1428 Mhz birdie in alfa
Project a2010 (alfalfa) does
600 second drift scans dumping once a second using 100 Mhz centered
at 1385 Mhz. A birdie a 1428 Mhz appeared on 18may05 and remained
until 23may05. It was not seen after that (looking at data through 09jun05).
The a2010 data was examined to characterize the birdie.
The rms/mean by channel was computed for each strip
taken by a2010 (998 strips) between 01may05 and 09jun05 (this is
part of the monitoring
of a2010 data). If there is rfi in a frequency channel then the rms/mean
will be larger than the radiometer equation dictates. The rms/means for
each strip were put together into an image to see when the 1428 Mhz birdie
The 1428Mhz problem started around strip number
370 (18may05) and lasted till strip number 610 (about 23may05). After this
it did not appear again (at least through 09jun05). There is a narrow birdie
(not resolved in 50Khz channels after hanning smoothing) and a wider birdie
(about 250 khz wide) centered at 1427.8 Mhz. When looking at the narrow
birdie on the spectrum analyzer, it looked like two birdies separated by
about 30 khz. The birdies appear in all pixels and both polarizations.
Using the rms/mean for each strip to see when the birdie
The next set of plots takes the rms/mean for the
frequency channel of the narrow birdie and divides it by the adjacent channels
(so that 1 --> the expected rms/mean for the bandwidth , integration time).
rms/mean of the narrow band birdie (1428.24 Mhz) for all strips(.ps)
The top figure shows the normalized rms/mean of this channel for 01may05
thru 09jun05 (note that the major tick marks are placed at 12:noon
of the day since they are julian days). On the 18may05 the rms/mean increases
and stayed elevated till 23may05. On 21may05 one strip had the rms/mean
for this channel 1400 times the expected value.
The middle and bottom figures are blowups of the strength so you can see
when the problem started and ended.
The 1428Mhz biridie was strongest on 21may05 in the
strip with scannumber 514184839. The images below are dynamic spectra for
this strip. The image is for pixel 3b where it was strongest.
in the birdie frequency channel versus time (.ps) (.pdf)
for the strip with the strongest birdie.
Looking at the strip where the 1428 mhz birdie is strongest.
The top figure plots the power in frequency channel 1428.24 for scan
514184839. The birdie increased to about 240 times Tsys near second
370 of the strip.
The bottom figure blows up the vertical scale. The birdie has a value larger
that Tsys for the entire strip (tsys=1). The images of the dynamic
spectra do not show this since they normalize each frequency channel of
the image to the median value.
A birdie appeared at 1428.24 Mhz on 18may05 and stayed until 23may05. It
was less than 50 Khz wide. It's strongest value was 240 times tsys on 21may05.
A wider birdie (about 200Khz) was also present at 1427.76 Mhz.
The birdie did not affect the spectra more than a mhz away.
When looking at the 1428.24 birdie on the spectrum analyzer it looked like
there were two narrower birdies separated by about 30khz. On the radio
there was a hum (i'm not sure what frequency) but no other identifiable
The narrow birdie was relatively stable in frequency.
The birdie disappeared before we could do any more troubleshooting.