Why the cals were remeasured.
The model used to compute the cal values
The parameter values used in the computation
Taking the data
Processing the data
Results
    rms/mean spectra.
     calRatio (caldif/caloff) vs freq
     Computed Tcal
     compare Tcal with 2" shims and Tcal with 4"shims
Summary

Why the cals were remeasured.

Parameters used for Tcal computation

    The values i ended up using for the computations (see below for explanations) were:
 

	value	Notes
Tabsorber	297.8K	The tabs was 75.3F when we started and 78.1F when we finished.mp.
Tsky	5k-11K	From paul Goldsmiths memo. 5K at 1GHz, 11K at 12GHz using a 2nd order fit. These values should be for zenith. I made no correction of the elevation
Tscatterd	10K	best guess. brought Tsys Yfactor and Tsys abs closer together.
Treceiver	20k	from values measured at the test range.
g	.0	horn cooled
alpha	0	horn cooled
misc

 Taking the data:

• We used 4 filter settings to measure the cal
• wb3:2775-3475 MHz
• covered 2775 to 3475 MHz
  
• wb1:2500-14000 straight thru mode
• covered 3470 to 4600 MHz
  
• wb5: 4600 MHz hipass
• covered 4600 MHz to 12000 MHz
  
• wb6: 8050- 9200 bandpass  
  
• covered 8050 to 9200 MHz
• This redid the 8050-9200 MHz that was done with wb5.
  
• The straight thru (wb1) and 4600hipass (wb5) have problems with
• wb1 - saturation
• wb5 - we can't include enough gain at the hi freq end (or we saturate). So the freq about say 9GHz is adding unwanted noise to Tsys (orders of 10's of K)
• So the valid cal values will be for wb3 and wb6
  
• the mock spectrometer was used to take the data
• 7x172MHz  bands.
  
• 1024 channels (should have been 8192 by i screwed up
• integration time of .39 seconds per dump (should have been 1 sec but the 16 bit integration counter iimited this.
• polA and polB recorded.
  
• The following procedure was done for on sky and on abs.
• set the receiver to use wb3 filter
• use 5 mock boxes with cfr: 2846  2988  3130  3272  3414 MHz
• calOn 3 secs, cal Off 3 seconds, loop three times
• set the receiver to wb1 straight thru
• use 7 mock boxes.
• cover 3500 to 4600 MHz
• calOn 3 secs cal off 3 secs, loop 3
  
• set the receiver to wb5 hipass
• take 8 sets of data with 7 mocks each covering 4600 to 12000
• calOn 3 secs cal off 3 secs, loop 3 for each set.
• set the receiver to wb6 bandpass
• use up to 7 mock boxes to cover 8050 to 9200 MHz
• calOn 3 secs cal off 3 secs, loop 3 for each set.
• this is covering part of the same freq range done with wb5.
• At each setting the attenuation was set to give around 30cnts for 1 sigma in the mocks
  

• The scan sequence was:

• Loc	ra,dec	az,el	mockFile #	Notes
  abs	..		39100-46800
  sky	07:43:40/25:50:05	(320-306,68-62)	46900-54600

Processing the data:

    For  each 3calon,3caloff's at a given freq range

• input the data for the n mock bands used.
  
• separate the calon from caloff
• compute the median spectra for the 3 sec x 3 loop data for cal on,caloff
• compute the rms/mean by mock band for the 3secx3loop data for calon, caloff.
• compute the ratio calOnSpectra / calOffspectra (after the average)
• 
  
• Input the winking cal spectra and compute the ratio of calDeflection(freq)/Caloff(freq)
  
• Average the 20 ms of calon, caloff, giving bCalOn[25*120], and bcalOff[25*120]
  
• exclude the 2 ms spectra before and after each cal transition. This gives 16ms of data after the average
• Compute the rms by chan for the bCalon and bCalOff data (for debugging)
  
• Compute the  avg calOn, calOff: bonAvg,boffAvg using the median for averaging
• compute the caldeflection and then the calRatio (by Tsys)
• bcalDef= bcalOn - bcalOff
• bcalRatio=bcalDefl/ bcaloff
• this removes the IF bandpass.
  
• This is then the cal value in units of Tsys (with cal off).
• interpolate the calratio across the entire band to a fixed 167khz channel width
• do this for wb3,wb1,wb5
  
• and separately for wb6
  
• Compute the cal In kelvins using the Tatm,Tscat, Trcv, Tabs values above
• Do this using the yfactor method:
• calRcold, calRhot is the calDeflection / hot-cold loade
• prodR=calRhot*calRcold
• diffR=calRCold - calRhot
  
• calK=(prodR/diffR *( 1-alpha*g2*(thot - tcold)
• g2=1 - g*g =1 since g set to 0
• alpha set to0
• Also compute calK using just the absorber.
  
• calK=calRhot * Thot
• To generate the cal lookup table
• smooth the computed calK to 10MHz and output the values.
  

The Results

rms/mean for 3sec*3loops of data

    The rms/mean for the measured data (.png)

• The rms/mean is computed for each 3sec*3 loops of data.
• This is done for the data on abs and on sky
• white is PolA, red is PolB
• The top frame is on absorber, the bottom frame is on sky
• the yellow vertical lines show where the filters change.
• the green line is the expected rms =1/sqrt(bw*tau) 
  
• this gives a good idea of where the rfi is
• Even on absorber we get rfi leakage into the receiver.
  
• the 3.5 to 4.5 elevated values may also come from the straight thru mode going into saturation.
  
• outside of the rfi ,the rms/mean is close to the expected value.
  

Calratio (calDif/calOff) vs frequency

    The 2nd image shows the cal ratio (calOn-calOff)/caloff (.png)

• top frame: on absorber
• bottom frame: on sky
• White is polA
• red is polb
• green, purple: polA,B  measured ratio using the wb6 filter (instead of the wb5 hipass filter)
• The yellow lines separate the filters.
• 2800-3200mhz
• on sky 50% drop  in ratio. not sure whether this is Tsky or Tcal deflection.
  
• this ends up being a 10% calK drop after computing the cal value
• 8050-9200
• the green, purple ratio computed with the xb6 filter has a ratio 10 to 20% higher than that measured with wb5 (the hipass filter)
• This is a problem with the high pass filter. the wide band does not allow us to put in enough gain at the higher freq.
• So the if/lo after the filter is adding 10 to 20K at the higher freq to Tsys
  
• A larger Tsys gives a smaller ratio.

cal value

    The plots show the computed cal values in degK (.ps)  (.pdf)

•  Page 1: calK vs freq for entire band
• Top polA, bottom PolB
• Black: Tcal using the yfactor and filters: wb3,wb1,wb5
• Red   : Tcal using the absorber and filters wb3,wb1,wb5
• green: Tcal using yfactor and wb6 (xband filter)
• blue :   Tcal using absorber and wb6.
• the purple lines show where the filter edges sit.
• If everything worked correctly, the different color lines should overlap one another.
• The black,red diversion above 9 GHz may  be from extra rcvTemp added after the filters.
• Page 2:  3GHz cal values
• top polA, bottom polb
• black: using yfactor to computer calK, red uses absorber. The green line is the 10MHz smoothing and interpolation.
• 3200 has a large bump, probably from rfi.
• the bump is larger for yfactor (which includes the sky)
• Page 3: xband cal values
• top polA, bottom polB
• black: from yfactor, red from absorber, green smoothed and interpolated to 10MHz.
• The ripples are probably coming from the cal injection (we saw this on the s parameters).

Comparing Tcal with 2" and 4" shims

    The Tcal values measured on 30may23 with 2" shims are compared to the 20jun23 4" shim values (.ps) (.pdf)

• Black: Tcal measured 30may23 with 2" shims
• Red: Tcal measured 20jun23 with 4" shims
• Page 1: 3Ghz cals values
• frame 1-2: polA Tcal using Yfactor and Absorber
• Frame 3-4: polB Tcal using Yfactor and Absorber
• Page 2: 8 Gh cal values
  

Summary:

• the 12m cal values where measured on 20jun23 using sky and absorber for the cold and hot loads.
• the horn had 4 1inch shims installed.
• The 2750 -3450 cal values were computed and will be used with filters wb3 and wb4
• the  8050-9200 MHz cal values were computed and will be used with the wb6 filter
• The cal  values outside this freq range could not be accurately measured with the current filter configurations.
• the cal values used are those from the yfactor computation
• They were installed 27jul23 and backdated to be used by any data taken after 20jun23.
• the cal values are installed in /share/megs/phil/svn/idl/data/
• cal12m.datR23  wb3 values
• cal12m.datR24  wb4 same as wb3
• cal12m.datR26   wb6 xband values
• cal12m.datR20   wb6 xband values (for routines that have been updated to no rcvr names)
  

12 meter page
home_~phil