CHIME (Canadian Hydrogen Intensity Mapping Experiment)
It sounds almost too apt to be true. An observatory shaped like the half-pipes used by snowboarders, and dependent on technology originally designed for gaming and mobile phones, will soon be tasked with plugging a crucial gap in the cosmological record: what the Universe did when it was a teenager. 1)
The information will allow cosmologists to gauge whether the strength of dark energy — the force accelerating the Universe’s expansion — has changed over time, an unresolved question that governs the fate of the cosmos.
Whereas typical radio telescopes have round dishes, the Canadian Hydrogen Intensity Mapping Experiment (CHIME) comprises four 100-meter-long, semi-cylindrical antennas, which lie near the town of Penticton in British Columbia.
From 2016, CHIME’s half-pipes, which are scheduled to be completed this week, will detect radio waves emitted by hydrogen in distant galaxies. These observations would be the first measurements of the Universe’s expansion rate between 10 billion and 8 billion years ago, a period in which the cosmos went “from being a kid to an adult”, says Mark Halpern, the leader of CHIME and an experimental cosmologist at the University of British Columbia in Vancouver. Straight after the Big Bang 13.8 billion years ago, the rate of the Universe’s expansion slowed. But somewhere during the ‘adolescent’ period, dark energy — which eventually turned the Universe’s slowing expansion into the acceleration observed today — began to be felt, he says.
It is a window in time that has, until now, been closed. Cosmologists measure the Universe’s past expansion rate using ancient objects, such as supernova explosions and the voids between galaxies, that are so distant that their light is only now reaching Earth. Over the past few decades, such objects have revealed that the cosmos has been expanding at an accelerating rate for more than 6 billion years. And surveys of quasars — mysterious, super-bright objects that outshine the entire galaxies they lie in — have shown that until 10 billion years or so ago, the Universe’s expansion was slowing down.
But cosmologists have struggled to measure the expansion rate in the interim, leaving open the question of whether the strength of dark energy’s repulsive force may have varied over time.
CHIME is designed to fill the gap, says Kendrick Smith, an astrophysicist at the Perimeter Institute for Theoretical Physics in Waterloo, Canada, who will work on analysing CHIME’s data. The half-pipe antennas will allow CHIME to receive radio waves coming from anywhere along a narrow, straight region of the sky at any given time. “As the Earth rotates, this straight shape sweeps out the sky,” says Smith.
To sort out where individual signals are coming from, a custom-built supercomputer made of 1,000 relatively cheap graphics-processing units — the type used for high-end computer gaming — will crunch through nearly 1 terabyte of data per second. The team will also use signal amplifiers originally developed for mobile phones. Without such powerful consumer-electronics components, CHIME would have been prohibitively expensive, says experimental cosmologist Keith Vanderlinde of the University of Toronto, Canada, who is co-leading the project.
CHIME’s supercomputer will look specifically for radio waves with a wavelength that suggests an age of 11 billion to 7 billion years, emitted by the hydrogen in the interstellar space inside galaxies (at their source, such emissions have a wavelength of 21 cm). Researchers will then subtract the ‘radio noise’ in the same wavelength range that comes from the Milky Way and Earth.
Although CHIME will not be able to distinguish individual galaxies in this way, clumps of hundreds or thousands of galaxies will show up, says Vanderlinde. This will allow researchers to map the expansion rate of the voids between the clumps, and in turn to calculate the strength of dark energy during that time.
If the results imply that the strength of dark energy then was the same as it has been in the past 6 billion years, it could suggest that galaxies will eventually lose sight of each other. But if the strength of dark energy has changed over the eons, all bets are off: the Universe could collapse in a ‘big crunch’, for example, or be ripped apart into its subatomic components.
As well as mapping the adolescent Universe, CHIME could also detect hundreds of the mysterious ‘fast radio bursts’ that last just milliseconds and have no known astrophysical explanation. And it will help other experiments to calibrate measurements of radio waves from rapidly spinning neutron stars, which researchers hope to use to detect the ripples in space-time known as gravitational waves. 2)
CHIME is part of a growing trend in astronomy. A number of experiments that are now active or in the planning stages, including the hotly anticipated SKA (Square Kilometer Array) — to be built on sites in Australia and South Africa — are designed to look at hydrogen emissions with 21cm wavelengths. These emissions are an untapped trove of cosmological information, says Tzu-Ching Chang, an astrophysicist at the Academia Sinica Institute of Astronomy and Astrophysics in Taipei who helped to pioneer the hydrogen mapping of galaxies in a 2010 study . 3) She likens the boom in hydrogen mapping today to the trend in the 1990s of studying the relic radiation of the Big Bang, which revolutionized cosmology.
Figure 1: The CHIME telescope incorporates four 100-meter long U-shaped cylinders of metal mesh that resemble snowboard half-pipes, with total area equivalent to five hockey rinks. CHIME reconstructs the image of the overhead sky by processing the radio signals recorded by thousands of antennas. Its signal processing system is the largest of any telescope on Earth, allowing it to search huge regions of the sky simultaneously (image credit: CHIME collaboration) 4)
Development of CHIME
• September 7, 2017: CHIME, the revolutionary Canadian radio telescope inaugurated today in British Columbia, incorporates creative technology that will enable it to simultaneously tackle major astrophysics and cosmology topics — including studying the nature of “dark energy” by making unprecedented maps of the distant Universe, and determining the origin of the mysterious phenomenon known as Fast Radio Bursts. 5)
Unlike traditional telescopes, CHIME (Canadian Hydrogen Intensity Mapping Experiment) has no moving parts. The telescope, produced by 50 scientists from the University of British Columbia, the University of Toronto, McGill University, and the National Research Council of Canada "sees" the radio sky in a novel way: It uses over a thousand antennas to record the information from all the radio waves falling across its surface.
The radio signals are converted to digital data at a rate of 13 Tbit/s - comparable to the entire world’s mobile data rate. This amount of information can’t be stored, so it is processed in real time. With these techniques, CHIME is able to digitally “point” anywhere within a wide stripe of sky overhead, making traditional telescope steering unnecessary. More importantly, it is able to look in many different directions simultaneously.
Low-cost, high-powered signal processing
The innovation that has led to CHIME was made possible by creative new technology developed at universities across Canada, allowing this telescope to be built with $16 million (Canadian) in federal and provincial funding -- a fraction of the cost of other world-class radio telescopes.
At the heart of CHIME is a sophisticated digital network and signal-processing "brain," known as the correlator, where the massive amount of information from radio waves is processed into an image of the overhead sky. The correlator developed by the team is unparalleled in the world in terms of raw networking and processing rate, and low cost. To achieve this, the team had to develop new, custom innovations integrated with market-driven commercial technologies.
“You can’t just go out and buy the world’s largest radio correlator”, explains McGill University's Matt Dobbs, “it needs to be carefully developed by a team who knows the science objectives in detail and is able to engineer the technology.” In doing this, Dobbs explains, “we’re building both a novel instrument and a team of experts in Canada with the know-how to innovate world-class scientific instruments. Working with the amazing team of engineers, postdocs and students who together have realized this project has been inspiring.”
The custom-designed-and-built circuit boards convert radio signals detected by the antennas into digital data. They split these data into about a thousand different radio frequencies, much like a glass prism separates different colors of light. The data are then passed through a giant custom-built data network.
Figure 2: CHIME graduate student researcher Juan Mena with the custom electronics from the CHIME correlator (photo: Keith Vanderlinde)
A supercomputer built from computer-gaming gear
The second crucial component of the CHIME correlator was built by assembling a huge array of Graphics Processor Units (GPUs) — technology driven by the computer gaming industry — into a specialized supercomputer built for this application. By cleverly reprogramming the GPUs, researchers at the University of Toronto were able to use this supercomputer to construct cosmic sky maps while simultaneously enabling a sensitive search for FRBs (Fast Radio Bursts).
"My university wondered why I was using grant money to buy so many gaming cards," laughs University of Toronto researcher Keith Vanderlinde. "By building this supercomputer with commercially available GPUs, and developing a software that images the sky, we were able to build a world-leading, yet affordable, telescope. This lays the foundations for building even more powerful telescopes in the future."
CHIME's foundation is an array of four large cylinders resembling snow-board half-pipes, with total area equivalent to five hockey rinks. Each cylinder focuses radio waves onto a line of 256 radio antennas per cylinder that turn celestial signals into electrical signals. "CHIME's radio signals are sent through custom-designed analog circuitry that is based on commercially available and inexpensive filters and amplifiers developed for the cell-phone industry, says University of British Columbia researcher Gary Hinshaw. “With CHIME, we have made market forces work for game-changing scientific progress.”
A search engine for Fast Radio Bursts
A sensitive instrument to detect Fast Radio Bursts is also being developed for CHIME. Consisting of 128 compute nodes, this instrument can search continually for fast radio transients in the entire area of sky viewable by CHIME. Sophisticated algorithms and software pipelines have been developed to sift through 130 billion bits of data per second in real time. “By making smarter computer algorithms we have prepared CHIME to be the world’s most efficient Fast Radio Burst search engine,” says researcher Kendrick Smith of the Perimeter Institute, one of several other institutions contributing to the project.
Figure 3: Nolan Denman (UofT PhD student) at the CHIME telescope, inspecting one of the 256 processing nodes which make up the signal processing backend (photo: Keith Vanderlinde)
CHIME, now in its commissioning phase in preparation for science operations, was built at the Dominion Radio Astrophysical Observatory (DRAO) near Penticton, B.C., with funding from the Canada Foundation for Innovation, the governments of British Columbia, Quebec and Ontario, and the Canadian Institute for Advanced Research.
Figure 4: In recent years, scientists have discovered fast radio bursts, energetic single pulses of radio emission arriving in random directions from unknown sources well beyond our galaxy, the Milky Way. The origin of these pulses is a major puzzle in high energy astrophysics. With its huge field of view and broad frequency coverage, the CHIME telescope is a nearly ideal instrument for finding and studying many of these bursts, enabling scientists to tackle one of the most mysterious new areas of astrophysics (video credit: McGill University)
Figure 5: The new CHIME radio telescope will act as a time machine allowing scientists to create a three-dimensional map of the universe extending deep into space and time. CHIME’s unique “half-pipe” telescope design and advanced computing power will help scientists explore the frontiers of modern astronomy better understand the shape, structure and fate of the universe (video credit: McGill University)
A Canadian effort to build one of the most innovative radio telescopes in the world will open the universe to a new dimension of scientific study. The Honorable Kirsty Duncan, Minister of Science, today installed the final piece of this new radio telescope, which will act as a time machine allowing scientists to create a three-dimensional map of the universe extending deep into space and time. 6)
CHIME is an extraordinarily powerful new telescope. The unique “half-pipe” telescope design and advanced computing power will help scientists better understand the three frontiers of modern astronomy: the history of the universe, the nature of distant stars and the detection of gravitational waves.
By measuring the composition of dark energy, scientists will better understand the shape, structure and fate of the universe. In addition, CHIME will be a key instrument to study gravitational waves, the ripples in space-time that were only recently discovered, confirming the final piece of Einstein’s theory of general relativity.
CHIME is a collaboration among 50 Canadian scientists from the University of British Columbia, the University of Toronto, McGill University, and the National Research Council of Canada (NRC). The $16-million investment for CHIME was provided by the Canada Foundation for Innovation and the governments of British Columbia, Ontario, and Quebec, with additional funding from the Natural Sciences and Engineering Research Council and the Canadian Institute for Advanced Research. The telescope is located in the mountains of British Columbia’s Okanagan Valley at the NRC’s Dominion Radio Astrophysical Observatory near Penticton.
• The CHIME telescope incorporates four 100-meter long U-shaped cylinders of metal mesh that resemble snowboard half-pipes. Its overall footprint is the size of five NHL (National Hockey League) hockey rinks.
• CHIME collects radio waves with wavelengths between 37 and 75 cm, similar to the wavelength used by cell phones.
• Most of the signals collected by CHIME come from our Milky Way galaxy, but a tiny fraction of these signals started on their way when the universe was between 6 and 11 billion years old.
• The radio signal from the universe is very weak and extreme sensitivity is needed to detect it. The amount of energy collected by CHIME in one year is equivalent to the amount of energy gained by a paper clip falling off a desk to the floor.
• The data rate passing through CHIME is comparable to all the data in the world’s mobile networks. There is so much data that it cannot all be saved to disk. It must first be processed and compressed by a factor of 100,000.
• Seven quadrillion (7 x 1015) computer operations occur every second on CHIME. This rate is equivalent to every person on Earth performing one million multiplication problems every second.
Figure 6: New Canadian telescope will map largest volume of space ever surveyed (image credit: CHIME collaboration)
CHIME is an interferometric radio telescope at the Dominion Radio Astrophysical Observatory (DRAO) in British Columbia, Canada, operating at frequencies between 400 and 800 MHz. This corresponds to a redshift range of 0.8< z <2.5 for the21 cm line. CHIME consists of four large (20 m x100 m) cylindrical reflectors, each fitted with256 dual-polarization feeds. The cylindrical design of CHIME enables a large instantaneous field of view of~100º in the north-south direction, and a narrower field of view of ~1ºto2º in the east-west direction. For cosmology observations, the feeds are individually digitized and correlated with each other, resulting in a synthesized beam that splits up the instantaneous field of view into pixels with width ~0.25º to 0.5º. The telescope is operated as a drift-scan telescope with no moving parts, providing some much-needed stability for the control of systematics. As the sky drifts overhead, CHIME essentially maps out the entire northern sky. The telescope's low-noise amplifiers are built with components adapted from the cellphone industry and its data are processed using a custom-built FPGA electronic system and 1000-processor high-performance GPGPU (General-Purpose Computing on Graphics Processing Units) cluster. 7)
The CHIME Fast Radio Burst Project(CHIME/FRB)will search beam-formed, high time and frequency resolution data in real time for FRBs in the CHIME field of view. Here we describe the CHIME/FRB back end, including the real-time FRB search and detection software pipeline, as well as the planned offline analyses. We estimate a CHIME/FRB detection rate of 2–42 FRBs sky–1day–1normalizing to the rate estimated at 1.4 GHz by Van der Wiel et al. Likely science outcomes of CHIME/FRB are also discussed. CHIME/FRB is currently operational in a commissioning phase, with science operations expected to commence in the latter half of 2018. 8)
The CHIME Telescope Structure, Feeds, and Analog-signal Path
The CHIME telescope (Figure 7) is located on the grounds of the Dominion Radio Astrophysical Observatory(DRAO)near Penticton, British Columbia. The choice of operating frequency, collecting area, and angular resolution for the CHIME telescope was driven by the original motivation for the project:hydrogen intensity mapping of the entire northern hemisphere to probe the accelerating expansion of the universe over the redshift range where dark energy began to exert its influence,z=0.8–2.5. Since the BAO (Baryon Acoustic Oscillation) signal is weak, and large sky coverage is needed to overcome sample variance, exceptionally fast mapping speed is required, driving the instrument to a design with many hundreds of feeds to achieve the mapping goal in a reasonable amount of time. As 100 m class telescopes are expensive, and no positions on the sky are favored, a transit telescope with no moving parts is the preferred option. Table1provides a summary of the telescope’s key properties. Detailed telescope performance metrics will be provided in a future publication.
Figure 7: Photograph of the CHIME telescope on 15 September 2016, looking northwest. The shipping containers housing the X-Engine and CHIME/FRB back endcan be seen adjacent to the rightmost cylinder. The receiver huts containing the F-Engine are beneath the reflectors and cannot be seen here. The DRAO SynthesisTelescope(Kothes et al.2010) can be seen in the background (image credit: CHIME Collaboration)
The CHIME telescope consists of four 20 m wide and 100 m long cylindrical paraboloidal reflectors. The intercylinder gaps are 2 m. The cylinders are stationary, aligned north–south. The focal length was chosen to be 5 m (f/D=0.25) to place the feeds in the plane of the aperture to minimize cross-talk and ground pickup. With no moving parts, the telescope structure could be built at low cost. The structure is steel, built of conventional sections and components, to the dimensional accuracy common for commercial buildings. The reflecting surface is galvanized steel mesh with 16 mm openings, a compromise between allowing snow to fall through and minimizing ground noise leaking through to the feeds. The measured surface roughness is ~9mm(<2% of the observing wavelength), adequate for operation in the CHIME band.
Table 1: Key properties of the CHIME Telescope relevant to the CHIME/FRB Project
A schematic diagram of the CHIME telescope signal path is shown in Figure 8 and further explained in the following sections. A total of 256 dual-polarization feeds, spaced by30 cm, are placed along 80 m of the focal line of each of the four cylinders, giving a total of 2048 signal paths. Digital signal processing of these signals generates the multiple beams that make this telescope a powerful instrument for FRB research. The feed design(Deng et al.2014)achieves nearly equal beamwidths in both polarizations and excellent matching over the octave operating band of the receiver. Mutual coupling between adjacent and nearby feeds was considered in the design of the baluns and matching networks. The entire feed is constructed from printed-circuit materials, appropriate for mass production.
The focal line is deliberately kept simple, housing only feeds and low-noise amplifiers. Two low-noise amplifiers accept the signals from each feed, and those signals are carried through coaxial cables of equal length(50 m)to receiver huts(shipping containers that have been modified with radio frequency shielding and liquid cooling). One receiver hut serves two cylinders and is placed between them. Within each of the two receiver huts, signals from 1024 inputs are further amplified in a stage that includes a bandpass filter and are digitized and split into 1024 frequency channels by the custom F-Engine electronics. Temperatures inside the receiver huts are controlled with a liquid-cooled system, but no attempt is made to control the temperature of focal-line components.
Upgraded CHIME Correlator
The CHIME correlator was originally designed to handle the2048 inputs from the CHIME antennas to map the CHIME-visible sky in red shifted 21 cm emission. The CHIME correlator, of hybrid FX design, uses custom FPGA boards to digitize and channelize the data(the“F-Engine”), while a GPU cluster provides the spatial correlation(the“X-Engine”).The CHIME correlator was designed to be capable of recording visibilities across 400 MHz of bandwidth divided into 1024 frequency channels at~20 s cadence, sufficient for BAO mapping and radio frequency interference(RFI)excision.
FRB detection, however, demands significantly higher time and frequency resolution, as well as spatially localized sky beams, albeit with less stringent calibration requirements. First,typical FRBs have durations of at most a few milliseconds, so FRB surveys require at least comparable time resolution. Additionally, FRBs are dispersed by free electrons along the line of sight; without correction by dedispersion, their signals are rendered undetectable at the CHIME operating frequencies. Dedispersion of channelized intensity data, and hence detection, demands high-frequency resolution but leaves a residual dispersive smearing within each frequency channel. Minimizing this intrachannel smearing requires very narrow frequency channels, especially at CHIME’s low operating frequencies. Figure 9 shows the predicted intrachannel dispersive smearing time versus DM for a variety of FRB surveys. The CHIME/FRB project has opted for 1 ms cadence and 16,384(hereafter16k)frequency channels(see Table1)in order to minimize dispersion smearing in the most relevant part of phase space. The baseline CHIME correlator required upgrades to provide the independent high-cadence, high frequency resolution, and spatially discrete data streams necessary for FRB detection. The raw data output rate from the CHIME F-Engine is 6.5 Tbit/ s,making it difficult to duplicate or distribute beyond the X-Engine. The additional processing must therefore take place inside this system. The next section is a description of the correlator system,with emphasis on the modifications required for FRB detection.
Figure 8: Schematic of the CHIME telescope signal path. The four cylinders (black arcs), the correlator (F- and X-Engines), and the back-end science instruments are shown. The dashed orange segments depict analog-signal-carrying coaxial cables from the 256 feeds on each cylinder to the F-Engines in the corresponding East or West receiver huts beneath the cylinders. The black segments depict digital data carried through copper and fiber cables. Networking devices are not shown. The X-Engine is housed in two shipping containers(labeled North and South)adjacent to the cylinders. The HI intensity map making(CHIME/Cosmology)and CHIME/Pulsar back ends are housed in a shielded room in the DRAO building, and the CHIME/FRB back end(hatched red)is in a third shipping container adjacent to the cylinders(Figure 7). Note that the total input data rate into the F-Engine is 13 Tbit/s. The data rate into the CHIME/FRB back end is 142 Gbit/s (image credit: CHIME Collaboration)
Figure 9: Intrachannel dispersive smearing timescale as a function of DM for a variety of FRB searches, including CHIME/FRB. The black dotted line would be the CHIME/FRB smearing if the native CHIME frequency resolution(1024 channels)were used. Measured widths at 1.4 GHz of detected FRBs (color-coded by detection survey) are indicated. Observed FRBs having widths far below the black dotted line would be artificially broadened to the black dotted line at the native resolution, rendering them difficult if not impossible to detect. To mitigate this problem, rechannelization to 16k frequency channels is performed.The smearing for the current pipeline having 16k frequency channels is shown by the solid black line. FRBs with widths that lie above the CHIME/FRB line would be detected by the system with no flux degradation. FRBs with widths below the line may be detected, but they would suffer decreased significance. The colored lines are the intrachannel smearing of other FRB surveys, as labeled (image credit: CHIME Collaboration)
Digitizer, F-Engine, and Corner-turn
The custom electronics that perform the digitization,frequency channelization, and“corner-turn” are housed in two 20 ft steel shipping containers outfitted with radio-frequency-shielded enclosures that provide>100 dB of shielding of the focal line from the high-speed electronics within. These enclosures(labeled the East and West receiver huts)are located midway between the first and second cylinders and the third and fourth cylinders, halfway along their lengths,to minimize and equalize the coaxial cable lengths that connect to the feed lines.
The F-Engine system consists of 128“ICE”motherboards(Bandura et al.2016a)housed in eight rack-mounted crates and interconnected with custom high-speed, full-mesh backplanes. The system is described in detail by Bandura et al.(2016b).Briefly, 16 amplified and filtered analog sky signals are digitized in the second Nyquist zone on daughter cards attached to each motherboard at a data rate of 800 MHz with 8-bit accuracy. This information is transmitted to a Field Programmable Gate Array (FPGA) located on each motherboard. The total data rate digitized by the F-Engine is 13.1 Tbit/s for the2048 time streams. The 400 MHz bandwidth of each time stream is channelized into 1024 frequency bins, each 390 kHz wide, using a polyphase filter bank. A programmable gain and phase offset are applied to each frequency channel, and the data are rounded to 4+4 bit complex numbers, providing a totaldata rate of 6.5 Tbit/s.
The“corner-turn”modules in the F-Engine reorganize the channelized data from all the motherboards in a crate in order to concentrate the data for a subset of frequencies into a single FPGA. A second corner-turn module reorganizes the data between a pair of crates located in the same enclosure before offloading the data to the X-Engine GPU nodes through 1024 optical high-speed transceivers using standard 10 Gbit/s ethernet protocol. Data from all eight ICE motherboard crates are assembled in each GPU node,forming the last stage of the corner-turn. At this point, data for four frequency bins originating from all 2048 digitizers are assembled in a single GPU node, ready to be spatially correlated.
The X-Engine consists of 256 processing nodes, eachreceiving 25.6 Gbit/s of frequency-channelized data on4 x 10 Gbit E ports. An Intel Xeon E5-2620v3 CPU handles data transfer, using the Intel Data Plane Development Kit(DPDK)to reliably achieve high throughput. Each node has64 GB of DDR4 RAM, sufficient to buffer resampled data for up to 20 s; an upgrade to at least 96 GB(for a 31 s buffer)is planned for the near future(see Section5.3). The nodes occupy custom 4U rack-mount chassis and have no persistent local storage, but instead, a local set of file servers boot the nodes over the network.
Each node processes four input frequency channels of~390 kHz bandwidth each. These are processed by two dual-chip AMD FirePro S9300x2 GPUs, with each chip operating independently to perform all spatial and in-band processing on a single frequency channel. The processed data, including there sampled stream for FRB searching, are exported to the back ends over a pair of Gbit E links.
The X-Engine is entirely liquid cooled, using direct-to-chip cooling on the CPUs and GPUs and coupling the coolant to the ambient outside air with a 3 x 120 mm radiator in the front of each node, without any active chilling. Each rack operates a sealed and independent coolant loop, which is coupled to externally circulating coolant through a CoolIT CHx40 liquid-handling unit. The external coolant exhausts heat to the outside air through a high-capacity dry cooler. The X-Engine is housed in two 40 ft shipping containers directly adjacent to the cylinders(see Figure 7). To limit self-generated interference,a custom-built Faraday cage in each container provides>100 dB of shielding from 1 MHz to 10 GHz.
L0: Beam-forming and Frequency Channelization
In order to take full advantage of the large instantaneous FOV of CHIME while maintaining the full sensitivity, we have developed a hybrid beam-forming pipeline(hereafter referred to as the“Level-0”or L0 process; see Figure4)to be employed in the X-Engine correlator. Details of this pipeline are described in Ng et al.(2017). In summary, we synthesize256 formed beams via a fast Fourier transform(FFT)algorithm along the north–south direction. The formalism of FFT beam-forming in the context of a radio interferometer can be found in, e.g., Tegmark & Zaldarriaga(2009)or Masui et al.(2017).We zero-fill the FFT by a factor of two and resample the result to improve spatial alignment of the 256 formed beams across the 400 MHz observing bandwidth. The final N–S beams are tiled across a runtime-defined range of angles(nominally ~110°), evenly spaced in sin θ space, where θ is the zenith angle. In the east–west direction, we form four beams via exact phasing. These give a total of 1024 discrete formed static beams that are closely spaced—with the exact spacing a tunable parameter—and tile the entire primary beam continuously. For beams spaced evenly in sin θ from θ=-60°to θ=+60°, beam centers are separated by 0.4º, to be compared with the beam FWHM of 0.5º at 400 MHz and 0.25º at 800 MHz. Thus, there is significant beam overlap at the lower frequencies.(For reference, forθ=-90°to θ=+90°, which is not obviously desirable in spite of the larger FOV because of primary beam fall-off and occultation by nearby mountains, the beam separation would be 0°.46.)Simulations show that this planned beam spacing results in detections in 6–10 beams for bright(several janskys or more)events that are near the center of a beam, and typically just one for faint events. For bright sources in between beams, as many as 12 beams may detect it.
To increase spectral resolution, 128 successive 2.56 µs voltage samples are collected and Fourier-transformed. The square of the magnitude of this spectrum is downsampled infrequency by a factor of 8. Three successive downsampled transforms are averaged together, and the two orthogonal polarizations are summed, producing a Stokes I data stream with a cadence of 0.983 ms and a spectral resolution of 24.4 kHz. The output data thus consist of 1024 total intensity beams, with 16k frequency channels at ~1 ms cadence. These data are scaled, offset, and packed into 8-bit integers for transmission. In principle, it is possible to perform coherent dedispersion to a selected DM on phase data in the X-Engine;however, this feature has yet to be implemented.
CHIME/FRB Instrument and Software Pipeline
The CHIME/FRB instrument is the system built to search for FRBs in real time, after receiving the 16k frequency channels at 1 ms cadence for the 1024 CHIME telescope beams from the upgraded X-Engine correlator. The pipeline is split into four further stages or levels, named L1 through L4.Figure 10 schematically represents the different components of the pipeline and the flow of data. Each correlator node calculates the intensities for four frequency channels for all beams, and each L1 node runs FRB searches on the full frequency data for eight independent beams. Thus, the data from the L0 nodes to the L1 nodes are cross-distributed across the clusters via the networking described in Section 4.2. L1 performs per-beam RFI rejection and dedispersion using a highly optimized tree algorithm(K. M. Smith et al. 2018, in preparation), identifying candidate events in the DM/time plane. L1 processing is performed on each of the 1024 formed beams by a dedicated cluster of 128 compute nodes, and candidate events are consolidated at L2. L2 groups events seen simultaneously in different beams at the same DM and improves localization based on the strength of the signal in multiple beams. L3 classifies the detection and selects among different actions—including alerting the community within a few seconds of the detection—based on source properties. L4performs the selected actions and hosts a database that stores astrophysical events, including individual pulses from radio pulsars and RRATs, for further offline analysis. The L0 and L1stages include buffers so that baseband(i.e., voltage)and intensity data, respectively, can be retrieved upon request and further analyzed offline when an event is detected.
For further details, the reader is referred to Ref. 8).
CHIME mission events and status
• June 17, 2020: A team of astronomers, including researchers at MIT, has picked up on a curious, repeating rhythm of fast radio bursts emanating from an unknown source outside our galaxy, 500 million light years away. 9)
- FRBs (Fast Radio Bursts) are short, intense flashes of radio waves that are thought to be the product of small, distant, extremely dense objects, though exactly what those objects might be is a longstanding mystery in astrophysics. FRBs typically last a few milliseconds, during which time they can outshine entire galaxies.
- Since the first FRB was observed in 2007, astronomers have catalogued over 100 fast radio bursts from distant sources scattered across the universe, outside our own galaxy. For the most part, these detections were one-offs, flashing briefly before disappearing entirely. In a handful of instances, astronomers observed fast radio bursts multiple times from the same source, though with no discernible pattern.
- This new FRB source, which the team has catalogued as FRB 180916.J0158+65, is the first to produce a periodic, or cyclical pattern of fast radio bursts. The pattern begins with a noisy, four-day window, during which the source emits random bursts of radio waves, followed by a 12-day period of radio silence.
- The astronomers observed that this 16-day pattern of fast radio bursts reoccurred consistently over 500 days of observations.
- “This FRB we’re reporting now is like clockwork,” says Kiyoshi Masui, assistant professor of physics in MIT’s Kavli Institute for Astrophysics and Space Research. “It’s the most definitive pattern we’ve seen from one of these sources. And it’s a big clue that we can use to start hunting down the physics of what’s causing these bright flashes, which nobody really understands.”
- Masui is a member of the CHIME/FRB collaboration, a group of more than 50 scientists led by the University of British Columbia, McGill University, University of Toronto, and the National Research Council of Canada, that operates and analyzes the data from the Canadian Hydrogen Intensity Mapping Experiment, or CHIME, a radio telescope in British Columbia that was the first to pick up signals of the new periodic FRB source.
- The CHIME/FRB Collaboration has published the details of the new observation today in the journal Nature. 10)
• February 7, 2020: A periodic flurry of radio waves from some unknown object in deep space could help astronomers figure out what’s triggering similar radio bursts in other galaxies. 11)
- Since 2007, researchers have cataloged over 100 fast radio bursts, or FRBs, coming from every direction in the sky. But it’s unknown what causes these radio bursts. Only 10 have been seen to repeat (SN: 8/14/19), and none of those had exhibited any sort of steady tempo — until now.
- One of the known repeaters has a relatively brief window of activity about every 16 days, researchers report January 28 at arXiv.org. That means something about the source or its environment is reliably controlling the burst activity, a potential clue to the true nature of these enigmatic objects.
- Dongzi Li, an astrophysicist at the University of Toronto, and colleagues found the pattern in data from the CHIME radio telescope in British Columbia. They determined that the FRB blasts out about one to two radio bursts per hour for four days and then goes silent for just over 12 days before usually repeating the cycle.
- “This is very significant,” says Duncan Lorimer, an astrophysicist at West Virginia University in Morgantown and co-discoverer of the first FRB (SN: 7/25/14). “It’s potentially going to take us in an interesting direction to get to the bottom of these repeaters.”
- One possible explanation for the periodicity is that the FRB is orbiting something else, perhaps a star or black hole. In that case, the 16-day period might reveal how often the source of the radio waves is pointed toward Earth.
- Alternatively, stellar winds from a companion might periodically boost or block the radio pulses. Winds might also explain why not every 16-day cycle produces bursts: If the companion occasionally belches out more material than usual, that could mask the FRB’s signal.
- Either explanation implies that repeating FRBs — or at least, this one — might come paired with something else.
- Li and her colleagues aren’t ready to rule out stand-alone objects, where the 16-day period might arise from the FRB rotating or wobbling. But that scenario is a bit tougher to reconcile with the data. For example, one popular FRB culprit is a type of highly magnetic neutron star known as a magnetar. But known magnetars in our galaxy spin around once every 12 seconds or less, the team notes, a far cry from the fortnight needed for this FRB.
- This particular radio burst was also recently traced to a star-forming region in a spiral galaxy nearly 500 million light-years away from Earth (SN: 1/6/20). Future scans of its home with telescopes sensitive to other electromagnetic radiation, such as X-rays or gamma rays, might whittle down the list of suspects and move astronomers closer to solving this cosmic mystery.
- There’s also hope that this find is just the first of many periodic FRBs to be detected. “There’s nothing particularly special about this repeater,” Lorimer says. “The fact that they detected periodicity on this one hints that other ones will have periodicity as well.”
• January 9, 2019: A Canadian-led team of scientists has found the second repeating FRB (Fast Radio Burst) ever recorded. FRBs are short bursts of radio waves coming from far outside our Milky Way galaxy. Scientists believe FRBs emanate from powerful astrophysical phenomena billions of light years away. 12)
- The discovery of the extragalactic signal is among the first, eagerly awaited results from the Canadian Hydrogen Intensity Mapping Experiment (CHIME), a revolutionary radio telescope inaugurated in late 2017 by a collaboration of scientists from the University of British Columbia, McGill University, University of Toronto, Perimeter Institute for Theoretical Physics, and the National Research Council of Canada.
- In a resounding endorsement of the novel telescope’s capabilities, the repeating FRB was one of a total of 13 bursts detected over a period of just three weeks during the summer of 2018, while CHIME was in its pre-commissioning phase and running at only a fraction of its full capacity. Additional bursts from the repeating FRB were detected in the following weeks by the telescope, which is located in British Columbia’s Okanagan Valley.
Discovery of second repeating FRB suggests more exist
- Of the more than 60 FRBs observed to date, repeating bursts from a single source had been found only once before – a discovery made by the Arecibo radio telescope in Puerto Rico in 2015.
- “Until now, there was only one known repeating FRB. Knowing that there is another suggests that there could be more out there. And with more repeaters and more sources available for study, we may be able to understand these cosmic puzzles—where they’re from and what causes them,” said Ingrid Stairs, a member of the CHIME team and an astrophysicist at UBC (University of British Columbia).
- Before CHIME began to gather data, some scientists wondered if the range of radio frequencies the telescope had been designed to detect would be too low to pick up fast radio bursts. Most of the FRBs previously detected had been found at frequencies near 1400 MHz, well above the Canadian telescope’s range of 400 MHz to 800 MHz.
- The CHIME team’s results – published January 9 in two papers in Nature and presented the same day at the American Astronomical Society meeting in Seattle – settled these doubts, with the majority of the 13 bursts being recorded well down to the lowest frequencies in CHIME’s range. In some of the 13 cases, the signal at the lower end of the band was so bright that it seems likely other FRBs will be detected at frequencies even lower than CHIME’s minimum of 400 MHz. 13) 14)
FRB sources likely to be in ‘special places’ within galaxies
- The majority of the 13 FRBs detected showed signs of “scattering,” a phenomenon that reveals information about the environment surrounding a source of radio waves. The amount of scattering observed by the CHIME team led them to conclude that the sources of FRBs are powerful astrophysical objects more likely to be in locations with special characteristics.
- “That could mean in some sort of dense clump like a supernova remnant,” says team member Cherry Ng, an astronomer at the University of Toronto. “Or near the central black hole in a galaxy. But it has to be in some special place to give us all the scattering that we see.”
A new clue to the puzzle
- Ever since FRBs were first detected, scientists have been piecing together the signals’ observed characteristics to come up with models that might explain the sources of the mysterious bursts and provide some idea of the environments in which they occur. The detection by CHIME of FRBs at lower frequencies means some of these theories will need to be reconsidered.
- “Whatever the source of these radio waves is, it’s interesting to see how wide a range of frequencies it can produce. There are some models where intrinsically the source can’t produce anything below a certain frequency,” says team member Arun Naidu of McGill University.
- “[We now know] the sources can produce low-frequency radio waves and those low-frequency waves can escape their environment, and are not too scattered to be detected by the time they reach the Earth. That tells us something about the environments and the sources. We haven’t solved the problem, but it’s several more pieces in the puzzle,” says Tom Landecker, a CHIME team member from the National Research Council of Canada.
1) Davide Castelvecchi, ”‘Half-pipe’ telescope will probe dark energy in teen Universe,” Nature News, Volume 523, pp: 514-515, 29 July 2015, URL: https://www.nature.com/news/half-pipe-telescope-will-probe-dark-energy-in-teen-universe-1.18088
3) Tzu-Ching Chang, Ue-Li Pen, Kevin Bandura & Jeffrey B. Peterson, ”An intensity map of hydrogen 21-cm emission at redshift z ~ 0.8,” Nature, Volume 466, pp: 463–465(2010), Published: 22 July 2010, https://doi.org/10.1038/nature09187
4) ”Canada's CHIME telescope detects second repeating fast radio burst,” McGill University, 9 January 2019, URL: https://phys.org/news/2019-01-canada-chime-telescope-fast-radio.html
5) ”Canadian ingenuity crafts game-changing technology for CHIME telescope,” McGill Newsroom, 7 September 2017, URL: https://www.mcgill.ca/newsroom/article/canadian-ingenuity-crafts-game-changing-technology-chime-telescope
6) ”New Canadian telescope will map largest volume of space ever surveyed,” McGill University, Press Release, 7 September 2017, URL: https://www.mcgill.ca/newsroom/channels/news/new-canadian-telescope-will-map-largest-volume-space-ever-surveyed-270107
7) A. Liu, S. Foreman, H. Padmanabhan, H. C. Chiang, S. Siegel, D. Wulf, J. Sievers, M. Dobbs, K. Vanderlinde, ”Low-redshift 21cm Cosmology in Canada,” Submitted to LRP2020 panel on 30th September 2019, URL: https://arxiv.org/pdf/1910.02889.pdf
8) M. Amiri, K. Bandura, P. Berger, M. Bhardwaj, M. M. Boyce, P. J. Boyle, C. Brar, M. Burhanpurkar,P. Chawla, J. Chowdhury, J.-F. Cliche, M. D. Cranmer, D. Cubranic, M. Deng, N. Denman, M. Dobbs,M. Fandino, E. Fonseca, B. M. Gaensler, U. Giri, A. J. Gilbert, D. C. Good, S. Guliani, M. Halpern, G. Hinshaw,C. Höfer, A. Josephy, V. M. Kaspi, T. L. Landecker, D. Lang, H. Liao, K. W. Masui, J. Mena-Parra,A. Naidu, L. B. Newburgh,C.Ng, C. Patel, U.-L. Pen, T. Pinsonneault-Marotte, Z. Pleunis,M.Rafiei Ravandi,S. M. Ransom, A. Renard, P. Scholz, K. Sigurdson, S. R. Siegel, K. M. Smith, I. H. Stairs, S. P. Tendulkar,K. Vanderlinde, and D. V. Wiebe, ”The CHIME Fast Radio Burst Project: System Overview,” The Astrophysical Journal, Volume 863:48(16pp), 10 August 2018, https://doi.org/10.3847/1538-4357/aad188, URL: https://iopscience.iop.org/article/10.3847/1538-4357/aad188/pdf
9) Jennifer Chu, ”Astronomers detect regular rhythm of radio waves, with origins unknown,” MIT News, 17 June 2020: http://news.mit.edu/2020/astronomers-rhythm-radio-waves-0617
11) Christopher Crockett, ”This is the first fast radio burst known to have a steady beat,” Science News, 7 February 2020, URL: https://www.sciencenews.org/article/first-fast-radio-burst-steady-beat
12) ”Canada's CHIME telescope detects second repeating fast radio burst,” McGill Newsroom, 9 January 2019, URL: https://www.mcgill.ca/newsroom/channels/news/canadas-chime-telescope-detects-second-repeating-fast-radio-burst-292994
13) The CHIME/FRB Collaboration: M. Amiri, K. Bandura, M. Bhardwaj, P. Boubel, M. M. Boyce, P. J. Boyle, C. Brar, M. Burhanpurkar, P. Chawla, J. F. Cliche, D. Cubranic, M. Deng, N. Denman, M. Dobbs, M. Fandino, E. Fonseca, B. M. Gaensler, A. J. Gilbert, U. Giri, D. C. Good, M. Halpern, D. Hanna, A. S. Hill, G. Hinshaw, C. Höfer, A. Josephy, V. M. Kaspi, T. L. Landecker, D. A. Lang, K. W. Masui, R. Mckinven, J. Mena-Parra, M. Merryfield, N. Milutinovic, C. Moatti, A. Naidu, L. B. Newburgh, C. Ng, C. Patel, U. Pen, T. Pinsonneault-Marotte, Z. Pleunis, M. Rafiei-Ravandi, S. M. Ransom, A. Renard, P. Scholz, J. R. Shaw, S. R. Siegel, K. M. Smith, I. H. Stairs, S. P. Tendulkar, I. Tretyakov, K. Vanderlinde & P. Yadav, ”Observations of fast radio bursts at frequencies down to 400 megahertz,” Nature Letter, Volume 566, pp: 230–234(2019), https://doi.org/10.1038/s41586-018-0867-7, Published: 09 January 2019.
The information compiled and edited in this article was provided by Herbert J. Kramer from his documentation of: ”Observation of the Earth and Its Environment: Survey of Missions and Sensors” (Springer Verlag) as well as many other sources after the publication of the 4th edition in 2002. - Comments and corrections to this article are always welcome for further updates (firstname.lastname@example.org).