Observatoire de Paris-Meudon Département de Radio-Astronomie CNRS URA 1757 5, Place Jules Janssen 92195 MEUDON CEDEX " " Vincent CLERC and Carlo ROSOLEN E-mail adresses : Carlo.rosolen@obspm.fr Vincent.clerc@obspm.fr Summary Goals of this study CNET s survey system Day night influence Evolution of average power over 4 years The Nancay s decameter phased array Evolution of average power over 7 years Impact of resolution enhancement with digital receivers in decametric band
The situation in the Decametric Band: Interest for radio astronomy at decameter wavelengths is increasing New projects are under study (LOFAR,DEK2A, ) One of the biggest challenges for the future, will be : To make astronomical observations in between very strong RFI. Compensate ionospheric phase and amplitude effects. Goals of the Study Our goals with this study : Characterize the activity in this band over a long period in term of power, evolution. Set warnings on dynamic considerations and calculations to determine the characteristics of new generations of digital receivers Data Sources We are studying data from two different sources: CNET s survey system* Omnidirectionnal monopole 10 years automatic survey 3 khz resolution survey 3 MHz to 30 MHz frequency coverage Decametric array of Nançay 12 years of routine observations for solar and Jovian emissions Phased array 10 MHz to 100 MHz frequency coverage New generation of digital receiver * Patrick.LassudrieDuchesne@enst-bretagne.fr Roland.Fleury@enst-bretagne.fr
CNET s Survey System In 1991, the National Center for Telecom Studies (CNET) started a program of spectral occupation measurement on the Losquet Island. The system is composed of a calibrated omnidirectionnal antenna and a radio electric field analyzer interfaced to a computer. The receiver measures the power of the signal in each transmission channel from 3 MHz to 30 MHz, with a 3 khz resolution step. One acquisition is started every hour, and lasts 12 minutes. We have used a first set of data collected over years. The goal was to study day/night alternation impact and long term influence of the ionosphere. Ionospheric Effects Over 24 Hours Ionization and concentration in electron has a direct influence on RFI propagation. Two main effects : Cut-off frequency Short and long term phase and amplitude variation
Ionospheric Effects Overview of the Situation The D layer : Lowest layer of the ionosphere Exists at altitudes around 60 to 90 km Is present during the day when radiation is received from the sun Tends to attenuate signals that pass through it Density of the air at this altitude means that ions and electrons recombine relatively quickly After sunset, electron levels fall and the layer effectively disappears. Typically produced as the result of X-ray and cosmic ray ionization. D layer behaves as an attenuator for the radio waves. Ionospheric Effects Overview of the Situation The E layer : Exists at an altitude of between 100 and 125 km Instead of acting chiefly as an attenuator, this layer reflects radio signals although they still undergo some attenuation. In view of its altitude and the density of the air, electrons and positive ions recombine relatively quickly. This occurs at a rate of about four times that of the F layers that are higher up where the air is less dense. After nightfall the layer virtually disappears although there is still some residual ionization There are a number of methods by which the ionization in this layer is generated. It depends on factors including : The altitude within the layer. The state of the sun. The latitude. X-rays and ultraviolet produce a large amount of the ionization light, especially that with very short wavelengths.
Ionospheric Effects Overview of the Situation The F layer : Is the most important region for long distance HF communications During the day, it splits into two separate layers (F1 and F2) At night these two layers merge to give one layer called the F layer. The altitudes of the layers vary considerably with the time of day, season and the state of the sun. Like the D and E layers, the level of ionization falls at night, but in view of the much lower air density, the ions and electrons combine much more slowly and the F layer decays much less. Accordingly it is able to support communications, although changes are experienced because of the lessening of the ionization levels. Most of the ionization in this region of the ionosphere is caused by ultraviolet light, both in the middle of the UV spectrum and those portions with very short wavelengths. Ionospheric Effects Overview of the Situation Recombinaison is proportional with density of molecules and presence of energy The electronic concentration dramatically fluctuate with sun exposure.
Day-Night influence 30 dbm -10-30 -50 Frequency (MHz) 20-70 -90 10-110 -130 3 Example of a dynamic spectrum : july 1997 Day - Night Influence 30 dbm -10-50 -30-70 Frequency (MHz) 20-50 -70-90 10-110 -90-130 3 Zoom on one week of observation : Average power fluctuation.
Ionospheric Effects Level in dbm -10-30 -50-70 -90-110 -130 Level in dbm -10-30 -50-70 -90-110 <85dB> 3 6 9 12 15 18 21 24 27 30 Frequency (MHz) <70 db> -130 3 6 9 12 15 18 21 24 27 30 Frequency (MHz) Amplitude variation are essentially caused by propagation attenuation in open space and absorption (or not) by the D layer of the ionosphere. The disappearance of the D layer during night creates a level raise in the lower frequencies due to reflections on other layers. Wide band receivers have to handle an additional 15 db variation at input levels. 12bits are not sufficient anymore... Future system should have 14bits A/D converters! Evolution of power over 4 Years Dynamic spectrum of average day spectra over 4 years in the region of the Losquet Island. Even if seasons have an influence on average levels, the average level in the band remains stable over the last 4 years in this region.
Evolution of power over 4 Years Evolution of average power in the 3-30MHz band 0 Average power (dbm) - 10-20 - 30-40 - 50 The Decametric array of Nancay Two type of analysis are currently in progress with the data of the DAM array: Long term monitoring analysis of Swept frequency receiver: Determine the average power evolution over 10 years Wave form analysis with ROBIN (ROBust Receiver INterference tolerant) : Fine analysis on spectrum occupation (up to 160Hz resolution) Impact of resolution enhancement on digital processing
The DAM Array Developed in 1975 The Nançay Decameter Array consists in two filled aperture, phased antenna sub-arrays, made of 72 conical helix antennas each, in opposite senses of circular polarization. Characteristics of the Array and its Associated Instruments The main Telescope and receiving system characteristics are : full bandwidth: 10-120 MHz instantaneous bandwidth : one octave antenna gain : 26 db in each polarisation maximum effective aperture at 25 MHz : 2 x 4000 m2 declination coverage: -20 < d < 50 tracking time : ± 4 hr from meridian transit computer controlled pointing and calibration system remote observing capability (through Internet) digital observations data base (since 1990), availability from the Web. Set of high resolution, wideband spectrum analysers: "survey" swept frequency receiver: 10-40 MHz, 1sec time resolution, two polarisation channels, daily operated AOS: 24 MHz bandwidth, 2048 channels, 35 khz frequency resolution, 3 ms time resolution; operated on request. DSP polarimeter: >65 db dynamic range, 12.5 MHz bandwidth, full polarisation, 2 x 1024 frequency channels, 1 ms time resolution; operated on request
New Generation of Digital Receiver For this study, we used the ROBIN receiver. ROBIN : Gigital Interference Tolerant Receiver Wave form and Spectrum acquisition digital receiver 65 db of dynamic range 20MHz, 2 channels 18 Gflops Data Acquisition: Recording of 2 seconds of signal waveform with the 12 bit A/D converters on 10MHz bandpass for the 10-20, 20-30, 30-40 and 40-50MHz bands Spectral analysis with 10 different frequency resolutions and statistical analysis with MATLAB. Vincent.clerc@obspm.fr RFI Mitigation workshop. Bonn, March 28 30 2001 Impact of Resolution Enhancements Resolution = 600 Hz Vincent.clerc@obspm.fr Resolution = 30 khz RFI Mitigation workshop. Bonn, March 28 30 2001
Computational power requirements Fast Fourier Transform (FFT) : OPS FFT = 5.N.Log 2 N Number of FFT/s needed : FFT/s = f sampling / (OV.N) with : N number of frequency bins, OV overlap ratio Computational power rapidly cost a lot of time processing... Impact of Resolution Enhancements The absolute number of free channel is reached for a very high resolution. This threshold depends on the band. 48k 24k 12k 6k 3k 1.5k 750 325 162 In this example, with a 3 khz resolution, we have 95% of free channels in the band!!! While, a 12 khz resolution gives only 82% of free channels... Resolution (Hz)
Impact of Resolution Enhancements Max = 97% Max = 62% Absolute percentage of free channels versus resolution. 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 Max = 21% Max = 33% 1 = 256 pts = 50 khz 2 = 512 pts = 25 khz 9 = 65536 pts = 160 Hz 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 Impact of Resolution Enhancements Resolution needed to recover at least 90% of the free available channels. 35-45 MHz : 6.25 KHz 25-35 MHZ : 1.6 KHz 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 15-25 MHz : 160 Hz 05-15 MHz : 320 Hz 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 1 = 256 pts = 50 khz 2 = 512 pts = 25 khz 9= 65536 pts = 160 Hz
Impact of Resolution Enhancements The situation is summarized in this figure. Therefore, the compromise between resolution and number of free channel to reach can be obtain on this diagram. Consequently, calculation power in digital receivers can be adapted by looking at this diagram... Concluding Remarks Analysis of DAM array data over 10 years must be completed. RFI situation at decameter wavelengths is very particular. Our current approach is : To live with interferences That means : Ultra high dynamic range receivers (>80dB) High time and spectral resolutions (ms and >1kHz) Very High Computational power for real time RFI excision