Newer results can be found here (20sep22)

Intro

    Tsys vs elevation was measured on 24sep21. The measurements  consisted of:

• The telescope was driven between 7 and 87 degrees elevation at 1 deg/second at 5 different azimuths:
  
• az=270,0,90,180, and 270 again
• The test was repeated 3 times.
• While driving the telescope the mock spectrometers took data:
• 7 172 MHz bands were taken. cfr: 8221.00 8363.00 8505.00 8647.00 8789.00 8931.00 9073.00 MHz,
• 1024 channel across the 172 MHz
  
• data was sampled at .1 seconds

Processing the data:

• The total power for each .1 second spectrum was computed. 100 channels on each edge of the band were excluded.
• no rfi removal was attempted.
• pola and B were added
  
• The az,el,ra,dec,mjd were interpolated to the .1 sec spectra (they are only recorded once a second).
• For each strip, the total power was scaled to the median value between elevation 80 and 85 deg.  
  

Plotting tsys vs elevation

The plots show the  tsys vs elevation (.ps) (.pdf)

• Page 1: Tsys vs el for all 3 sets of observations using the 8221 MHz band.
  
• The total power plots have been normalized to the median value 80 to 85 deg elevation.
• Each color is an elevation strip at a different azimuth.
  
• Each observation has 5 elevation strips at az=270,0,90,180,270
• Top: first set starting at 14:40:54
• the bumps at az 180,270 are probably sidelobes/scattering from the sun
  
• Middle: 2nd set starting at 14:51:40
• Bottom: 3rd set starting at 15:39:28
• Page 2: Tsys vs el for the 7 freq bands when az=0.
• each color is a different frequency band
• Top: tsys vs elevation
• Bottom: tsys/tsys8221 band vs el
• The tsys vs el for each freq  is normalized to the 8221 band values
• There is some periodic rfi in the 8931 and 8791 MHz bands
• The spikes are every 4 seconds.
• The 7 bands show a variation of up to 1.5%

Fitting Tsys vs elevation

    Most of the tsys vs elevation is probably coming from the amount of atmosphere the signal traverses at each elevation.
Three different fits were tried

• fit1:Tsys=A0 + A1*sin(el)^A2
• fit2:Tsys=A0 + A1*sin(el)^-1
• fit3:Tsys=A0 + A1*el^A2
• Page 3: fit tsys vs elevation.
• az=0,freq=8221, set=1 was used for the fits.Pa
• Top:  tsys vs el with fit
• black: measured data
• red: fit1 A0 + A1*sin(el)^A2
• green:fit2: A0 + A1*sin(el)^-1
• blue: fit3: A0 + A1*el^A2
• Bottom: fit Residual
• plot Measured - fit
• the residuals varied between .08% to .2% of Tsys.
• The residuals increase at low elevation.
  
• This could be ground radiation getting into the beam.
• Page 4,5 How well does the fit work at other azimuths
• fit # 1 was used.
  
• Page 4: tsys vs el for the 5 azimuths and the az=0 fit evaluated
• Black: measured data, red: az=0 fit evaluated at the corresponding elevations
• Page 5: fit residuals
• Measured data - fit(from az=0)
• the fit gives residuals less that 1% of tsys  (excluding the sun,source variations).
• there is a small slope from el = 60 to 0 degrees  when the azimuth was 180 and 270 degrees.
• this may a residual from the sun.

---

SUMMARY

• elevation strips 7 to 87 degrees were done at 5 different azimuths (270,0,90,180,270) degrees.
• The set of 5 strips was repeated 3 times.
• 7 172 MHz freq bands were recorded and then the total power was computed.
  
• The tsys vs elevation for the 7 freq bands varied by up to 1.5%tsys.
• Going from el=88 to el=7 tsys increased by about 24%
  
• The  tsys variation was modeled with  A0 + A1*sin(el)^A2 at az=0 degrees.
• each tsys vs el was normalized to the median value  el=80-85 degrees
  
• The results:
• for the same azimuth , the residuals were < .5% of Tsys for the 7 bands
  
• for different azimuths the residuals less than 1% (excluding sun, sources).

Tsys=A0 + A1*sin(el)^A2

	A0	A1	A2
az=0 fit	.95231	.04765	-.85115

• If you wanted to use the fit and you didn't  cover el=80-85 deg  to normalize your data:
• just pick another elevation range and normalize your data  to  that part of the fit.

To do:

•  repeat this at night to  get rid of the sun.
  

processing:x101/210924/tsysvsel.pro