Birdies in the pdev device (narrow band)
Previous data taken with the pdev spectrometer used
the "sp1" fpga code and a temporary mixer chassis. The "sp1" code gave
20Khz channel widths (hires is not yet installed) while the mixer
chassis cut out a few hundred Khz about dc.
On 03jul07 data was taken with the "radar device"
that tamara is working on. It uses the pdev spectrometer with fpga code
that filters the data and then writes the time samples out to disc. The
version used was still in the development stages. This data set allowed
us to look at birdies close to dc that may be coming from the switching
The setup was:
- pulsar noise src used to generate the noise.
- The noise was passed through at 25 Mhz lowpass filter. This
filter less than .5 db drop from 0 to 10 Mhz.
- Data was sampled at 156Mhz (internal clock) using a single
- The data was filtered and decimated by 8. It was symmetric since
we only took I rather than I and Q.
- We were limited to 1 seconds worth of data per file (this will be
corrected shortly). 120 files were taken.
- The spectral density was computed from the data and then averaged
for the 120 seconds of data. Resolutions of 595 and 19 Hz were used.
The plots show the
averaged spectral density (.pdf):