lbw polB cal jumps 04dec05
The cal value for lbw polB changed on 04dec05. The
value of the system temperature changed as well as the gain (which is measured
using the cal value). The sefd (which is independent of the cal) remained
unchanged. The Tsys jump in polB occurred with both diode 1 and diode
2 so the problem was after the switch that selects which diode to send
Looking at cal on/offs taken by users:
All of the cal on/offs taken using the interim
correlator between 23nov05 and 15dec05 were used to compute Tsys. I limited
the scans to those taken in linear polarization since the circular mode
has a hybrid which mixes the two channels. The plots show
tsys versus date (.ps) (.pdf):
Top: The frequency of the cal on/offs versus date. Many asterisks are over
plotted on top of each other. The date tick marks are for 12 noon utc on
Center: Tsys versus date. The black * are polA and the red * are
polB. Part of the spread in Tsys is coming from the frequency variation.
PolA stays close to the same level for the entire period. The polB
Tsys begins to rise around 13 Hrs (utc) 04dec05 (project a2121 data).
Bottom: The ratio TsysA/TsysB is plotted versus date. The dashed vertical
lines are a2121 data taken on 03dec05, 04dec05, and 05dec05. Tsys polB
was ok on 13 hr 03dec05, it had increased by 10% on 04dec05, and on 05dec05
it stabilized at a value 20% larger than the 03dec05 value. For this
to happen the cal power getting to the amps had to decrease by about
20%. Note that the data was sampled with 24 hour resolution so we
can't pinpoint exact times.
The daily tsys measurements taken at 1400/1405 Mhz.
The lbw system temperature is measured (almost) daily
using all the cal combinations. It was taken at 1400 Mhz until 05dec05
when it was switched to 1405 Mhz (to get away from a resonance in polA
at 1400 Mhz).
The plots show lbw
Tsys for jan01 through feb06 (.ps) (.pdf):
Top: Tsys polA vs date. The black * are from diode 1. The red * are from
diode 2. The dashed lines mark where things occurred:
The PolA Tsys wandered around between jun05 (when the polA resonance
was created) until 06dec05 (when the measurements were moved
to 1405 Mhz). This wandering occurred with both diodes.
01jun05 (dashed black). Dewar warmed up and then cooled down. After the
cool down, a
resonance appeared at 1400 Mhz in polA. It had not been there previously.
10/13jun05 (dashed purple): cal values remeasured.
21jul05 (dashed red) : diode 1 breaks. All
tsys measurements switched to diode 2.
23nov05 (dashed green): new diodes installed.
23/25nov05 (dashed purple): cal values remeasured.
06dec05 (dashed blue): tsys measurements moved
1400Mhz->1405Mhz to get off of the resonance in polA at 1400Mhz.
2nd: Tsys polB vs date. It was relatively stable except for the jump on
01jun05 and on 04dec05.
3rd/4th: These are blowups of the 01nov05 to 01jan06 time range. PolB
jumped up prior to the 06dec05 change of frequencies.
Tsys, gain from the calibration scans:
The lbw calibration scans between jul05 and feb06 at
1415 Mhz were inspected to see if the change in tsys was a real increase
in tsys or a result of the cal values changing.
The plots show lbw
calibration scans jul05 to feb06 at 1415 Mhz (.ps) (.pdf):
Top: TsysA and TsysB were each divided by their average value over this
period to show their variation with time. The black polA jumped down after
the frequency was moved from 1400 to 1405 Mhz. PolB red jumped up after
2nd: Ratio of TsysA/TsysB. The major jump occurred around 04dec05.
3rd: The ratio of TsrcA/TsrcB. This is the ratio of the measured source
strength in polA and polB (polarized sourced were excluded). This also
showed a jump in early december. These measurements include the cal value.
bottom: The ratio (TsysA/GainA)/(TsysB/gainB) is constant over the entire
period. This value does not use the cal. If TsysPolB had actually jumped
then this data would not be flat. So the problem is in the cal value.
TsysPolB jumped on 4/5 dec05. It took about 2 days to change. It changed
by about 20%.
The change was caused by the value of the cal changing. The actual tsys
value did not increase.
The increase was independent of the diode used to measure it. So the problem
is after the switch that chooses the diode to use (the diodes are ok).
This is not the first time that the polB cal value has jumped. It also
jumped in apr04.
The cal values were remeasured on 23feb06 and then back dated to be valid