Next Article in Journal
Magnetic-Free Nonreciprocal Multifunction Device Based on Switched Delay Lines
Previous Article in Journal
Ensemble Bagged Tree Based Classification for Reducing Non-Technical Losses in Multan Electric Power Company of Pakistan
Open AccessArticle
Peer-Review Record

Modernized Solar Radio Spectrograph in the L Band Based on Software Defined Radio

Electronics 2019, 8(8), 861; https://doi.org/10.3390/electronics8080861
Reviewer 1: Anonymous
Reviewer 2: Anonymous
Electronics 2019, 8(8), 861; https://doi.org/10.3390/electronics8080861
Received: 28 June 2019 / Revised: 24 July 2019 / Accepted: 1 August 2019 / Published: 3 August 2019
(This article belongs to the Section Microwave and Wireless Communications)

Round 1

Reviewer 1 Report

Extensive revision of the English is required - missing indefinite articles, wrong verbs used, Many corrections ~20 per page are needed. e.g. 'Sun' used as a adjective - should be solar.. As it is the paper is difficult to follow. I would recommend it being read by someone fluent in English.

The discussion on noise factors is unnecessary and not relevant. A statement to say that the effective noise temperature of the receiver is increased by the presence of a flare would suffice.

The introduction could be reduced in length by careful editing. Also the acronyms may not be familiar to he general reader and should be explained.

It would be useful if the caption to fig 1 had details of the diameter, frequency range and sensitivity (rms noise) in sfu.

The arrangement  shown in figure 7 would produce identical spectra in each branch, unless the splitter introduces a 90 degree phase shift in one arm producing I and Q branches. This could be achieved by an offset clock. If so this should be stated.

Figure 9 does not show the colour scale (even if only approximate).  Is the scale in figure 10 in dB?  - that should be indicated in the caption.

The colour scale in figure 11 is different to that in fig.10 - this should be made clearer.

Comments on the RFI should be made - is this a radio quiet site, how close is the nearest city? It is reassuring that the protected band at 1430 MHz is clear.

There should be a more quantitative comparison between the analogue and digital systems - it is not clear from the data presented that the digital system brings out more details and is lower noise as stated.

The amount of comparison data is small - Is there a reason why the data shown is only 40 seconds in length? Data on the sun in quiescence should also be shown,

What are the noise levels (in sfu, or even degrees K) for both systems? How does it vary with frequency?

Importantly many references are incomplete - with volume and page numbers not shown This must be corrected.


Author Response

Dear Sir/Madam,

The responses to your valuable comments are presented below.

We have also reworked the content of the paper according to the comments and suggestions of all reviewers and the new version is included in the attachment (MS Word with revisions).

With regards,

Pavel Puricer


Q1. Extensive revision of the English is required - missing indefinite articles, wrong verbs used, Many corrections ~20 per page are needed. e.g. 'Sun' used as a adjective - should be solar.. As it is the paper is difficult to follow. I would recommend it being read by someone fluent in English.

R1:

The text was consulted with a language department of our faculty and the suggested changes were applied as much as possible and with the best intent.

*************************************************************************

Q2. The discussion on noise factors is unnecessary and not relevant. A statement to say that the effective noise temperature of the receiver is increased by the presence of a flare would suffice.

 

Q3. The introduction could be reduced in length by careful editing. Also the acronyms may not be familiar to he general reader and should be explained.

R2+R3:

The introduction was reduced mainly in the section of impact of radio bursts to radio systems performance. The discussion on noise factor was appropriately reduced according to the recommendations. The text was clarified and the relevant acronyms and abbreviations were explained.

 

*************************************************************************

Q4. It would be useful if the caption to fig 1 had details of the diameter, frequency range and sensitivity (rms noise) in sfu.

R4:

The antenna diameter is 9.5 m, frequency range 0,8 – 2,0 GHz (added to figure caption), the sensitivity of both systems is then discussed in the paper in Test Results commentary.

 

*************************************************************************

Q5. The arrangement  shown in figure 7 would produce identical spectra in each branch, unless the splitter introduces a 90 degree phase shift in one arm producing I and Q branches. This could be achieved by an offset clock. If so this should be stated.

R5:

In order to meet the timing constraints, the sampled signal data stream is  in the splitter block separated to two consecutive half sized sub-blocks that are put into two parallel branches and FFT is therefore calculated only with a half of processed data block and can work on half of sampling frequency.

The clarification of this approach was added to the paper. 

 

*************************************************************************

Q6. Figure 9 does not show the colour scale (even if only approximate).  Is the scale in figure 10 in dB?  - that should be indicated in the caption.

R6:

The image from the old spectrometer was reworked to show colorbar scale. The Old spectrograph produces images with values relative to quiet Sun radiation, so values in colorbar scale represent multiples of QSR value.  The description of the reworked image was added to the paper.

The caption of Figure 10 was completed with an information about scale in dB.

 

*************************************************************************

Q7. The colour scale in figure 11 is different to that in fig.10 - this should be made clearer.

R7:

The colour scale for Figure 10 was changed to respect the colour scale in 3D spectrum in Figure 11.

 

*************************************************************************

Q8. Comments on the RFI should be made - is this a radio quiet site, how close is the nearest city? It is reassuring that the protected band at 1430 MHz is clear.

R8:

The measurements for a quiet Sun state were also evaluated to determine main sources of radio interference. The antenna placement is on the south-western slope of the hill in an elevation of about 510 m. The nearby Ondrejov village is about 1 km from observatory and with elevation of about 450 m which helps a little to reduce the impact of a possible close interference. Main sources of radio interference are a cellular phone network transmitter in south-west in distance 1.8 km at elevation 451 m  and azimuth 242 degrees and a digital terrestrial TV transmitter in distance 4.9 km at elevation 386 m and azimuth 168 degrees. The impact of these interference sources is suppressed by use of narrow beam antenna and preventing the antenna pointing directly to the transmitter locations. Another possible source of the RFI is an ADS-B signal at 1090 MHz, during the tests it was not observed but it can be relevant since it is produced by aircrafts that can fly through antenna beam.

The notable interference was during the test operation observed at frequencies 1.458 GHz with relative power 36 dB and 1.467 GHz with relative power 27 dB. Nevertheless, the observed levels are within designed dynamic range of the system and therefore do not cause a saturation of ADC. They will be digitally filtered in the processed signal data. Of course, the next interference analysis has to be also made for the final installation to the location of current analog system.

The related discussion about possible RFI is now included in the paper. 

 

*************************************************************************

Q9. There should be a more quantitative comparison between the analogue and digital systems - it is not clear from the data presented that the digital system brings out more details and is lower noise as stated.

R9:

The time resolution of digital spectrometer is 1 ms and frequency bin width is better than 303 kHz. The time resolution of analog instrument is 10 ms shared by 256 frequency channels and the width of frequency bin is 4 MHz. The summary of parameters for both devices was included to the new version of the paper in the Table 3. The noise suppression improved by longer averaging during signal processing was however degraded by the noise of sub-optimal antenna system (the quality antenna that will be used for final installation is currently in the operation with the analog RT5 spectrometer). Still, the observed background level is about 3dB lower than quiet Sun level. The described fact is now included in the paper and the statements about noise are corrected.

The noise parameters are in the text expressed in the form of quiet Sun radiation (QSR) to noise ratio, there were made using comparisons of signal form the Sun and from the clear sky and signal obtained during night operation (almost pure receiver noise). The ratios came QSR/N 3 dB for Sun vs. clear sky and 10 dB for Sun vs. night. The QSR/noise ratio of analog system is 3.7 dB. If we should express it in the terms of SFU, considering that the QSR value during the measurements was 70 SFU (obtained from RT5), the related noise levels would be 30 SFU for the analog system and 35 SFU for the digital one. The digital system however suffers from lower performance antenna. For the comparison in the text we have chosen QSR/N ratios.

 

*************************************************************************

Q10. The amount of comparison data is small - Is there a reason why the data shown is only 40 seconds in length? Data on the sun in quiescence should also be shown,

R10:

The limited frequency range is a result of using data from only one realized spectrograph channel with a bandwidth of 250 MHz. The frequency band around 1.5 GHz was chosen as first one to be put in the operation because of the research focus on bursts impacts on satellite navigation systems that in such band operate.

Selected time range was chosen with the respect to cover a significant change in spectrum with full time resolution and still a reasonable size of data to be processed. The selected time represents about 200MB of data processed in Matlab. Longer depicted data record would also be paid by loss of detail in the images.

The longer time intervals are kept with lower resolution of one second averages and are now included as an additional picture in the paper.

The image of spectrum for a quiet Sun is added to the paper as an additional picture with appropriate comments.

*************************************************************************

Q11. What are the noise levels (in sfu, or even degrees K) for both systems? How does it vary with frequency?

R11:

The tests with the new installations were rather affected by non-optimal condition of the used antenna. Instead of expected noise temperature of about 10000 K for quiet Sun (based on values in reference [9] – “ITU Recommendation ITU-R P.372-13 (09/2016) Radio noise” for our frequency and antenna size) we obtained levels about 3 dB worse. So, the observations of the impact of the antenna pointing to the Sun were hard to precisely evaluate, it was about 2-3 dB of level increase compared to a clear sky orientation of the antenna. Based on these findings we can then estimate for the current testing configuration the noise level of 35 SFU (including an internal noise, a clear sky noise, and an industrial noise background) as written to the response to comment No.9 above. It corresponds with the estimated noise figure of 9 dB combining a flat gain amplifier, an antenna cable, and an antenna feed.  The results can be improved in a change of the final installation, that will use a LNA (low noise amplifier) directly at the antenna feed output, the current testing configuration uses a lossy cable between antenna feed and the LNA input.

The frequency dependence of the noise level in the realized one channel band is mainly determined by the used low-pass filter characteristics. Besides this impact, there were observed two spurs on frequency bins 1.436 GHz and 1.615 GHz caused by internal circuits of the receiver. These single bin spurs will be compensated during the signal processing and excluded from the final data.

This discussion is now added to the new version of the paper.

 

*************************************************************************

Q12. Importantly many references are incomplete - with volume and page numbers not shown This must be corrected.

R12:

The references information was completed.


Reviewer 2 Report

I  would like to see more scientific results, not only technical

details.  How Authors will use  the dynamical spectra of the solar radiations. Have they a clear criteria of the flaring activity and predictions of the future flares?

Could they understand where (on the solar disk) the bursts happened?   

Author Response

Dear Sir/Madam,

The responses to your valuable comments are presented below.

We have also reworked the content of the paper according to the comments and suggestions of all reviewers and the new version is included in the attachment (MS Word with revisions).

With regards,

Pavel Puricer

Q1. I  would like to see more scientific results, not only technical details.

R2:

The availability of more scientific results is strongly dependent on the solar activity that is (unfortulatelly) currently in its long-term minimum. The main aim of the work was to develop and put in the operation the modernized equipment capable to register possible future high dynamics of solar busts with sufficient resolution both in time and frequency for “Commissioning and Science Verification”. We would like to use this instrument for future (hopefully) expected solar bursts investigation and publication if more oriented astrophysical results in the appropriate magazine (e.g. Solar Physics).

However, we made test measurements for quiet Sun spectrum levels and the result was added to the paper as an additional picture. Moreover, the test measurements were also used for a determination of possible sources of radio interference as added in the corrected version of the paper.

*************************************************************************

Q2. How Authors will use  the dynamical spectra of the solar radiations. Have they a clear criteria of the flaring activity and predictions of the future flares?

R2:

The flaring activity can be classified by several aspects. A climax of the signal level in comparison to the quiet Sun level can be expressed in terms of solar flux units (here presented in decibels of relative value of quiet Sun level) and this quantitative value and a time information can be then correlated with the performance of other systems (This is focus of the group of authors from Czech Technical university) So, we just want to register the time of occurrence and relative power of solar burst and then correlate it with the navigation systems precision and a quality of service.

 

The frequency and time distribution of the solar burst from the qualitative point of view is then an object of investigations of our colleagues from the Astronomical Institute. The high-resolution spectroscopy of solar radio burst holds as an essential diagnostic tool for studies of magnetic reconnection in the solar flares. Namely, the fine structures in radio dynamic spectra frequently represent unique signatures of the fast plasma processes in magnetic reconnection, which are anticipated by our numerical models but which cannot be – because of their relatively small dimensions – seen directly in optical, UV/EUV or X-ray images. Typical usage of the radio spectra in connection with numerical modelling works via forward-fit models of the radio emission built on top of our plasma simulations (magnetohydordynamic / MHD, or Particle-In-Cell / PIC) and comparison of the the modelled radio spectra with our observations. Examples of such approach are shown in the papers

http://adsabs.harvard.edu/abs/2008SoPh..253..173B ,  http://adsabs.harvard.edu/abs/2001A%26A...379.1045B, and

http://adsabs.harvard.edu/abs/2005ApJ...631..612B

 

 

The designed dynamic range was set to 50 dB which shall be according to the statistical analysis of values in the past sufficient to prevent saturation of ADC and therefore keep the ability to observe both low levels and top extremes with full resolution. Many of current spectrograph have a limitation in the dynamic range and therefore they are either set to register low changes of solar flux but the extreme values cause saturation of the instrument, or they are set to be able to register top values but low changes are then overlaid by background noise of the system.

Although the prediction of future bursts is however not practically possible with the current state of the art, the detailed research of the burst drivers can aid to a development of the prediction mechanisms in the future.

 

*************************************************************************

Q3. Could they understand where (on the solar disk) the bursts happened?

R3:

Since the diameter of used antenna is 8 m which corresponds in the operating frequency band with the 3 dB beamwidth of about 1.8 degrees, the current equipment is not capable to locate the origin of the solar burst on the solar disk. The realized measurements are focused primary on the spectral and time parameters of investigated radio emissions. However, the context images of the flaring region are available from the space-born instruments onboard, e.g., Solar Dynamics Observatory (SDO) or Hinode satellites and form the ground-based optical telescopes. We combine those multiwavelenght observation in order to acquire as much as possible to complete a picture of the flare. Exceptionally, the imaging information at our frequency range is available form the Chineese Radioheliograph MUSER.


Round 2

Reviewer 1 Report

A useful paper on SDR use. The authors might be aware of the AirSpy and other dongle devices fro radio spectrum measurement, though they do not have the same time resolution,

Back to TopTop