The problem seen were:

• Very little power from he antenna getting through the system
• Tsys was not afffected
• The load switch portion of the Jennings relay was still working..
  

History:

• 02apr12 a2599 used 430 dome and it worked ok
• 15may12 through 14oct12 x102 gain for gr430 was 6 K/Jy
• 18oct12: removed Jennings relay, replaced with elbow. Relay brought down to the lab to be adjusted
• 19oct12: after adjustment of Jennings relay in lab, it was reinstalled in the dome.
  
• Relay still not working, elbow reinstalled, relay back to the lab
  

X102 calibration runs may12 -> oct12  (top)

• source B0618+145 (3C158)
  
• 7.45 Jy at 430 MHz. size~7"
• .2% polarized at 1405 (via nvss)
• 16 patterns done. 
• Each pattern consits of 4 strips each +/- 3 beamwidths
• az strip,za strip, +45 deg strip, -45 deg strip. There are 60 samples/strip.
• Each pattern has a 10 second cal on ,cal off.
  
• freq/bw: 430/25MHz, 430/12.5MHz, 429/1.MHz,431/1MHz
• the 430MHz 25 and 12.5 MHz bandwidth have some rfi.

• The cal deflection  for each pattern was used to scale the data to Kelvins.
• You can see the source deflection for the 4 strips.
  
• Each frame is a different freq/bandwidth dataset
• the 16 patterns have been over plotted.
• polA is black, polB is red.
• The tsys for the two polarizations is within a few kelvin
• The source deflection varies by a factor of 10.
  
• the  jumps in the 1st two frames are rfi
• The rfi jumps in polB are much smaller than the rfi jumps in polA
• 430gr is circularly polarized so this difference is probably not a polarization problem.
• looks like there is not much signal getting in from the sky

• Page 1: 14oct12 x102 data.
• I processed  stokesI (a+b), polA, and polB separately.
• Top: gain vs za in K/Jy
• stokesI = 6K/Jy. This is 1 about 1/2 of the expected value
• polA= 11 K/Jy. this is the expected value
• polB= 2 K/Jy. This is much less that 5x less than expected.
• 2nd frame: tsys vs za
• PolA is around 70K while polB is about 65K. This difference could be explained by inaccuracies in the cal values.
• 3rd frame: SEFD = Tsys/Gain
• this is independent of the cal values (since both gain and Tsys use the cal value)
• polA is around 6 Jy (expected) while polB is 30Jy. stokes I gain is around 10Jy.
  
• if the gain of one pol went to 0, the sefd for stokes I would get 2 time bigger.
• Bottom frame: average beam width
• The average beam width is 650 asecs. this is the expected value.
  
• The beam widths for stokesI, polA, polB are all the same
• This says that the optics and telescope gain are probably ok.
• Page 2-4 show the 430 dome x102 data for jul-oct12
• page 2: StokesI gain,tsys,sefd, and average beam width
• The data agrees with the stokes I data from 14oct12.
• page 3: stokesI pointing error
• The top 2 frames are za Error vs za,az, the bottom 2 frames are az error vs za,az
• there is a 17 asec za offset and 8 asec az offset. the rms  are 10.8 and 7.5 asecs
  
• page 4  gain,Tsys,sefd, and SEFDPOLb/SEFDpOLA for stokesI, polA, and polB
  
• the 3 freq bands are shown with different colors
• + is polA, * is polB
• Gain. i've used a log axis for y. polA is around 10K/Jy, polB is around 1K/Jy
• tsys: The tsys for polA and polB are about the same
• SEFD: polA is around 6Jy while pol B is 20 to 100Jy
• sefdB/sefdA the ratio is 10 to about 60.

When did the problem start?

• x102 430data was taken on:
  
• low Gain  (5-6 K/Jy) on 15may12,01jul12,14aug12, 07sep12,14oct12.
• normal gain (11 K/Jy) on 24oct09
  
• There was no 430 x102 data between these two dates
• A2599 on/off position switching on continuum sources mar/apr2012 (.ps) (.pdf)
  
• This show source strength in degK vs freq.
  
• Black is polA, red is polB
• PolA and polB have close to the same value for the source in degK (using the cals).
• The values were ok on 02apr12.
• x111 430dome  rfi mar-jul 2012 (.ps) (.pdf)
• The x111 (rfi monitoring data) for mar-jul12 was used to see any gain differences
• The processing was:
• each 60 second scan was averaged.
• the average spectra was converted to deg K using the cals
• The Peak rfi for polA and polB was located in each spectra
• The plots show the Peak rfi in polA and polB with the other pols value over plotted.
• The peaks for polA and polB did not have to be at the same channel.
  
• Black in polA, red is PolB. The dotted green line is the end of each month: mar,apr,may,june.
  
• Top: PolA max Value with corresponding polB value from this channel.
  
• Bottom: polB max value with corresponding polA value from this channel.
• There are measurements where polA is larger than polB, but it is not consistant.
  

x102 Summary:  (top)

• The gain in polB is about 10% of the expected gain.
• The gain in polA is correct
• the system temperature for polA and polB are about the same value 65-70K.. this is the expected value
• This means that any problems are occurring before the cal injection.
• If there is a 90% gain loss after the cal injection then the Tsys from the dewar would go up by a factor or 10. This would probably be noticeable.
  
• RFI coming into polB is much weaker than the rfi coming into polA (and 430gr is circularly polarized).
• The SEFD also shows a factor of 10 difference between the two pols
• the sefd does not use the cal values, so it is not a cal value problem.
• The beam widths are identical for polA, polB  and stokesI.
• this means the telescope gain is ok
• The pointing error has
  

• 	offset err asecs		rms err asecs
  	az	za	az	za
  polA	.6	19.0	5.8	9.9
  polB	46.9	19.8	10.4	10.4

  
  
• So the beam squint is about 45 asecs. This is what it has historically been. This is not enough (with an 11 amin beam) to cause the large drop in source deflection.
  
• The problem must be before the cal injection and it is not the optics. There is either a bad cable (capacitively coupled?) or the omt is not working for polB
• if the omt is bad, it has to be bad only for polB leaving polA functioning correctly?
• How long has the problem been present.
• The problem is in the x102 measurements since 15may12 - 14oct12
• The problem was not seen in the 24oct 09 data. there is no x102 data between 24oct09 and 15may12
• a2599 took data on 02apr12 and the problem was not present.
• Looking a the strength of rfi in polA,B does not give a consistent measure of the gain (probably because the rfi is coming in through a far out sidelobe and is not being passed equally between the two pols.
  

18oct12: Replacing Jennings relay with an elbow. (top)

• Before the relay was removed.
• the turret was moved to 125.6 deg (the 430 is over the tertiary at 251.7 degrees). Tsys at this location should be around 250 K.
• We used the winking cal to normalize the electronic gain of the system (calA=95.5, calB=102.5... we set the attenuation so the oscilloscope caldeflection was the same.
• After adjusting the gains, Tsys polA was about 3.5 times higher than Tsys polB.. no signal was getting into polB

• black is polA, red is polB.
• The 1 second cal toggle was used. The interim cor sampled the data at 5 hz.
• Tsys for the two pols are now similar.

• the 25 hz hardware cal was sampled at 1khz rate by the mock spectrometers. The data was then converted to kelvins using the winking cal values.
• Tsys polA and B are similar. The rfi is now the same amplitude in polA and polB.
• When the polB Jennings relay was in, polB rfi was much smaller than polA.
• It looks like the signal for polA is varying more than polB.
  

19oct12: readjust Jennings relay.. still trouble

    After the 18oct12 tests, the polB Jennings relay was brought down to the lab and it was readjusted. On 19oct12 the polB Jennings relay was reinstalled for some tests.

Test 1:rcvr moved off and then back on the sky to change Tsys. Jennnigs relay polB installed.

• The telescope was stationary (za=11,az=277deg). The sky was drifting by,
  
• The turret was moved off of the tertiary so the 430 feed was no longer connected to the sky. This should increase the power coming in by 50 - 100 K or more.
  
• the hcorcal was toggled once a second with the hardware toggle line
• The interim correlator sampled the data at 5 hz. with a 25 Mhz bw centered at 430 Mhz
• The Jennings relay was switch to load and antenna a few times

• 0-50 seconds 430 off the sky
• temp polA =130K
• temp polB =80K
• 50-132 secs: Jennings relays switched to load
• temp polA =300K
• temp polB =350K
• 132 - 192  secs: back to antenna. start moving turret so 430 points to sky at 160 secs
• There is a large change  in Tsys polA but very little change in tsyspolB (not much is getting into the recevier
  
• 192-250 secs: on sky(ant)
•  tsysA=55K, tsysB=60K
• 250-300 back to load.
• This shows that power from the horn is not getting into the Jennings relay polB very well. The load switch seems to work.

Test 2: let sky drift through beam. Measure polA,B source strength with Jennings relay polB in and out.

    The setup:

•  The telescope sat at az=277, za=11. The 430 dome was left looking at the sky (over the tertiary).
• cfr 430 MhZ, 25Mhz bandwidth, 1 second samples were taken with the interim correlator.

• No cals were fired. the total power was computed for each 1 sec spectra and then the entire dataset was normalized to the median value. This made the units for the vertical axis Tsys,.
• The spikes every 120 seconds are rfi at 433.9 Mhz.
  
• Page 1: total power vs time. Jennings relay polB in.
• Top: all 1900 secs of data. black is polA, red is polB
• A source takes about 45 seconds to drift through the 11 arcminute  beam
• The bump at 1150 seconds is a source drifting through the beam
• The dashed green lines are blown up in the 2nd frame.
• 2nd frame,. Blowup around the source at 1150 seconds.
• a linear baseline was removed from this data.
• there is not much source in polB (assuming Tsys is the same.)
• The data between the dashed green lines are used to compute the source strength ratio TsrcA/TsrcB in frame 3.
• bottom: TsrcA/TsrcB when source drifts through the beam
• TsrcPolA is about 6 times larger that TsrcPolB
• The TsysA/TsysB where the rfi spike occurs has a different ratio. .
• Page 2: total power vs time. Jennings relay polB  replaced with an elbow.
• These are the same plots as page 1 after the Jennings relay polB was replaced with an elbow.
• top: The tsys values are now similar, even when sources drift through the beam.
• The rfi spikes have different amplitudes.
• 2nd: blowup showing a source drifting through the beam (actually it looks like it is drifting through the sidelobes).
• The amplitudes remain similar for polA and polB
• bottom: TsrcA/TsrcB at source
• The ratio of TsysrA/TsrcB is very close to 1 (with 15%).
  
• the rfi spikes do not have the same strengths:

Test3:  rcvr moved off and then back on the sky to change Tsys. Jennnigs relay polB replaced by elbow.

    This repeated test1 with the polB Jennings relay removed. 

The plot shows   Tsys as the turret moved the 430 off and then on the sky (.ps) (.pdf)

• Turret started with the receiver off of the tertiary.
• At 200 secs it started to move the receiver toward the tertiary (sky)
• The receiver arrived on the sky at 310 seconds.
• The Tsys for polA,B were identical off of the tertiary. On the sky they differed by 10K (this may be real)l.

Summary 19oct12:

• adjusting the polB Jennings relay on 18oct12 in the lab did not solve the problem.
• On 19oct12 the ratio of TsrcPolA/TsrcPolB was about 6.
• The system seems to work correctly when we bypass  the jennings relay with an elbow
• This should work ok until the next 430 dual beam transmitter experiment. At that point we'll need to reinstall the relay.
  
• the ratio of rfi differed in polA and polB when the relay was in, and also when the relay was replaced by an elbow,
• So rfi is probably not a good ruler to use. The different may come from:
• rfi does not come in through the main beam.
• There may be some compression somewhere in the system (interim correlator?).