Pulsar Timing

Back

Eight week report for IASc-INSA-NASI Summer

Research Fellowship

Pulsar Timing

Name: Pavan U A

Application No.: PHYS253

Guide: Prof. Yashwanth Gupta

National Centre for Radio Astrophysics,

TIFR,

Pune

Pune Univerisity Campus, Post Bag 3,

Ganeshkhind Pune 411007, India

Contents

1 Introduction 4

1.1 Overview on Pulsar Timing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

1.2 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2 Introduction to GMRT 6

2.1 Array Conﬁguration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

2.2 Receiver Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

2.3 Digital Backends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2.3.1 GMRT Antenna System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

2.3.2 Design Speciﬁcations of GMRT Antennas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

2.4 The GMRT beam mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

2.5 Coherently vs Incoherently Phased Array . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

2.5.1 Voltage Mode and Coherent Dedispersion Pipeline . . . . . . . . . . . . . . . . . . . . . . . . . 10

3 Radio Frequency Interference 11

4 Introduction to Presto - Pulsar Search Tool 11

4.1 Working with Presto Software . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

4.2 Analysis using presto . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

4.3 The Time Domain and Pulse Proﬁle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

4.4 Frequency and Sub-bands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

4.5 Other plots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

5 GPTool 17

5.1 Running GPTool . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

5.2 Example GPTool.in ﬁle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

5.3 GPTool Output ﬁles and plots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

6 Automation of Analysis 20

7 Pulsar Timing Observation 20

7.1 TOA measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

7.2 Forming Pulse Emission Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

8 Delays encountered in Pulsar Timing 22

8.1 Atmospheric Delay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

8.2 Einstein delay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

8.3 Romer Delay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

8.4 Shapiro delay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

8.6 Tempo2 plot window . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

8.5 Pulsar Timing Observation with Tempo2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

8.7 Results of Timing analysis for the pulsar J1857+0943 . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

8.7.1 Pre-ﬁt residuals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

8.7.2 Post-ﬁt residuals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

1 Introduction

Pulsars are what is known as the ”lighthouses” of the universe they are the rotating neutron stars which emits a

focused beam of electromagnetic radiation that is only visible in line of sight. Known as pulsars, these mysterious

objects get their name because of the way their emissions of electromagnetic waves which appear to be pulsating out

into Universe. Pulsars are highly compact objects that are about the size of a small city but contain more mass than

the any main sequence star.Pulsars are used to study extreme states of matter, search for planets beyond Earth’s

solar system and measure cosmic distances. Pulsars can also be used to indirectly ﬁnd gravitational waves, which

could point the way to energetic cosmic events like collisions between supermassive black holes. Discovered in 1967,

pulsars are fascinating and mysterious members of the cosmic regime.

1.1 Overview on Pulsar Timing

The process of using observations of radio pulsar pulse proﬁles to calculate pulse arrival times which are then

used to study the rotational history, kinematics and orbits of pulsars, is called pulsar timing. This is usually carried

out at radio wavelengths, the mean pulse proﬁles of radio pulsars are usually remarkably stable, and when compared

with the average proﬁle can yield highly accurate times of arrival or ”TOAs”. In the case of millisecond pulsars,

these can be highly accurate in case of the millisecond pulsars.

The process of timing has to be carried out with considerations of the delays which are present in their arrival

times, for example PSR J1909-3744 Pulse arrival times as a function of orbital phase for the binary pulsar PSR

J1909-3744. This pulsar is orbiting a white dwarf every 1.5 days and is nearly edge-on to our line of sight. The

curvature of space-time leads to the distortion of the arrival times seen above after subtracting a simple Keplerian

model of the pulsars orbit from the arrival times. In the absence of the space-time curvature the arrival times would

be ﬂat.

These TOAs can be used to deduce the entire rotational history of a given pulsar. This practice has been used

at many telescopes around the world to examine how pulsars rotate allowing the discovery of pulsar glitches (sharp

discontinuities in their rotation periods) and the study of their neutron superﬂuid interiors. When a pulsar is in orbit

about another body, pulsar timing can be used to make highly accurate observations that can be used to deduce

companion star masses, and in extreme cases, test theories of relativistic gravity such as Einsteins general theory of

relativity. Observations over many years can provide accurate proper motions and even parallax measurements.

Pulsars can be designated as cosmic clocks or ”clo cks in space”. On long timescales they can match the precision

of atomic clocks. Using a process called pulsar timing, the exact calculation of TOAs can allows a number of

applications, from theories of gravity to detecting gravitational waves. Also universal navigation reference system

suitable for autonomous space navigation can be created by pulsars, using them as natural navigation beacons, and

one which does not require the costly setup of navigation system satellites which can function only up to a certain

designated period and only conﬁned to one planet. By comparing pulse arrival times measured on-board a spacecraft

with predicted pulse arrivals at a reference location, the spacecraft position can b e determined autonomously and

with high accuracy everywhere in the solar system and beyond. There are currently many on-going experiments to

use the clock-like nature of pulsars to produce a huge gravitational wave detector for low-frequency gravitational

waves.

1.2 Motivation

Though much research is already been carried out at various parts in the world in the study of pulsars, through

its discovery in the year 1967 by accident while Jo celyn Bell and Antony Hewish were looking for twinkling sources

of radio radiation, many aspects still remain a large mystery such as the emission mechanism of the pulsars and their

potential use in the future in the Galactic positioning (see Becker Lange et al., 2013) made me choose this topic for

my Summer Research Fellowship, in the two Nobel prizes which have been conferred to the pulsar related ﬁeld, one

Nobel prize was presented for the indirect detection of the Gravitational waves by Princeton University astronomers

Russell A. Hulse and Joseph H. Taylor in 1974 for their work on PSR 1913+16 by the Pulsar Timing. PSR 1913+16

emits radiation every 59 milliseconds. It orbits another star, which is likely another neutron star. These stars orbit

each other at super-high speed every eight hours.

Four years after ﬁrst discovering PSR 1913+16 and after some careful timing measurements of the pulsar, Hulse

and Taylor found that the two stars move closer to each other by about three millimeters per orbit. That could only

happen if something was pulling energy out of the system which was precisely the detection of the Gravitational

waves.

The pulsar timing also has many other advantages with respect to reducing the time residual in the observation

of the pulsar in GMRT, which is currently been carried out by calculating the Time of Arrival (TOAs) of the pulses.

2 Introduction to GMRT

The Giant Metrewave Radio Telescope (GMRT) consists of an array of 30 antennas. Each antenna is 45 m in

diameter, and has been designed to operate at a range of frequencies from 50 MHz to 1450 MHz. The antennas have

been constructed using a novel technique (nicknamed SMART) and their reﬂecting surface consists of panels of wire

mesh. These panels are attached to rope trusses, and by appropriate tensioning of the wires used for attachment the

desired parabolic shape is achieved. This design has very low wind loading, as well as a very low total weight for

each antenna. Consequently it was possible to build the entire array very economically.

2.1 Array Conﬁguration

GMRT has 30 parabolic radio dishes each of diameter 45m. Owing to the interferometric arrangement GMRT

has achieved enough sensitivity to study the proﬁle of single pulses. The antennae dishes are arranged in a hybrid

conﬁguration. 14 of the dishes are packed in a region called the central square, which is 1 Km across and they are

arranged randomly. This random arrangement is intentional to avoid grating lobes. The antennas in the central

square are numbered as C00 to C14, (C07 being absent) . Then the remaining of the antennas are arranged in a

‘Y’ conﬁguration the length of each arm being 14 Km. The arms are caller the East, West and South arms and are

numbered as E02 to E06, S01 to S06 (S05 not present), W01 to W06. This makes up the total 30 dishes

2.2 Receiver Systems

The GMRT currently operates at 5 diﬀerent frequencies ranging from 150 MHz to 1420 MHz. Some antennas

have been equipped with receivers which work up to 1750 MHz. Above this frequency range however, the antenna

performance degrades rapidly both because the reﬂectivity of the mesh falls and also because of the rapidly increasing

aperture phase errors because of the deviations of the plane mesh facets from a true parabola.

The GMRT feeds, (except for the 1420 feed), are circularly p olarized. The circular polarization is achieved by means

of a polarization hybrid inserted between the feeds and the RF ampliﬁers. No polarization hybrid was inserted for

the 1420 MHz feed, in order to keep the system temperature low. None of the receivers are cooled, i.e. they all

operate at the ambient temperature. The feeds are mounted on four faces of a feed turret placed at the focus of the

antenna. The feed turret can be rotated to make any given feed point to the vertex of the antenna. The feed on one

face of the turret is a dual frequency feed, i.e. it works at both 233 MHz as well as 610 MHz.

After the ﬁrst RF ampliﬁer, the signals from all the feeds are fed to a common second stage ampliﬁer (this ampliﬁer

has an input select switch allowing the user to choose which RF ampliﬁers signal is to be selected), and then converted

to IF. Each polarization is converted to a diﬀerent IF frequency, and then fed to a laser-diode. The optical signals

generated by the laser-diode are transmitted to a central electronics building (CEB) by ﬁber optic cables. At the

central electronic building, they are converted back into electrical signals by a photo-diode, converted to baseband

The array was originally meant to have 34 antennas but due to escalating costs it was reduced to 30 antennas.

frequency by another set of mixers, and then fed into a suitable digital backend. Control and telemetry signals

are also transported to and from the antenna by on the ﬁber-optic communication system. Each antenna has two

separate ﬁbers for the uplink and downlink.

2.3 Digital Backends

There are a variety of digital backends available at the GMRT. The principle backend used for interferometric

observations is a 32 MHz wide FX correlator. The FX correlator produces a maximum of 256 spectral channels for

each of two polarizations for each baseline. The integration time can be as short as 128 ms, although in practice

2 sec is generally the shortest integration time that is used. The FX correlator itself consists of two 16 MHz wide

blocks, which are run in parallel to provide a total instantaneous observing bandwidth of 32 MHz. For spectral line

observations, where ﬁne resolution may be necessary, the total bandwidth can be selected to be less than 32 MHz.

The available bandwidths range from 32 MHz to 64 kHz in steps of 2. The maximum number of spectral channels

however remains ﬁxed at 256, regardless of the total observing bandwidth. The GMRT correlator can measure all

four Stokes parameters, however this mode has not yet been enabled. In the full polar mode, the maximum number

of spectral channels available is 128. Dual frequency observations are also possible at 233 and 610 MHz, however in

this case, only one polarization can be measured at each frequency. The array can be split into sub-arrays, each of

which can have its own frequency settings and target source. The correlator is controlled using a distributed control

system, and the data acquisition is also distributed. The correlator output, i.e. the raw visibilities are recorded in a

GMRT speciﬁc format, called the LTA format. Programmes are available for the inspection, display and calibration

of LTA ﬁles, as well as for the conversion of LTA ﬁles to FITS.

The ﬁrst block of the GMRT pulsar receiver is the GMRT Array Combiner (GAC) which can combine the signals

from the user-selected antennas (up to a maximum of 30) for both incoherent and coherent array operations. The

input signals to the GAC are the outputs of the Fourier Transform stage of the GMRT correlator, consisting of 256

spectral channels across the bandwidth being used, for each of the two polarization from each antenna. The GAC

gives independent outputs for the incoherent and coherent array summed signals, for each of two polarizations. For

nominal, full bandwidth mode of operation, the sampling interval at the output of the GAC is 16µsec.

Diﬀerent back-end systems are attached to the GAC for processing the incoherent and coherent array outputs. The

incoherent array DSP processor takes the corresponding GAC output signals and can integrate the data to a desired

sampling rate (in powers of 2 times 16 microsec). It gives the option of acquiring either one of the polarizations or

the sum of both. It can also collapse adjacent frequency channels, giving a slower net data rate at the cost of reduced

spectral resolution. The data is recorded on the disk of the main computer system.

The coherent array DSP processor takes the dual polarization, coherent (voltage sum) output of the GAC and can

produce an output which gives 4 terms the intensities for each polarization and the real and imaginary parts of the

cross product from which the complete Stokes parameters can be reconstructed. This hardware can be programmed

to give a sub-set of the total intensity terms for each polarization or the sum of these two. The minimum sampling

interval for this data is 32 microsec, as two adjacent time samples are added in the hardware. Further preintegration

(in powers of 2) can be programmed for this receiver. The ﬁnal data is recorded on the disk of the main computer

system. There is another independent full polarimetric back-end system that is attached to the GAC. This receiver

produces the ﬁnal Stokes parameters, I,Q,U & V. However, due to a limitation of the ﬁnal output data rate from this

system, it it can not dump full spectral resolution data at fast sampling rates. Hence, for pulsar mode observations

the user needs to opt for online dedispersion or gating or folding before recording the data (there is also a online

spectral averaging facility for non-pulsar mode observations). In addition, there is a search preprocessor back-end

attached to the incoherent array output of the GAC. This unit gives 1-bit data, after subtracting the running mean,

for each of the 256 spectral channels. Either one of the polarizations or the sum of both can be obtained.

Most sub-systems of the pulsar receiver can be conﬁgured and controlled with an easy to use graphical user interface

that runs on the main computer system. For pulsar observations, since it is advisable to switch oﬀ the automatic

level controllers at the IF and baseband systems, the power levels from each antenna are individually adjusted to

ensure proper operating levels at the input to the correlator. The format for the binary output data is peculiar to

the GMRT pulsar receiver. Simple programs to read the data ﬁles and display the raw data - including facilities for

dedispersion and folding - are available at the observatory and can be used for ﬁrst order data quality checks, both

for the incoherent mode and coherent mode systems.

2.3.1 GMRT Antenna System

A radio telescope in its simplest form consists of three components, (i) an antenna that selectively receives

radiation from a small region of the sky, (ii) a receiver that ampliﬁes a restricted frequency band from the output of

the antenna and (iii) a recorder for registering the receiver output. In this chapter we focus on the antenna, and in

particular the antennas used for the GMRT. The GMRT antennas are parab olic reﬂector antennas. The ﬁrst reﬂector

antenna was invented by Heinrich Hertz in 1888 to demonstrate the existence of electromagnetic waves which had

been theoretically predicted by J.C.Maxwell. Hertzs antenna was a cylindrical parabola of f/D = 0.1 and operated

at a wavelength of 66 cm.(450 MHz). The next known reﬂector antenna was that constructed in 1930 by Marconi for

investigating microwave propagation. After that, in 1937, Grote Reber constructed the prototype of the modern dish

antenna - a prime-focus parabolic reﬂector antenna of 9.1 m. diameter, which he used to make the ﬁrst radio maps

of the sky. During and after World War II, radar and satellite communication requirements caused great advances

in antenna technology

2.3.2 Design Speciﬁcations of GMRT Antennas

The f/D ratio for the GMRT antennas was ﬁxed at the value 0.412 based both on structural design issues as

well as preliminary studies of various feeds radiation patterns. Since the antennas are to work at meter wavelengths

prime focus feeds were preferred. Cassegrain feeds at meter wavelengths would result in impractically large secondary

mirrors (the mirror has to be several across) and concomitant large aperture blockage. Six bands of frequencies had

been identiﬁed for the GMRT observations. It was deemed essential to be able to change the observing frequency

rapidly, and consequently the feeds had to mounted on a rotating turret placed at the prime focus. If one were to

mount all the six feeds on a rotating hexagon at the focus, the adjacent feeds will be separated by 60 . If one wants

to illuminate the entire aperture, then one has to have a feed pattern that extends at least up to the subtended

angle of the parabola edge, which is 0 = 62.5

◦

(Note that cot(

◦

/2) = 4f/D). Hence this arrangement of feeds would

cause the one feed to see the feeds on the adjacent faces. It was decided therefore to mount the feeds in orthogonal

faces of a rotating cube. Since one needs six frequency bands, this leads to the constraint that at least two faces

of the turret should support dual frequency capability. For astronomical reasons also dual frequency capability was

highly desirable. One speciﬁc aspect of GMRT design is the use of mesh panels to make the reﬂector surface. Since

the mesh is not perfectly reﬂective, transmission losses thorough the mesh have to be taken into account. Further,

the expected surface errors of the mesh panels was 5 mm. This implies that the maximum usable frequency is (see

Section 19.2) 3000 MHz, independent of the transmission losses of the mesh. (Incidentally, since the mean-spacing

of feed-support legs, L = 23.6 m, the lowest usable frequency is around 6 MHz).

2.4 The GMRT beam mode

The GMRT Software Backend (GSB; Roy et al., 2010) consists of a beam former, wherein the signals from the

antennas can be appropriately combined for high time resolution studies of radio transients like pulsars (Gupta et

al., 2000). Depending on the speciﬁc science requirement, the signals can be combined to form a incoherently phased

array or a coherently phased array for transient observations.

2.5 Coherently vs Incoherently Phased Array

Normally, the signals from an n-element phased array are combined by adding the voltage signals from the diﬀerent

antennas after proper delay and phase compensation. This summed voltage signal is then put through a square-law

detector and an output proportional to the power in the summed signal is obtained. For identical elements, this

phased array gives a sensitivity which is n times the sensitivity of a single element, for point source observations. The

beam of such a phased array is much narrower than that of the individual elements, as it is the process of adding the

voltage signals with diﬀerent phases from the diﬀerent elements that produces the narrow beam of the array pattern.

For some special applications, it is useful to ﬁrst put the voltage signal from each element of the array through a

square-law detector and then add the powers from the elements to get the ﬁnal output of the array. This corresponds

to an incoherent addition of the signals from the array elements, whereas the ﬁrst method gives a coherent addition.

In the incoherent phased array operation, the beam of the resultant telescope has the same shape as that of a single

element, since the phases of the voltages from individual elements are lost in the detection process. This beam width

is usually much more than the beam width of the coherent phased array telescope. The sensitivity to a point source

is higher for the coherent phased array telescope as compared to the incoherent phased array telescope, by a factor

√

n. The incoherent phased array mode of operation is useful for two kinds of astronomical obervations. The ﬁrst

is when the source is extended in size and covers a large fraction of the beam of the element pattern. In this case,

the incoherent phased array observation gives a better sensitivity. The second case is when a large region of the sky

has to be covered in a survey mode (for example, in a survey of the sky in search for new pulsars). Here, the time

taken to cover the same area of sky to equal senstivity level is less for the incoherent phased array mo de. Only for

a ﬁlled aperture phased array telescope are these times the same. For a sparsely ﬁlled physical aperture such as an

earth rotation aperture synthesis telescope, this distinction between the coherent and incoheret phased array modes

is an important aspect of phased array operation.

2.5.1 Voltage Mode and Coherent Dedispersion Pipeline

The PA mode has provision for recording the raw voltage stream from a single phased array beam on a disk,

where the data is stored with the real and imaginary parts from the output of a Fast Fourier Transform (FFT) on the

sampled voltage time series, with 256 or 512 channels. The voltage recording mode has many powerful applications,

such as coherent dedispersion of pulsar signals, but is often limited due to the large disk storage requirements for

Nyquist sampled voltage data.

The prohibitive requirements for disk storage of voltage data necessary for oﬄine coherent dedispersion can be

avoided by implementing a real- time coherent dedispersion pipeline. Such a pipeline has recently been developed

at the GMRT for the currently available maximum bandwidth of 32 MHz (De & Gupta, 2016). The pipeline uses a

hybrid CPU-GPU based signal processing system for real-time co- herent dedispersion of dual polarisation voltage

signals from a single phased array beam, followed by recording of the power signals (with desired integration) on

disks.

Independent samples are obtained if sampled at Nyquist Rate see Shannon Sampling Theorem

3 Radio Frequency Interference

Pulsars emit much of their radiation in the radio region. The radio waves emitted by the Pulsar has to travel

through interstellar space to reach us, because of this the power of the signal will be already very low. These signals

are then collected by our telescopes, more over there is tremendous amount of man made signal in todays world,

from the small smart watch to that of the large planes and Satellites, all of them create humongous amount of

problems in the pulsar observation. Even the spark plugs in a gasoline-powered car give oﬀ radio waves! There are

cosmic sources of radio waves too, like our very own Sun. Many sources such as our FM radios, radio broadcasts

which are highly important for a human being in todays world are from the point of view of a radio astronomy is

all very bad. These sources of radio waves get in the way, and make some observations almost impossible. For this

reason, astronomers refer to these man-made sources of radio waves as Radio Frequency Interference, or RFI. RFI

is similar to light pollution. In large cities, light pollution can make it impossible to see faint stars, and RFI can

make it similarly diﬃcult to see faint sources of radio waves. There are many methods for RFI mitigation such as

using a ﬁlter to block the RFI and to build the Telescopes outside the cities to point out some.Filters can be used

when ever we deﬁnitely know about the sources but even that has certain draw backs such as , if a radio source is

present in that particular band width we are eﬀectively removing the whole band! And sometimes computers can

be used to try to identify obvious sources of RFI and then ignore it in a given set of data, which is what the below

mentioned software does. RFI is especially problematic for pulsar searches because it can often look like a pulsar

signal in the Fourier transformed data because the the Fourier transform is essentially helping us ﬁnd the repeating

signal of the pulsar. But these man-made sources of RFI often repeat as well. For example, the alternating current

(AC) that supplies your home with power changes polarity at a frequency of 60 Hz. Power lines that carry this

current therefore give oﬀ RFI at a frequency of 50 Hz if they are not shielded. For this reason, we always see a large

signal in our Fourier transformed data at 60 Hz. Luckily, we can simply tell a computer to ignore that frequency.

Other sources of RFI may not be as easy to handle, because we dont always know what frequency they will show up at.

4 Introduction to Presto - Pulsar Search Tool

PRESTO is a software package for pulsar search and analysis written by Scott Ransom of NRAO. It was primarily

designed to eﬃciently search for binary millisecond pulsars from long observations of globular clusters although it is

in use for various surveys with short integrations and to process a lot of X-ray data. It is written primarily in ANSI

C, with many of the recent routines in Python such as the get TOAs routine. The ﬁles generated by presto is used

for estimation of the time of arrival of a pulse using the get TOA routine which is also written by Scott Ransom.

4.1 Working with Presto Software

The input to the presto is the binary raw ﬁle which is generated by observation carried out by the Observatory

along with certain other header ﬁle (For GMRT it has the extension of . gmrt hdr) and a parameter ﬁle which is

obtained from the ATNF Catalog maintained by the Australia Telescope National Facility. The sample .gmrt hdr

ﬁle is as shown below.

#DATA FILE HEADER #

Site : GMRT

Observer : Gupta

Proposal : 30_043

Array Mode : PA

Observing Mode : Timing

Date : 25/04/2017

Num Antennas : 13

Antenna List : C00 C01 C03 C04 C05 C06 C08 C09 C10 C11 C12 C13 C14

Num Channels : 2048

Channel width : -0.09765625

Frequency Ch.1 : 300.1

Sampling Time : 81.92

Num bits/sample : 16

Data Format : integer binary, little endian

Polarizations : Total I

MJD : 57868.63887927306

UTC : 15:19:59.169192001

Source : PSRJ1022+1001

Coordinates : 10:22:58.0007, +10:01:52.77

Coordinate Sys : J2000

Drift Rate : 0.0, 0.0

Obs. Length : 2621440

Bad Channels : 2:1-20,2020-2048

Bit shift value : 8

The hdr ﬁle consists of parameters such as the Sampling time , No of bits per sample, MJD, UTC etc. which are

explained below.

Site: Gives the name of the observatory.

Array Mode: In Phased Array arrangement the voltage outputs from the antennas are combined by adding phases

to obtain a total power signal.

Frequency Ch.1 : It is the lower limit in the frequency of observation where the 300.1MHz represents the observation

in the band 300MHz - 500MHz with 200Mhz bandwidth by the GMRT Observatory.

Antenna List: Denotes the names of the antenna which were used for t he observation.

Sampling Time: The time interval in which the data is collected for digitalization.

Source: It gives the name of the Pulsar of which the observation was.

Coordinates: This gives the coordinates of the Pulsar in the J2000 coordinate system.

Bad-channels: This used to specify the presto to ignore certain channels during the folding process.

4.2 Analysis using presto

The work carried out in the last 8 week period was primarily on Presto, GPTool and Tempo2, using the data

generated in the uGa-10 ,cycle 30 047, 31 057 and 32 032. The various pulsars observed in these cycles were as

follows

• J06130200

• J0645+5158

• B0740-28

• J0751+1807

• B0950+08

• J1012+5307

• J1022+1001

• 1024-0719

• 10454509

• B1133+16

• J14553330

• J1614-2230

• J1640+2224

• B1642-03

• J1713+0747

• J1730-2304

• J1738+0333

• J1857+0943

• J19093744

• J2302+4442

• JB1508+55

• J16003053

The data obtained from the pulsar observation will be in the binary format which had to linked to dat ﬁle in order

to run the prepfold command for the Presto analysis which also requires a parameter ﬁle containing the details of

the various parameters of the pulsar such as the Barycentric period of the pulsar (s),Time derivative of barcycentric

period, Barycentric rotation frequency (Hz),Time derivative of barycentric rotation frequency,Second time derivative

of barycentric rotation frequency etc which is obtained from the ATNF catalog

A typical result form the presto analysis of the pulsar J1909-3744 is shown in the ﬁgure 1.

http://www.atnf.csiro.au/people/pulsar/psrcat/

Figure 1: A prepfold plot or a presto output plot of J1909-3744. This pulsar has a period of 0.002947108069160717

seconds.

4.3 The Time Domain and Pulse Proﬁle

Figure 2 above shows one of the most important plots that is generated through the prepfold routine for which the

analysis was carried out during this last 8 weeks. The time domain plot shown above gives details regarding the

strength of the signal collected by any telescope which varies during the course of the observation. The darker gray

segments represents the strength of the signal in some small piece of the data, which is also know as bin in the

terminology. The darker the bin indicates higher signal strength where as the white parts indicates that no signal

was detected in that place or bin. Each of the row on the Time vs Phase plot is a sub-fold, i.e. a fold of a small part