7 SETI Program at the Medicina INAF Radioastronomy …
53
Fig. 7.3 Block diagram of the NewSpec hardware design
The first approach to developing such a spectrometer was to use a system based on
a high-performance PC with an Intel 3.6 GHz four-core i7 3820 CPU hosting on the
PCI Express bus a 1,5 GS/s, 8-bit resolution DAQ board Ultraview AD8-1500 × 2.
All the data computing tasks were carried out by the CPU that also managed Direct
Memory Access (DMA) internal data handling, operating system control, and data
storage and visualization (see Fig. 7.3). The algorithm used for spectrum analysis
was the FFT implemented in the Fastest FFT in the West (FFTW) library [16–18].
In this configuration, the system was able to analyze the entire Medicina RT-32
IF bandwidth with a maximum of 4 million channels at variable resolutions. The
spectrometer was also able to reach the Hz resolution, but not in real-time because
it required a substantial decimation of the sampled signal due to the impossibility of
the acquisition board to work at bandwidth lower than 250 MHz.
The spectrometer had a cost of about e14,000 (a significant part of which was
devoted to the acquisition board) and the CPU capabilities limited its performances.
Moreover, the system did not have high scalability and reliability mainly because the
DAQ board software drivers were available by very few Operating Systems (OS).
The spectrometer operated adequately only for a few months because an OS
update was enough to put the system out of order [20].
7.2.2 Spectrometer’s Design Final Requirements
After the NewSpec experience, Dr. Montebugnoli carried out another global analysis
of the project to refine guidelines and spectrometer final requirements.
The final goal of the project was to develop a low cost, long term, very open
system that should be available for:
• SETI applications,
53
Fig. 7.3 Block diagram of the NewSpec hardware design
The first approach to developing such a spectrometer was to use a system based on
a high-performance PC with an Intel 3.6 GHz four-core i7 3820 CPU hosting on the
PCI Express bus a 1,5 GS/s, 8-bit resolution DAQ board Ultraview AD8-1500 × 2.
All the data computing tasks were carried out by the CPU that also managed Direct
Memory Access (DMA) internal data handling, operating system control, and data
storage and visualization (see Fig. 7.3). The algorithm used for spectrum analysis
was the FFT implemented in the Fastest FFT in the West (FFTW) library [16–18].
In this configuration, the system was able to analyze the entire Medicina RT-32
IF bandwidth with a maximum of 4 million channels at variable resolutions. The
spectrometer was also able to reach the Hz resolution, but not in real-time because
it required a substantial decimation of the sampled signal due to the impossibility of
the acquisition board to work at bandwidth lower than 250 MHz.
The spectrometer had a cost of about e14,000 (a significant part of which was
devoted to the acquisition board) and the CPU capabilities limited its performances.
Moreover, the system did not have high scalability and reliability mainly because the
DAQ board software drivers were available by very few Operating Systems (OS).
The spectrometer operated adequately only for a few months because an OS
update was enough to put the system out of order [20].
7.2.2 Spectrometer’s Design Final Requirements
After the NewSpec experience, Dr. Montebugnoli carried out another global analysis
of the project to refine guidelines and spectrometer final requirements.
The final goal of the project was to develop a low cost, long term, very open
system that should be available for:
• SETI applications,