Links:
     intro
    wc background
    wc find files
    wc input and process
    wc example
    wc notes

Other pages
Why use a winking cal on the 12meter.
The mapradec tcl routine used to take the data
datataking example: winking cal map of SGRA* at xband

---

Intro.

    The data should  have been taken with the tcl routine mapradec() with the code variable having the winking cal bit set (0x100). The hardware winking cal has a 25 Hz cycle and is hardware synchronized to the 1 second tick.

The processing described below uses idl.  It is broken up into two steps:
1. find all the files that belong to the map
2. input the map, and compute total power.

Note:

• The current winking cal is 30-40K with an xband Tsys of around 110K
• If you average the CalON, calOff, your tsys will go 100 to 120K
• If you only use the calOff, the deltaT/T will go down by  sqrt(2)
• If you have 5 secs/bm (4.secs of data with  the 4 dropped samples per cal cycle), and assume 150Mhz bw (ignore edges of band).
  
• the rms is a  little better if you used Calon and calOff, than if you just used calOff.
  

Some background on the map and the terminology used.

    The map will be made of multiple strips at constant dec and driven in ra. Each strip will be a separate scan. All of the scans that belong
to the map are called a pattern  (we also have on,off position switch patterns, etc..). The first scan number of the map is used as an identifier
for all of the scans in the map. It is called the pattern_id. It is included in the header of every scan of the map.

    There can be up to 7 separate freq bands (also called beams.. a holdover from alfa) included in each strip. Each of these bands
is stored in a separate file (but they all have the same scan number since they were recorded at the same time). So a map of
25 strips, and 3 bands will have 3x25= 75 files.

Find the files that belong to your map

    You need the following info:

•  project name
• date of observation (yyyymmdd AST)
  
• warning this currently does not span ast midnite very gracefully.. (future fix:)
• pattern_id.
• The pattern_id is the first scan number for the map (1st strip)
  
• You can find this in the log file. It will look something like:
• 2022-04-10_02:13:11 MAPRADEC
  2022-04-10_02:13:11 src: ra,dec:174540.04 -290028.20 rate:119.67 strip1:1 numStrips:25 code:0x180
  2022-04-10_02:13:48 mockadjpwr:
  b0s1g0  rms:A  30.3   28.8 B  29.4   30.8 mn:A   0.7    0.7 B   1.0   -1.6 gn:  6  5
   b1s1g0  rms:A  31.2   30.2 B  30.0   28.7 mn:A  -5.2   10.8 B  -7.7   -2.2 gn:  1  1
   b2s1g0  rms:A  30.1   30.1 B  31.4   29.1 mn:A   0.4    0.5 B  -0.4    0.1 gn:  9  5
   b3s1g0  rms:A  32.0   31.3 B  30.0   31.1 mn:A  -0.2   -0.2 B  -0.6   -0.2 gn:  4  3
   b4s1g0  rms:A  32.1   29.8 B  30.4   29.5 mn:A   0.4    2.1 B  -0.1   -1.0 gn: 11 11
   b5s1g0  rms:A  31.8   31.4 B  28.6   28.3 mn:A   1.0    0.8 B  -0.9    1.5 gn: 10 10
   b6s1g0  rms:A  33.3   29.4 B  31.2   31.2 mn:A  -0.6   -0.5 B  -0.7    0.5 gn: 14 15
  2022-04-10_02:13:50   1 210000001  offsets:  -437.150 -10000.000 start:21354
  2022-04-10_02:13:51 Mock scan starting filenumber:00000
  2022-04-10_02:13:54 Observation begins Sun Apr 10 02:13:54 2022. scan:210000001
  In this example the pattern  id is 210000001

    Scan for the files using the idl routine masfindmap()

• I'm using a test map made on 10apr22
• yyyymmdd=20220410
• proj='x101'
• patId=210000001
• nstrips=masfindmap(proj,yyyymmdd,patId,mFI,band=1,grp=0)
• (band=1,grp=0 is not needed, it just speeds up the scanning)
• The File info is returned in mFI (map FileInfo) structure.
• help,mFI,/st
  
•   PATTERN_ID      LONG         210000001
   SRC             STRING    'SGRA'
   NSTRIPS         LONG                25
   NBMS            LONG                 7
   CALPERSTRIP     INT              0    ; there is not separate cal at start,end of each strip.
   STRIPCODE       INT              2    ;2 -> this uses the winking cal
   STRIPI          STRUCT    -> <Anonymous> Array[25]

   mFI will then be passed to the processing routine.

Input the files and compute the total power.

    The routine masgetmapwc() is used to input the data. The call is:

• nstrips=masgetmapwc(mFI,mI, edgeFract=edgeFract, median=median,degGenVar=degGainVar,verb=verb)

    The processing steps done for each strip and freq band are:

• input the files for this strip (masgethwcal()).
  
• split the calon,calOff spectra.
• when splitting, exclude spectra adjacent to each  cal transition.
• the 25 Hz cal has calOn for 20ms and calOff for 20ms (1/40ms = 25hz)
• If you integrated spectra for 2ms each then there are 10 spectra in calon and 10 spectra in caloff.
• excluding spectra next to each transition leaves you with 8calOn, 8 calOff spectra.
• Average the spectra for each cal transition
  
• this leaves you with 25 calon,25calOff per second. Using the example above they are 8*2=16ms averages.
• Compute the frequency channels to include in the totalPwr computation.
• compute the rms/mean by channel for the calOn and calOff spectra over the entire strip.
• Do a robust linear fit to the rms/mean by channel. Exclude all points that are 3 sigma from the fit.
• Also exclude all points to the left and right of edgeFract. So we don't include the steep fall off of the filters.
• For a freq channel to be included it must pass the test for the calOn,calOff, polA, and polB spectra.
  
• Compute the calTotPwr deflection for each 40ms cal cycle.
  
• Compute the calDeflection spectra (calOn - calOff) for each 40ms cal cycle
  
• Compute the total power calDeflection using the good freq channels
  
• Fit a polynomial of order degGainVar  (the default is a 3rd order fit).
•  Compute the average measured cal value over freq channels used.
• Compute the totalPower for each calOn,calOff spectra using the good channels.
• Scale to Kelvins
• let Scl=(the average calValue/( by the fit to the Caldeflection)
• scl the computed tpCalOn,tpCalOff by this scale factor.
• We divide each 40ms cal cycle total power by the 40ms fit to the calDeflection. This should help correct for any electronic gain variations after the cal injection (assuming the fit is good).

    Return the results in the mI structure (MapInfo):

    help, mI,/st
    NSTRIPS         LONG                 25           ; number of strips
   NRECSSTRIP      LONG                60           ; number of rows in file (may have multiple samples/row)
   NBMS            LONG                 7           ; number of beams,freqbands
   NCALCYCLESSTRIP LONG              1500           ; number of total power samples (calcycles) in strip)
   HDRAR           STRUCT    -> MASFHDR Array[60, 25];fits header from bm[0] for each row
   BMI             STRUCT    -> <Anonymous> Array[7] ; this holds the beaminfo,total power samples.
   RACEND          DOUBLE           266.41683       ; center ra for map (degrees)
   DECCEND         DOUBLE          -29.007833       ; center dec for map (degrees)
   RAPOSD          DOUBLE    Array[1500, 25]        ; ra position (deg) for each sample of map
   DECPOSD         DOUBLE    Array[1500, 25]        ; deg position (deg) for each sample of map

    The mI.bmI struct has info for each beam:

    help,mI.bmI,/st

   TPARONK         FLOAT     Array[1500, 25, 2]       ;totPwr calOn. samples/strip, strips,pols(a,b)
   TPAROFFK        FLOAT     Array[1500, 25, 2]       ;totPwr calOff.   ''
   BRMSAR          STRUCT    -> <Anonymous> Array[2, 25]; output masrms for calOn,caloff data this strip
   MASKAR          INT       Array[512, 25]            ; mask for good freq channels 1-> used, 0-> notused (nchan,nstrips)
   EDGEFRACT       FLOAT     Array[2]                  ; fract spectra excluded left,right edge of spectra.
   NCHANUSED       LONG      Array[25]                 ; number of freq chan used each strip
   BWUSED          FLOAT     Array[25]                 ; bwused (MHz) each strip (nchnUsed*chanWidth)
   CNTSTOK         FLOAT     Array[1500, 25, 2]        ; conversion spectrometer counts to degK for each sample
   CALKUSED        FLOAT     Array[25, 2]              ; avg cal value used each strip (nstrips,npol)

Example:

    map of Sagitarius A* taken 10apr22 using xband with 7 172.032 Mhz input bands.

In idl:

proj='x101'
yyyymmdd=20220410
patId=210000001

band=1
grp=0
;
;  find the files
;
    nstrips=masfindmap(proj,yyyymmdd,patid,mFI,band=band,grp=grp)
;
; input and process
;
    nstrips=masgetmapwc(mFI,mI,/verb)
;
; look at total power polA all strips bm 0 for calOn and calOff
;
    nsmp=mI.ncalCyclesStrip
    x=findgen(nsmp)
    ibm=0
    ipol=0
    !p.multi=[0,1,2]     ; 2 frames/page

    hor
    ver,140,150
    stripsxy,x,mI.bmI[ibm].tpArOnK[*,*,ipol],0,0,/step,$
        xtit='smp',ytit='DegK',tit='10apr22 sgra*  totPwr vs smp CalOn polA'

    ver,95,105
    stripsxy,x,mI.bm[ibm].tpAroOffK[*,*,ipol],0,0,/step,$
      xtit='smp',ytit='DegK',tit='10apr22 sgra*  totPwr vs smp CalOff polA'

Notes:

• The total power strip data is in the order it was taken so each strip starts at the opposite ra side.
• If you want the data in correct ra order, plot it vs the mI.RaPosd[*,istrip]
• No baseline removal has been done

---