Pulsars


The beginning of the story


Back in 2013, I had completed my 7.3m homebuilt offset dish, and the 5m dish I had used before became obsolete. Mario, I0NAA, came to my QTH to pick it up. He mentioned that beside EME he plans to use the dish for radio astronomy, especially for receiving pulsars. That time I only knew that pulsars exist.
When I met Mario later from time to time, personally or on internet chat, pulsars were often a subject of our talks. He was insisting in a nice way that I should try to receive pulsars myself. He also sent me the software Murmur he has written, this is a very good tool for pulsar receives prediction, and I use it a lot now.
But I had no real idea how to start. Looking on internet, I came across the Neutron Star Group webpage, and was surprised about the reports of Andrea, IW5BHY, who regularly can receive the pulsar B0329+54 on 420 MHz with a corner reflector antenna, equivalent to a 2.5m dish. On occasion I had an EME QSO with Andrea, and decided to ask him by email about what I would additionally need, how to start and so on.
Andrea was (and is) extremely helpful, he told me what hardware to use and let me use the software he has written, which performs excellent. He guided me through the whole installation process, analyzed my first recordings and is still a helping hand for giving me hints for improvements. He spent many hours and wrote many emails. That way I learned a lot, and I am now able to do the measurements myself.


Equipment

Antenna:  7.3m homemade offset dish, OE5JFL tracking system
Feeds: 70cm (424 MHz) dual-dipole with solid reflector, 23cm (1294 MHz) RA3AQ horn
Preamplifiers: 23cm cavity MGF4919, 70cm 2SK571 (30 years old!)
Line Amplifier: PGA103+
Interdigital filter: designed with VK3UM software, 70cm 4-pole, 23cm 3-pole
Receiver: RTL-SDR (error <1ppm), 2 MHz bandwidth
Software: IW5BHY, Presto, Tempo, Murmur

The following 23 pulsars were received with this setup on 70cm until end of May 2017, 11 of them also on 23cm.
Signal/Noise ratio measured with IW5BHY software

Pulsar 70cm (424 MHz) 23cm (1294 MHz)
B0329+54
110 85
B0531+21 (Crab) *
10 ---
B0823+26 18 9
B0834+06 10 ---
B0950+08
32 14
B1133+16
24 11
B1237+25 6 ---
B1508+55 9 ---
B1642-03
26 9
B1749-28
21 ---
B1818-04 8 ---
B1859+03** 11 ---
B1911-04 12 ---
B1919+21 14 ---
B1929+10
33 9
B1933+16
20 31
B1946+35 6 ---
B2016+28 ***
14 12
B2020+28 ***
9 6
B2021+51
16 17
B2111+46 6 ---
B2217+47 15 ---
B2310+42 11 6

note *: The Crab pulsar was a challenge, 30 rotations/sec and high dispersion. Received at the first attempt!
note **: The B1859+03 has a very high dispersion: DM=402!

note ***: The B2016+28 and the B2020+28 are only about 1deg apart from each other. 424 MHz profiles for both pulsars were obtained by analyzing the same recorded file.
dish_7m3_night

 


In June 2017 the RTL-SDR was replaced by an Airspy SDR for receiving with 10MHz bandwidth.

 It is difficult to find 10 MHz with an acceptable RFI situation, especially on 70cm.


The receiving software is based on Gnu-Radio, all developed by Andrea (IW5BHY) and his son.


A more detailed description can be found at the webpage of Nando, who is using the same system: I1NDP Pulsar II

He has big success detecting pulsars on 23cm, have a look!


Finding new pulsar candidates for observation is more a challenge now. Background noise from the sky is unfortunately high on 70cm for most pulsars. Also RFI at low elevation (declination) is an issue.


Pulsar 70cm (420 MHz) 23cm (1290 MHz)
B0355+54 detected (S400=46) ---
B0525+21 detected (S400=57) ---
B0450-18 detected (S400=82) ---
B0450+55 detected (S400=59) ---
B0626+24 detected (S400=31) ---
B0740-28 detected (S400=296) ---
B0809+74 detected (S400=79) ---
B0818-13 detected (S400=102) ---
B0919+06 detected (S400=52) ---
B1540-06 detected (S400=40) ---
B1541+09 detected (S400=78) ---
B1604-00 detected (S400=54) ---
B1706-16 detected (S400=47) ---
B1915+13 detected (S400=43) ---
B2000+40 detected (S400=53) ---
B2045-16 detected (S400=116) ---
B2154+40 detected (S400=105) ---


 A screenshot summary of all detected pulsars, containing outputs by IW5BHY software and Presto software you can download here:

PULSAR_SCREENSHOTS_BHY_PRESTO.ZIP



RECEPTIONS:


B0329+54

70cm 3 pulses 

For receiving this strong pulsar on 424MHz, I need normally only one minute or even less.
Sometimes it is even possible to get single pulses. With an integration time of several hours, the S/N of course improves a lot.
   

 scintillation23

Scintillation can be a problem, especially on 23cm. Sometimes nothing for hours, then suddenly the pulsar is showing up. The whole plot is 5 hours.

 

 

 b0329_pre_pot_pulse In this 23cm reception the pre-pulse and post-pulse of the B0329+54 normal mode can be seen very well.

 

 

b0329dispersion

These graphs of a 424MHz reception illustrate the effect of dispersion. The bandwidth of 2 MHz is divided into 4 subbands, 500 kHz each.

The upper graph shows the subbands separately and zoomed. The blue line is offset by slightly more than 4 ms from the white line. The difference of the center frequencies is 1.5MHz. This is a good coincidence to the 2.9ms/MHz dispersion of the pulsar on 420 Mhz.

The lower graph shows a beautiful pulse after de-dispersion.

 

b0329_70_presto

Another good software is Presto

See the 424 MHz profile of B0329+54....

b0329_23_presto .....and the 1294 MHz  profile of B0329+54
FMD_scan_0329

IW5BHY has programmed a nice tool to confirm if the pulses are likely to come really from a pulsar and not RFI. The folding time is shifted (useful on 70cm and 23cm) and the dispersion time (useful only on 70cm).

For an example I made these tests for the 424 MHz recording of the B0329+54. The left picture shows S/N depending on shifting the folding frequency, in the right picture the dispersion time is shifted.

Andrea, IW5BHY, has found 50 single pulses in a one hour recording I made on 424 MHz. With a special written program he put the single pulses in a row, and generated an audio file from that. So you can even listen to the sound of the pulsar:   sound pulsar B0329+54  (The tone is very low frequency, it is good to use headphones and put level to maximum)


This 3D plot displays 50 consecutive periods at a peak of positive scintillation.

It is from one piece of observation of 36 seconds containing many single pulses.

50 single pulses

Also produced by IW5BHY, harmonic analysis of the dedispersed timeseries showing the fundamental and two harmonics:

spectrum_b0329





.......other pulsars


B0531+21 (Crab pulsar)



b0531_70_crab

The reception of the Crab-pulsar B0531+21 was one of the highlights!

This young pulsar exists since a supernova explosion in 1054, which was observed on earth as a star even visible at daylight for about two years.

It rotates 30 times per second, is highly dispersed ( 3 ms per channel is the same as the pulse width) and its slowdown in rotation speed is remarkable.

Nevertheless, the pulsar reception was positive even on the very first attempt! The observation time in that case was two hours.

Beside the main pulse the Crab pulsar has an interpulse which is varying in strength. In this reception it is not present.

jfl_crab70_pr_s
424 MHz, S/N=10


The following receptions of the Crab pulsar were made with my new setup: Airspy 10 MHz bandwidth


Crab main pulse and interpulse (no frequency tracking, curve bent)

After an 6 hours observation of the much weaker pulsar B0525+21 I was looking for the Crab pulsar in this file also, because their positions are close to each other.
There was low RFI during that time, so both pulsars signals came through nicely.
Because of the long observation time the Doppler effect was noticable and the curve was somewhat bent.
With the frequency tracking option of Presto the curve is straightened and S/N even better.

It was also nice to see the interpulse of the Crab on this occasion.



same as on the left side, but with frequency tracking, curve straight


Crab giant pulses

The Crab pulsar is known for its giant radio pulse emission. IW5BHY has written a very good software to search for giant pulses in the recorded files. I analyzed several observations, sometimes found nothing because separating the giant pulses from RFI spikes is not so easy. Sometimes found between 1...3 within 2 hours.
The 6 hours observation mentioned above was extremely good also in that respect. Depending on the threshold, up to more than 20 giant pulses are displayed with Andreas software.
Even one giant pulse at the phase of the interpulse was found, which is a very rare event.
I estimate the peak flux level of the best observed giant pulse (S/N=26) to be between 1000Jy and 2000Jy, means an increase of around 25dB above average!

Best 3 giant pulses during a 6 hours observation.
Threshold for S/N=20, best S/N=26



Best 11 giant pulses during a 6 hours observation.
Threshold reduced to S/N=10






 One giant pulse was found at the phase of the interpulse.
This is a very rare event!



Best 25 giant pulses during a 6 hours observation.
Threshold reduced to S/N=9
The phase of the giant pulses is not constant, because during 6 hours of observation Doppler shift is significant.



B0823+26


b0823_70_23

left side:   424 MHz, S/N=18                                                                      right side:  1294 MHz, S/N=9











b0823_70_23_pr
left side:  424 MHz                                                                                                                         right side:  1294 MHz

B0834+06





b0834_70

424 MHz, S/N=10










jfl_0834_pr


B0950+08


b0950_70_23



left side:  424 MHz, S/N=32                                                 right side:  1294 MHz, S/N=14






b0950_70_23_pr

                left side:  424 MHz                                                                                                                             right side:  1294 MHz

B1133+16

b1133_70_23

left side:  424 MHz, S/N=24                                                 right side:  1294 MHz, S/N=11

especially on 70cm the typical double pulse can be seen nicely





b1133_70_23_pr

               left side:  424 MHz                                                                                                                       right side:  1294 MHz

B1237+25


b1237_70

424 MHz, S/N=6










b1237_70_pr


B1508+55


b1508_70

424 MHz, S/N=9










b1508_70_pr


B1642-03




left side:  424 MHz, S/N=26                                                 right side:  1294 MHz, S/N=9






b1642_70_23_pr
              left side:  424 MHz                                                                                                                       right side:  1294 MHz


B1749-28

b1749_70

424 MHz, S/N=21


The declination of this pulsar is so low, that my antenna can ‘see’ it only for 2 hours just before culmination, maximum 2deg above a building. Because of the additional noise pickup the S/N is low.

On 23cm I see no chance because of the much lower flux level.




b1749_70_pr


B1818-04





b1818_70

424 MHz, S/N=8









b1818_70_pr_s



B1859+03


b1859_70

424 MHz, S/N=11

This pulsar has a very high dispersion: DM=402
Therefore the four 500kHz channels are separated by 22ms to each other, this is twice the W50 of 11ms! Because of the low pass filtering the pulse looks broader but became visible.


Thanks to Andrea IW5BHY for the extra analyzing work




jfl_1859_70_presto



B1911-04

b1911_70

424 MHz, S/N=12









b1911_70_pr



B1919+21

b1919_70

424 MHz, S/N=14









b1919_70_pr



B1929+10

b1929_70_23

left side:  424 MHz, S/N=33                                                 right side:  1294 MHz, S/N=9







b1929_70_23_pr



               left side:  424 MHz                                                                                                                      right side:  1294 MHz


B1933+16

b1933_70_23

left side:  424 MHz, S/N=20                                                 right side:  1294 MHz, S/N=31

b1933_waterfall_23 b1933_70_23_pr

This pulsar has a very high dispersion, making things difficult on 424 MHz. The channels are separated by slightly more than 8ms to each other, this is nearly as much as the pulse width itself!

If you look carefully, the dispersion can be even seen on 23cm, but it is only in the range of 1ms over all channels.

The waterfall graph of a 6 hours reception of the B1933+16 pulsar on 1294 MHz shows that the frequency is not constant over this long period.

This can result in a lower S/N when making long time observations without using special software.

                      left side:  424 MHz, S/N=20                                                                                                            right side:  1294 MHz, S/N=31



B1946+35


b1946_70

424 MHz, S/N=6










b1946_70_pr


B2016+28

b2016_70_23


left side:  424 MHz, S/N=14                                                 right side:  1294 MHz, S/N=12









b2016_70_23_pr


               left side:  424 MHz                                                                                                                     right side:  1294 MHz



B2020+28

b2020_70_23

left side:  424 MHz, S/N=9                                                 right side:  1294 MHz, S/N=6



especially with weak signals it is good to confirm the positive reception with a second software.


 

b2020_70_23_pr

                left side:  424 MHz                                                                                                                         right side:  1294 MHz 


B2021+51

b2021_70_23_bhy

left side:  424 MHz, S/N=16                                                 right side:  1294 MHz, S/N=17


Reception of this pulsar showed me how important patience can be, especially on 23cm.

After 3 attempts, each 5 hours, with completely negative results, I decided to try a 4th (and last) time.

During the last 2 hours the signal came up. So 18 hours nothing, 2 hours signal….10%


 

b2021_70_23_pr

                left side:  424 MHz                                                                                                                         right side:  1294 MHz 



B2111+46


b2111_70

424 MHz, S/N=6










b2111_70_pr



B2217+47


b2217_70

424 MHz, S/N=15










b2217_70_pr



B2310+42

b2310_70_23

left side:  424 MHz, S/N=11                                                 right side:  1294 MHz, S/N=6







jfl_2310_70_23

               left side:  424 MHz                                                                                                                       right side:  1294 MHz




The following pulsars were additionally detected with the new setup, Airspy 10MHz bandwidth

During some observations, especially for pulsars with low or negative declination resulting in low elevation, I suffered from extra RFI coming from buildings around.
Therefore sometimes around 50% of the channels were useless.

B0355+54




B0450-18



B0450+55



B0525+21

The pulse profile of this pulsar shows a double pulse. It was nice to see the coincidence with the EPN pulsar profile catalogue.





B0626+24




B0740-28



B0809+74



B0818-13



B0919+06


B1540-06



B1541+09



B1604-00



B1706-16



B1915+13



B2000+40



B2045-16





B2154+40







Reception of pulsar B0329+54 on 1294MHz with 3m dish ( 2MHz bandwidth)

oe5jfl_3m_dish
After maybe 10 observations with negative result, I finally  detected the pulsar also with my 3m dish at home, which I use mainly for terrestrial contacts.
The antenna is placed on top of a building with no obstacles around. This is good for operating tropo, but worse for observing pulsars, because I pick up a lot of RFI.
It was a nice experience to see that it is possible to detect a strong pulsar even with moderate sized antenna without large bandwidth.

b0329_3m_dish

b0329_23cm_3m_dish


return to radio astronomy