Monitoring alfa stability

How the data is processed:
Description of the plots:
The plots by month:

    The stability of the alfa receiver is monitored by comparing TsysA/TsysB for each 1 second data sample. The assumption is that if a jump occurs, the two pols  will jump differently. Taking the pol ratio removes continuum sources (at least the unpolarized part). The processing is done at the end of each month. Plots are then made and entered in the monthly table (see below).
 

---

How the data is processed:  (top)

• Use the a2010 drift scan data. It has 600 1 second samples per drift (followed by a cal). There have been 300 to 500 drifts/month.
• For each strip compute Tsys over the 100 Mhz band for each 1 second sample using the cal. Do  this for each pol and beam (more info on how tsys is computed).
• Compute the ratio of TsysPolA/TsysPolB. This removes unpolarized continuum sources. What's left will be any gain drifts caused by temperature or jumps.
• Normalize each strip of ratios to their median value and then look for any outliers (say 2 to 4% of Tsys).
• For the strips that have outliers, plot out the original tsys data to see if it is polA or polB that is jumping.
• Using the ratio of TsysPolA/TsysPolB compute the rms/Median for each strip. Plot these versus stripnumber and day of month.

---

Description of the plots:  (top)

1. date
2. Ratio A/B img:
• Data: For each strip of the month compute TsysA/TsysB at the 1 second samples (this removes the continuum sources).
• Horizontal axis:  600 samples/strip for each of the 7 beams.
• Vertical axis: are the strips of the month.
• Notes:
• vertical dashed lines are drawn between each beam.
• You can see all the beams at once and the continuum sources are gone. You can't tell if the problem is polA or polB (see next image).
3. Beam 0,1 TsysA,B:
• Data: Tsys by strip for beam 0 (top image) and beam 1 (bottom image).
• horizontal axis: 600 points of polA followed by 600 points from polB
• Vertical axis: strips of the month.
• Notes:
• Vertical dashed line drawn between polA and polB (at sample 600).
• The ratio image shows which beam is bad. Use the TsysA/B images to see if it was polA or polB.
• You occasionally see dots across a strip spaced by about 12 samples. This is when the radar compressed the data and it took a long time to recover.
4. Beam 2,3 TsysA,B:
• Same as beam0,1
5. Beam 4,5 TsysA,B:
• Same as beam0,1
6. Beam 6 TsysA,B:
• Same as beam0,1
7. BadRatios,rms:
• Page 1: The strips with large TsysA/TsysB ratios are plotted versus sample in strip. The threshold (see col 10) is typically 2 to 4% of Tsys.
• Data: TsysA/TsysB. Each strips had been normalized to its median value.
• horizontal axis: 600 samples per strip times the 7 beams.
• Vertical axis: the bad strips in the month. An offset is included for plotting. The vertical scale is Tsys.
• Notes: vertical lines are drawn between beam boundaries.
• Page 2: For each strip the ratio TsysA/TsysB is computed and then the rms/median of these 600 points are computed. When there are jumps within a strip the Ratio rms will be larger.
• top plot: data: rms/mean of tsys ratio. horizontal scale: strips in the month. Vertical scale the rms/median. Each beam in plotted in a different color (with an offset for display).
• bottom plot: The rms/mean for the Tsys ratio is plotted versus day of month. Each beam in plotted in a different color (with an offset for display). This plot shows which days had the jumps.
8. total Strips: Total strips a2010 took for this month
9. bad strips: The number of strips who have at least one point with abs(TsysA/TsysB-1) greater than the bad threshold (typically 2-4% of Tsys).
10. bad days: this lists some of the days with the largest jumps. It does not include all days with jumps, just the bad ones.

---

The plots by month:  (top)