APEX Calibration and Data Reduction Manual
Contents
1. Determine the observing mode for this Scan A basic ScanInfo entity is filled from the file which containes observing mode used frontend backend subscans etc Make a list of all unique subscan numbers present within the FITS file that has just been read Check that the supplied subscan number exists within this list Make a list of all unique FeBes present within the FITS file these are determined from the ScanInfo entity Reset SciDic This is the internal dictionary where reduced science data are stored Read the required subscan s from the FITS file and perform observing mode specific data reduction for each FeBe and BaseBand using any Reference and Calibration data previously stored Make a basic display of the results Write the calibrated data to CLASS lFeBe frontend backend combination BaseBand is an integer which specifies a group of backend sections physical inputs that share the same frequency setup Create Date May 3 2013 Page 4 Contact author H Hafok APEX APEX Calibration and Data Reduction Manual Observing Scan Number Engine Subscan Number Get Observing mode from FITS file Check Subscan number exists in this scan Open FITS file a RE Loop over FeBe BaseBands ere A N ee ee Calibration Science Data Skydip ONOFF SCANTYPE CAL x ae SKYDIP MAP CALs in MAPs FOCUS OBSTYPE an REF ON ON SUBSTYPE2. APEX MPI MAN 0012 Atacama Pathfinder oiga EXperiment Release May 3 2013 Category 4 User Manual Author H Hafok APEX Calibration and Data Reduction Manual Edward Polehampton Heiko Hafok Keywords Calibration APEX Author Signature H Hafok Date May 3 2013 Approved by D Muders Signature Institute MPIfR Date Released by Signature Institute Date APEX APEX Calibration and Data Reduction Manual Change Record Revision Date Author Section Remarks Page affected 0 0 2005 06 20 E Polehampton All Initial Version 0 1 2005 09 14 E Polehampton All transfer to official format 0 2 2005 09 16 E Polehampton All minor updates add script examples 0 3 2009 02 02 H Hafok All revisited including several changes 1 0 2009 04 27 H Hafok All minor changes concerning the gain array release version 1 1 2010 03 08 H Hafok All added description of fsw modes 1 2 2010 03 15 H Hafok All added preset of water 1 2 1 2013 05 03 D Muders Title changed document category Create Date May 3 2013 Page 2 Contact author H Hafok APEX APEX Calibration and Data Reduction Manual Contents 1 Introduction 4 1 1 Calibrator a IM ao pen ads Desc So tt oe Ooh hme op nee eh eee oe Arh rece eh cen at 4 1 2 Onlin casera La Beat Gp oe GR es a ee ee SS 4 1 3 Offline 0 00 LA ate A ek ANA Denn te i a il be hehe Ate d
3. APEXCAL gt obsEntity plotContinuum feedNum 2 numPlots 1 renew 0 4 this adds the second Feed to the previous plot in the lower half of the page For pointing observations a Gaussian fit can be made to the data from a single subscan by including the extra keyword dofit DOFIT This only works for pointing data as focus and skydip fits require data from multiple subscans The special plot methods for these observing modes must be used instead see Sect 4 3 2 To obtain the pointing result from a single subscan contained in myEntity APEXCAL gt pointingResult myEntity plotContinuum feedNum dofit DOFIT APEXCAL gt print pointingResult 1 23 100 0 20 0 0 2 1 2 0 3 where the result gives centre amplitude width of the Gaussian fit and their respective errors 4 2 2 Writing to CLASS In order to specify the filename and directory to be used for writing data to CLASS the following method can be used APEXCAL gt setClassName filename June2004 apex directory The defaults if one of these keywords are omitted are a filename based on the project ID of the scan in the current directory To check the current values for CLASS filename and directory APEXCAL gt showClassFile Although the reduced data are written automatically to CLASS during the reduce method this can also be done manually for single entity objects using APEXCAL gt writeClass obsEntity obsEntity is either a single
4. APEXCAL gt scanInfo Entities ScanInfoEntity After it is defined and initialized it can be filled with the dataset APEXCAL gt scanInfo fill dataset The data within this object can then be accessed using getter methods APEXCAL gt scanInfo getBasebandList FLASH460 FFTS1 APEXCAL gt scanInfo getFeBeList APEXCAL gt scanInfo getInstrument APEXCAL gt scanInfo getNumPhases APEXCAL gt scanInfo getObsMode APEXCAL gt scanInfo getObsTypeList FLASH460 FFTS1 APEXCAL gt scanInfo getScanMode APEXCAL gt scanInfo getSubscanNum APEXCAL gt scanInfo getSubscanNumList APEXCAL gt scanInfo getSubscanTypeList APEXCAL gt scanInfo getSwitchMode However for easier use within Python individual subscans for each FeBe BaseBand combination can be filled into an ObsEntity object This object contains attributes equal to the MBFITS keywords and is the basic data structure used in all further processing methods The ObsEntity object is filled from the dataset object using APEXCAL gt myEntity ObsEntity APEXCAL gt myEntity fill dataset subscan FLASH460 FFTS1 baseBand Monitor points must be filled individually APEXCAL gt myEntity fillMonitorPoint dataset MonitorPointKey subscan The resulting ObsEntity object can be worked with as described in Section 4 2 4 5 Atmospheric model features the following method s can be called to use the atmospheric model features in the offline
5. APEXCAL gt plotSkydip FeBe BaseBand numPlots 1 renew 0 oplotFit 1 fitType ATMfixTS If there is a spike or bad value in the data that is not removed automatically the skydip fit includes a 20 clip the data can be edited by hand before calling one of these plot routines Data points set to a blanking value of 999 will not be included in the fit 4 4 Reading a FITS file by hand FITS files can be read into the apexOfflineCalibrator in several stages The first involves reading all the data in the file and creating a dataset object The second stage involves creating ObsEntity objects for single subscans within the FITS file These ObsEntity objects are used in the further processing of the data and have associated methods and attributes that are easier to manipulate than the full FitsFile object The directory containing the FITS file can be set using APEXCAL gt setFitsDir home FitsFiles Create Date May 3 2013 Page 22 Contact author H Hafok APEX APEX Calibration and Data Reduction Manual The file is found automatically within this directory from the scan number assuming the standard APEX convention for file names and read using APEXCAL gt dataset internal getDatasetForScan scanNum Now we can define a ScanInfo entity which has methods to rerieve basic information from the dataset like the scan mode the obs mode the scan geometry used frontend backends basebands etc APEXCAL gt import Entities
6. Before a reference is used the FeBe BaseBand and frequency are checked with the ON subscan The Reference closest in time to the ON is chosen In the online case the RefDic is reset for every new Scan ie only References within a scan are considered In the offline case it is only reset by calling the resetRefs method and References from other scan numbers can be applied internal SciDic Reduced science data are stored in SciEntity objects These combine the information in the raw data ObsEntity with the calibration that was applied CalEntity ie all the information needed to write the CLASS header This is a dictionary of dictionaries The first level has keys labeld by lt FeBe gt lt BaseBand gt The second level has keys labelled by lt subscan number gt It is reset every time the CalibController is called as after every call all the reduced data is written to CLASS and so is not needed any more The resulting Calibration entity stays valid until a new calibration was reduced successfully There are also other internal dictionaries to store the data from Pointing Focus and Skydip observations and for raw data from CalCold observations Raw calibration data are reset for every new calibration scan processed internal RawCalDic internal PointingDic internal FocusDic internal SkydipDic Create Date May 3 2013 Page 6 Contact author H Hafok APEX APEX Calibration and Data Reduction Manual 2 Reduction of calibra
7. baseband For continuum observations the data are usually plotted against integration number except for several special cases e g pointings where offset in arcseconds is used APEXCAL gt obsEntity plotContinuum feedNum There are also several methods to plot the final results of pointing focus and skydip observations see Sect 4 3 2 All the plot methods have several extra keywords that can be used to get multiple plot windows on the same page The default behaviour is to start a new page for each plot In order to obtain multiple plots per page specify the number of panels using e g numPlots 2 To ensure that a new page is not started for the second plot renew must be set to be 0 To close the ppgplot device more important for postscript output the last plot panel must have numPlots 1 to show that no more plots are expected For example APEXCAL gt obsEntity plotContinuum feedNum 1 numPlots 2 renew 1 this plots first Feed in top half of a new page Create Date May 3 2013 Page 19 Contact author H Hafok APEX APEX Calibration and Data Reduction Manual Table 2 Plotting and fitting combinations SCANTYPE SCANDIR x axis Fit Result POINT ALAT DATAPAR_LATOFF Gaussian centre amplitude width amp errors POINT ALON DATAPAR_LONGOFF Gaussian centre amplitude width amp errors FOCUS FOCUSPOS Quadratic x offset SKYDIP Airmass Exp lin or ATM depends on skydip fit mode all other INTEGRATIONS none none
8. FeBe BaseBand combination if available Loop over Subscan Numbers If subscan is a REF sum the individual Integrations save to internal reference dictionary RefDic If subscan is an ON prepare an On object and a BestOff object taking account of the observing mode send these to the Reducer save the result to the SciDic Write the contents of the SciDic to a CLASS file The observation definition used by the HeterodyneReducer is based on the following keywords from MBFITS Observing mode gt SCANTYPE POINT FOCUS CAL SKYDIP OTF ONOFF RASTER RASTER can also be called MAP Switching mode gt SWTCHMOD TOTP FSW CAL WOBSW NONE BEAMSW HORNSW LOADSW cannot be processed Create Date May 3 2013 Page 11 Contact author H Hafok APEX APEX Calibration and Data Reduction Manual Subscan type gt OBSTYPE ON REF or SKY HOT COLD The BestOff is selected by comparing the mid time in the ON object for the subscan for otf the mid time of the full subscan with all the REF objects that match the FeBe BaseBand The REF with closest time is selected as the BestOff In the online case the reference dictionary is emptied when a new scan number is started This means only REFs from the same scan as the ON are considered As the subscans also in offline mode are sequencially added allways the closest REF before the observation is selected Of course it is possible to process in offline mode the REF observation after the subscan In
9. The array is rotated in such a way that it compensates for the nasmyth rotation and earth rotation for each subscan The array is kept constant for a subscan The pixel coordinates are con stant in the equitorial system with azimut elevation and time They rotate slightly within a subscan During the reduction of a scan with a array receiver The used derotation mode is read out and compared to actually used coordinate system We use the commanded mode So if an observation was taken in the equatorial system and the derotation mode was set to equatorial we take the positions as there are Tests have shown that the error of the taken positions is less then one arcsec for the used on the fly scans If the observation was taken in equatorial coordinates the derotator was set to CABIN mode so the array was fixed during the observation the calibration pipeline takes into account the nasmyth rotation and the paralactic rotation for a receiver located in the nasmyth focus e g CHAMP Create Date May 3 2013 Page 13 Contact author H Hafok APEX APEX Calibration and Data Reduction Manual 3 4 Pointings All Pointing observations are first processed using the science data reduction method described in the previous section This takes care of switched observations and any references and calibrations In case of wobbler switched pointing observations the off phase is treated as a reference If a reference is not present the first 10 of the scan will be us
10. frequency and an offset frequency The big advantage is that no reference position is needed so that one can stay always on source Nevertheless it is possible to observe a reference position which enhances the baseline which is normally rather bad in frequency switch but so the noise is limited to the one of the reference position At APEX both modes are supported Depending on the observing mode with without a reference a switched calibration and the chosen divMode it is possible to reduce frequency switch observations in different ways First I will summarize the online modes In the online case divMode is always set to divideoff Offline it is possible to set the divMode to normal The calibration observation can be done either switched with two phases or in total power divmode divideoff unswitched calibration no reference position Ts Teal GAIN x ONphase1 ONphase2 ON ase divmode divideoff switched calibration no reference position mode Tx Teal phasel Teal phase GAINphase1 GAIN phase2 x ONphase1 ONphase2 ONphase2 25 Create Date May 3 2013 Page 12 Contact author H Hafok APEX APEX Calibration and Data Reduction Manual divmode divideoff switched calibration reference position OF Fphase1 OF Fphase2 Th Teal phase1 GAIN phase1 x ONphase1 OFF phase1 OFF phase1 26 Teal phase2 GAINphase2 x ONphase2 3 OFF phase2 OFFphase2 In the offline mode
11. mode 4 5 1 plotATMspectrum This command plots the results of the atmospheric model for a certain frequency range APEXCAL gt plotATMspectrum minFreq maxFreq Airmass 1 0 Tamb 270 0 Pamb 550 0 Alt 5 098 Hum 20 0 water 1 0 oldATM 0 newATM 1 write 0 numberPoints 100 where minFreq and maxFreq are in GHz The other optional parameters can be used to tailor the model the defaults are shown above If write 1 then it writes the results to an ascii file The routine runs the model between minFreq and maxFreq in numberPoints steps an dplots the atmospheric emission temperature Create Date May 3 2013 Page 23 Contact author H Hafok APEX APEX Calibration and Data Reduction Manual References Archibald E N Jenness T Holland W S et al 2002 MNRAS 336 1 Chamberlain R A Lane A P amp Stark A A 1997 ApJ 476 428 Delgado G Ot rola A Belitsky V et al 1999 The Determination of Precipitable Water Vapour at Llano de Chajnantor from Observations of the 183 GHz Water Line ALMA Memo 271 1 Delgado G Rantakyr F T P rez Beaupuits J P amp Nyman L A 2003 Some Error Sources for the PWV and Path Delay Estimated from 183 GHz Radiometric Measurements at Chajnantor ALMA memo 451 Giovanelli R Darling J Henderson C et al 2001 PASP 113 803 Lampton M L Margon B amp Bowyer S 1976 ApJ 208 177 Mangum J 2002 Load Calibration at Millimeter and Submillimeter Wavelen
12. on an evenly spaced grid Spectral line pointing observations are also possible at the telescope However the fit to determine pointing offsets is not carried out in the apexCalibrator Individual spectra are written to CLASS and the Ipoint class script is used to carry out the fit 3 5 Focus All Focus observations are first processed using the science data reduction method This takes care of switched observations and any references and calibrations However no reference observation is used and Create Date May 3 2013 Page 14 Contact author H Hafok APEX APEX Calibration and Data Reduction Manual so if a calibration is applied the result is in units of the system temperature and not antenna temperature Focus data are stored internally in a dictionary containing focus objects for each FeBe BaseBand combination These are updated with each new focus subscan that is processed from the current scan only subscans within a single scan are used The internal data is automatically reset when reduction of a new scan is started The plot display is updated after each Focus subscan is processed with a fit performed after 3 or more subscans have been processed The focus positions are read from the MONITOR table DFOCUS_X_Y_Z keyword but will eventually be read from the DATAPAR table The normal way of measuring the focus is doing a continuum observation using the APEX pocket backend For spectral line data all channels are summed to produce
13. one can switch to divmode normal unswitched calibration no reference position Tx Tea GAIN x ONphase1 ONphase2 27 divmode normal switched calibration no Reference pos T Teal phase1 GAINphasel x ONphase1 Tealphase2 GAINphase2 x ONphase2 28 3 3 Position Calculation One important issue is the calculation of positions Normally position information is retrieved from the DATAPAR table see MBFits documentation For multi feed pixel receivers we have to take array rotation during an observation into account Here two effects must be mentioned The rotation of the equatorial coordinate system with respect to an azimuth elevation system change of the paralactic angle and in the case for receivers located in the nasmyth foci of the telescope a nasmyth rotation due to differen elevation angle Both effects are calculated in the beamRotation py module which is used to set up the observations and to reduce the observations At APEX an array receiver like champ can be operated in three modes CABIN The array receiver dewar is fixed according to the CABIN coordinate system HORIZ The array is rotated in such a way that it compensates for the nasmyth rotation in the horizontal Az El system The mirror is moved to a fixed position at the beginning of the subscan The pixel offset coordinates true angles are constant in the horizontal system with elevation They change slighty within the subscan EQUA
14. that case the nearest one maybe after the selected ON subscan is selected No interpolation of REF objects is currently carried out but this could be added later The Reducer module is then supplied with On and BestOff In addition calibration information is supplied if a CalEntity exists in the CalDic with matching FeBe BaseBand and reference frequency The On and Off are then divided by integration time to obtain counts per second There are two modes to combine the On and Off These are divMode normal or divMode divideDff In the normal mode If no calibration information for this observation setup present gt ON OFF If CalEntity present gt Toa Xx 0N 0FF GAIN gt Tx where GAIN HOT SKY The result is stored in ARRAYDATA DATA of a new SciEntity The remaining attributes of the SciEntity are filled from the On and CalEntity objects In the divideOff mode If no calibration information for this observation setup present gt ON OFF OFF If a suitable CalEntity is present gt Tea Xx ON O0FF OFF GAIN gt Tx In this mode the calibration observation must also have been processed using divMode divideOff Then the Gain Array is formed by HOT SKY SKY This divMode divide0ff is the default mode to calibrate at the APEX telescope 3 2 Frequency Switch Observations Frequency Switch is an observing mode in which the receiver switches between an observing
15. total power All Feeds are stored but in the online case only the reference feed displayed and fitted The fit is a quadratic to determine the best offset in mm which is returned and stored in the internal focus dictionary Focus data are not written to CLASS 3 6 Skydip Skydip measurements are made by measuring the total power on sky at different elevations The relationship between measured power on sky Pe and airmass A can be determined from equation 14 The atmospheric power in this equation can either be found using the ATM model or by assuming a simple single slab atmosphere with constant temperature e g Chamberlain et al 1997 In the single slab model Pe m 1 exp TA Patm 1 m Pepit 30 where Pas is measured by comparing counts on sky with those on hot and cold loads see equation 15 and Patm and Pspiy contain contributions from both sidebands This results in a single average value for the opacity over both sidebands By default skydips are fitted using this exponential function with the sky opacity forward efficiency and atmospheric temperature as free parameters Spillover temperature is kept fixed with its value set by equation 11 During offline processing several other options are available and these are listed in Table 1 Each of the fit methods makes various assumptions and approximations and may give better results depending on which parameter is required to be determined In the exponential fit the
16. 00 3 2048 in this case 200 integrations 3 feeds and 2048 channels APEXCAL gt print myEntity DATAPAR_INTEGTIM sum lt sum all the integration times to find the total 20 00000000 APEXCAL gt print myEntity ARRAYDATA_DATA type lt find the python type of the data Float64 Calibration information is stored in an extended version of the ObsEntity called a CalEntity The extra information contained in a CalEntity can be printed using APEXCAL gt calEntity printCal The final reduced data are stored in science entities SciEntity which are a combination of the ObsEntity and CalEntity The actual data are contained in the ARRAYDATA_DATA attribute This can be plotted using ppgplot For multi channel data a simple bandpass spectrum can be plotted by specifying the integration number and feed number default values are integNum 1 and feedNum 1 APEXCAL gt obsEntity plotSpectral integNum feedNum Postscript plots can be obtained by adding the desired filename APEXCAL gt obsEntity plotSpectral integNum feedNum psName myPath myFile ps In these plots the x axis is always velocity For calibrated data the y axis is Tx otherwise it is counts On the fly data are plotted as 2D spectra by default It is possible to plot other data with multiple integrations in 2D by using APEXCAL gt obsEntity plotAsOTF feedNum To plot the reduced and calibrated data stored in internal SciDic APEXCAL gt plotSci FeBe
17. H460 PBE_A 1 2 Reset any internally stored reference objects APEXCAL gt resetRef Reset any internally stored calibration objects APEXCAL gt resetCal O 4 2 Working with data entity objects 4 2 1 Displaying information plotting and printing The raw and processed data within the apexCalibrator are stored as Entity Objects each of which contains data from a single subscan The data within one of these objects can be examined using several methods to plot or print its contents A simple method to determine the details of the object attributes names types and sizes is the standard Python print statement APEXCAL gt print obsEntity These attributes and their format shape follow the standard MBFITS definition However this does not print the value of individual attributes that are arrays This must be done individually for each attribute APEXCAL gt print obsEntity ARRAYDATA_DATA Create Date May 3 2013 Page 18 Contact author H Hafok APEX APEX Calibration and Data Reduction Manual An empty entity object can be created using APEXCAL gt myEntity ObsEntityQ The attributes of the ObsEntity are either numarray arrays python lists of strings or single strings This means that standard python numerical methods can be applied to the data This includes methods to examine the contents of an attribute for example APEXCAL gt print myEntity ARRAYDATA DATA shape lt find the shape of the data array 2
18. SciEntity or a Python dictionary of SciEntities If calibration information is not included it will be set to default values in the CLASS file On the fly observations are written as individual spectra rather than using the CLASS OTF data format The current CLASS file is always closed after writing if the above methods are used 4 3 Further details for individual observing modes 4 3 1 Calibration The CalEntities that are currently stored in internal CalDic can be examined using APEXCAL gt showCals 2 internal Calibration Objects stored internal CalDic FLASH810 PBE_A 2 FeBe name FLASH810 PBE_A Baseband 2 Frequency 461 040768GHz Scan 6885 Create Date May 3 2013 Page 20 Contact author H Hafok APEX APEX Calibration and Data Reduction Manual internal CalDic FLASH460 PBE_A 1 FeBe name FLASH460 PBE_A Baseband 1 Frequency 461 040768GHz Scan 6885 The results show the content of the calibration dictionary and the keys needed to retrieve the data In order to allow the effect of the different parameters on the calibration to be investigated the fol lowing method allows one or more parameter to be fixed by hand before reducing a calibration observation APEXCAL gt setCalParam attrName fixedValue The method works by fixing one of the attributes in the CalEntity To fix more than one parameter the method should be called multiple times Some of the attributes that would be reasonable to change are T
19. al 4 3 2 Pointing focus skydip The results of processing pointing focus and skydip observations are written to special objects and stored in internal dictionaries with key names equal to lt FeBe gt lt BaseBand gt These dictionaries and their attributes are internal PointingDic internal FocusDic internal SkydipDic scan scan scan result Az off El off in arcsec result offset in mm result 7 T Tatm previousResult previousResult previousResult amplitude Az El object feed width Az El feed useFeeds Nfeeds object useFeeds Nfeeds power Nsubs Nfeeds Create Date May 3 2013 Page 21 Contact author H Hafok APEX APEX Calibration and Data Reduction Manual feed counts Nsubs Nfeeds dataUnit for power longCounts Nsubsx Nints dataUnit for counts airmass Nsubs longOffsets NsubsxNints degrees offsets Nsubs mm fitType see Table 1 latCounts Nsubsx Nints axis e g Z X TILT calEntity latOffsets NsubsxNints degrees dataUnit for longCounts and latCounts An overview of the data that is currently saved is given by APEXCAL gt showPointing APEXCAL gt showFocus APEXCAL gt showSkydip The internal data storage can be reset using APEXCAL gt resetPointing APEXCAL gt resetFocus APEXCAL gt resetSkydip The ObsEntity method plotContinuum can be used for plotting individual pointing subscans see Sect 4 2 1 but there are also special methods available to plo
20. asymptote at infinite airmass is set by both the forward efficiency and atmospheric emission A small change in one can be compensated by changing the other and their errors are highly correlated This means that it is important to get a good value for Patm in order to accurately determine the efficiency In stable conditions good results can be achieved by allowing it as a free parameter in the fit However the resulting value should be examined to see if it makes qualitative sense before trusting the efficiency this is done by inverting the Planck equation to determine an effective blackbody temperature In order to check the atmospheric temperature it can either be fixed to a reasonable value in the fit e g assuming a blackbody with temperature equal to the measured ambient temperature or the efficiency can be fixed to a reasonable value and the data re fitted with just 2 parameters 7 and Patm This is done in the reduce command using the fixedTatm parameter as a physical Temperature in K e g reduce scanNumer subscanNumber skydipMode exp2free fixedTatm 270 0 In general fixing the atmospheric temperature to the ambient value gives similar forward efficiencies to the fit with Patm free but with a larger scatter when many datasets are analysed This indicates that Tamp is not always a good estimator of Tatm It should be emphasized that this works quite well for frequencies 345 GHz For higher frequencies one sees numerical instabi
21. b Cain 23 equivalent to setting Tea Tamb in equation 1 Note this is for DSB data yet to implement proper treatment for other cases The CSO mode can be selected using calibration mode CSO see Sect 4 1 Frequent observations show that at APEX the CSO mode is only a very rough estimate for the calibration We see deviations of the order of 30 60 Another possibility which leads to much better results is to predefine a receiver gain factor and a physical cold load temperature using setColdLoadParams frontend yfactor tCold Then the Calibrator creates a fake COLD load measurement from the observed HOT Using this schema it is possible to perform Create Date May 3 2013 Page 10 Contact author H Hafok APEX APEX Calibration and Data Reduction Manual a standard calibration reduction using the ATM This is available in the online mode via a CORBA function and in the OFFLINE mode see 4 1 2 5 Water vapour radiometer The measured sky emission temperatures from the water vapour radiometer are included as a monitor point in the MBFITS file For each calibration the precipitable water vapour column PWV is calculated from these measurements using the apexRadiometer software based on the same version of the ATM model as the calibration schema for astronomical measuerments Pardo et al 2001 We use also the same site specific Model input parameters as e g the tropospheric lapse rate As the radiometer measure
22. e 5 1 4 Internalidata storage ee ee A ER eee a eee a ee ESS 6 2 Reduction of calibration observations 7 2 1 Calibration method es s 2 40604 26 422 ee Awe A Ee ee a ee ba 7 2 2 Atmosphiericimodel 00 20206 225 i AY OA Nets ln ee a en E h 9 2 3 Channelwise calibration e a E o a a a a a a a a a a E 10 2 4 Alternative calibration methods a 10 2 5 Water vapour radiometer 2 oa a a 11 3 Reduction of other heterodyne observing modes 11 3 1 Science data reduction toi a A Saag AS a Sd Be a 11 3 2 Frequency Switch Observations o o 12 3 3 Position Galculation 2 0 ae a Byte eho aaa a a ee Boe A 13 3 4 POINCINGS we ete ed ere Ae a ea dae E O Be BA a Ee ee 14 3 5 FOCUS A oe cae ts ant ae eo A We ie ee Mile aya tee i ta da pein et Bh ah aches e e 14 3 6 Skydi prem E SN Be tae ee cal 15 4 Offline Data Reduction 17 4 1 Basic FUNECIONS gt tsio s eee ok be ea Ge A EA ote ck A ol eee the amp amp 17 4 1 1 help methodName oe m eiee ey pee ee Pad A PR ERAS Ae eG A aoe Ree GS 17 4 1 2 setint ractive Ol a sss aa oh tm BB lee ee bP vee A ee 17 4 1 3 setFitsDir directory sarea ba ewe RP ea De OE RM ee ee 17 4 1 4 setColdLoadParams frontend yfactor tCold 17 4 1 5 show and resetira iei GP ae eK Ble eee ee ak oe ee ee AS 17 4 1 6 educa ScanNuim ss a Be eG Rh aoe eed Geek ete en ek da amp E 18 4 2 Working with data entity objects 2 2 0 e 18 4 2 1 Displaying information plotti
23. e command The preset is valid until a real SKY HOT COLD calibration for a given frontend backend combination is processed 4 1 5 show and reset Results of the processing are stored in internal dictionaries The contents of these can be summarised and reset using the following methods APEXCAL gt showCals APEXCAL gt resetCal O APEXCAL gt showRefs APEXCAL gt resetRefs APEXCAL gt showSci contains only data from previous call to reduce APEXCAL gt showPointing APEXCAL gt resetPointing APEXCAL gt showFocus APEXCAL gt resetFocus APEXCAL gt showSkydip APEXCAL gt resetSkydip 2Python dictionaries are used as the basic internal data storage medium This allows many objects to be stored in a single named structure Access to individual data objects is achieved by specifying the dictionary key in square brackets after the dictionary name e g internal SciDic FLASH460 PBE_A 1 where FLASH460 PBE_A 1 is the key giving the FeBe and baseband Create Date May 3 2013 Page 17 Contact author H Hafok APEX APEX Calibration and Data Reduction Manual 4 1 6 reduce ScanNum The simplest way to use the apexCalibrator offline is with the method APEXCAL gt reduce ScanNum In this case the whole scan is reduced looping over all FeBe BaseBand combinations and the result stored in internal SciDic or internal CalDic for CalColds and written to CLASS for everything except Focu
24. ed to subtract the continuum Therefore the result of a pointing observation is always ON OFF OFF If a CalEntity is available the result is Tx Currently only continuum on the fly pointing reduction is possible and line pointing fits must be carried out in CLASS using the lpoint class script For continuum observations the result is fitted with a Gaussian profile plus baseline and the offset width and amplitude are displayed Errors on each of these quantities are also calculated Detailed reduction The scan direction is read from the MBFITS keyword SCANDIR ALAT or ALON located in the DATAPAR table If this keyword is not present the program attempts to determine the scan direction from the offsets in the DATAPAR table Only the Feed listed as the Reference Feed in the FEBEPAR table is used to calculate the pointing correction Offsets are read from the LATOFF and LONGOFF keywords depending on whether direction is ALAT or ALON and stored internally This allows any number of subscans with the same scanning direction to be combined In the online case the subscans are supplied one at a time and after every one the program attempts a fit to the current Azimuth and Elevation data The graphical display is divided into 2 with the latest fit displayed for Az El scans If there are multiple FeBe BaseBands then the data for each are reduced in turn and displayed together in online mode When a new subscan is received with the same scan
25. esults of processing pointing focus and skydip observations are written to special objects see Sect 4 3 2 for details of the attributes These contain only the results of fitting the reduced data and are used for both heterodyne and bolometer processing The CalibController contains several attributes to store the reduced data entities so that they can be used with subsequent subscans In the offline case these appear as attributes of the internal object internal CalDic Calibration information is stored in CalEntity objects These are filled when a Calibration scan is sucessfully reduced and saved to the CalDic which is a Python dictionary with keys equal to lt FeBe gt lt BaseBand gt Only one CalEntity is stored for each FeBe BaseBand combination This means that if the observing frequency for an FeBe BaseBand combination changes a new CalCold must be measured and reduced When Science data is reduced the CalDic is consulted to find a matching CalEntity If none is found then no calibration is applied and the resulting data are saved to CLASS as ON OFF OFF per default In the offline mode it is possible to store the resulting data also as ON OFF internal RefDic References are stored in ObsEntity objects These are stored in the RefDic which is a dictionary of dic tionaries The first level is labelled by lt FeBe gt lt BaseBand gt with each FeBe BaseBand combination containing references labelled with keys REF1 REF2
26. eve the optimum combination of calibrations Ons and Reference subscans There is also the possi bility to refit pointing focus or skydip data e g using different fitting modes for skydips and there are basic tools to plot bandpasses spectra Offline processing is run from a dedicated Python command line interface from which specific methods and help are available provided by the apexOfflineCalibrator This is a user friendly interface to the methods contained in the CalibController module It makes use of the same underlying code as for online reduction ensuring that exactly the same results can be achieved as during the actual observations It is also possible to create reduction scripts using the offline commands Create Date May 3 2013 Page 5 Contact author H Hafok APEX APEX Calibration and Data Reduction Manual 1 4 Internal data storage Raw and reduced data are stored internally as Python entity objects These consist of the data them selves plus relevant methods needed to work on plot the data The basic data object is the ObsEntity This contains attributes equal to all the header and binary table entries in an MBFITS file from SCAN FEBEPAR ARRAYDATA and DATAPAR tables see MBFITS definition Polehampton et al 2005 Calibration data and reduced science data are stored in extended objects which also contain the calibra tion parameters The actual data raw or reduced is always stored in the ARRAYDATA DATA attribute The r
27. gths ALMA memo 434 Pardo J R Cernicharo J amp Serabyn E 2001 IEEE Trans on Antennas and Propagation 49 1683 Pardo J R Serabyn E amp Wiedner M C 2005 Icarus in press Pardo J R Wiener M C Serabyn E et al 2004 ApJSS 153 363 Polehampton E T Hatchell J amp Muders D 2005 Multi Beam FITS Raw Data Format APEX IFD MPI 0002 Ulich B L amp Haas R W 1976 ApJSS 30 247 van der Tak F 2005 The APEX Water Vapour Radiometer APEX report Waters J 1976 Absorption and emission by atmospheric gases in Methods of Experimental Physics M Janssen Ed 12B 142 Create Date May 3 2013 Page 24 Contact author H Hafok
28. hot Tcold Tamb Pamb humidity LapseRate Bandwidth Feff Beff GainImage Tspill note that the units of Bandwith are GHz It is important to preset the parameters with the required numerical type and shape e g Feff is an Numerical array with the size of the number of feeds APEXCAL gt setCalParam Feff Numeric array 0 85 In addition this method can be used to adjust the channel range that is used to calculate the total power on SKY HOT and COLD In this case the command is for example APEXCAL gt setCalParam ChanRange 20 2028 It is possible to preset the precipitable water vapour column PWV using APEXCAL gt setCalParam Water 0 5 If the water value in mm precipetable water vapor is preset the ATM Library is only used to calculate the corresponding opacities To change the frequency at the centre of the band used in the ATM model attrName can be set to Freq_LO to change the frequency of the signal band or to Freq_IF to change the frequency of the image band Note that these frequencies should be entered in GHz The effect of the fixed parameter is limited to the next time a calibration scan is processed and after that is forgotten The changes show up in the attributes of the CalEntity that is produced These attributes can be checked using the printCal method APEXCAL gt internal CalDic FLASH460 PBE_A 1 printC
29. how that the actual value varies during the day Giovanelli et al 2001 quote a value of 1 13 km for the median atmosphere over Chajnantor and Delgado et al 2003 quote an average value of 1 5 km Therefore we have adopted 1 3 km top of atmospheric profile 45 0 km The tropospheric lapse rate is read from the telescope system a value of 6 5 K km is used this is also used to calculate the refraction correction 2 3 Channelwise calibration Internally all relevant properties of the calibration are vectors of the size of the spectral channels of the used spectrometer band In the used calibration algorithm the water although also a vector of the same size is treated as one value for the whole band as we expect the same water for each spectral channel in the band From this value the opacity is calculated for the signal and the image band These are always vectors but there are set to a constant value as a default It is possible to calculate these vectors channelwise by setting the calResolution to atmRes or full In atmRes mode the opacities are calculated in chuncks of 128 channels which is the highest pos sible resolution of the used ATM FORTRAN library In the full mode the opacity is calcu lated channel by channel This part of the algorithm is very time consuming a few minutes to an hour for 16000 channels and therefore only for offline use One can select it with the command setCalM
30. it is much more accurate to remove both the bandpass and atmosphere using hot sky although the atmospheric model is accurate enough to calibrate the overall level across the band it is not accurate enough to remove the small scale spectral shape of the atmospheric emission without very careful optimization The method used to obtain the calibration is based on the emitted power of the sky and calibration loads The counts on sky hot and cold can be defined in terms of the emitted power e g Pardo et al 2005 taking into account both sidebands Csky Klm 1 exp 7 A Pato 1 m Pepin m exp 7 4 P5g1 1 G KG m 1 exp 7 A Pato 1 m Pepin m exp 7 A Pig 1 G K Prec Cort 3 Chot K Pf GP G K Prec Coz 4 Coa K Pia GP ig 1 G K Piec Cor 5 where K proportionality constant between counts and power A airmass Patm power emitted by the atmosphere forward efficiency rear spillover blockage scattering and ohmic efficiency Create Date May 3 2013 Page 7 Contact author H Hafok APEX APEX Calibration and Data Reduction Manual Pspin spillover power received from the rearward hemisphere optics etc Pyg power received from cosmic background at sub mm wavelengths this term is negligible P receiver noise power Co offset in counts before proportionality with power T zenith opacity and G sideband gain ratio image signal band gain ratio The superscript
31. lities It may be not possible to describe the atmosphere with one single temperature in the submillimeter domain This may also explain the following At the JCMT for SCUBA skydips they use a slightly more sophisticated variant of the slab model by setting up an exponential temperature profile for the atmosphere and integrating to determine the Create Date May 3 2013 Page 15 Contact author H Hafok APEX APEX Calibration and Data Reduction Manual Table 1 Skydip fitting modes Fixed parameters are read from the CalEntity object and can therefore be set by hand in the offline reduction using the method setCalParam see Sect 4 3 1 For the 2 parameter exponential fit it is possible to set the atmospheric temperature to a fixed value in the reduce command reduce scanNumer subscanNumber skydipMode exp2free fixedTatm Skydip Mode Fit description Free parameters Fixed parameters Limitations Comments lin Linear T M Tspill Thot Fails if E gt Tanib exp2free 2 parameter exponential fit T m Patm Tamb Tatm could be set in the reduce method exp3free 3 parameter exponential fit 7 m Tatm expfixeff 2 parameter exponential fit 7 Tatm M Tspin expsimple Free exponential fit 4 free parameters only determines T atm3free ATM model fit Tspill m NH20 gives Ts Ti atmfixts ATM model fit m Nupo Tspill gives Ts Ti atmfixeff ATM model fit Nyo Tspills M gives Ts Ti atmfixwater ATM model fit Mm Tspin NHzO T fixed atmfreets ATM m
32. lose in time to the on source observation and the sky measurement is made as part of a load calibration observation At APEX we use a slighly modified formula with a normalization to account for changes over the bandpass with the gain array hot sky sky Con lt lt Ea 2 Ca Tx Tear 2 de i Cref Csky The standard method used to calculate Tea at APEX uses a SKY HOT COLD calibration measure ment with the Atmospheric Transmission at Microwaves model ATM Pardo et al 2001 The first stage of the process is to determine the calibration factor either channelwise or a scalar for the whole band using HOT COLD and SKY total power measurements Each subscan is first summed over the spectral channels as a default we remove 4 of the edge channels to account for edge effects and divided by total integration time to obtain total counts per second over the band These values are used to determine the measured power emitted by the sky which is then used as the input for the atmospheric model to estimate an average water value for the band this is used to calculate the opacities For the scalar mode which is the default for the online case the opacities are set to the same value for each channel Using this the absolute calibration factor is then calculated individually for each channel in the spectrum The gain array cancels from the equation and the bandpass shape is removed by hot cold However in practice
33. meters can be specified to further contrain the processing e g FeBe BaseBand Calibrations are stored internally in Python dictionaries so that they can be applied tho the following scans Pointing focus skydip and calibration information is also stored and the latest value can be retrieved by the online system so that these can be applied to the telescope All observing mode specific decisions and data storage that is independent of the frontend type heterodyne or bolometer is taken care of within the CalibController module Heterodyne and bolometer specific reduction is then carried out by the HeterodyneReducer and BolometerReducer respectively The bolometer reduction makes use of the Bolometer Analysis BoA software Heterodyne processing makes use of the Reducer for science data or Calibrator for calibrations to perform standard reduction 1 2 Online case The interface of the apexCalibrator modules with the main control software is via the apexOnlineCali brator Methods within the apexOnlineCalibrator module call those with the same name within the CalibController The online reduction is started when the scan number and subscan number are sent to the CalibController via apexOnlineCalibrator The sequence of operations for different observing modes are shown in Fig 1 The basic method within the CalibController is Find the corresponding FITS file in MBFITSDirectory with name APEX lt ScanNum gt x fits using getFitsFile
34. ng and printing 0 18 4 2 2 Writingto CLASS a do ee eek 20 4 3 Further details for individual observing modes oaa aaa a 20 4 3 1 Calibration etc da fe A whew ah aoe a 20 4 3 2 Pointing focus skydine s ai e ee BA ew ae Oe a 21 4 4 Reading a FITS file by hand 2 N E ETES A a e E AE E N 22 4 5 Atmospheric model features a 23 4 5 1 PLOCATMS PSC UUM pose Hai sen p IRE o a Gault nes at 23 Create Date May 3 2013 Page 3 Contact author H Hafok APEX APEX Calibration and Data Reduction Manual 1 Introduction 1 1 Calibrator aim The APEX Calibrator is the online and offline reduction pipeline software for observations at the APEX telescope It will read raw data MBFITS files and reduce these depending on the content It will analyse automatically pointing focus skydip calibration and astronomical observations in various modes for heterodyne receivers and bolometers taking proper care of different observing modes and scan types For spectral heterodyne obeservations the calibration pipeline will produce CLASS files containing calibrated data For Bolometer observations the APEX Calibrator will produce quicklook results by using certain BoA routines For the latter the results may differ compared to a precise interactive analysis For the online case the subscan number is also supplied so that the reduction can proceed sequentially for each subscan whilst the scan is observed For offline processing more para
35. number the data are included and the appropriate plot is updated looping through each FeBe BaseBand combination in turn It is not possible to combine data from more than one scan When a new pointing scan is received the internal data are initialised and a new set of correction coefficients is determined The fit results are saved for each FeBe BaseBand combination The results from one previous pointing scan are also saved Repeated subscans in the same scanning direction are combined to improve the fit this is achieved by simply using the points from both in the fit no averaging or rebinning is done Pointing coefficients are determined by fitting a Gaussian with straight baseline to the data ie a least squares fit with 5 free parameters z B C2 F A exp 41n2 D Ex 29 The results for the centre of the Gaussian B are stored for Azimuth and Elevation in a Python dictionary containing pointing objects for each FeBe BaseBand combination This can be accessed by the system via the CalibController method getPointingResults FeBe BaseBand Errors are calculated for each of the fitted parameters by examining Ay as each parameter is varied away from its best fit value re optimising the other parameters at each point see Lampton et al 1976 The RMS in the baseline calculated gt 2 FWHM away from Gaussian centre is used as a measure of the data errors Pointing data are written to CLASS but note that this sets the data
36. nvolve the assumption that Tamb Thot This is a good assumption for the CSO where the hot load is located in the open dome and probably close to ambient temperature but not for APEX where the hot load is inside the cabin and usually hotter than ambient temperature The method used here only works if pe lt Pamb In most of the allowed skydip fitting modes the spillover temperature is kept fixed This is because tests have shown that in general it does not significantly affect the results of the fit although this may not be true at higher frequencies If the observation is made in spectral line mode the total power is calculated by summing over all channels If a valid CalEntity is present the counts on hot and cold load are used to calculate Py or In Pamb P for the linear mode This is entered into an internal dictionary SkydipDic where an array is built up as the subscans at different airmasses are made If there is no valid measurement of hot and cold load only counts on sky are stored The airmass is calculated from the elevation using a curved atmosphere In the online mode a fit is attempted when there are 3 points or more with the best fitting values of efficiency opacity at zenith and atmospheric temperature saved in SkydipDic Only subscans within the current scan are included in the calculation when the scan number changes the skydip object for that FeBe BaseBand is reset to start a new fit Outlying points are ignored f
37. ode calMode newATM calResolution atmRes full scalar 2 4 Alternative calibration methods Several alternative less accurate methods are available to calculate the atmospheric opacity used in the calibration The simplest of these is to assume that the atmosphere is a single slab with constant temperature This makes use of equation 14 with the assumption that the effective blackbody temperature of the atmosphere is equal to the measured ambient value This requires a SK Y HOT COLD calibration measurement to determine Pee For online reduction the value obtained is always printed with the output of the ATM model as a simple check of the opacity For offline reduction it can actually be used in the calculation of Tea by setting the calibration mode to noATM see Sect 4 1 If only counts on SKY and the HOT load are measured and as a first estimate after the SKY and HOT subscans in a regular SKY HOT COLD calibration scan the CSO mode can be used This method is based on the assumption that Thot Tatm Tspin Although at APEX Thot is not equal Tatm Therefore we set Tamb Tatm Tipin If this is substituted into equations 3 4 and 5 the following quantities can be defined Csource Voff K mT source exp 7TA Baek 7 20 CSO Ge Cog Chot Csky K mThot exp 7A C ain 21 gt Csky Csky 22 It is then clear that the antenna temperature on source can be calculated from Coso Tx Tam
38. odel fit Tsp NH20 Mm gives Ts Ti atmfixed ATM model fit none Tsp m Nuzo fixed to calcold values effective temperature Comparison of the results from these skydips shows a good correlation with the opacity measured by the CSO 225 GHz tipping radiometer although they discarded 20 of the results at 350 GHz and 50 of the results at 650 GHz due to bad fits Archibald et al 2002 Alternatively the atmospheric temperature can be set using the atmospheric model In this case the model is used to set up a temperature profile for the atmosphere based on the measured weather conditions The opacity in both signal and image bands is then calculated at the observing frequency and a radiative transfer calculation used to predict TaY for each airmass Forward efficiency and precipitable water vapour Ny 0 are varied to obtain best fit The atmospheric model can be used with varying numbers of free parameters see Table 1 but always takes longer to calculate than the exponential fits because a full model has to be calculated for each point in every iteration of the minimisation routine If a calibration observation has been made and pS can be calculated there is the further option to carry out a linear fit using mode lin This mode assumes that Pamb Patm Pepin in order to simplify equation 30 to linear form In Pamb PFY In 71 Pamb TA 31 Note that this is slightly different to the technique used at the CSO because it does not i
39. or the final fit by clipping data above 20 from an initial fit By default the fit is carried out to the reference feed only Create Date May 3 2013 Page 16 Contact author H Hafok APEX APEX Calibration and Data Reduction Manual 4 Offline Data Reduction 4 1 Basic functions 4 1 1 help methodName Help for all of the methods used in the apexOfflineCalibrator can be obtained at the command line by typing APEXCAL gt help methodName Note without quotation marks General help including a list of all available methods can be obtained using APEXCAL gt help apexcal Note that help with no method name enters the standard Python help environment to exit from this use ctrl D 4 1 2 setInteractive 0l1 This command toggles between interactive mode setInteractive 1 and batch mode setInteractive 0 In interactive mode one has to press return for each fronend backend combination per subscan 4 1 3 setFitsDir directory The raw data directory containing APEX MBFITS files is set by default to be APEXRAWDATA In order to change this use APEXCAL gt setFitsDir directory If directory is not specified it will use the current directory 4 1 4 setColdLoadParams frontend yfactor tCold With this comand it is possible to set a cold load parameter by hand After setting this a fake COLD in a SKY HOT Calibration is created from the HOT measurement using the yfactor and a physical tCold temperature given in th
40. ore Sum Integrations If switched observation Sum Integrations Save to separate and combine phase Sum Channels RawCalDic Sum Integrations Save t A ave to WS Calculate GAIN ARRAY EN If CalEntity else late Su Use CalEntity to get Sum Channels y counts on hot and cold loa Calculate Airmass Separate phases for switched observations Calculate power on sky Calculate sky and reciever temperature Sum Integrations if not OTF J When SKY and HOT present When more than 2 points Use CSO mode Find best referencg Fit with exponential When SKY HOT and COLD present Apply Calibration using best reference Save to Use ATM model to calculate opacity matching Calibration SkydipDic Calculate calibration and system temperature SciEntity containing 7 z Return opacity CalEntity ONOFF EOIN A If spectral line observation RASTER MAP OTF Return forward efficiency and opacity For continuum obs Save to Save bandpasses to CalDic SciDic Save to Combine subscans Sum Channels SciDic from same scan direction Fit with quadrati Write to CLASS Write to CLASS Fit with Gaussian Save to Save to FocusDic PointingDic Return Offset Return Offset Figure 1 Sequence used for calibration of heterodyne data 1 3 Offline In offline mode the user can recreate CLASS files from the raw MBFITS data This may be necessary to achi
41. rom the SKY HOT and COLD measurements The basic method works by optimising the precipitable water vapour column so that the sky emission produced by the model matches the power calculated from the SKY HOT and COLD measurements equation 15 The tolerance in Pmodel Psky is set to 1x10718 W m sr approximately 0 01 of the measured power Both signal and image sidebands are taken into account The model is run using 20 points across the band taken into account existing lines which are then averaged to obtain the final result The optimisation algorithm is the same as the one used at the IRAM 30m telescope and usually converges within 3 or 4 iterations The obtained water value is used to retrieve the final opacity at zenith It is a combination of H20 lines and continuum Oz lines and continuum plus minor contributions from O3 CO and N20 either channelwise or for the whole band The power spectrum of the atmosphere differs significantly from a blackbody and is given by the solution to the radiative transfer equations Pardo et al 2005 However the model can output an effective blackbody temperature at the observing frequency as well as the equivalent power level We use the power level directly Model input parameters Create Date May 3 2013 Page 9 Contact author H Hafok APEX APEX Calibration and Data Reduction Manual scale height of water distribution 1 3 km This has been measured with radiosonde observations which s
42. s PWV every minute the values can be used with the atmospheric model to determine real time variations in opacity at the observing frequency However this can not entirely replace the calcold method as the counts on sky hot and cold loads are still required to calculate Tea equation 18 At the moment the radiometer water vapour is only used as a backup for calibrations when the optimization for the water vapour is not converging which is quite often the case for good atmospheric conditions PWV lt 1mm at 230 GHz In that case the water obtained from the radiometer is used to calculate the opacities in the observed backend band 3 Reduction of other heterodyne observing modes 3 1 Science data reduction Science data reduction involves the treatment of ON and REF subscans in combination with calibration observations to produce a final calibrated product Observing mode specific reduction is carried out in the HeterodyneReducer module with a final on observation On and an off observation BestOff object produced This pair is then passed to the Reducer module which performs a standard subtraction and uses equation 1 to apply the calibration to Tx if possible The detailed procedure is as follows A list of subscan numbers within the current scan is supplied can be one or many If a subscan cannot be processed it will be skipped and will not appear in the final CLASS file A matching CalEntity is selected from CalDic for this
43. s and Skydip A final fit is plotted for Pointing Focus and Skydip observations By default the CLASS file is named with the project ID of the scan and written to the current directory This can be changed using APEXCAL gt setClassName filename directory To restrict the processing several keywords can be defined e g APEXCAL gt reduce 7572 subscan 1 FeBe FLASH460 PBE_A BaseBand 1 The mode used to calibrate the data can be set using APEXCAL gt setDivMode lt can be normal or divide0ff see Section 3 APEXCAL gt setCalMode calmode newATM oldATM noATM calResolution scalar atmRes full see Section 2 The ability to restrict the processing using the keywords in the reduce method means that the processing is very flexible Weird combinations of scan numbers and subscans can be combined e g to process a calcold using different scan numbers APEXCAL gt reduce 7574 subscan 1 FeBe FLASH460 PBE_A BaseBand 1 lt SKY APEXCAL gt reduce 7575 subscan 2 FeBe FLASH460 PBE_A BaseBand 1 HOT APEXCAL gt reduce 7577 subscan 3 FeBe FLASH460 PBE_A BaseBand 1 lt COLD To reduce an ONOFF with ON and OFF in different scans APEXCAL gt reduce 7578 subscan 1 FeBe FLASH460 PBE_A BaseBand 1 OFF APEXCAL gt reduce 7579 subscan 2 FeBe FLASH460 PBE_A BaseBand 1 lt ON gt leads to a final calibrated ObsEntity in internal SciDic FLAS
44. s s and denote power spectra in the signal and image bands These equations are an extension of the standard millimetre submillimetre method which approximates the received power by equivalent Rayleigh Jeans brightness temperatures e g Mangum 2002 When an astronomical source is introduced into the beam with a spectral line located in the signal band the counts on source can be defined as Csource Usky UN K m Psource exp 7 A 6 The power spectrum of the hot and cold loads can be considered to follow the Planck distribution for a blackbody with constant temperature This allows the power in signal and image bands to be calculated from their measured temperatures 2hv3 c B T _____ A ups 7 B Troaa V GB Toaa v Proad Frond v CB Bont 1 8 To match with standard radio astronomical definitions the source power and receiver noise must then be converted to brightness temperatures using the Rayleigh Jeans approximation T P v 9 BY P 9 However it is important to make this conversion only once the final power has been calculated because in the submillimetre and teraHz domain the Rayleigh Jeans approximation breaks down and the relationship between power and brightness temperature is not linear Equations 3 4 and 5 can then be combined to determine the absolute calibration factor Tea The method used is as follows Sum over channels to obtain total counts per second Chot Ccola C
45. sky per default ignoring first and last 4 of channels in offline mode it is possible to give a validity range in channels Estimate spillover power using a empirical determined formula for the flash receiver which is in fact receiver depended due to different optics and losses Pai v 0 2B Thot v 0 8B Tamb v 10 Popin vt GPapin v Popin 14G 2 11 Calculate receiver noise power and convert to brightness temperature P co C ot R Oo Coo p ldChot hot Ccold 12 Coola Chot A receiver dark signal must be nulled during hardware comissioning otherwise Cog then Prec will show an offset of Cog Phot Peoia Chot Ceold C2 Tree Precio 2ku2 13 where y v Gr 1 G Calculate sky temperature PX m 1 exp 7 A Pom Gm 1 exp 7 A Phn 1 G 1 m Papin 14 Chot _ Cex PY Ph gat Pi Poa ee 15 aot pu 1a 2 s Tiy PA 1 Pepin 16 Create Date May 3 2013 Page 8 Contact author H Hafok APEX APEX Calibration and Data Reduction Manual Calculate airmass from measured elevation in SKY measurement using curved atmosphere Use the ATM model to determine the water content of the atmosphere taking the sideband gain ratio into account Calculate opacity in signal and image bands either as a single value for the whole band set CalMode calResolution scalar or channelwise set CalMode calRe
46. solution atmRes full which is very time consuming Calculate absolute calibration factor For spectral line observation with line in only the signal band Phot Pa G A Teal 17 o m exp TsA 2kv Se For continuum observation assuming source power is equal in signal and image bands Phot PAY HG 2 R a m exp T A Gm exp 7 A 2kv where v v Gv 1 G Calculate system temperature accounting for both sidebands Csk Ts s tea lt 19 y i E EVIR ve where Teal is defined by equation 18 The airmass calculation uses a curved atmosphere using an Earth radius of 6370 km and atmospheric height of 5 5 km The receiver dark signal Cog only affects the receiver and system temperatures as it cancels from the other equations Currently no correction for Tre Or Tsys is implemented if a measured value of Cog exists This must be adjusted during hardware installation and commissioning 2 2 Atmospheric model The atmospheric model ATM Pardo et al 2001 is run using a Python wrapper to access 4 FORTRAN subroutines We use a version of ATM from 2003 the current version of ATM being developed for ALMA uses a C interface The 4 FORTRAN subroutines were written by J Pardo and D Muders specially for APEX and give access to the main modules of ATM for which the source code is not freely available The ATM model is used to determine the average sky opacity across the band f
47. t the results that have already been calculated and are stored in the internal dictionaries described above These are APEXCAL gt plotPointing FeBe BaseBand APEXCAL gt plotFocus FeBe BaseBand APEXCAL gt plotSkydip FeBe BaseBand Pointing data are only stored for the reference feed but for focus and skydip the data for other feeds can be accessed using the feedNum keyword in these methods feedNum Ref for reference feed For skydips the fitting method can be changed using the fitType keyword However the fit type must be compatible with the original processing of the data for example if the data were not originally processed with the linear fit option selected a linear fit cannot be performed here The possible values for fitType are given in Table 1 The additional keywords for multiple plot panels described in Sect 4 2 1 are also accepted by these methods and hardcopy can be obtained by using psName filename For skydips there is an additional option to overplot results from the different fitting methods using the keyword oplotFit The fits to overplot should include renew 0 oplotFit 1 and numPlots should be set to zero until the final call when numPlots 1 will close the ppgplot device For example APEXCAL gt plotSkydip FeBe BaseBand numPlots 0 renew 1 oplotFit 0 fitType gt exp3free APEXCAL gt plotSkydip FeBe BaseBand numPlots 0 renew 0 oplotFit 1 fitType gt exp2free
48. tion observations 2 1 Calibration method The calibration of heterodyne spectral line and continuum data is carried out using an extension of the chopper wheel method used for millimetre observations Ulich amp Haas 1976 This involves using the difference in temperature between one or two standard loads and the blank sky to calibrate the absolute temperature scale of the data and remove the spectral variations of the atmosphere across the bandpass The following sections describe how these observations are processed to determine the calibration factors The method is based on the following equation to obtain the observed source antenna brightness temperature corrected for atmospheric attenuation radiative loss and rearward scattering and spillover Tx Teal Con Cref Chot jid Csky 1 This contains 3 steps 1 subtract the atmosphere and receiver offsets per channel using the difference between receiver counts on source and on a reference position Con Cref 2 divide out the spectral shape of the bandpass and atmosphere using the difference between counts on a hot load and the blank sky Chot Csky and 3 correct the absolute temperature level using a single calibration factor Tca1 The quantity hot sky is known as the gain array This calibration assumes that the spectral shape of the atmospheric emission does not change significantly between the measurements of Cref and Csky where the ref measurement is made c
Download Pdf Manuals
Related Search