NenuFAR Open Time Cycle #4: call for proposals

January 15th, 2024
We invite the worldwide community to apply for observing with NenuFAR during Cycle #4
Cycle #4: 1st June 2024 to 30 November 2024
Submission deadline : Monday 11 March 2024, 12 UT

Proposals to be sent to nenufar-proposals@obs-nancay.fr




NenuFAR is a large low-frequency radiotelescope, partly constructed and in commissioning, but already operational in « Early Science » phase. It is located in Nançay, 200 km south of Paris, France, in the Nançay Radioastronomy Observatory (NRO).
NenuFAR will gather at completion 1938 dual polarization antennas hierarchically distributed (Figure 1), of which 1064 are presently in operation.
It observes in the 10-85 MHz range, between the Earth’s ionospheric cutoff and the radio broadcast FM band. Most of the antennas are located within a core of diameter ~400 m, except for 6 groups of 19 antennas that will be at distances up to ~3 km.
The antennas are connected in parallel to several receivers, that allow the instrument to operate in 4 distinct modes (if needed simultaneously) :

  • as a standalone beamformer, hereafter the NenuFAR Beamformer mode
  • capturing waveform snapshots from selected antenna signals, hereafter the Transient Buffer mode
  • as a standalone imager, hereafter the NenuFAR Imager mode
  • as a giant low-frequency (LBA) station of the LOFAR array, hereafter the LOFAR super station (LSS) mode

The operating modes and data products are further described below.

NenuFAR Science Case

Science Management Plan

NenuFAR Family Meeting 2024: 

NenuFAR is about to enter Cycle #4 of its operation. Observing time is allocated to long-term programs and regular programs over a 6-month observing periodicity. We propose to gather in a hybrid “family meeting” during April 3rd-4th, 2024  in Nançay to present and discuss the science and current results obtained since the last workshop covering the end of the “Early science” phase and during the first cycles carried in open time.
We want to offer to users and new potential users, the opportunity to get the latest up-to-date information about the instrument, data processing methods and facilities. Participants will get the possibility to do practical sessions around various projects and tools (Beamforming, pulsar, imaging, etc.) with NenuFAR data.

 See details and workshop material on https://nfm2024.sciencesconf.org/

NenuFAR users workshop #2: 

After nearly 2 years of conducting observations in the “Early Science” phase, NenuFAR users gathered again for a workshop in Nançay on November 17-19, 2021. This workshop was dedicated to users and potential users, the opportunity to highlight results obtained from the various teams, and get the last up-to-date information about the instrument, data processing methods and facilities. Several practical sessions were organized around various projects and tools (Beamforming, pulsar, imaging, etc.) with NenuFAR data.

 See details and workshop material on https://nenufaruser2021.sciencesconf.org/

NenuFAR users workshop #1: 

NenuFAR will be a hopefully powerful but certainly complex instrument. In order to help the teams to prepare their activities, and to introduce observers to NenuFAR capacities, operating modes, and data, a a first NenuFAR workshop was held in Nançay on March 18-20, 2019. 

 See details of programme and attendance on https://nenufar2019.sciencesconf.org/

NenuFAR  Key Programmes & Early Science phase

NenuFAR (New Extension in Nançay Upgrading LOFAR) is a new radioastronomy instrument, operating in the 10 – 85 MHz frequency band, which is in the last stages of its construction at the Station de Radioastronomie in Nançay (France). At completion, the instrument will be composed of a compact core of 96 mini-arrays of 19 antennas each plus 6 distant mini-arrays. Four operating modes will be implemented: standalone beamformer, standalone imager, Transient Buffer mode, and LOFAR super-station.

The scientific questions that can be adressed with NenuFAR are very diverse. A call has been issued early 2019 by the Comité Scientifique de NenuFAR (NenuFAR Scientific Committee, CSN) to the French community, plus foreign colleagues already involved in the project, for proposing Key Programmes (KP) using NenuFAR for the Early Science phase of the instrument. Key Programmes are structured teams led by a KP coordinator.

During the Early Science phase (July 2019 to December 2021), approximately 8000 hours of observations was available for KPs. The KP proposals included the various standalone operating modes of NenuFAR, or a combination of these modes. The LSS mode will ultimately be accessed through regular LOFAR calls.

Key Programmes have been originally proposed during the “Early Science” phase, gathering ~140 scientists (~50% French). Some programs evolved into “Long-Term” programs (LT), “Special” programs (SP) and “Regular” programs (RP).

The current list (as of Cycle #4) are listed below:

LT01: Cosmic Dawn (Koopmans et al.)

LT02: Exoplanets & Stars (Zarka et al.)

LT03: Pulsars (Theureau et al.)

LT04: Transients (Corbel et al.)

LT05: Fast Radio Bursts (Decoene et al.)

LT06: Planetary Lightning (Grießmeier et al.)

LT07: Joint Jupiter studies (Lamy et al.)

LT09: Cluster filament & Cosmic Magnetism (Bonnassieux et al.)

LT10:  Radio recombination lines (Gusdorf et al.)

LT11: Sun (Briand, Morosan et al.)

LT12: Radio detection of Gamma showers (Dallier et al.)

LT13: SETI (Hellbourg et al.)

SP16: Formation of students (Girard et al.)

SP17: Radio amateur community (Maintoux et al.)

RP1A: Faraday Tomography in the 3C196 field (Bracco et al.)

RP1B: NenuFAR Low-Frequency Sky Survey (Girard et al.)

RP1C: Low-Frequency measurements on Cas A (Stanislavsky et al.)


Individuals wishing to join a KP should contact the CSN at nenufar-csn@listes.obs-nancay.fr (or via the contact form at the bottom of this page) who will transmit to the relevant KP P.I.


NenuFAR elementary antennas are crossed dipoles in inverted-V shape erected 1.6 m above a 3 m x 3 m ground plane (metallic grid). The antenna radiators are those developed for the LWA (USA). They are coupled to a custom-designed low noise preamplifier (developed in Nançay and Nantes) that ensures a relatively flat response across the entire 10-85 MHz band with a level of sky background several dB above the preamplifier noise at all frequencies. As a consequence, the sensitivity is sky-limited at all frequencies

The effective area of the instrument can be estimated by taking an effective area per dipole antenna Ae-dipole(l) » l2/3 and summing over the total number of antennas taking into account effective area overlaps for nearby antennas. Effective areas of the NenuFAR dipole are listed in Mini-arrays description (Table 1 in Mini-array section)

All dipoles across NenuFAR are (as for LOFAR) aligned at 45° from the meridian, either along the NE-SW direction (so-called NE or X dipoles) or along the NW-SE direction (so-called NW or Y dipoles). Combination of the two linearly polarized antenna signals allows to retrieve the full polarization of the incoming waves (4 Stokes parameters).

NenuFAR antennas (dipole + grid) have a quasi-isotropic response in the E-plane (perpendicular to the dipole) whereas the gain drops below an elevation ~20° in the H plane (containing the dipole).

Antenna response at 20 & 80 MHz in E and H planes.


NenuFAR antennas are grouped in hexagonal tiles of 19 crossed-dipoles, hereafter called Mini-Arrays (MA). The elementary mesh of the hexagon is an equilateral triangle of side length 5.5 m (inter-antenna distance within an MA), thus the net diameter of a MA is 25 m (Figure 4). Effective areas of the NenuFAR MA are listed in Table 1.

Within each MA, the signals of the 19 antennas (in each polarization) are combined through an analog phasing+summation system in order to form a steerable beam, that restricts the instantaneous Field of View (FoV) but provides higher sensitivity within it. The analog MA phasing system is formed of delay lines (switchable cables lengths) and thus essentially achromatic (the philosophy is the same as that of the square 4×4 tiles of LOFAR-HBA or MWA antennas, except that the hexagonal shape provides – for the same number of elementary delay lines, 10 per MA or per 4×4 tile – more effective area and order-6 grating beams of lower amplitude than the order-4 grating beams of 4×4 tiles). The delay lines are 7-bit systems of switchable cables lengths, permitting to point 27 = 128 discrete directions in the EW direction and 128 in the NS direction, hence a total of 128×128 = 16384 pointable directions in the sky, from horizons to zenith. After analog phasing+summation, each MA delivers two linearly polarized signals (NE and NW) via coaxial cables to the receivers.

The analog-phased steerable MA beam has a FWHM of order of λ/D with D~25 m, i.e. from ~8° at 85 MHz to ~46° at 15 MHz (Figure 5). Beam widths and instantaneous FoV at several frequencies are listed in Table 2.

Frequency (MHz)

Wavelength (m)

Ae-dipole (m2)


Ae-core (m2)


























Table 1 : Effective areas Ae-dipole, Ae-MA, and Ae-core and Ae-all as a function of the frequency and wavelength.


Frequency (MHz)

Wavelength (m)

Ɵ-Ma (°)

FoV-Ma (°2)


Ɵ-all (’)

























Table 2 : θMA and FoVMA are the beam width (FWHM) and instantaneous Field of View of NenuFAR’s MA as a function of the frequency and wavelength. θcore is the beam width (in beamformer mode) or angular resolution (in imager mode) of the core. θall is the angular resolution of the imager when all MA (core+remote) are used together.

Sketch of a Mini_array with 0° rotation

Analog-phased MA beams at 20  pointed to the zenith, in linear and dB scales. Cuts are shown at azimuth = 30°, where grating lobes are maximum.

Analog-phased MA beams at 80 MHz pointed to the zenith, in linear and dB scales. Cuts are shown at azimuth = 30°, where grating lobes are maximum.

Global distribution of Mini-Arrays

NenuFAR’s antenna field consist of a dense core of 96 MA distributed within a disk of diameter ~400 m (Figure 6), plus 6 remote MA placed to the North-West, East/North-East and South of the core. The distribution of core MA ensures a quasi-perfect coverage of the uv plane for baselines between 25 m and 400 m, with Gaussian radial density and uniform azimuthal density, ensuring a Gaussian main beam in Beamformer mode (Figure 7). Remote MA are placed at distances in geometrical progression from the core up to ~3 km, resulting in a good coverage of the uv plane for baselines up to 3 km when combined with Earth rotation during a few hours (and multi-frequency synthesis if allowed by the scientific objectives).

The MA are rotated with respect to each other by angles multiple of 10°, in order to reduce the contribution of the side (and grating) lobes when MA signals are summed together. The antennas are back-rotated by the same angle to remain aligned throughout NenuFAR (same polarizations as LOFAR antennas). The rotation angle 0° corresponds to the 5 lines of antennas forming the hexagon all along the EW axis. Rotation angles from 0° to 350° are attributed, that will correspond to rotated local coordinates for each MA. However, due to the hexagonal symmetry, MA rotated by multiples of 60° have the same analog beam responses in the sky. For deconvolution of MA response during imaging, 6 rotation angles only need to be taken into account : 0°, 10°, 20°, 30°, 40°, 50°. All remote MA have the same rotation angle of 0°.

Figure 6 : Global MA distribution within NenuFAR, with deployment plan indicated.


The 96 MA x 2 polarizations analog signals from the core are split in order to feed simultaneously the LOFAR FR606 station backend and the main NenuFAR standalone receiver LaNewBa.

The outputs from the NenuFAR core MA are connected to the 96×2 LBL inputs (not used in LOFAR stations and thus available) of the LOFAR FR606 station, digitized, beamformed, and sent to the LOFAR central correlator as standard LOFAR signals. For the use of this LOFAR super station mode, please refer to the LOFAR astronomers page.

LaNewBa digitizes each of the 96×2 input signals at a sampling frequency of 200 MHz. The data streams are then channelized in subbands of 200 MHz/1024 = 195.3125 kHz bandwidth each. Beamforming then consists in coherently summing up the complex data streams from all MA for a given subband, polarization and direction (θ,φ) in the sky, after application of adequate phase shifts. A beamlet consists of one subband x 2 polarizations beamformed to one direction in the sky, i.e. it is defined by a triplet (fc,θ,φ) with fc the central frequency of the subband. The raw data stream corresponding to one beamlet consists of 195312 pairs of complex X (NE) and Y (SW) signals values per second. LaNewBa computes and delivers 768 such beamlets at any given time, that correspond to an instantaneous usable bandwidth of 150 MHz. These 768 beamlets can be distributed in (fc,θ,φ) as desired, from 2 beams (θ00) and (θ11) with 75 MHz bandwidth (i.e. the full band of the instrument) to 768 beams paving the analog beam (and possibly more) at a single frequency fc, through any intermediate scenario.

LaNewBa also computes and delivers in parallel 3 types of statistical data :

  • Spectral STatistics (SST) are X and Y spectra per MA per polarization per second, that permit to monitor the response of each MA ;
  • Beamlet STatistics (BST) are the beamlet outputs per polarization averaged at 1 s resolution, that provide observation quick-looks ;
  • Cross-correlation STatistics (XST) are interferometric visibilities computed from all MA signals at a rate of one full set of cross-correlations per second within 16 subbands. These statistics data are described in details in the next section (data products).

UnDySpuTed (Dynamic spectroscopy & pulsars modes)

The beamformed raw data from LaNewBa are distributed to several calculators via virtual local area network (VLAN). At present, the 2 identical CPU/GPU calculators UnDySPuTeD are operational. Each of them can process the raw data from 384 beamlets in one of 3 modes :

  • A Pulsar mode, in which the signals in each beamlet can be coherently dedispersed and folded, either in known pulsar mode (pre-determined DM and period), or in search mode (exploration of DM-period space) ;
  • A time-frequency (or dynamic spectrum) mode, in which the signals in each beamlet can be further channelized (up to 64 channels / beamlet, i.e. a spectral resolution ~3 kHz) and integrated (1 ms to 1 s integration), with full polarization information (auto-correlations and cross-correlations of the beamformed NE and NW signal are computed and recorded).

A raw waveform mode, in which the raw data stream (195312 pairs of complex X and Y signals values per second per beamlet) is written to disk.

A single mode can be used on each UnDySPuTeD calculator at a given time (possibly the same on both if 768 beamlets need to be processed simultaneously in the same mode).

Data products:

  • Time-frequency files (*.spectra)
  • Waveform
  • Pulsar

Lanes limitations:

  • The stream of data are routed through 4 separate lanes (1 lane is a real-time compute node with dedicated CPU, GPU and RAM)
  • There could be only a maximum of 192 beamlets per lan.
  • If multiple digital beams are set up, their configurations should be identical within a given lane. Otherwise, only the first configuration is applied to every beam within the given lane.


NICKEL Correlator (Interferometric mode)

The NenuFAR correlator (the selection of which is ongoing) will fed by the signals from the 96 MA x 2 polarizations, digitized and channelized by LaNewBa, plus signals from the remote MA digitized and channelized directly at the foot of each MA and send to the correlator without passing through LaNewBa. The correlator will further channelize the signals and compute interferometric visibilities from all MA signals at a rate of one full set of cross-correlations per second within 75 MHz of bandwidth at 3 kHz resolution (384 subbands x 64 channels per subband = 24576 frequency channels). Note that this is the only receiver that take into account distant Mini-arrays.

Data products:

  • “Measurements Sets” (.MS) following LOFAR standards
  • NICKEL can currently produce a maximum of 244 Subbands. The number of selected subbands should be a multiple of 4.
  • Raw MS (“L0” – level 0) are recorded on the COPPER storage and pre-processig nodes. By default, pre-processing steps (Flagging & averaging) are performed only on COPPER nodes (parameters can be defined at programming time). Pre-processed data (“L1”) can then be processed further on the offline cluster (access must be requested).


SETI (piggy-back receiver)

A dedicated SETI spectrograph with very high spectral resolution and possibly more advanced functionalities is under study. It will be connected to the VLAN and process in parallel the beamformed data delivered by LaNewBa, enabling piggyback SETI observations parallel to NenuFAR beamformed observations.


Transient Buffer Board (TBB)

The digitization boards of LaNewBa include transient buffers that store the last 5 seconds of waveform for each MA (or selected antenna within each MA) and each polarization. These buffers can be frozen and written to disk in response to an external trigger.

Report to Data Product for more details.

The NenuFAR receivers and data distribution are sketched below


NenuFAR pointing is done in 2 steps : (1) analog pointing of the MA and (2) digital pointing of the whole array.

Analog pointing is performed via the MA delay lines permitting to point a grid of 128×128 = 16384 directions in the sky. The grid step is <1° at zenith, corresponding to a gain variation <<1%, and increases toward the horizons (Figure 9). The working interval of NenuFAR corresponds to elevations ≥20°. Not all azimuths can be pointed for an elevation <20°. Furthermore, the instantaneous MA gain pattern is the product of the array term depending only on MA topology, by the antenna response. As a result, the real maximum of the MA gain pattern (when the MA does not point to the zenith) is located at a larger elevation than the maximum of the array term, this elevation is frequency-dependent. This is called the beam squint effect. In order to partly correct for it, the array term of the MA actually points to a pre-defined lower elevation than the desired one if the squint correction is enabled (a reference frequency must be specified). This lower elevation cannot be lower that 20°, due to the above limitation, so that the beam squint cannot be completely corrected for a target (desired) elevation ≤30°.

Digital pointing is continuous (grid-free). It is achieved by digital phasing of MA signals before summation (in beamformer mode) or correlation (in imager mode). Analog pointing, and digital pointing in beamformer mode, are locked on round multiples of 10 s, i.e. they occur at min:00 s, :10, :20, :30, :40, and :50 s. This ensures that any RFI or anomaly related to pointing can be easily identified and mitigated. Analog re-pointing of each MA occurs only if the nearest grid position has changed since the previous pointing. Not all MA need to be repointed together, due to their different rotation angles, implying that the grid of Figure 9 is rotated accordingly for each MA, so that the source trajectory through the grid is different for each different rotation angle. Digital re-pointing occurs every 10 s, and causes a gain variation ≤0.5% (≤0.02 dB) as a function of time.

Analog pointing is hardware-dependent (delay lines), which implies that a single pointing direction per MA is possible at any given time. Sub-arrays of NenuFAR can be defined with different analog pointing for each sub-array (the default mode is an analog-pointing of all MA in the same direction). On the contrary, digital pointing is software-defined per beamlet, so that up to 768 directions are in principle pointable simultaneously.

NenuFAR can work in transit or tracking mode. In tracking mode, the target can be defined by its coordinates (RA, DEC), its name (NenuFAR’s pointing is interfaced with IMCCE ephemeris service for solar system objects, and for the CDS/Simbad database for all other astronomical objects), or a list of coordinates in (Azimuth,Elevation, that follow the Geographic convention – AzElGeo – with origin of azimuths to the local North). When working in transit mode, the analog pointing is fixed to the nearest grid point and the digital pointing is directed exactly to the desired coordinates. When working in source tracking mode, the direction pointed at each re-pointing time (t = min:00, 10, …50 s) is the direction of the source at t+5 s, so that the source path crosses the centre of the beam in each consecutive 10 s interval.

Figure 9 : Grid of analog-pointable positions in local coordinates.

Summary of NenuFAR characteristics

  • Number of antennas : 1938 = 96 core + 6 remote MA of 19 antennas each
  • Dimensions : MA diameter = 25 m, core diameter ~400 m, remote MA up to ~3 km
  • Number of baselines : 5151
    = 4560 baselines from 25 m to 400 m + 591 baselines from 400 m to 3 km
  • Frequency range : 10-85 MHz (λ = 3.5 – 30 m)
  • Time-Frequency resolutions : δf = 195.3125 kHz x δt = 5.12 μs
    Channelization down to 3.05 kHz x time integration down to 1 ms
    Waveform at 5 ns time resolution
  • Full polarization (4 Stokes)
  • Collecting area : from ~220 λ2 to ~650 λ2 (function of effective area overlap within MA)
    • Core:  80 MA (current) up to 96 MA (complete)
    • Distant Mini-arrays: 4 MA (current) up to 6 MA (complete)

      Collecting area (m2)

      Frequency (MHz)



      Core only

      15 MHz

      69000 m2

      83000 m2


      50 MHz

      18000 m2

      22000 m2


      85 MHz

      6300 m2

      7500 m2

      Core + Distant Mini-arrays

      15 MHz

      72000 m2

      88000 m2


      50 MHz

      19000 m2

      23000 m2


      85 MHz

      6600 m2

      8000 m2

Table 3: NenuFAR collecting area as a function of frequency and telescope completion

NenuFAR Effective area as a function of frequency


  • Pointing : from declination δ = -23° to δ = +90°
  • Field of View : ~46° (1650°2) at 15 MHz to ~8° (51°2) at 85 MHz
  • Angular Resolution : from 2.9° at 15 MHz to 0.5° at 85 MHz for the core
    from 23’ at 15 MHz to 4’ at 85 MHz for the core+remote MA   (down to 0.4″ in LSS mode)

Angular resolution (°)

Frequency (MHz)



Core only

15 MHz




50 MHz




85 MHz



Core + Distant Mini-arrays

15 MHz




50 MHz




85 MHz



Table 4: Maximum angular resolution as a function of frequency and telescope completion

NenuFAR angular resolution as a function of frequency

  • Sensitivity (thermal noise) :
    from ~130 mJy at 15 MHz to ~9 mJy at 85 MHz with Δf=10 MHz x Δt=1 h
    (similar values for the core & core+remote MA)
  • Confusion at zenith :
    from ~120 Jy at 15 MHz to ~2 Jy at 85 MHz for the core
    from ~5 Jy at 15 MHz to ~100 mJy at 85 MHz for the core+remote MA

Confusion noise (Jy)

Frequency (MHz)



Core only

15 MHz

105.2 Jy

86.9 Jy


50 MHz

7.1 Jy

5.8 Jy


85 MHz

2.2 Jy

1.8 Jy

Core + Distant Mini-arrays

15 MHz

5.7 Jy

4.8 Jy


50 MHz

383 mJy

323 mJy


85 MHz

116 mJy

98 mJy

Table 5: Estimated confusion noise as a function of frequency and telescope completion

NenuFAR confusion noise as a function of frequency

Data products

The suite of NenuFAR receivers delivers/will deliver the following data products (names of data products are underlined.

Beamformer mode :

  • Pulsar data (PSRFITS format, processed by the PRESTO pulsar pipeline) :
    LaNewBa beamformed data is either coherently dedispersed within each 195 kHz beamlet and folded (or not) in known DM and period pulsar mode, or either saved in filterbank/search mode (to enable future exploration of DM-period space). Full polarization is computed within each beamlet.
  • DynSpec data (proprietary format) :
    LaNewBa beamformed data in each beamlet is further channelized by UnDySPuTeD (up to 64 channels / beamlet, i.e. a spectral resolution ~3 kHz) and integrated (1 ms to 1 s integration), with full polarization information (auto-correlations and cross-correlations of the beamformed NE and NW signal are computed and recorded). DynSpec data are fully mapped by a quadruplet (t, beamlet # = (θ,φ,f), channel #, Stokes parameter).
  • Waveform data (proprietary format) :
    LaNewBa raw beamformed data stream (195312.5 pairs of complex X and Y signals values per second per beamlet) is written to disk by UnDySPuTeD.
    similar to DynSpec data product with very high spectral resolution, and possibly more advanced functionalities. The NenuFAR-SETI receiver is under study.
  • BST (Beamlet STatisics) data (FITS format) :
    LaNewBa beamformed data in each beamlet is integrated over 1 s in X and Y polarizations without channelization narrower than 195 kHz. BST data are fully mapped by a triplet (t, beamlet # = (θ,φ,f), X or Y polarization). NenuFAR BST data are similar to BST data from LOFAR stations. They are systematically recorded in parallel to any other data product during all observations, with the purpose of (i) providing a quick-look of the observation and (ii) providing the observation context in their header (complete NenuFAR setup, active MA, beams, etc.). BST data can be used for science requiring low time-frequency resolutions (1 s x 195 kHz) and no polarization information (X & Y intensities only). They provide the same sensitivity as DynSpec data integrated in the same t-f bins.

Transient Buffer mode :

  • PARTLY AVAILABLE– TBB (Transient Buffer Board) data (proprietary format):
    • TO BE DONE: last 5 seconds of waveform from each MA (or selected antenna within each MA) and each polarization (X, Y), frozen in response to an external trigger and written to disk.
    • AVAILABLE: data from all MA are captured on external software trigger synchronized to GPS time. 10 uS of data are acquired without interrupting the acquisition. 

Imager mode :

  • Visibilities data (MS format) : the correlator will produce LOFAR-compatible Measurement Sets (MS) containing full-polarization cross-correlation matrices in up to 384 subbands (75 MHz total instantaneous bandwidth) channelized at up to 3 kHz resolution (64 channels / subband) once per second. If necessary, up to 4 sets of visibilities per second can be recorded, with reduced bandwidth. 1 MS per subband is written to disk. The phase centre is pointed to by application of digital delays to the signals from each MA. The FoV will be limited by the analog MA beam. NenuFAR Visibilities are pre-processed using the standard tools used for LOFAR pipelines.
  • XST (Cross-Correlation STatistics) data (MS format) : LaNewBa computes interferometric visibilities from all MA signals at a rate of one full set of cross-correlations per second within 16 subbands (i.e. 3 MHz bandwidth). No channelization is performed within subbands (spectral resolution of 195 kHz). No pointing is applied (the phase centre is at the zenith). Data is recorded in FITS format. A python program nenums.py is run immediately after acquisition that converts the recorded data in MS format. During the conversion, the data can be rephased to any phase centre in the sky, and MS can simulate the tracking of a (RA, DEC) position, or transit (fixed pointing) at the zenith. The 16 subbands / s can sweep across NenuFAR’s frequency range. 1 MS is written per subband and per 2-hour slice of observation, padded with zeroes at time steps when the corresponding subband is not recorded (in case of frequency sweeping of the observed subbands during the observation). NenuFAR XST data are similar to XST data from LOFAR stations, except that the latter are recorded at the rate of one full set of cross-correlations per second per subband. XST data are recorded primarily for inter-MA phase calibration (permitting to build the calibration table to be loaded in LaNewBa that ensures fully coherent beamforming), but they can also be used for science requiring low spectral resolution (195 kHz) and narrow instantaneous bandwidth (3 MHz / s). The default use of XST data for imaging is deprecated since the default imaging receiver is NICKEL.
    XST stream are now used in the NenuFAR-TV (snapshot image of the sky diaphragmed by the analog MA beam, at a rate of ~1 image / min).

LOFAR Super Station mode :

  • TO BE AVAILABLEVisibilities data (MS format) : they will be produced by the LOFAR correlator, with NenuFAR replacing the LOFAR FR606 LBA antennas to increase the sensitivity of long baselines, giving access to more calibrators and better SNR. The MA will be analog pointed to the target during the observation. The data, in MS format, will be identical to LOFAR long baselines imaging data, and processed similarly. NenuFAR Imager Visibilities can be recorded simultaneously with LOFAR Visibilities in LSS mode, in order to give access to baselines down to 25 m, shorter than the shortest LOFAR baselines of 68 m. The LOFAR and NenuFAR Imager MS can be combined during the analysis.

Other  products :

  • SST (Spectral STatisics) data (FITS format) : they are recorded systematically by LaNewBa during all observations. They consist of spectra of all core MA in X and Y polarizations without channelization narrower than 195 kHz, recorded each second. SST are recorded primarily for engineering purposes (monitoring the response of all MA in both polarizations). They are kept at 1 s resolution for 6 months, and then integrated at 10 s resolution (one full set of spectra per 10 s) before being archived.
  • TO BE AVAILABLE – Incoherent summation data (format tbd) : it consists in the summation of the intensities from each MA (by opposition to the coherent beamformed mode where the amplitudes are summed). LaNewBa can perform this incoherent summation but only at subband spectral resolution (195 kHz). Incoherent summation in narrower channels cannot be performed by the UnDySPuTeD receivers because they are fed by the coherently summed beamlet signals from LaNewBa, while channelization must be performed before incoherent summation. It may be performed by the NenuFAR correlator if the selected hardware allows for it.

Identified features under study

  • Pointing of the instrument is shifted from ideal pointing by a 10’ to 20’ shift dependent in the position in the sky. Pointing calibration is ongoing and expected to bring the pointing accuracy down to a few arcmin.
  • Sawtooth level variations of 0.1 to 0.2 dB amplitude may be observed during beamformed tracking observations, related to the above pointing shift. They should be strongly reduced or eliminated after pointing correction.
  • Intermittent broadband RFI can appear around 40 MHz. Its origin is being investigated (windmills ? other ?).
  • NenuFAR measurement set produced via conversion of XST data are compatible with most of CASA suite, most NDPPP tasks, Pyrap/TaQL, WSClean. They are not (yet) compatible with NDPPP SmartDemix, BBS, DDFacet, CASA listobs() & gaincal().The column « corrected data » is not created by default (to save space). It is created automatically at calibration (e.g. with NDPPP).
  • Intensity blobs at a few % level may appear in beamformer time-frequency data, generally above ~55 MHz. They are due to strong sources (A-team, Galaxy) in grating lobes. They are largely predictable via instrument+GSM simulations and show smooth variations that can be mitigated. The frequency dependence comes from the relation between inter-antenna distance (d=5.5 m) and the observed wavelength (λ) that governs the presence of grating lobes :
    • grating lobes are always present in the sky for λ ≤ d, i.e. f ≥ 54.5 MHz
    • grating lobes are present when pointing is not at zenith for λ ≥ d ≥ λ/2, i.e. 27.3 MHz ≤ f ≤ 54.5 MHz
    • no grating lobe is present in the sky for d ≤ λ/2, i.e. f ≤ 27.3 MHz




LOFAR’s Black Board Self-cal software


LOFAR Super Station


Beamlet Statistics


Long Wavelength Array


Common Astronomy Software Applications package


Mini-Array (of 19 crossed dipoles antennas)


Centre de Données de Strasbourg


Measurement Set


Central Processing Unit


Murchison Widefield Array


Dispersion Measure


LOFAR’s New Default Pre-Processing Pipeline


Flexible Image Transport System


New extension in Nançay upgrading LOFAR


Frequency Modulation radio broadcast band, 87-108 MHz


Nançay Radioastronomy Observatory


Field of View


PulsaR Exploration and Search TOolkit


Full Width at Half-Maximum


Right Ascension


Graphic Processing Unit


Radio Frequency Interference


Global Sky Model


Search for Extra Terrestrial Intelligence


High Band Antennas from LOFAR (110-250 MHz)


Signal to Noise Ratio


Institut de Mécanique Céleste et de Calcul des Éphémérides (Observatoire de Paris)


Spectral STatistics


LSS advanced New Backend


Transient Buffer Boards


Low Band Antennas from LOFAR (30-80 MHz, down to 10 MHz with reduced sensitivity)


Unified Dynamic Spectrum Pulsar and Time Domain receiver


Low Band Low antenna inputs unused in LOFAR stations


Virtual local area network


LOw Frequency ARray


Cross(X)-correlation STatistics

Sections in preparation

  • Publication policy, builders list
  • Sensitivity, FoV, Resolutions : plots and calculator in Python and IDL will be provided.
  • Graphic User Interface for the programmation of the observations and monitoring of the data.
  • Calibrators : A-team sources are used for phase calibration of the beamformer mode and flux calibration of the beamformer and imager modes.
  • Instrument simulations : Python and IDL tools will be provided. Coupled to a Global Sky Model, they allow to simulate and interpret the expected response of the beamformer.
  • Data volume : a calculator in Python and IDL will be provided.
  • Data storage and processing : data will be stored by default on the Nançay data centre, that will provide long-term archiving as well as some processing capabilities (nancep computer nodes). IDL and Python software packages will be provided for reading and basic processing of beamformer data. The LOFAR pipelines & tools (NDPPP, BBS, KillMS/DDfacet …) can be used to process imager data.

Contact NenuFAR team