10oct12

Intro

    Alfalfa data (a2010) taken during sep12 showed long recovery times with alfa beam1b when the radars were present.

Some History:

• oct09: Mitsubishi amplifiers outfitted with leds to solve radar long recovery times
• apr12: alfa amplifiers replaced with nrao amplifiers. The amps have led's installed
• sep12: alfalfa runs and we see long recovery times on beam1b when the radars are present.
• 10oct12: checked led currents. 3milliamps (as they were supposed to be).
  

Data setup and processing:

• a2010 does 600 second drift scans at the meridian using the wapps.
• 100 Mhz bandwidth centered around 1380 Mhz.is taken.
  
• Data is sampled at 1 sec intervals.
• The total power for each one second is computed over the 100 mhz. The computation tries to not include frequency channels with rfi.
  

The plots show the total power vs time for alfalfa (.ps) (.pdf):

• 600 seconds of data from 1 drift scan. The sample rate is once per second.
  
• Each color is a different beam.
  
• The vertical display units are Tsys. A vertical offset has been added to each beam for display.
  
• Top frame is polA, bottom frame is polB.
• Page 1:  600 seconds of data from 1 drift scan.
• The positive going spikes are the 12 second radars.
• beam 1b (bottom frame, red trace) has negative going spikes.
  
• This band has gone into compression
• You can see that it takes many seconds for the value to recover.
• Page 2: a blowup showing seconds 180 to 380.
• You can see the long recovery time for beam1b
• beam 0b looks a like it may have a little delay.

Summary:

• long recovery times were seen on 10sep12 with the a2010 data.
• The long recovery can also be seen on 12,22,29,30 sep12.
• 10oct12: denis checked the led currents. They were 3 milliamps (as the notes say they should be).
• Previously the long recovery times came because the Mitsubishi amps didn't have leds.
• We don't know whether the current problem is in the amps, or somewhere else.
• galfacts a2130 cal deflection data has shown pol1b cal deflection jumping around a lot more than the other beams.
• This occurred  during:
• jul-oct12
• may-jun12
• nov11-dec11.. this is before the new amps were installed.
• So maybe this problem is separate from the slow recovery
• or maybe the slow recovery in 1b is not happening in the amps?
• After the new amps were installed in apr12, i never checked for the slow recovery time of the new amps on the telescope with the radars present. Looking back at the a2048 data may-sep12 shows:
• baseline variations in bm1b: 31may12, 3/4/14jun,14/17/19jul, 27aug
• long recovery from radars: 02jun,25aug,
• So it looks like bm1b has been acting up for awhile.
  

processing:x101/121010/alfabm1b.pro

---