Intro:

A2812 data:

• lbw linear pol
• redshifted H1.. 1360-1420
  
• interim correlator
• sbc1 polA H1
• sbc2 polB H1
• sbc3 PolA OH
• sbc4 polB OH
• 1 second dumps

• each frame contains the total power vs time for a nights run.
• black is pola, red is polB
• They usually do on, off position switching with a cal
• 300 secs on source
• 300 secs off source
• then fire the cal
  
• Each narrow spike is the cal firing
• between the cal spikes, you should see a power variation (on za change) and then a similar repetition (on off source over same za range).
• The vertical scale is linear in power
• You can see the red trace (polB) drifting during most night's runs.

  19aug15:Total power data taken while sitting:

• lbw,linear pol, cfr=1405 MHz
• 1370 hipass filter and 1360 to 1470 band pass filter
• (i also tried it with the default 1100 to 1800 MHz filter with similar results).
• interim correlator,
• 25 MHz polA,B in each of 4 sbc (to make sure the problem wasn't in the iflo or interim correlator.
• 1 second dumps for 910 seconds. telescope sitting at az=285, za=11 deg.

• Black is polA, red is PolB
• There are 4 polA traces and 4 polB traces.
• the traces within each set of 4 are identical
• This means that the problem is before the 4 mixers downstairs that splits up the band.
• the bumps at 310 and 400 seconds are sources drifting through the beam
• Each trace was normalized to it's median value and then 1 was subtracted.. so the vertical scale is in units of Tsys.
• You can see that polB is drifting around by up to 1% of Tsys.

150831: totPwr data taken while dewar warmed then cooled.

• the telescope was at: az=284.7, za=11
• lbw was setup with cfr=1405 Mhz, 1370hipass and 1360-1470 Mhz filters in
• interim correlator took 25Mhz centered at 1405 Mhz using 4 separate boards, and 1 second dumps
• the extra boards were to check if there was an instability in the if/lo or interim cor.
• all 4 were identical..
• The total power for each 1second dump was computed over the 25 Mhz bw.
• i averaged the 4 separate boards, since they showed no variation.

• Black is polA, red is PolB. The vertical scale is Tsys (i removed the average)
• The green line is the 16K stage dewar temp (divided by 100).
• The green line shows when we started warming up the dewar, cooling it back down, and the warming it up again.
• PolA Tsys is stable during these changes
• PolB reacts to changes in the 16K stage.
  
• around 9.3 hours, the 16K stage went from 20K to 28Kthen back down to 20k.
• polB power output power went from tsys to 1.38 Tsys.
• It looks like polB is reacting the changes in the 16K stage.

150917,18 : Data taken after dewar worked on.

• lbw center freq of 1405, 1370 MHz hipass filter installed as well as the 1360-1470 MHz filter
• interim correlator using 25MHz bandpass, 1 second dumps, 4 boards
  
• Total power was computed for each 1 second dump.

    The plots show the total power  vs time for the 2 days (.ps) (.pdf)

• black in polA, red is polb
• Top: 17sep15
• the narrow spikes are from punta salinas transmitting in all frequencies.
• PolB starts to drift up at 16.2 hours.
• This may be polarized power (north polar spur) or it could be a problems with lbw polB
• Bottom: 18sep15
• both traces follow one another
  
• The rise at 11.8 hours is the sun sidelobes coming into the beam.
• For the measurements on both of these days, the dewar temperature was stable.
  

150922:totpwr data while we warm up dewar

    The data from 18sep15 showed that lbw polA and polB tracked one another for over an hour. During this observation the dewar temperature was stable. Back on 31aug15 we saw that as we warmed up the dewar a little, polB signal increased rapidly.
    On 22sep15 we turned off the lbw refrigerator for a few minutes while we took some total power data.
The setup was:

• the telescope was sitting at az=21.5, za=1.8998
  
• lbw linear pol, center freq 1405Mhz, 1370 hipass and 1360-1470  filters in.
• interim correlator took 25 Mhz bw of data using 4 boards, 9  level sampling, 1 second integrations.
• we computed the total power from the 25Mhz band.
  
• We turned off the lbw refrigerator at 9:57:49 (9.96361 hour).

• Black is pola, red is polB. The vertical scale is in units of Tsys ( 1 has been removed).
• green is the the 16 K dewar temperature. The scale is degK/100. So the temp started at 15.5K.this data is measured once every 32 seconds.
• Top plot: 499 seconds of data.
  
• You can see polB power head as soon as the Tsys starts to rise
• We have undersampled the 16k Temp. The refrigerator was turned off at the blue dashed line.
• Bottom: blow showing totpwr vs seconds from refrig turn off.
• The polB started to rise about 3 seconds after the shutdown (each + for the polA,B is a 1 second sample).
• The fall in the 16k temp was when the refrigerator was turned back on.
  

processing: x101/150922/lbwchk.pro

Summary:

• a2812 reported unstable baselines for their aug15 data.
  
• The problem was with polB (not polA as they mentioned..)
• data taken on 19aug15 showed that polB is drifting by up to 1% of tsys.
• The problem occurred with the 1100-1800 MHz filter as well as the 1370 hipass and1360-1470 band pass filter
• so the problem is not in the filter box.
• On 31aug15 we warmed up the dewar a little and the power level for polB shot way aup
  
• After working on the dewar in sept, data was taken on 17,18sep15.
  
• On 17sep15 there was a slow rise in polB (relative to polA)
• On 18sep15 both pols  tracked on another.
• The 22sep15 data shows that polB is  still extremely sensitive to 16K dewar temperature variations
• The signal increased 35% in 50 seconds.. as the 16k stage went from 15K to 23K.
• We still have a problem with polB.
  

 home_~phil