REU program-Summer 2006
Univ. of Wisconsin - Madison
Madison, WI 53706


Research projects of other REU students
Useful links
Obs Cos Logo

The Wisconsin Small Telescope Array for Radio-Waves


From June to August 2006 I worked, as part of the Astrophysics REU program at the University of Wisconsin-Madison, as an undergraduate researcher at the Observational Cosmology Laboratory at the Department of Physics in UW-Madison. My project was the Wisconsin Small Telescope Array for Radio-Waves (WSTAR), and my work largely involved software and hardware modification and instrumentation for the first telescope constructed for the array. This website provides a general description of WSTAR use, instrumentation, and theoretical relevance, as well as a summary of my contributions to the project during that summer. My final REU presentation can be viewed here.

I have also had the opportunity to give a group talk this summer about the Gamma-Ray Large Area Space Telescope (GLAST) and gamma-ray detection. A brief summary of my presentation and a link to its .ppt file is also found on this website.

Above: One of the three small radio telescopes (SRTs) comprising WSTAR

1. Introduction
2. Dectection of Anisotropies in the Cosmic Microwave Background Radiation (CMB)
3. 21-cm Emission Line from Neutral Hydrogen
4. WSTAR Design and Instrumentation (inc. my contributions to the project)
5. The Gamma-Ray Large Area Space Telescope (GLAST) and Gamma-Ray Detection

1. Introduction

The Wisconsin Small Telescope Array for Radio-Waves (WSTAR) is a three-dish interferometer designed to investigate and experiment with alternate beam-combination techniques in interferometry in order to minimize systematic errors in the detection of anisotropies in the cosmic microwave background radiation (CMB). To reach definitive conclusions about these techniques, WSTAR will map in different configurations the 21-cm radiation emitted from neutral hydrogen in the Galaxy and in the environment of nearby extragalactic sources. Additionally, WSTAR will serve as a tool in undergraduate education and training in radio astronomy.

Construction of WSTAR began in June 2005. Currently, one of the three small radio telescopes (SRTs) that will comprise the array is nearing completion. It has been built almost exclusively by undergraduate students working under the supervision of Prof. Peter Timbie at the Observational Cosmology research group at the University of Wisconsin-Madison. The design for the individual SRTs follows the design of the Haystack Observatory of the Massachusetts Institute of Technology, as outlined on the Haystack Small Radio Telescope support website. The completion target date for WSTAR has not yet been determined.

This website starts with an introduction to the CMB and the significance of the WSTAR mission for the future of CMB surveys. This section includes information about the beam-combination techniques in interferometry that we will be comparing. Section 2 provides an overview of the 21-cm emission line from neutral hydrogen, including the emission mechanism and the motivation for testing WSTAR with 21-cm radiation. Section 3 contains information about the design, instrumentation, and software of WSTAR, and also includes a general description of my personal contributions to the project.
Back to contents

2. Detection of Anisotropies in the Cosmic Microwave Background (CMB)

2.1 Background

One of the greatest questions in cosmology concerns the formation and evolution of large-scale structure in the universe and the initial conditions and processes that have led to the current observed large structures. The largest gravitationally-bound structures that we know of in the universe are galaxy clusters and superclusters, millions of light-years across, containing massive amounts of matter in the form of stars, planets, black holes, neutron stars, brown dwarfs, and other bodies. Yet according to the Big Bang model, at its earliest stages, the universe was a hot, extremely dense plasma of photons, electrons, and baryons (protons, neutrons, and some heavier particles), constantly interacting with each other with tremendously high energies, and thus unable to form neutral atoms. Consequently, the universe was opaque to radiation; due to the constant collisions of photons with other particles, a given emitted photon would almost instantaneously be reabsorbed (the mean free path of the photons was very close to zero). In other words, radiation was “coupled” to matter. As the universe expanded, the wavelength of the photons stretched, decreasing the energy of the photons and therefore the temperature of the universe down to a threshold temperature, about 3000 K, in which conditions were conducive for electrons and protons to combine as neutral atoms (predominantly hydrogen, with some helium). As a result of this process, which took place about 380,000 years after the Big Bang, termed recombination, photons were able to travel more freely in space and in effect “decoupled” from matter. These photons constitute the cosmic microwave background radiation (CMB), an imprint of matter distribution and spatial geometry from the era of recombination.

In the first few fractions of a second after the Big Bang the universe was rather small and in full causal contact and therefore also in thermal equilibrium. Then, about 10-35 seconds after the Big Bang, the universe underwent a period of accelerated expansion known as cosmic inflation, after which many regions that were initially in causal contact lost contact with each other. However, because of the original thermal equilibrium, temperature and matter distribution across the universe would be expected to remain generally uniform even after inflation, resulting in an isotropic plasma. Yet because of the tremendous expansionary effects of inflation, tiny quantum fluctuations in the causally-connected universe were expanded to classical levels, resulting in slight variations in matter distribution and spatial geometry. Since radiation was coupled to matter prior to recombination, information about any such inhomogeneities in the early plasma would be carried by the CMB, which would be expected to be isotropic otherwise. Over time, the effects of these variations, or anisotropies, became amplified, as denser regions gravitationally attracted more matter and grew denser. This process eventually led, as the universe continued expanding, to the formation of larger structures, resulting in the large-scale structure that we see today.

The information that the CMB contains about matter distribution and spatial geometry can be extracted from temperature and polarization observations and calculations. Temperature is a clear indicator of matter density and distribution as hotter regions are also denser. Polarization contains information about both matter distribution (“E mode” polarization) and spatial geometry, specifically gravitational waves (“B mode” polarization). The variations in all these parameters are quite small: CMB temperature variations are on the order of less than a 100 micro-Kelvins, E-mode variations of under 1 micro-Kelvin, and B-mode variations of a few tens of nano-Kelvins. So far, only temperature and E-mode variations have been detected in the CMB. B-mode variations are currently below the detection thresholds and remain a theoretical prediction.

WMAP map of CMB Anisotropies
WMAP map of CMB anisotropies (courtesy of NASA)

2.2. Detection Methods

Currently, detection of CMB anisotropies is done either by imaging systems, like the space-based Wilkinson Microwave Anisotropy Probe (WMAP) or by interferometers, which work by combining signals from a number of antennas constituting an array. Each of these methods has different systematic effects and therefore each has its different associated benefits.

However, traditional beam-combination techniques of interferometry pose a number of problems when applied to the detection of the CMB. First, the computational algorithms used in traditional correlation (multiplying) interferometers are rather cumbersome, and become substantially more complicated as the size of the array increases. Given the small scale of the CMB anisotropies, it is requisite that arrays be as large as possible, yet current computing capabilities put a rather low upper limit on the size of the array if the algorithms associated with correlation interferometers are used. Additionally, because they use voltage dividers, which lead to signal loss, correlation interferometers necessitate the use of amplifiers to compensate for the signal loss. Unfortunately, at high CMB frequencies such low-noise amplifiers are very difficult to construct.

Therefore, a different technique that would allow detection of high CMB frequencies (up to 140 GHz) and make larger arrays possible is called for. An “adding” interferometer would involve a direct addition of signals from the different antennas and therefore would not need either voltage dividers or amplifiers. Furthermore, the computational algorithm involved is considerably less complicated, making larger arrays feasible without straining current computing capabilities.

correlation interferometer
Correlation (Multiplying) Interferometer
adding interferometer
Adding Interferometer

WSTAR will serve as an initial test of the potential of adding interferometry. The array will be configured first as a correlation interferometer and then as an adding interferometer, using an adding algorithm. Data and results from each of the surveys will then be compared for quality determination, allowing us to decide whether further alterations to the algorithm are needed and whether adding interferometer indeed has important potential in future CMB detection. Testing will be done at the 1.42 GHz frequency – the frequency of the 21-cm emission line from neutral hydrogen.
Back to contents

3. 21-cm Emission Line from Neutral Hydrogen

About 90% of the interstellar medium is in the form of neutral and ionized hydrogen gas. The 21-cm spectral line is the result of a change of energy state in hydrogen atoms and the consequent emission of photons at a wavelength of 21.1 cm (f = 1420.4 MHz), which is in the microwave part of the electromagnetic spectrum (hence in the CMB spectrum).

3.1. Emission Mechanism and Motivation for Testing

Emission of photons at the 21 cm wavelength occurs when a hydrogen atom, which consists of a single proton and a single electron, undergoes a hyperfine transition at the ground state. The energy of the atom is slightly higher when the spins of the electron and the proton are parallel, i.e. in the same direction, than when they are antiparallel. Therefore, if the electron has spin in the opposite direction to the proton’s spin, it will eventually change spin direction, releasing energy in the process and thus emitting a photon at the 21-cm wavelength. While the particular change of energy state that leads to the photon emission at the 21 cm wavelength is fairly rare, with the average neutral hydrogen atom taking 12 million years to undergo this transition, because of the large number of neutral hydrogen atoms in the universe the total emission is substantial.

The 21-cm line is a particularly appealing choice for testing adding interferometry because it is in the CMB range, allowing us to test the computational algorithm, yet it is a relatively low CMB frequency, which facilitates building the equipment. Additionally, the signals from 21-cm radiation are rather large and there is low atmospheric interference at that temperature. Finally, there is a host of data available from other 21-cm emission detection experiments to permit further image comparison and result confirmation.

3.2. Further Applications

Radiation from neutral hydrogen has extensive applications in radio astronomy; in particular, the 21-cm emission line can be used to map the distribution and density of neutral hydrogen in the Galaxy and to find the velocity of hydrogen clouds, which can then be used to track the distribution of mass in the Galaxy. Thus neutral hydrogen gives information about not only the physical shape of the Galaxy (its distribution can be used to estimate distances to the spiral arms of the Galaxy), but also about the amount of matter in the Galaxy. Notably, calculations of the amount of matter and mass in the galaxy have led scientists to predict the existence of dark matter in the universe. Neutral hydrogen is therefore of interest for various subfields of astronomy such as stellar astronomy, galactic astronomy, and cosmology. In addition to dark matter surveys, applications in cosmology of the 21-cm hydrogen line include furthering our understanding of the “dark ages” of the universe – the era between recombination and reionization – and investigation of theories about dark energy.
Back to contents

4. WSTAR Design and Instrumentation

This section is divided into a hardware overview, a software overview, and a general description of my main contributions to the project.

4.1. Hardware

The array will consist of three radio telescopes that will be combined as an interferometer. Initial spacing will be 30 feet, but this may change with time, depending on the particular experimental requirements.

Telescope spacing
Above: Telescope spacing diagram

The design of each individual radio telescope is based on MIT’s Haystack Observatory’s design of the Small Radio Telescope (SRT). The SRT has been designed by Haystack and is sold as a kit by CASSI Corp.; however, our SRT has been built from scratch using the Haystack documentation, schematics, and java files (for the software) only. As a result, our design somewhat differs from the Haystack design, and I have included in this website our own schematics for the receiver board and the ground controller (the Haystack schematics can be accessed on the Haystack website), as well as a description of the most important modifications that I have made to both our design and to the Haystack design.

SRT Block Diagram
Above: SRT block diagram

Signals enter the antenna feed and are subsequently amplified in the frequency range of interest by the low-noise amplifier (LNA). The signals are then transmitted through a coaxial cable for processing in the digital receiver board, which is primarily controlled by its digital signal processor (DSP). Pulses from and to the transmission and receiving pins of the DSP chip, along with DC power from the controller, are then sent through a single conductor coax connecting the receiver board to the ground controller. This two-way communication cable between the two boards provides voltage to the receiver board and allows data to flow from the ground controller to the receiver board and vice versa, hence permitting active communication of the software with the signal receiving and processing parts of the telescope.
The ground controller communicates with both the motor and the receiver board, transmitting signals from the motor and the receiver board to the computer and vice versa. The I/O connection on the ground controller consists of an RS232DX/RX chip that accepts serial connections. Currently, the serial port is connected directly via a serial cable to a computer on which the user-interface software is installed, but I have set the stage for remote control of the telescope from the lab, which, once set up, will involve running the serial cable from the RS232DX/RX on the ground controller to a serial-to-Ethernet adapter with an IP address rather than directly to a computer. The adapter will be connected to the Observational Cosmology local area network (LAN) via an Ethernet cable running from the roof to the main network control room. Communication between the software and the ground controller will thus be done remotely through the internet and port simulation software.

Antenna. The dish is constructed of four sections, each made from C/Ku band mesh supported by an aluminum frame. The mesh reflects all incident electromagnetic energy if the holes on the surface are less than one tenth of a wavelength (given the long 21-cm wavelength this is not a problem). The dish is 102 inches across and is 15.75 inches deep. The focal length was calculated to be 41.3 inches.
Mount and motor. Our SRT uses an az/el (azimuth/elevation) motor purchased from Alpha-Spid. The motor is capable of a 0-359 degree motion range in azimuth and 0-90 degree motion range in elevation. The travel limits are set through the software; there are, additionally, two built-in limit switches for the elevation rotation that prevent the dish from colliding with the telescope stand in case of a software malfunction. The motor’s rotation is accurate to about 1 degree, and can be instructed, through the software, to move in alternating az/el steps while slewing. There are two wiring blocks in the mount – an azimuth block and an elevation block, each with four incoming wires from the motor and three outgoing to the motor in addition to a grounding wire from each (attached to grounding post on mount through the elevation block and to ground on the ground controller through the azimuth block; there is a jumper between these connections on the two blocks). Motor I/O processing is done by the BS1 stamp chip on the ground controller, which was programmed with the PBASIC code provided by Haystack.
Low noise amplifier (for correlation configuration). We used the DEM 1420ULNA from Down East Microwave Inc. This LNA has a 17dB nominal gain (with a noise figure of <.0.4dB) and works on a voltage range of +7 to +16 V DC. The 1420ULNA amplifies frequencies in the range of 1350-1500 MHz.
Digital receiver board. Click here to view our schematic for the receiver board. The receiver board uses the power supply of the ground controller, with +16 V flowing from the controller to the receiver board through the connecting cable. The main signal processing and transmission unit on the receiver board is the DSP56F803 chip, which is flash-programmed, using Metrowerks Codewarrior for DSP56800, with C and Assembly language code provided by Haystack.
Ground controller. Click here to view our schematic for the receiver board. The brain of the ground controller is its BS1 stamp chip, which was programmed using the PBASIC code available on the Haystack website. Part of my work involved debugging and temporarily modifying this code to identify potential motor and controller hardware problems as well as make necessary modifications in the java-based user-interface software.

The parts list for both the receiver board and the ground controller is detailed on the Haystack website.

4.2. Software

The control software for the SRT is a java-based, user-friendly GUI. It enables user communication with both the motor and the receiver and accepts both real-time user instructions and command-file instruction sequences. To identify the hardware and sources, it reads a catalogue file (srt.cat) that may be changed by the user to include different coordinates. The software can be run on one of four modes: full simulation (i.e. both antenna and motor simulated), simulation of only the motor, simulation of only the antenna, and an actual run of all components. These options enable faster problem identification and will also be beneficial when used for training purposes. Click here to view a full description of the interface and problem capabilities of the SRT software.
Back to contents

4.3. My Main Contributions and Modifications

When I came to UW-Madison, the SRT was structurally complete, but neither the simulation nor the actual software were executing properly, and the telescope had a number of hardware problems. In the course of my project I made a number of modifications to the java code for the control software and to the environmental setting on our Windows OS. I have also used the code (java for the control software; PBASIC for the BS1 stamp chip on the ground controller; C, Assembly, and linker script for the DSP chip on the receiver board) to identify and locate hardware and electronics problems. I then made certain hardware modifications as well. By the end of my stay at UW-Madison, the simulator software was successfully executing, and the motor and ground controller were fully operational. With the receiver simulated, the actual software communicated correctly with the motor through the ground controller. The receiver, however, was not yet communicating with the ground controller due to a software compatibility issue relating to the Metrowerks Codewarrior IDE version used to write the code created for the receiver board’s DSP chip (the code, written by the Haystack team for an older version of Metrowerks Codewarrior, is no longer compatible with current available versions and therefore could not be flash-programmed to the DSP with tools available to us). I therefore took the receiver board with me to the Haystack Observatory in Massachusetts, where I received very valuable assistance with programming the chip from Dr. Alan Rogers. Once the chip was programmed, we were able to identify several other small glitches in the board and those were fixed in Haystack as well. I then sent the board back to my advisor, Peter Timbie, in UW-Madison. In the course of the next few weeks, I provided remote assistance with some other functional problems relating to the telescope that arose after I left. At this point, with the fixed receiver board, the telescope is fully operational and has been set up at its correct position, allowing the students returning to the project in the fall to proceed with building the remaining two telescopes in the array.
Finally, during the last few weeks of the program, I originated and devised a remote-control scheme for the array. Under this plan, an Ethernet connection simulating a remote serial port will be substituted for the current physical serial connection between the ground controller and the main computer from which the array is controlled. The serial port on the ground controller (located close to the array) will be connected to a serial-to-Ethernet adapter, which will in turn be connected to the Observational Cosmology Lab local area network (LAN), allowing an operator to control the telescope from the lab. I found and ordered the necessary components for this scheme, but, since the adapter that we received was defective, I was unfortunately not able to set up this arrangement myself. The part has been re-ordered and students returning to the lab in the fall will be involved in installing it.
I expect to continue providing remote assistance with the telescope software and hardware, and in particular with the installation of the remote control plan, to students returning to the laboratory in the fall.
Back to contents

5. The Gamma-Ray Large Area Space Telescope (GLAST)

Gamma rays have the highest energies, 10 keV and above, on the electromagnetic spectrum and are produced by a number of intriguing phenomena, including blazars, pulsars, solar flares, gamma-ray bursts, and possibly the mysterious dark matter. A notable feature of gamma rays – in contrast to, for example, the CMB – is the fact that they are not bent by magnetic fields and therefore carry precise information about their origin in the universe. However, given the extremely high energies of gamma radiation, gamma ray detection is rather difficult and often employs the techniques used in experimental high-energy particle physics. Currently, earth-based telescopes searching for gamma rays can only detect rays of energies higher than 1 TeV, while the most recent space-based telescopes have a high detection threshold of about 100 GeV (0.1 TeV). One of the main goals of current gamma-ray instrumentalists is to build equipment capable of detecting gamma rays in this energy gap. The Gamma-Ray Large Area Space Telescope (GLAST), scheduled for launch in 2007, is a multinational project partly aimed at that goal. Its other attributes include a large field of view, a full-sky scan every three hours, and an onboard instrument specifically dedicated to the detection of gamma-ray bursts. In my talk I discussed the history of gamma-ray detection, the main astrophysical phenomena relating to gamma-ray emission, and the details of the design and instrumentation of GLAST. My talk can be viewed as a PowerPoint presentation here.
Back to contents


Schematics and parts lists

Pictures and videos


Useful links

UW Madison Physics Department

Observational Cosmology Research Group at UW Madison

MIT Haystack Observatory SRT Homepage