorium Argon lamps The ThAr1 long term reference lamp illuminates fibre A the ThAr2 lamp inlluminates fibre B During the night only
31. f_G2_A fits 2D extracted spectrum one row per order HARPS 2003 11 01T02 40 09 824_e2ds_A fits 1D extracted full spectrum wavelength calibrated in the solar system baricentric frame HARPS 2003 11 01T02 40 09 824_s1d_A fits Other summary tables are produced at the end of each night cal BIAS result tbl bias table cal Joe ONE result tbl order localization table cal FF result tbl flat field table cal TH result tbl Thorium table for wavelength calibrations drift result tbl Th lines drift in m s measured on fiber B filled only in the simultaneous Thorium reference mode CCF _results tbl results of the CCF with measured RV and RV sigma filled only when the RADVEL field in the template is different by 99999 33 34 Page of viii 49 HARPS User Manual Issue 1 3 3P6 MAN ESO 90100 0005 All files with extension tbl are ASCII files not MIDAS tables For details and format description see the DRS user manual available in http www 1s eso org lasilla sciops 3p6 harps manuals html 7 2 Data archiving 7 2 1 La Silla and Garching archives HARPS raw data are compliant with the requirements of ESO s Data Interface Control Board DICB They are stored locally at La Silla and in the central Garching Science Archive however see 7 2 3 Since the amount of raw data generated by HARPS can be quite substantial exceed ing 20Gb night in case of asteroseismology HARPS will use the Next Generation Archive Syst
32. he vacuum vessel They are linked by a specially developed mechanical interface which is damping the CFC vibration Each of the two CCDs has 50 pre scan and 50 over scan columns A 1 HARPS User Manual Issue 1 3 3P6 MAN ESO 90100 0005 Page of viii 49 9 The two CCDs are read by two different amplifiers and a difference of up to 10 in the bias level of the two CCDs might be expected Following we present the CCD test results as they were obtained on the test bench in Garching before installation at the telescope Property Jasmin red Linda blue CTE vertical 50 kpx s 0 999992 0 9999991 CTE horizontal 50 kpx s 0 999991 0 9999990 CTE vertical 416 kpx s 0 9999997 0 99999991 CTE horizontal 416 kpx s 0 9999995 0 9999990 Non linearity not available lt 0 298 Read out noise 50 kpx sec 2 87 0 le 2 76 0 09e7 Read out noise 416 kpx sec 7 05 0 23e7 5 5 0 16e7 Conversion factor 50 kpx sec 0 63 0 02e7 ADU 0 62 0 02e7 ADU Conversion factor 416 kpx sec 1 42 0 04e7 ADU 1 4 0 04e7 ADU Dark current at 110C not measured not measured Quantum efficiency peak 82 at 440nm peak 85 at 460nm Cosmetics science grade grade 1 science grade grade 1 Readout time 4296x4096 px 50 kpx s 180s 416 kpx s 23 s Mosaic flatness peak to peak 15um CCD parallelism 12 Chip to chip gap 1215 45um The two CCDs are nicknamed Jasmin the red CCD and Linda the blue one
33. his works correctly for exposure times up to 2700 s For very faint stars which require even longer exposures this may lead to an overexposure of the calibration spectrum with contamination of the stellar spectrum As the ultimate accuracy of HARPS 1 m s will usually not be reached on such faint stars it is recommended not to use the simultaneous Thorium reference method but to rely on the excellent short term stability of HARPS and take separate wavelength calibration exposures immediately before and after the science exposure to interpolate and remove possible instrumental drift errors The additional time spent on this is negligible given the long science integration 6 6 Ateroseismology Asteroseismology observations are particularly demanding from the point of view of data transfer and storage Users are recommanded to use USB disks as a storage media The observatory can loan these disks that should be returned within 6 months after which they will be reformatted and recycled Asteroseismology observers need to pay special attention to the guiding parameters Guiding cor rections should be faster if the exposure time is short For example if the exposure time is less than 5 seconds guiding correction time should be no more than 2 seconds this is also the maximum correction frequency we can use for guiding due to the telescope reaction time The dynamic fibre centering correction time should be set to 10 15 minutes maximum 30 minute
34. ifferent from 99999 e cross correlation function CCF only if parameter TARG_RV is defined and different from 99999 The pipeline output is available immediately after the processing is finished see section 5 4 1 It can then be transferred to the offline workstation for further analysis It can also be saved to disk and CD DVD using the Data Archiving Unit see 7 2 2 available with HARPS This is typically done next morning by the telescope operator 8 2 High accuracy radial velocities The reduction concept applied by the pipeline for the calculation of high accuracy radial velocities using the Thorium reference method is described in the paper ELODIE A spectrograph for accurate radial velocity measurements by Baranne Queloz Mayor et al A amp AS 119 373 1996 In order to get the full performance of the pipeline with respect to the determination of accurate radial velocities the following items should be noted 1 to achieve an accurate solar system barycentric Radial velocity correction of 1 m s the target coordinates must be known to within 6 including proper motion 2 the RV of a star needs to be known to within 1 2 km s to give the pipeline a reasonable starting point for the RV computation 8 3 Iodine cell data On 12 cell data the DRS provides to the user a wavelength calibrated spectrum For further analysis the observer should consider the following input data 37 38 Page of viii 49 HARPS User M
35. ilt by the HARPS Consortium consisting of Observatoire de Gen ve Observatoire de Haute Provence Physikalisches Institut der Universitat Bern Service d A ronomie du CNRS and with substantial contribution from ESO La Silla and ESO Garching Its purpose is to reach a long term radial velocity accuracy of 1 m s for slowly rotating G dwarfs Such precision enables the detection of low mass Saturn like extra solar planets and low amplitude stellar oscillations The design of HARPS is based on the experience acquired with ELODIE installed at the 1 93m telescope at OHP and CORALIE at the 1 2m Swiss Euler telescope at La Silla during the past 10 years by the members of the HARPS Consortium The basic design of HARPS is therefore very similar to these instruments The efforts to increase the HARPS performance compared to its predecessors address mainly three issues e Increase of the instrumental stability The spectrograph is installed in an evacuated and temperature controlled vacuum enclosure This allows to remove to a very large extent all RV drifts which would be produced by temperature variations or changes in ambient air pressure and humidity e Increase of the signal to noise ratio SNR on single RV measurements The improvement is attained through different steps First HARPS is installed on the ESO 3 6 m telescope Second the spectral resolution is increased by a factor of about two The higher spectral resolution helps also to reduc
36. ion 5 3 and the HARPS Template Guide sec tion 1 1 3 3 4 Pipeline data reduction The pipeline performs the same reduction as for the simultaneous Thorium reference method sec tion 3 1 4 but it does not correct for the instrumental drift this one not being traced by the Thorium lamp as in the simultaneous Thorium reference method It does not perform sky subtraction For more details about the pipeline Data Reduction Software DRS please refer to the DRS user manual Pipeline execution times see section 5 4 1 18 Page of viii 49 HARPS User Manual Issue 1 3 3P6 MAN ESO 90100 0005 Chapter 4 RV accuracy The high RV accuracy obtainable with HARPS is a result of an extremely stable and strictly con trolled instrument and a data reduction software designed and optimized for the purpose The pipeline RV determination is optimized for data taken in the simultaneous Thorium reference method and nothing can be said at the moment about the obtainable RV accuracy with the iodine self cali bration method with HARPS However Jy data were taken recently and an accuracy for the method with the HARPS instrument will be quoted in the near future For reference the method could yield a long term 2 years accuracy of 2 65 m s with the UVES spectrograph with a S N of 66 per pixel and a resolving power of 100000 120000 similar to the HARPS one of 115000 M Kiirster et al A amp A 403 1077 2003 The short term RV accuracy of
37. l and contributes to stabilize the input point spread function PSF of the spectrograph The scrambler serves also as vacuum feed through for the fibres and in addition houses the exposure shutter The shortest useful exposure time supported by the shutter is 0 2 seconds A second fibre link connects the Calibration Unit CU section 2 4 next to the spectrograph with the HARPS Cassegrain Fibre Adapter HCFA on the telescope see figure 2 1 2 4 Calibration unit The Calibration Unit CU contains a Tungsten flat field lamp three Thorium Argon lamps for spectral calibration and a tungsten lamp illuminating an iodine cell for tests purposes It is con nected via two optical fibres to the HCFA which redirects the light of the calibration sources into the spectrograph fibres as required The two calibration fibres can be fed either by the same or independently by two different calibration sources Of the three Thorium Argon lamps the lamp dubbed THARI is the absolute reference and its use should be minimized Typically it is used for 5 minutes per day during the afternoon calibrations shining on fibre A The lamp named THAR2 is used to measure the instrument drift in parallel with the science observations THAR3 is a hot spare The lamp can be switch on at any moment It is particularly useful when one of the other lamps burns out in the course of the night 8 Page of viii 49 HARPS User Manual Issue 1 3 3P6 M
38. lank List of Tables 2 1 2 2 2 3 2 4 5 1 5 2 5 3 9 1 Level of diffuse stray light and ghosts as a percent of the flux in the order 5 Characteristic optical data of HARPS aoaaa aaa a 7 Exposure meter count rotes 10 Overall efficiency of the HARPS system See text o 11 Fiber illumination scheme for each template 22 Execution times overheads 23 Off line overheads E E 24 Comparison of the HAM and EGGS performances 204 39 vil viii Page of viii 49 HARPS User Manual Issue 1 3 3P6 MAN ESO 90100 0005 This page was intentionally left blank Chapter 1 Introduction 1 1 Scope This User Manual is intended to give all necessary information to potential users of the HARPS instru ment to help them decide on the opportunity to use the instrument for their scientific applications to be used as a reference when writing observing proposals and when preparing the observations For this purpose we give e an overall description of the HARPS instrument its performance and its observing modes e information on the preparation of the observations e information on the observing process e a description of the HARPS data and near real time pipeline data reduction The following documents are closely related to this manual and should be consulted as well e the P2PP User Manual e the HARPS Template Guide Both are available through the Informatio
39. lates The necessary acquisition and observing templates are available e HARPS ech acq thosimult for star acquisition and setup of simultaneous Th exposures e HARPS_ech_obs_all for taking simultaneous Th exposures For a detailed description of the templates see section 5 3 and the HARPS Template Guide 3 1 4 Pipeline data reduction The online pipeline does spectrum extraction wavelength calibration RV calculation using a template spectrum of ideally the same spectral type as the target star A comprehensive library of stellar spectral templates is being built up Currently it contains templates of the following spectral type others are in preparation e GV The pipeline applies the following corrections detector bias dark flatfield cosmic ray removal and rebins the spectrum according to the wavelength calibration obtained in the afternoon or in the closest succeful HARPS_ech_cal_thoAB exposures The drift correction is not done At present the drift is measured and inserted in the fits header but is not applied to the RV value The user can do tha if he she wishes by simple subtraction Radial velocity and Julian date correction are calculated in the solar system barycenter reference based on the Bretagnon amp Francou 1988 VSOP87E planetary theory Radial velocity computation is automatically done for all exposure types when a radial velocity value different from 99999 is provided by the TARG_RV parameter Conversely RV c
40. m those near the RITZ In particular the wind speed is a few meters per second higher and the humidity lower The TIO will make the decision to close the dome as necessary VAs should accept the decision since the reason is exclusively the protection of the telescope from damage At any rate in case of a HARPS User Manual Issue 1 3 3P6 MAN ESO 90100 0005 Page of viii 49 29 disagreement the dome should first be closed and subsequently the VA may take up the issue with the support astronomer and finally the La Silla shift leader though it is highly unlikely that the decision will be altered The dome may be re opened if weather conditions improve and stay below the operating limits for at least 30 minutes This waiting period is particularly important in case of humidity The TIO will further confirm that the condensation on the dome has completely evaporated 6 3 Night calibrations Night calibrations in addition to the calibration observations taken before the start of the science observations are only necessary when using the Iodine self calibration method see section 3 2 2 6 4 Real time display Raw data coming from the instrument are displayed on a FIERA Real Time Display RTD Both CCDs are displayed in the same RTD fig 6 2 6 5 Observing very faint stars As explained in section 3 1 a variable neutral density filter is used to balance the intensity of the Thorium Argon calibration spectrum depending on the exposure time T
41. me incl writing of FITS headers and transfer to IWS minimum time between successive exposures switching between ThAr and Iodine modes telescope focusing at the beginning of the night to be re peated 4 5 times during the night 5 min upper limit for large more than 180deg dome ro tation 2 min typical for new point ing within a few degrees from the previous position 23s with 416kpx s readout speed 87s with 104kpx s readout speed Table 5 2 Execution times overheads 24 Page of viii 49 HARPS User Manual Issue 1 3 3P6 MAN ESO 90100 0005 5 4 2 Off line overheads DRS pipeline for thosimul without RV computation 110 sec DRS pipeline for objA without RV computation 50 sec DRS pipeline for objAB without RV computation 70 sec DRS pipeline extra time for each RV calculation 30 sec spectrum ThAr lamp pre heating once at the beginning of the night 2 min minimum 10 min rec ommended 15 min maximum Iodine cell pre heating once at the beginning of the night 2 hours Table 5 3 Off line overheads The pipeline overheads are for reference only observations can proceed without the need for waiting the pipeline results ThAr lamp and iodine cell pre heating overhead have to be considered before the start of the afternoon calibrations 5 4 3 Fast time series observations asteroseismology The shortest exposure time possible with the HARPS shutter is 0 2 seconds while the shortest exposur
42. n Sources section of the HARPS web pages http www 1s eso org lasilla sciops 3p6 harps 1 2 Additional information The latest information updates about the HARPS instrument can be found on the HARPS web pages http www 1s eso org lasilla sciops 3p6 harps General information about observing at La Silla is available from the La Silla web pages http www 1ls eso org 2 Page of viii 49 HARPS User Manual Issue 1 3 3P6 MAN ESO 90100 0005 1 3 Contact information In case of specific questions related to visitor mode observations please contact the La Silla High Resolution Spectroscopy Team likewise for specific questions related to Service Mode observations and proposal preparation ls hires eso org 1 4 Acknowledgments Most of the contents of this manual is based on information from the HARPS Consortium Observa toire de Gen ve Observatoire de Haute Provence Universitat Bern Service d A ronomie ESO La Silla and Garching in particular by F Pepe and D Queloz and from La Silla Science Operations G Lo Curto and T Dall Releases of this documents are based on the original version edited by Gero Rupprecht Feedback on this User Manual from users is encouraged Please email to ls harpsQeso org Chapter 2 HARPS Characteristics 2 1 Instrument Overview HARPS High Accuracy Radial velocity Planetary Searcher is an instrument designed for the mea surement of Radial Velocities RV at highest accuracy It was bu
43. n of HARPS Data 37 8 1 The HARPS data reduction pipeline 0 0 20 200 0000002 eee 37 8 2 High accuracy radial velocities o oo aaa ee 37 8 3 Jodine cell data sussa 04 6 fhe ea a ee A AVE ee ee a ee 37 9 HARPS high efficiency mode EGGS 39 A Description of archived HARPS data Al A L Data naming rules en ama Re ee aoe a a Dee ka EA eee eG 41 2 Data formats a A A EE oe BSS 41 PAD Raw frames A E 41 B Use of the Data Archiving Unit 45 C List of acronyms 47 List of Figures 2 1 2 2 2 3 2 4 2 5 2 6 6 1 6 2 9 1 Al Optical layout of the spectrograph 4 Spectral format of HARPS Blue orders are down red areup 5 Table showing the spectral format of HARPS and the expected number of electrons for a 1 minute exposure on a G2V star of magnitude 6 A seeing of 0 8 airmass 1 and new moon are the values of the selected parameters In the table the spectral bin is defined as one pixel This table is obtained by the HARPS Exposure Time Calculator ETC aco pra Boe oe Git ae AA A deg 6 The main components of the HARPS system 8 Quantum efficiency of Jasmin 12 Quantum efficiency of Linda 12 The sky area accessible for HARD 28 HARPS spectrum inthe RED siis ss tes cereais a aa o do ot 30 HARPS EGGS comparison 40 Raw frame formats aerial AR he Pha ES Be ae AA ae 43 vi Page of viii 49 HARPS User Manual Issue 1 3 3P6 MAN ESO 90100 0005 This page was intentionally left b
44. nt between ESO and the HARPS Consortium the data taken by the Consortium during their Guaranteed Time are subject to special protection e Raw data and reduced spectra I f A in the Earth reference frame at the time of the obser vation will be made public one year after observations e All raw data and radial velocity measurements obtained by the Consortium will be made public one year after the end of the 5 year Guaranteed Time period In practice this means that data obtained by the Consortium can be requested from the Garching Science Archive as usual one year after the observations However in order to make recovery of precise radial velocities impossible the keywords containing information about the time of the observations HARPS User Manual Issue 1 3 3P6 MAN ESO 90100 0005 Page of viii 49 30 will be filtered from all file raw and reduced headers by the Archive during the de archiving process details are TBD This filtering will be applied until one year after the end of the 5 year Guaranteed Time period 36 Page of viii 49 HARPS User Manual Issue 1 3 3P6 MAN ESO 90100 0005 Chapter 8 The Reduction of HARPS Data 8 1 The HARPS data reduction pipeline Every HARPS frame is processed by the online pipeline Depending on the observation the pipeline uses different reduction recipes Results of the reduction are e extracted spectrum all modes e precise radial velocity only if parameter TARG_RV is defined and d
45. odine cell it is necessary to obtain the following spectra e target through the iodine cell S Napprox200 e target without iodine cell template S N 400 e B star with and without the iodine cell S N 400 The spectra of the B star are only needed when the template spectra is obtained In this case they must be taken in the same night during which the template spectra are recorded Sufficient time for obtaining these spectra must be foreseen HARPS User Manual Issue 1 3 3P6 MAN ESO 90100 0005 Page of viii 49 25 5 5 The HARPS Exposure Time and Spectral Format Calculator The HARPS Exposure Time Calculator ETC models the instrument and detector in their different configurations It can be used to compute the detailed spectral format wavelength and order number as function of position on the detector and the expected SNR for the specified target under given atmospheric conditions as a function of exposure time It is available via the HARPS web pages http www 1s eso org lasilla sciops 3p6 harps 26 Page of viii 49 HARPS User Manual Issue 1 3 3P6 MAN ESO 90100 0005 Chapter 6 Observing with HARPS 6 1 Before the night Depending on the observing method applied simultaneous Thorium reference Iodine self calibration classical fibre spectroscopy different sets of calibration exposures need to be taken before the start of the science observations For all three methods it is necessary to take a series of c
46. or ETC HARPS User Manual Issue 1 3 3P6 MAN ESO 90100 0005 Page of viii 49 7 System 2 fibres each 1 diameter on sky distance 114 spectral range 380 690nm collimated beam diameter 208mm Echelle grating R4 31 6 gr mm blaze angle 75 mosaic 2x1 on Zerodur monolith 840 x 214 x 125mm efficiency gt 65 in the visible Cross disperser grism FK5 grism 257 17 gr mm blazed at 480nm 240 x 230 x 50mm T 73 average Collimator mirror Zerodur with protected silver coating f 1560mm used di ameter 730mm triple pass Camera all dioptric 6 elements in 3 groups f 728mm f 3 3 T gt 85 Detector 2 2k x 4k EEV CCDs pixel size 15 um Beam focal ratio inside the spectrograph 7 5 Spectral format upper red CCD Jasmin orders 89 114 533 691nm lower blue CCD Linda orders 116 161 378 530nm Spectral resolution RS 115 000 measured Sampling per spectral element 3 4 px per FWHM Spectrum Separation 17 3 px fibres A and B Order separation Jasmin order 89 1 510mm 100 7px order 114 0 940mm 62 7px Order separation Linda order 116 0 910mm 60 7px order 161 0 518mm 34 2px Spectrograph stability 1 m s in one night under normal conditions Table 2 2 Characteristic optical data of HARPS fibres are mounted in the adapter 2 3 Fibre links The spectrograph is linked to the 3 6 m telescope via two optical fibres The fibre link incorporates an image scrambler which is fixed on the vacuum vesse
47. perations overview mu ica ta E a RE e as a SE a e E WEE links tad a A A AID 8 dete AA a 2 4 Calibration Unib pelas ae cd leas a a rr a be Ged tz 2 5 Fibreadapter sms sa ada hoe a ee TESE Rae e EE c 2 6 Detector and read out electronics ee 25 EXPOSure meter ee ee RE SE PARES ek eS e e o a 2 8 Data reduction software lt rroa u id a ee ae a 2 9 System efficiency ee 3 Observing modes 3 1 Simultaneous Thorium reference method ZEIL Performance cs gos feos fo he a Re al ge A ES E 31 2 Calibrations e EE EE eo a ee e SS 3 1 3 Observing templates em Se ee Se ee ee a 3 1 4 Pipeline data reduction 2 2 0 0000 a ee ee 3 2 Iodine self calibration method out of service since May 2004 3 2 1 Performance coo aa e a BB 3 2 2 NEEN 323 Templates sce 228 bh et Sud Ae Ra hoe Be Bae Oe Ben ele ee ales 3 2 4 Pipeline data reduction 00 000 eee ee ee 3 3 Classical fibre spectroscopy 1 a 3 21 PerfOrMance ui ee AR a Be hale So ee ha a a as 3 9 2 Calibration es ag hh dw th wa Oe a ee ek ee A Doo Templates uti A Ge hg See EE 3 3 4 Pipeline data reduction 2 2 a 4 RV accuracy 41 Photon noise su pgs a he ON a EE ADE E ER 4 2 Telescope focus oa e suis E ON a a A e a Za Centerino errors pta A a RT hE a ee BS Se 4 4 Thorium calibration errors iii 13 13 13 14 15 15 15 15 15 16 16 16 16 17 17 17 iv Page of viii 49 HARPS User Manual Issue 1 3 3P6 MAN ESO
48. quences are required e g repeated observations using the same instrument setting but different targets a series of OBs must be built Usually one OB consists of two separate entities the acquisition template and the observation template s For normal science observations HARPS uses four different acquisition templates dif ferent for the various observing modes and one common observing template They are described in section 5 3 and the HARPS Template Guide 53 P2PP P2PP is the standard tool for the building of observing blocks from the instrument specific templates A comprehensive description including the user manual is available from the ESO web pages at http www eso org observing p2pp Observers using HARPS in Visitor Mode should prepare their OBs in advance using the HARPS Instrument Package which is automatically downloaded once P2PP is started and the HARPS 21 22 Page of viii 49 HARPS User Manual Issue 1 3 3P6 MAN ESO 90100 0005 HARPS_ech_acq_thosimult star ThAr lamp 2 3 HARPS_ech_acq_12ce11 Star Dark HARPS_ech_acq_objA Star Dark HARPS_ech_acq objAB Star Sky ech_cal_thoAB ThAr lamp 1 ThAr lamp 2 3 ech_cal_tun exposure 1 Tungsten Dark ech_cal_tun exposure 2 Dark Tungsten ech_cal_tunAB Tungsten Tungsten ech_cal_tunAT2 HCFA 12 cell Dark ech cal tunI2AB CU 12 cell CU 12 cell Table 5 1 Fiber illumination scheme for each template program is selected from the list of the user s approved programs OBs p
49. repared at the observer s home institution can be quickly imported after having ftp ed them in the proper machine in La Silla in the P2PP running at the telescope console and be ready for execution Service Mode observers need to check in their OBs with ESO see the La Silla web pages http www 1s eso org lasilla sciops for details For a concise description of all HARPS templates and the parameters selectable with P2PP consult the HARPS Template Guide see section 1 1 In table 5 1 is shown the fibers illumination scheme as a function of the template used 5 3 1 Acquisition templates HARPS uses the acquisition template to preset the telescope and to set up the instrument configu ration for the selected observing mode The following acquisition templates are available e HARPS_ech_acq_thosimult for simultaneous Th exposures e HARPS_ech_acq_12cell for 12 cell exposures e HARPS_ech_acq_objA for fibre spectroscopy no sky e HARPS_ech_acq_objAB for fibre spectroscopy with sky A concise description of the acquisition templates is given in the HARPS Template Guide see section 1 1 5 3 2 Observing template HARPS uses one single observation template for all observing modes because all instrument setup is done by the acquisition templates e HARPS_ech_obs_all for exposures in all observing modes A concise description of the observation template is given in the HARPS Template Guide HARPS User Manual Issue 1 3 3P6 MAN
50. rvation the pipeline provides only the spectrum of the fibre A and uses fibre B order location to compute the CCD background objAB observation should be preferred when a sky background correction may be needed The data reduction pipeline provides an extracted spectrum for each fibre The sky correction is left to the user The high stability of the instrument makes wavelength drifts very small If the same calibration sequence than for the simultaneous Thorium reference method is run before the beginning of the night a global wavelength calibration with an accuracy better than 3 m s can be expected 3 3 1 Performance For estimates of the SNR with an accuracy of about 10 under given observing conditions the ETC available via the HARPS web pages http www 1s eso org lasilla sciops 3p6 harps can be used see section 5 5 HARPS User Manual Issue 1 3 3P6 MAN ESO 90100 0005 Page of viii 49 17 3 3 2 Calibration A calibration sequence similar to the Simultaneous Thorium Reference method is recommended before the beginning of the night 3 3 3 Templates The necessary acquisition and observing templates are available e HARPS_ech_acq_objA acquisition and setup for fibre spectroscopy with the object in fibre A e HARPS_ech_acq_objAB acquisition and setup for fibre spectroscopy with the object in fibre A and sky in fibre B e HARPS ech obs all for fibre spectroscopy exposures For a detailed description of the templates see sect
51. s however a quite long time interval which shall not be the same nor a fraction of the period of the star s oscillations which are being measured The rationale behind this choice is to diminish the low frequency noise in the RV power spectrum The minimum correction time for the dynamic fibre centering algorithm is 30s 30 Page of viii 49 HARPS User Manual Issue 1 3 3P6 MAN ESO 90100 0005 XA Rtd Real Time Display version 2 59 File View Graphics Real time harps 2105 0 23 57 39 733 4503 0 01 15 08 00 Y 8 210 Max BET vie up Zi ZS ality E image select object D scroll image measure WCS Control JH select region Figure 6 2 The Real Time Display of a HARPS spectrum taken with the tungsten lamp illuminating both fibers The lower part is the blue chip Linda the upper one is the red chip Jasmin Wavelength increases from lower left to upper right HARPS User Manual Issue 1 3 3P6 MAN ESO 90100 0005 Page of viii 49 31 As from June 2004 a periodic signal is detected in the RV power spectrum at 2 8mHz and aliased at about 6 and 9 mHz This is possibly due to a hard point in the right ascentions main gear or motor and we are trying to fix it The users however should be aware of it The amplitude of this signal can reach 30cm s This noise is only noticed in short lt 20 seconds exposure
52. s When using exposure times larger than 20 seconds the guiding is able to compensate the glitch introduced by the gear motor Observers should ask the operators proper setting of the guiding and dynamic centering parameters 6 7 End of the night No further calibrations are necessary after the end of the science observations To prolong the life of the calibration lamps HARPS is switched to the so called Dark mode All lamps still in use at the time are thereby switched off and the dust cover in the fibre adapter is put in place to protect the fibre entrance The Ig cell will be left on until it is turned off manually All electronics are in stand by all internal house keeping functions temperature and pressure control logging continue to operate 32 Page of viii 49 HARPS User Manual Issue 1 3 3P6 MAN ESO 90100 0005 Chapter 7 Data products and archiving 7 1 Data products HARPS writes FITS files with extensions containing the data of both CCDs The size of one raw data file is approximately 32Mb By default the data products of the online pipeline are archived as well Following is an example of the files that are included in the archiving of one exposure Raw data file HARPS 2003 11 01T02 40 09 824 fits Cross correlation function summary table with extracted RV per each order HARPS 2003 11 01T02 40 09 824_ccf_G2_A tbl Cross correlation function matrix in fits format HARPS 2003 11 01T02 40 09 824_cc
53. s of both fibres will not be balanced The Thorium spectrum which is recorded simultaneously with the stellar spectrum is used to compute the instrument drift from the last wavelength calibration usually done at the beginning of the night The calibration unit contains two identical ThAr lamps For the simultaneous reference method only the lamp ThAr2 can be used The lamp ThAr1 should be used as a reference only for the afternoon calibrations and switch off afterwards Ideally this should prolong the life time of this reference lamp 3 1 1 Performance For estimates of the SNR under given observing conditions the Exposure Time Calculator ETC available via the HARPS web page http www 1s eso org lasilla sciops harps can be used with an accuracy of about 10 see section 5 5 The relationship between photon noise induced radial velocity error and S N is given by the following formula 100 3 1 S N550nm rms m s 13 14 Page of viii 49 HARPS User Manual Issue 1 3 3P6 MAN ESO 90100 0005 As a rule of thumb an photon noise error of 1 m s or S N 100 can be achieved for a 6th magnitude G dwarf in 60 sec Note that due to the small fiber aperture on the sky 1 the performances are critically seeing dependent 3 1 2 Calibrations The Simultaneous Thorium Reference Method needs a sequence of calibration exposures to be taken before the beginning of the night No further calibration exposures are r
54. x C List of acronyms ADC AG BOB CCD CCF CES CFC CU DAU DFS DHS DICB DIMM DRS E2DS ESO ETC FIERA FITS HCFA HARPS ICS IWS ND NGAS NTT OB OG OS P2PP PSF RITZ RTD RV SA SNR TBC Atmospheric Dispersion Compensator Auto Guider Broker of Observing Blocks Charge Coupled Device Cross Correlation Function Coude Echelle Spectrograph Continuous Flow Cryostat Calibration Unit Data Archiving Unit Data Flow System Data Handling System Data Interface Control Board Differential Image Motion Monitor Data Reduction Software Extracted 2 Dimensional Spectrum European Southern Observatory Exposure Time Calculator name for ESO s standard CCD controller Flexible Image Transport System HARPS Cassegrain Fibre Adapter High Accuracy Radial velocity Planet Searcher Instrument Control Software Instrument WorkStation Neutral Density Next Generation Archive System New Technology Telescope Observing Block Observatoire de Geneve Observation Software Phase 2 Proposal Preparation Point Spread Function Remote Integrated Telescope Zentrum Real Time Display Radial Velocity Support Astronomer Signal to Noise Ratio To Be Confirmed AT 48 TBD ThAr TIO UVES VA VLT XTC Page of viii 49 HARPS User Manual Issue 1 3 3P6 MAN ESO 90100 0005 To Be Determined Thorium Argon Telescope and Instrument Operator Ultraviolet Visible Echelle Spectrograph Visiting Astronomer Very Large Telescope
Download Pdf Manuals
Related Search