August 11, 2014 | 1
Before they can see Sagittarius A*, the black hole at the center of the Milky Way, the astronomers of the Event Horizon Telescope (EHT) must complete an epic to-do list. The most important item on that list: Bring the Atacama Large Millimeter Array (ALMA) into the group.
It’s easy to see why. After all, ALMA is by far the world’s most powerful radio telescope. But there is a second, subtler reason why ALMA is essential for the EHT. ALMA is located right in the middle of the planet-spanning array the EHT is assembling—between Hawaii and Mexico and the South Pole, creating long north-south baselines stretching across the equator. (Click here for a basic explanation of how the EHT works.)
The addition of ALMA (plus Mexico’s Large Millimeter Telescope and the South Pole Telescope) should make it possible for the EHT to directly observe Sagittarius A*’s “shadow,” a phenomenon predicted by General Relativity. That’s a big payoff, and adding ALMA is an appropriately big job. To join EHT observations, ALMA needs to be able to function as a phased array. Roughly, that means all of ALMA’s 50-plus dishes will together act like a single dish. This, alas, is not something ALMA was originally designed to. Endowing the telescope with this new capability has required an effort that Shep Doeleman, lead investigator for the EHT, likens to “giving ALMA a heart transplant.”
An international group of researchers has been working on the Alma Phasing Project (APP) for several years. The work is almost finished, and if all goes well, ALMA could join EHT observations next year. Michael Hecht at MIT’s Haystack Observatory oversees the APP. I called him to get a status update and a basic explanation of what the project involves. An edited transcript of our conversation follows.
What is the Alma Phasing Project?
The way ALMA works is you have a set of 50-odd dishes sitting on a mountaintop looking at the sky and acting like a camera. [Note: It's been pointed out that ALMA already has 66 dishes at the high site. -SF.] It’s a connected-element array, meaning it takes the signal from all of those dishes, combines them a certain way, and a picture comes out. We want to use it differently.
Overall, the Event Horizon Telescope (EHT) works on the same principle as ALMA, except our dishes are spread all over the planet. Think of ALMA as a telescope the size of a mountaintop and the EHT as a telescope the size of the earth. Each element of the EHT has to be one dish. The problem is ALMA isn’t one dish—it’s 50 dishes, each its own camera. To have ALMA participate as one more element in the EHT—one more pixel, if you will—you have to combine all those signals differently. They have to all be in phase. Each one becomes indistinguishable. You’re collecting more signal from the same source, adding them together so that they look like a stronger signal from the same source.
To do that we have to get into the guts of the ALMA correlator and change the way the signals are combined. Normally, what comes out of the correlator is a small amount of data that is converted into a picture. We don’t want that final picture—we want the combined raw data. We’ll take that and the equivalent data from the other sites around the world, bring it all to a central site—MIT Haystack Observatory—and combine them to make a picture.
Getting this done has required an extremely broad collaboration at every level. The National Radio Astronomy Observatory, the National Science Foundation and the Gordon and Betty Moore Foundation are all supporting the project. We have international partners at the Max Planck Institute for Radio Astronomy, the University of Concepcion in Chile, the Academia Sinica Institute of Astronomy and Astrophysics in Taiwan (ASIAA), the National Astronomical Observatory of Japan (NAOJ). It’s been an enormous cooperative effort.
How much of a departure from ALMA’s current design is the phased array?
ALMA’s design was left with the hooks that would make it possible for someone to do this in the future. We’re taking advantage of the fact that we don’t have to change the fundamental structure. We just have to add to it.
What do you have to add?
At ALMA there’s the Array Operations Site, or AOS—the “high site,” which has all the dishes and the correlator—and the Operation Support Facility, or OFS, which is the low site. At the low site we’re installing new digital recorders, which were contributed by ASIAA. There’s the hydrogen maser, which is to become the new frequency standard for all of ALMA. A new optical fiber link, developed by NAOJ, will convert the mountain of data from the high site, code it as a bunch of different colors, then send it to the low site where it’s demultiplxed and recorded. The big thing is the modification to the correlator itself—a set of eight electronic boards we call PICs, for Phasing Interface Cards.
What do the PICs do?
The PICs intercept all the data coming through correlator, determine the phase offset between different dishes, apply corrections, then send the data down to the low site. Those were developed at NRAO by the same team that developed the ALMA correlator to begin with. That was an important part of getting permission to do this.
How much more data will ALMA put out when it’s operating as a phased array?
Orders of magnitude more data than normal. It’s equivalent to the raw data from one of the dishes. It’ll easily add up to a petabyte.
You guys have been installing hardware at ALMA for several months now, right? How much equipment is in place?
All the equipment is in place. The PICS are installed. The maser is up and running, and initial tests have been performed showing it can run without disturbing normal ALMA operations. And the optical fiber link is installed and running. There are a few bugs here and there, but nothing serious. Only thing remaining is to debug some of the firmware. That could take a couple weeks, maybe a month.
If you’ve ever watched a rocket launch, a lot of the time they get down to 40 seconds and suddenly there’s a “scheduled hold”—that’s a good analogy. Everything’s done. We’re just sitting back and fixing these bugs. Just have to let the guys go through the punch list.
Give a 1 year subscription as low as $14.99X