Intro

    This page has the 12meter sun observations and  tests to make sure things are linear on the sun.

• 07jul21: track sun for 1 hour with 2ms sampling. saw a micro burst. 
  
• 01nov21: measured dynamic range in udc while tracking the sun. polB compresses
• 09nov21: added 15db pad polA, 20db pad polB in front of udc, so it will no longer compress on sun
• 18nov21: make a map of the sun. Look at strips through the center looking for compression
  
• 22nov21: Measured power levels on, off sun in pedestal and fiber
  
• Current power settings used for sun:
  
• 24Nov21: see if the downstairs if/lo and mock spectrometers compress while on the sun.
• 06aug22:look at ratio of caldeflections for each sunmap to check electronic gain stability
  

---

Current settings to use on the sun

30nov21:current values I'm using. may not guarantee no compression.

• attn udc 10 10
• attn if2 0 0
• mockrms 100 counts

---

18nov21: make a map of the sun

    On 18nov21 2 maps of the sun were made  with the 12meter and the mock spectrometers.
The setup was:

• Use jpl ephemeris to track the sun
• Use mapradec  to take the data
• +/- 1deg map
• drives in ra at 2 arcmin/sec (beam is 10 arcmin).  data sampled once a second
• step in dec 5 arcm/step (1/2 a beam)
• End up with data array of 60x25
• If/lo setup:
• udc attn 10 10
• if2 attn 0 0 
  
• if2 pwrmeter on sun .. -28dbm
  
• mock rms set to 100 counts a/d on sun.
• 7 freq bands used: 8219,8363,8505,8647,8789,8931,9073MHz
• mock setup:
• 7x172 MHz bands, 8192 channels, 1 second integration
• rms set to 100 counts while on sun

Processing the data:

• input spectra
• compute total power excluding 6% from each each of a band pass.
• save the header Info for each sample of the map.

Looking at strips through the center of the sun:

Map 1 strips in ra and strips in dec through the sun (.ps) (.pdf)
Map 2 strips in ra and strips in dec through the sun (.ps) (.pdf)

• The strips were 2 degrees long in ra. sampled every 2 arc minutes (the sun is 30 arc minutes across)
• Each adjacent strip then stepped 1/2 beam:(5 arcminutes) in dec.
  
• I've plotted +/- .6 degree about the center
• Page 1: 8219 MHz band. plot central and adjacent strips.
• Pol A and B have been averaged.
• All strips have been normalized to the minimum value in the 2x2 deg map.
  
• Top: plot ra strips
• black: through the center
• red: .08 deg above(+), below(*) the center of the sun
• green: .17 deg above/below the center
• blue: .25 deg above/below the center
• the central strip (black) shows a dip at the middle of the sun
• bottom: plot strips in declination
• black: through the center of the sun
• red: .08 deg before,after the center of the sun
• green: .18 deg before/after the center of the sun
• blue :   .25 degrees before/after the center of the sun.
• Page 2: plot the ra strip through then center for 7 freq bands, and polA,B
  
• PolA, B are plotted separately
• There are 7 freq bands. each plotted in a separate color
• Each strip is normalized to its edge values.
• Top: PolA
• Bottom:PolB
• The dip in the center is only seen for the lower freq bands (8219,8363 MHz)
• It is more prominent in polB.
• The polA,B levels appear different.
  
• This could be caused by:
• TsysA is different than TsysB
• The sun power on the edges (+/- 1 deg) is circularly polarized.
  

Summary:

• a 2 deg x 2deg map was map in ra,dec about the center of the sun.
  
• data was sampled (60x25) in ra x dec.
• The central strips showed a dip in the middle
  
• It was seen in the 2 lowest freq bands.
  
• The dip was larger in polB then polA.
• The polA,B levels at center were different
• This could be from:
• TsysA != TsysB or
• The sun power +/- 1deg from the sun is circularly polarized.
• If  the cal was accurately calibrated, we might be able to solve this.
  
• Looking at the power levels on the sun, it is not obvious why it should be compressing.
• 22nov21: Measured power levels on, off sun in pedestal and fiber
• 24Nov21: see if the downstairs if/lo and mock spectrometers compress while on the sun.

processing: x101/211118/stripssun.pro

---

24nov21:does the if/lo, mocks compress while on the sun?

    On 24Nov21 a test was done to see if the control room if/lo and mock spectrometers compress while tracking the sun. The udc was not tested (see previous udc measurements).

The setup was:

• track the center of the sun
• setup the xband receiver
• udc:
• cfr: 8646, attn 10.0 10.0 db
• if/lo ctrl room:
• attn if2  0  0 db.. then step 0 to 11 db and back to 0 in 1 db steps
• Mocks
• 4x172MHz bands with cfr: (8219., 8363., 8505., 8647.) MHz
• 8192 channels, 2 pols, 1 second sampling.
• Observing Sequence:
• track source
• set if2 attn 0 0 db attenuation
• Set the mock rms level
• this was done by adjusting the attenuator in the mock mixer chassis.
  
• if2mp.. measure the if2 power
• start taking mock data
• start cycle
  
• 5 seconds after setting the attenuator, read the if2 power meter and record it.
• After 10 seconds of data  with an attenuator setting, increase the attenuation by 1db.
• cycle going 0 to 11 db attenuation
• then return going from 11 to 0 db
• stop the mocks

The table shows the 3 sequences that were done

source	mockRms with attn if2 0 0	if2pwr attn 0,0 polA , polB
off sun	50 a/d counts	-46.25, -48.15 dbm
On sun	100 A/D counts	-28.41, -29.71 dbm
On sun	200 A/D counts	-28.23 , -29.74 dbm

Previous  sun measurements used mock rms= 100counts and if2pwrMeter=-28dbm.

Processing the data

    We want a single total power values for each attenuator setting by pol and freq band.

• The dc spike in the spectra was interpolated across.
• The average of the channels was computed.
• the inner 8 seconds of each 10 second attn step was averaged.
• For each sequence the total power series was normalized to the tp measured with 11 db attenuation.

Looking at the compression.

    If the system if/lo or mocks compressed, then the measured power step  should be less than 1db as the power increased (the attenuation decreased).

•  This assumes that attenuator steps were exactly 1 db
• the if2 attenuation use a 1db,2db, 4db, and 4db set off attenuators for polA and B.
  
• By combining these, you can step from 0 to 11 db attenuation.
• Unfortunately we'll see that these steps are not all exactly 1db.
• The input power level should be constant.
• we tracked blank sky and the sun. the input power was not guaranteed to be stable.

The first plot show the power vs 1db steps (.ps) (.pdf)

• top frame is polA, bottom frame is polB
• Each color is the total power from a different frequency band.
  
• Page 1,2 On sun, mock rms 100 counts
• Page 1 measured power vs attenuator setting
• Page 2: Linear fit - Measured value vs attenuation value
• a linear fit was done: ydb(atten)=c0 + c1*attenSeting
  
• the fit was done to atten settings 6 to 11db (in case the lower atten settings were saturating).
  
• The difference  yfitdb - yMeasuredDb  was plotted vs attenuator setting.
• there is a slight decrease as we approach 0db atten.
  
• the plots repeat since we went 0db -> 11db and then back to 0db atten.
• There is a jump of about .2 db for 7db when we were going 11db back to 0db
• Page 3,4 on sun, mock rms 200 counts.
  
• Page 3 measured power vs atten setting
• Page 4:  linear fit - measured Power vs atten setting.
• you can see a decrease going to 5db to 0 db.
• Page 5,6: off sun, mock rms 50 counts
• Page 5 measured power vs atten setting.
• Page 6: linear fit - measured Power vs atten setting.
• Still see a decrease in step toward 0db atten.

This showed that the attenuator steps were probably not exactly 1db.
PolB showed a decrease in step size as the attenuation approached 0db, but it showed this even off the sun when the power  was 20db lower than on the sun.

Compare Each 1 db step for the 3 mock rms levels.

• If there is compression in the mocks then the 1db step closer to no attenuation should show more compression.
  
• You should also see a difference in compression  for the 1db to 0db atten step when you compare the rms's 50,100,200 counts (since the   200 count rms has 16 time more power than the 50rms measurement.

 The plots compare the 1db atten steps for the 3 different rms settings. (.ps)   (.pdf)

• The top frame is polA, the bottom frame is polB
• Each color is data from a different mock rms setting:
• black: rms= 50 counts, off sun
• red   : rms=100 counts, on sun
• green: rms=200 counts,on sun
• The y values in totPwr(attnN-1) - totPwr(attnN)    vs attnN-1
• So x=0 shows the power difference going from attn 0 to atten 1 db.
• If compression is occurring then the green (200 rms) values should be lower than the red (100rms) and black(50rms) measurements.
• especially when the atten value is close to 0.
• This is not seen, so the mocks (and probably the if/lo) is not compressing up to a mock rms of 200 counts and an if2pwrmeter reading of -28dbm.
• We do see a lot more scatter in the values at a particular atten step for polb.
• This is probably not the attenuator step difference (since a particular step uses the same atten combination for each rms setting).
• This may be variation in polB itself.

Summary:

• We looked at whether the if/lo and mocks compressed while data was taken with power levels when on the sun
• The if2 attenuators were stepped 0 to 11 db and the power was measured with the mock spectrometers
• Turns out the atten steps were not all exactly 1db
• But there was not a consistent compression of the step size at higher powers (lower attenuation).
• PolB showed a decreased step when approaching 0db  atten, but it did the same on the sum (-28bm) as off the sun (-47dbm) so this must be the attenuators.
  
• For each of the 11 steps in attenuation we compared the step size while changing the mock rms from  50,100, to 200 counts.
• this is  variation of (200/50)^2 = 16 in power.
• this is only checking the  mocks
  
• the larger rms measurements did not show a smaller step size.
• So for if2mp values of-28dbm, mock rms's of 100 to 200 counts, the system is not compressing.
  

processing: x101/211124/lintest.pro, lintest2.pro

---