Intro

    On 08nov21 the cal diode for the 12meter was updated to be  temperature stabilized with a peltier cooler. The noise diode, fet switch, and polA,B splitter are all stabilized. Prior to this date, cal power output would vary with temperature.

    On 10nov21 we started a test that lasted for 26 hours. The setup was:

• telescope sitting at az=5, el=75 degrees.
• The mock spectrometer was setup to run:
• 7 bands of 172MHz centered at: 8220,8362,8504,8648,8788,8930, and 9072 MHz were recorded.
• 4096 channels /band and 1 second dumps.
• Data was taken
• for 60 seconds with the cal On
• Then 60 seconds with the cal off
• This sequence was started around 17:00 10nov21. It finished at 19:00 11nov21.

Processing the data:

• each 60 second scan (for each 172Mhzband)was input and averaged over the 60 seconds.
• the total power was computed while excluding 100 channels on each edge of the band (where the filters dropped off).
• the tpCalOn and tpCalOff scans were isolated.
• The cal difference tpCalDif= (tpCalON - tpCalOff) was computed.
• Tsys (in units of the cal) was then computed:
• Tsys[calUnits]=TpCalOff/(tpCalDif)

Plotting the results:

    The plots show the results of the cal stability test (.ps) (.pdf)

• Page 1:Band power and cal deflection vs hour of day from 10nov21.
• 11nov21 starts at 24 hours. noon 11nov21 is at 36 hours.
• Each color is a separate 172 MHz band
• Each freq band has been normalized the median value in the cal Off,
• 1st,2nd frames: total power vs hour of day  by frequency band.
• There is a large jump in power near 27 hours (3am 11nov21). It lasted for about 2.5 hours.
• The AO weather station had 0 rainfall 27 to 30 hours.
  
• the relative humidity did reach a max of about 92% during this time.
• This may have been caused by clouds.
• After the jump in power, the power in the different frequency bands diverged for polA (this had been seen before)
• Pol A also showed several jumps (that were not seen in polB.            
  
• 3rd, 4th frames: calDeflection for polA and polB
• the cal deflection for each band as been normalized by the median power in the cal Off.
• Page 2: Band power and temperature.
• A temperature sensor in the 12meter pedestal records the air temperature every minute.
• We also have a weather station atop the lidar lab that records temperature and rainfall. (it is a few 1000 feet from the 12meter).
• Top: total power vs hour of day for 8648 MHz band.
• each plot has been normalized to its median value.
  
• black is polA, red is polB
• You can see that polA is jumping around a lot more than polB (around 33 hours)
• Middle: temperature vs hour of day.
• Black: temperature inside the 12meter pedestal.
• red:     Temperature at he weather station (above the lidar lab)
• blue: weather station temp offset by 11 degF.
• The weather station and pedestal temp track each other pretty well with the 11F offset.
• Bottom: BandPwr vs  Pedestal temperature with linear fit.
• Black is polA, red is polB
• the jump in pwr around 83F (27Hours) was not included in the fit.
• Looking at the weather station:
• There was no rainfall recorded around 27 Hours, but the relative humidity did reach a peak of around 92%.
• The fits:
• polA 1.6 % decrease in power per degF
  
• polB 2.1 % decrease in power per degF
• Page 3: Tsys in cal units
  
• Top: BandPwr/calDeflection for the 7 bands polA
• Middle: BandPwr/Caldeflection for the 7 bands (polB)
• These two frames are Tsys in units of the cal.
  
• Bottom: Tsys [calunits] versus pedestal  temp with linear fit.
• the fits:
• polA : 1.98/2.4 = .8% tsys/degF
• polB:  2.8/3.     = .9% Tsys/degF

    If we assume there is a single Electronic gain (Gel) then:

•  PwrCalOn=(tsys + Tcal)*Gel
• PwrCalOff=(Tsys)*Gel
• PwrCalDif=pwrcalOn - PwrcalOff= Tcal*Gel
• pwrCalOff/PwrCalDif= Tsys*Gel/(Tcal*Gel) = Tsys/Tcal

    If the temperature variation is only in Gel then Tsys/Tcal should not have a temperature dependence.
The slope that we are seeing mean either:

• The cal pwr is changing with temperature , or
• Tsys changes with air temperature, or
• There is some temperature dependent gain that Tsys sees that is not seen by Tcal (horn?)

processing: x101/211110/moncalonoff.pro