Galfa pixel 4a/6a switched may05
The cables for pixel4a and pixel6a (count from 0)
from the distribution amplifiers in the receiver room to the galfa spectrometer
were switched sometime in early may05 (or late apr05). This meant that
when looking at the data offline:
This occurred when the distribution amplifier chassis for polA was removed
to install some fans for cooling. When the chassis was reinstalled, the
cables leaving the chassis were switched.
pixel 4a spectra would be in the location of pixel6a
pixel 6a spectra would be in the location of pixel4a.
Continuum sources drift/move through the alfa beams
at different times since the beams point at different locations in the
sky. 600 second total power scans showed that pixel 6a and pixel
4b had continuum sources moving through the beams at the same time (6b
and 4a also showed this). The last "good" data set that showed know cable
switch was 12dec05. The next data set available after this was 02may05.
The second data set showed the cables being switched (i didn't find any
data between these two dates to test).
The plots show the total
power drifts versus time after and before the switch (.ps) (.pdf).
Each pixel is plotted in a different color. For each color there are two
strips: polA and then polB over plotted:
When inspecting the cables that left the back of the polA rack in the receiver
room, we found that it was the polA cables that were switched. This problem
only affected the galfa spectrometer. The wapp data was never switched.
Fig 1 Top: 02may05. The purple strips are pixel4(a,b) while the light blue
is pixel6(a,b). The dotted lines around 400 show that the same source is
moving through pixel 4 and pixel 6. The dotted line at 465 shows another
source that is transiting pixel 4 and pixel 6 simultaneously.
Fig 1 Bottom: These are calibration scan data taken on 12dec05. The double
traces of each color track each other.
The galfa cables for pixels 4a,6a were switched sometime after 12dec04
and before 02may05 (probably in late apr05).
The problem was corrected on 31may05 at 11am ast.
The wapp alfa data was not affected by this problem.