The Omniscope Project!

We are a group based at the Massachusetts Institute of Technology Kavli Institute for Astrophysics and Space Research.

The Idea

What better way to show our love for science than to try our hand at being universe cartographers! Mapping our universe in 3D by imaging the redshifted 21 cm line from neutral hydrogen has the potential to overtake the cosmic microwave background as our most powerful cosmological probe, because it can map a much larger volume of our Universe, shedding new light on the epoch of reionization, inflation, dark matter, dark energy, and neutrino masses. We report on MITEoR, a pathfinder low-frequency radio interferometer whose goal is to test technologies that greatly reduce the cost of such 3D mapping for a given sensitivity. MITEoR accomplishes this by using massive baseline redundancy both to enable automated precision calibration and to cut the correlator cost scaling from N^2 to N log N, where N is the number of antennas. The success of MITEoR with its 64 dual-polarization elements bodes well for the more ambitious HERA project, which would incorporate many identical or similar technologies using an order of magnitude more antennas, each with dramatically larger collecting area. 

The actual Omniscope instrument has several components as illustrated below.

Hover over a component to see what it looks like!


This shows data flow in a very large omniscope. A hierarchical grid of dual-polarization antennas converts the sky signal into volts, which get ampli fied and filtered by the analog chain, transported to a central location, and converted into bits every few nanoseconds. These high-volume digital signals (thick lines) get processed by field-programmable gate arrays (FPGAs) which perform a temporal Fourier transform (using polyphase filter banks). The FPGAs (or GPUs) then multiply by complex-valued calibration coefficients that depend on antenna, polarization and frequency, then spatially Fourier transform, square and accumulate the results, recording integrated sky snapshots every few seconds and thus reducing the data rate by a factor  10^9. They also cross-correlate a small fraction of all antenna pairs, allowing the redundant baseline calibration software to update the calibration coefficients in real time. Finally, software running on regular computers combine all snapshots of sufficient quality into a 3D sky ball representing the sky brightness as a function of angle and frequency in Stokes (I,Q,U,V), and subsequent software deals with foregrounds and measures power spectra and other cosmological observables.

The Team

As with any project this large, one of the most important components is our awesome team!


Hauxuan (Jeff) Zheng

Max Tegmark

Victor Buza

Joshua S. Dillon

Hrant Gharibyan

Jack Hickish

Eben Kunz

Adrian Liu

Jon Losh

Andrew Lutomirski

Scott Morrison
Sruthi Narayanan

Ashley Perko

Devon Rosner

Nevada Sanchez
Katelin Schutz
Shana M. Tribiano

Matias Zaldarriaga

Kristian Zarb Adami
Ioana Zelko
Kevin Zheng
Richard Armstrong
Richard F. Bradley
Matthew R. Dexter
Aaron Ewall-Wice
Alessio Magro
Michael Matejek
Edward Morgan
Abraham R. Neben
Qinxuan Pan
Courtney M. Peterson
Meng Su
Joel Villasenor
Christopher L. Williams
Hung-I "Eric" Yang
Yan Zhu
Bob Penna
Michael Valdez

The Evolution

X0: Expedition 1


This was the first expedition. The antennas are arranged on the roof of the Kavli Institute at MIT.

X1: Expedition 2


The second expedition took place in Greenbank, West Virginia. Pictured above is the arrangement of the 16 antennas.

X2: Expedition 3


On the third expedition we started with the Roach system. We had 8 antennas and this was the first expedition in West Forks, Maine.

X3: Expedition 4


The fourth expedition was actually done in the snow in West Forks, Maine. It was again using 8 antennas instead of 16.

X4: Expedition 5


We reverted back to 16 antennas on the fifth expedition, which was in West Forks, Maine in the summer time.

X5: Expedition 6


For the last expedition we improved on the calibration as well as making it possible to use 64 antennas! This took place this past July in West Forks, Maine.

The Instrument

Let's walk through all the components of the Omniscope.

1. The Analog System

The Analog System consists of all the actual hardware and circuitry that we use to obtain signals before sending them to computers to be parsed together.



We start with the antennas. MITEoR contains 64 dual-polarization antennas, giving 128 signal channels in total. The 64 antennas are arranged in 8 rows of 8. They are responsible for receiving the signals from the universe. The signal picked up by the antennas is first amplified by two orders of magnitude by the low noise amplifiers (LNAs) built-in to the antennas. Each antenna then has two channels which we connect to a line driver island.

Swappers, Line Drivers and Amplifiers

line drivers

This past expedition we tested connecting some antennas to swappers and then to line drivers, while the others did not use the swappers at all. The signals, after being amplified in the antennas, are then phase switched in the swapper system, which greatly reduces cross-talk downstream. Following this, the signal is amplified again by about five orders of magnitude in our line-drivers before being sent over 50 meter RG6 cables to the receivers.

Receivers and ADC

Receiver connections

The receiver boards are collected and stored in the racks that contain part of the digital system. All the wires from the antennas and amplifiers connect to these boards. The receivers perform IQ demodulation on a desired 50 MHz band selected between 100 MHz and 200 MHz, producing two channels with adjacent 25 MHz bands, and sends the resulting signals into our digitization boards containing 256 analog-to-digital converters (ADCs) sampling at 50 MHz.

ADC boards

From there we move on to the digital system.

2. The Digital System

We designed MITEoR's digital system to be highly compact and portable, with the entire system occupying 3 shock-mounted equipment racks on wheels, each measuring about 5 ft by 5 ft by 3 ft. It takes in data from 256 ADC channels, Fourier transforms the data into the frequency domain, reconstructs IQ demodulated channels back to 128 corresponding antenna channels, computes the cross-correlations of all pairs of the 128 antenna channels with 8 bit precision (rather than 4 bits seen in most correlators of similar specifications), and then time-averages these cross-correlations. The digital hardware is capable of processing an instantaneous bandwidth of 12:5 MHz with frequency bin size 5 kHz, and the time interval for averaging and subsequent output to our data server is usually con figured to be either 2.6 or 5.3 seconds.


Photo Gallery


Video Gallery

On our last expedition we had some fun with time-lapse photography. This is a movie of the building of two of our antennas!


Starter Kit for Aspiring 21-cm Tomographers

Popular Level, No Math -- Time Magazine article on high redshift observations. -- Scientific American article on 21-cm tomography science with a focus on the EOR. -- Science magazine article on the Omniscope.

Engineering Summary -- M. Eng. thesis proposal written by Nevada Sanchez.  Great summary of the Omniscope project from an engineering perspsective.

Theses -- Ashley Perko’s undergraduate thesis.  The appendices are particularly useful for people who want to use the actual instrumentation of the Omniscope. -- Nevada Sanchez’s M. Eng. thesis.  An excellent introduction to the entire Omniscope. --    Andy Lutomirski’s Ph.D. thesis.  -- Eben Kunz’s M. Eng. Thesis.  Description of the Omniscope analog chain.

Big Picture Summaries, Some Astro Vocab Required -- Astrophysics with the 21-cm line. -- Cosmology with the 21-cm line. -- Big picture view of reionization.

Some Serious References -- THE reference for anything in 21-cm tomography up until 2006. -- An update of the previous paper written by some different authors, focusing on the recent interest in low redshift baryon acoustic oscillation measurements. -- A more compact (and therefore more readable) review with a lot of nice diagrams. -- A more theoretically-oriented review.

General Introduction to Cosmology, Suitable for Undergraduate Physicists

An Introduction to Modern Astrophysics by Carroll and Ostlie – Chapter 29 (of the 2nd edition) is an excellent introduction to cosmology for those familiar with basic physics but perhaps not astrophysics or cosmology.

Introduction to Cosmology by Ryden – fantastic book for the same audience, with lots of intuitive explanations of the physics to go along with the math.

Introduction to Radio Astronomy

Interferometry and Synthesis in Radio Astronomy by Thompson, Moran, and Swenson – probably the definitive reference on radio astronomy.

An Introduction to Radio Astronomy by Burke and Graham-Smith – chapters 5 and 6 the most relevant.

Tools of Radio Astronomy by Wilson, Rohlfs, and Huttemeister --  a popular reference.

Omniscopes the first Omniscope paper. the second Omniscope paper.


We've publicly released the key Omniscope data products and you can download them here.

  1. We've made the largest-area sky-map to date around 150 MHz and you can download it either as a gzipped tarball, as individual files or from github. The data product consists of the following four ASCII files in HEALPIX nest format with nside=64, in equatorial coordinates: For a detailed description of these maps, please see our paper arXiv:1605.03980.
  2. We've made a global sky model from 10MHz to 3 THz using other data and compared it with our MITEoR map in our paper arXiv:1605.04920. You can download this global sky model from github
  3. .
  4. For those of you who want to create your own maps or other analyses, we've released the raw visibilities for the MITEoR experiement in this gzipped tarball. The file format is explained here.
  5. Publications

    Our publications on the Omniscope project are listed below (starting from most recent):

    An Improved Model of Diffuse Galactic Radio Emission from 10 MHz to 5 THz

    Brute-Force Mapmaking with Compact Interferometers: A MITEoR Northern Sky Map from 128 MHz to 175 MHz

    It's Always Darkest Before the Cosmic Dawn: Early Results from Novel Tools and Telescopes for 21 cm Cosmology

    Mapmaking for precision 21 cm cosmology

    MITEoR: A Scalable Interferometer for Precision 21 cm Cosmology

    Global 21cm signal experiments: A designer's guide

    A Fast Method for Power Spectrum and Foreground Analysis for 21 cm Cosmology

    How well can we measure and understand foregrounds with 21 cm experiments?

    A Method for 21cm Power Spectrum Estimation in the Presence of Foregrounds

    Precision Calibration of Radio Interferometers Using Redundant Baselines

    Solving the Corner-Turning Problem for Larger Interferometers

    Omniscopes: Larger Area Telescope Arrays with only N log N Computational Cost

    An Improved Method for 21cm Foreground Removal

    Will point sources spoil 21 cm tomography?

    The Fast Fourier Transform Telescope

    How accurately can 21 cm tomography constrain cosmology?

    This web page was designed by Sruthi A. Narayanan (