1 Introduction

Observations with radio spectrographs help to understand transient activity such as flares and coronal mass ejections (CMEs) in the solar atmosphere through their associated radio bursts, which appear either as a patch of continuum or a drifting emission from high to low frequencies as a function of time. Although bursts can occur over a wide range of frequencies, ≈ 10 kHz – 10 GHz, in ground-based observations most of them are observed with good contrast at frequencies<500 MHz (Benz 1993). Radio emission at different frequencies originates at different levels in the solar atmosphere because of the decrease in the electron density with increasing radial distance. The appearance of bursts also changes from one frequency to the other. To obtain data on the radio emission related to flares and CMEs, near-simultaneous observations over a wide range of frequencies are therefore required.

There are numerous radio antennas with distinct transmission/reception properties. Conventional dipole, Yagi–Uda antennas are limited to observations over a very narrow range of frequencies (Kraus, Marhefka, and Khan 2006). It was pointed out by Rumsey (1957) that the geometry of a structure specified entirely by its angles will be ideal for operation over a wide range of frequencies. This is consistent with the Babinet principle, i.e. the impedance of a structure that is self-complementary is independent of frequency. The combined principle of angle-only specified geometry and self-complementary structures led to the development of Log-Periodic Structures (LPS). Their characteristics such as the input impedance, the Voltage Standing-Wave Ratio (VSWR), radiation pattern is a periodic function of the logarithm of the frequency. Initially, the LPS were designed as planar structures. But later they were modified by folding two planar structures to obtain a non-planar Log-Periodic Dipole Antenna (LPDA) (Duhamel and Isbell 1957; Isbell 1960). The elaborate mathematical treatment for designing a LPDA was later developed by Carrell (1961a). The LPDAs are the typical radio-frequency (RF) signal-receiving elements in low-frequency radio astronomy (< 1 GHz), particularly where observations over a wide range of frequencies are required (Erickson and Fisher 1974; Boischot et al. 1980; Ebenezer et al. 2001; Ramesh et al. 2008; Benz et al. 2009; Maan et al. 2013).

Using LPDAs, we have recently constructed a new spectrograph capable of observing in the frequency range≈40 – 440 MHz at the Gauribidanur Radio ObservatoryFootnote 1 to investigate 1) the U-type bursts in which the drift changes gradually from low to high frequencies some time after their initial drift from high to low frequencies. These bursts are interpreted as signatures of radio emission associated with closed magnetic-field lines in the solar atmosphere (Benz 1993). 2) Type IV bursts (start frequency range≈300 – 400 MHz) and their association with X-ray, microwave, white-light emissions, as well as their possible role as coronal magnetic-field probes (Aurass et al. 2003; Ramesh et al. 2013a). 3) Quasi-periodic type III radio bursts and their usefulness for coronal seismology (Mangeney and Pick 1989; Aschwanden, Benz, and Montello 1994; Ramesh et al., 2003, 2005). 4) The fundamental (F) – harmonic (H) type II bursts and their association with the flares and the CMEs (Claßen and Aurass 2002; Gopalswamy et al. 2005; Nindos et al. 2008; Ramesh et al. 2010b; Magdalenić et al. 2010; Cho et al. 2013). 5) Type I bursts (occurrence frequency range≈100 – 400 MHz range) and their association with CMEs (Kerdraon et al. 1983; Böhme 1993; Chertok et al. 2001; Ramesh and Shanmugha Sundaram 2001; Kathiravan, Ramesh, and Nataraj 2007; Iwai et al. 2012). 6) Absorption bursts and their association with the onset of CMEs and coronal dimming (Ramesh and Sastry 2000; Ramesh and Ebenezer 2001). 7) Radio bursts associated with energy releases≲1027 erg (Kundu et al. 1986; Kliem and Bastian 1996; Ramesh et al., 2010a, 2013b; Saint-Hilaire, Vilmer, and Kerdraon 2013; Iwai et al. 2013). This article describes the antenna array, the receiver system, and the first results.

2 Design of the LPDA and Test Results

Our idea was to design an LPDA that can work in the frequency range 40 – 440 MHz with a gain of ≈ 7 dB (which is nearly the same as that of most of the commercial LPDAs), broad response/radiation pattern (beam) so that the Sun can be observed for a longer period of time without mechanically or electrically steering the antenna, and a VSWR≲2 so that there is a good impedance match and the incident signal on the antenna is maximally transmitted to the subsequent stages of the receiver system. We designed and fabricated the LPDAs at the Gauribidanur Radio Observatory following the standard procedure mentioned in the literature (Carrell 1961b; Wakabayashi et al. 1999; Balanis 2005; Kraus, Marhefka, and Khan 2006; Sasikumar Raja et al. 2013). Figure 1 shows the VSWR of the LPDAs used in the GLOSS. The values are ≲ 2 over the entire frequency range 40 – 440 MHz. The far-field radiation pattern of the LPDAs designed for the GLOSS were measured in a manner similar to that described in Balanis (2005), Sasikumar Raja et al. (2013). Note that for any LPDA, response patterns are usually measured in two orthogonal directions: 1) in the direction of the arms of the LPDA (i.e. the E-plane) and 2) in the direction perpendicular to the arms of the LPDA (i.e. the H-plane). Figure 2 shows the E-plane and H-plane patterns at a typical frequency of 80 MHz. For the E-plane, the 3 dB or the half-power beam width (HPBW) of the pattern is ≈ 80∘. The corresponding width for the H-plane pattern is ≈ 110∘. The E-plane and H-plane widths were found to be approximately the same at different frequencies in the range of 40 – 440 MHz. The measured values of the HPBW along the E-plane (θ) and H-plane (ϕ) at 80 MHz correspond to a solid angle of Ω=θϕ≈2.7 sr. From this we calculated the directional gain of the antenna, i.e. \(G=10 \operatorname{log}_{10}(4\pi/\Omega) \approx 6.7\) dBi. The gain is approximately constant in the frequency range 40 – 440 MHz (Figure 3). We calculated the effective collecting area from the gain using the relationship A e=(G/4π)λ 2≈0.4λ 2.

Figure 1
figure 1

The Voltage Standing-Wave Ratio (VSWR) of one of the Log-Periodic Dipole Antennas (LPDAs) used in the GLOSS at different frequencies.

Full size image
Figure 2
figure 2

E-plane (asterisks) and H-plane (circles) patterns at 80 MHz for one of the LPDAs used in the GLOSS. The angular separation corresponding to E1–E2 and H1–H2 indicates the HPBW along the E-plane and the H-plane, respectively.

Full size image
Figure 3
figure 3

Gain of one of the LPDAs used in the GLOSS at different frequencies.

Full size image

3 Array Setup and Receiver System

A group of eight similar LPDAs were mounted vertically pointing toward the local zenith (≈ 77 ∘E 14 ∘N), and combined as an adding interferometer on a North–South baseline (Figure 4). The spacing between the adjacent LPDAs is 5 m. The arms of all the LPDAs were oriented in the East–West direction. This implies that the E-plane of the all the LPDAs are along the East–West direction and the H-plane is along the North–South direction. The above arrangement results in a half-power beam width of ≈ 80∘×6∘ (where the first value corresponds to the right ascension (RA) and the second one to the declination (δ)) at a typical frequency of 80 MHz for observations at the zenith. The output from each LPDA is passed through a high-pass filter with a 3 dB cut-off at ≈ 40 MHz (Figure 5). This is to attenuate the spurious and unwanted RF signal at lower frequencies. The filtered signal is then amplified to ≈ 28 dB with a low-noise amplifier. This is called the pre-amplifier and has a uniform gain in the range 0.5 – 500 MHz with a noise figure of ≈ 3 dB. The amplified outputs from the LPDAs are combined using two-way power combiners and analog beamformer network circuits, i.e. coaxial cable delay-system, in three stages DSB I, DSB II, and DSB III (Figure 5). Note that because of the array geometry with respect to the declination of the Sun on any given day, there will be a geometrical phase difference, \(\phi_{\mathrm{g}} = \frac{2 \pi}{\lambda}d \sin\theta\), between the RF signal incident on the LPDAs in the array; here d is the distance between the adjacent LPDAs or pairs of LPDAs, and θ is the zenith angle of the source in the North–South direction (i.e. along the declination). This phase difference should be compensated for before the signals are combined. The analog beam-former network takes care of this. Each of them consists of coaxial cables of different lengths that can be selectively included in the RF signal-path via control signals transmitted from the receiver room (Landecker 1984; Ramesh et al. 1998). These cables provide the instrumental phase ϕ i to compensate for ϕ g over the zenith angle range −37∘≲θ≲+9∘. Note that the HPBW of the beam along the declination is ≈ 100∘ (see Section 2). Since the local latitude at Gauribidanur is ≈ 14∘ North, this spread in zenith angle covers the declination range δ=−23∘ to +23∘ along which the Sun moves in a year. Care must be taken that the residual phase error Δϕ=ϕ g−ϕ i is as small as possible. For GLOSS, the Δϕ i≈12.5∘ at highest operation frequency i.e. 440 MHz. The error is expected to be smaller at lower frequencies.

Figure 4
figure 4

The GLOSS antenna system.

Full size image
Figure 5
figure 5

Radio frequency (RF) signal path from the individual LPDAs in the GLOSS.

Full size image

The output from the final beam-former DSB III is amplified to ≈ 28 dB in the group amplifier, similarly to the pre-amplifier mentioned above. The amplified output is transmitted to the receiver room (≈ 200 m away) through a low-loss cable. The latter is buried about ∼ 1 m below the ground to minimize diurnal variations. The signal attenuation in this cable is ≈ 2 dB at the highest operating frequency of the GLOSS, i.e. 440 MHz. Note that the spectral observations of the solar radio bursts at low frequencies are usually carried out using only a single LPDA (Benz et al. 2009). The eight LPDAs in the GLOSS cover a comparatively larger effective collecting area (A e≈3λ 2) and better directivity, particularly along the declination. The lowest detectable flux density (ΔS min) with the GLOSS at 5σ level is ≈ 0.5 sfu (sfu=solar flux unit=10−22 Wm−2 Hz−1) at 40 MHz and ≈ 3 sfu at 440 MHz. The comparatively larger A e and hence the lower ΔS min of the GLOSS facilitates observations of radio bursts from the Sun with better contrast and strong cosmic radio sources for calibration purposes. The frequency dependence of ΔS min is due to the change in both the system temperature T sys (dominated by the background sky temperature at low frequencies) and the effective aperture (A e) as a function of frequency.

In the receiver room the RF signal is passed through a high-pass filter followed by an amplifier similar to the antenna end, and then connected to a spectrum analyzer. We use the commercially available Agilent Spectrum Analyzer (Model E4401B, frequency range 9 kHz – 1.5 GHz) to record the spectrum of solar radio bursts. The instrument is used in the continuous-sweep mode in the frequency range≈40 – 440 MHz. The number of samples acquired in one sweep (≈ 250 ms) is 401 with a frequency resolution of ≈ 1 MHz. This sweep time is reasonable since the duration of the majority of solar radio bursts, particularly at low frequencies, is ≳ 1 s (McLean 1985). The interface to the computer is made with the commercially available GPIB-PCI card (Agilent 82350B). Data acquisition is carried out using Visual C++ (Ebenezer et al. 2007). Table 1 lists the characteristics of the GLOSS.

Table 1 Characteristics of the GLOSS.
Full size table

We stress that the spacing of 5 m between the adjacent LPDAs in the GLOSS is greater than the observation wavelengths at frequencies≥60 MHz. Then, the array has grating lobes in the corresponding frequency range, i.e. 60 – 440 MHz (Kraus 1950). The separation between the main lobe and grating lobe is ≈ 8∘ at 440 MHz and ≈ 90∘ at 60 MHz. However, any contribution to the GLOSS output from the possible presence of strong radio-wave emitting solar/cosmic sources that simultaneously occur at the location of the grating lobes as well as the main lobe in the sky is expected to be small because 1) the size of the Sun in the radio range increases with decreasing frequency. The average sizes are ≈ 0.6∘ and ≈ 0.8∘ at 440 and 60 MHz, respectively (Sheridan and McLean 1985; Avignon et al. 1975). Compared with this, the grating lobes at the corresponding frequencies are far away, so that there will not be multiple lobes present on the Sun at the same time. 2) The flux densities of most radio bursts from the Sun are higher than those of radio emission from cosmic radio sources (Nelson, Sheridan, and Suzuki 1985). Their temporal and spectral characteristics are different as well.

4 Observations

The GLOSS has been operating regularly for almost the past two years. The daily observing period is ≈ 2 – 11 UT. Several interesting radio bursts have been observed since then. The HPBW of the GLOSS in the East–West direction (≈ 80∘) is centered on the local meridian at Gauribidanur (see Section 3). This width corresponds to ≈ 5.5 hr in time. But in practice we carry out observations for longer periods (≈ 9 hr) since the solar radio bursts, particularly at low frequencies, are usually very intense (Nelson, Sheridan, and Suzuki 1985). Figure 6 shows the background-subtracted GLOSS dynamic spectrum of a type III burst followed by an F–H type II radio burst observed on 17 November 2011 during the interval≈07:28 – 07:31 UT. The frequency range corresponding to the F and H components of the type II burst in Figure 6 are ≈ 55 – 85 MHz, and ≈ 110 – 155 MHz. Both components show a split-band structure with distinguishable upper and lower bands, i.e. FU, FL, HU, and HL (Nelson and Melrose 1985). To observationally verify the effect of the comparatively larger A e and hence lower ΔS min of GLOSS as compared with a single LPDA (see Section 3), we carried out simultaneous observations with the GLOSS and a single LPDA. The signal transmission and data acquisition in the two cases were similar. Figure 7 shows the observations of the same radio bursts as in Figure 6, but with a single LPDA. The enhanced contrast of the spectral features in GLOSS observations is clearly evident.

Figure 6
figure 6

GLOSS observations of type III radio bursts followed by an F–H band-splitting type II burst on 17 November 2011. The horizontal dashed lines of intense emission near 125 MHz are due to local radio frequency interference (RFI).

Full size image
Figure 7
figure 7

Same observations as in Figure 6, but with a single LPDA. The vertical stripes centered at 65 MHz are instrumental artifacts.

Full size image

Figure 8 shows the spectral profile of the type II burst in Figure 6 at ≈ 07:30 UT. The total bandwidth of the F and H components, peak frequencies of FL and HL and FU and HU have a ratio of ≈ 1:2. The intensity of the harmonic component is higher than that of the fundamental component (Wild, Murray, and Rowe 1954; Roberts 1959). The event was associated with a C 6.0 class soft X-ray flare (07:16 – 07:30 UT) with peak at ≈ 07:27 UT and an SF-class Hα flare (07:19 – 07:29 UT) with a peak at ≈ 07:22 UT from AR 11346 located at S19 E08.Footnote 2 The Large Angle and Spectrometric Coronagraph C2 (LASCO C2, Brueckner et al. 1995) onboard the orbiting Solar and Heliospheric Observatory (SOHO) observed a CME around the same time as this flare and radio burst. The CME was first noticed in the coronagraph field of view at ≈ 07:48 UT when its leading edge was at a radial distance r≈2.6 R⊙. Its central position angle and angular width were ≈ 81∘ and 97∘, respectively.Footnote 3 The projected speed of the CME in the plane of the sky was ≈ 458 km s−1. It was a decelerating event (≈ −38 m s−2). The speed between the initial two height–time measurements was ≈ 580 km s−1.

Figure 8
figure 8

Spectral profile of the type II burst in Figure 6 at ≈ 07:30 UT.

Full size image

Figure 9 shows the location of the fundamental component of the type II burst at 80 MHz, observed with the Gauribidanur Radio Heliograph (GRAPH, Ramesh et al. 1998; Ramesh, Subramanian, and Sastry 1999; Ramesh, Sundara Rajan, and Sastry 2006; Ramesh 2011) around ≈ 07:28:30 UT. An inspection of Figures 6 and 9 indicates that GRAPH observations correspond most likely to the upper frequency FU of the split-band structure in the fundamental component of the burst shown in Figure 6. The radial distance of the burst is ≈ 1.4 R⊙, nearly the same as the extrapolated location of the leading edge of the CME (backward toward the Sun) around the same time, obtained using a quadratic fit to the height-time measurements.Footnote 4 This spatio-temporal association between the type II burst and the CME is consistent with the statistical results reported recently by Ramesh et al. (2012).

Figure 9
figure 9

Composite of the radio heliogram (black contours) of the type II burst observed with the GRAPH at 80 MHz on 17 November 2011 around ≈ 07:28:30 UT and the SOHO/LASCO C2, SDO-AIA (193 Å) images obtained around ≈ 08:00 UT on the same day. The peak brightness temperature (T b) of the burst is ≈ 2.7×108 K.

Full size image

We followed the methodology described in Cho et al. (2007) to calculate the magnetic-field strength (B) associated with the type II burst. Assuming the hybrid coronal electron-density model of Vršnak, Magdalenić, and Zlobec (2004), we found that B≈2.7 – 1.8 G in the radial distance range r≈1.4 – 1.6 R⊙, nearly corresponding to the fundamental component in Figure 6. This density model gives a drift speed of ≈ 700 km s−1 for the type II burst, which agrees reasonably well with the estimated speed of the CME close to the Sun using its height-time measurements and deceleration mentioned above. We remark that the B values are 1) about a factor of three higher than the estimates of the magnetic field in the ‘undisturbed’ corona in the similar height range (Ramesh, Kathiravan, and Satya Narayanan 2011), and that they are 2) consistent with similar estimates for the ‘disturbed’ corona using other types of radio bursts as tracers in the same height range (Dulk and McLean 1978; Ramesh et al. 2003; Sasikumar Raja and Ramesh 2013).

5 Summary

We have reported the commissioning of a new radio spectrograph for dedicated observations of the Sun from the Gauribidanur Radio Observatory near Bangalore, India and the first observations. The instrument operates in the frequency range≈40 – 440 MHz. Daily solar observations with dedicated radio facilities such as the GLOSS and the GRAPH (for positional information of the bursts), particularly over a wide range of frequencies, are expected to be useful to understand the solar corona because 1) it has been suggested that reconnection sites at larger radial distances (higher altitudes in the solar atmosphere) are more prolific in coherent radio emission (Benz 2004; Benz et al. 2005), 2) simulations indicate that the spectral characteristics of type III radio bursts are related to the coronal and electron heating conditions near their source region (Li, Cairns, and Robinson 2009), 3) observations indicate that the Sun can be a source of significant radio emission even when the activity level is weak at other regions of the electromagnetic spectrum (Kliem and Bastian 1996; Ramesh et al. 2010a). The scientific community will soon have free access to the daily observations of the facilities just discussed at our institute website.