Sunday, November 18, 2012

How To: Observe with APEX


Something I find exciting, interesting about a thesis on observational astronomy is the experience to learn how to use a variety of telescopes, often in stunning locations.  One purpose of this blog is to document the diversity in observing experiences, since each telescope, location, team has a story to tell.

Getting ready for observations at APEX.

Last week at APEX, circumstances obliged me to take the role of operator and astronomer at the telescope, and I had the privilege and responsibility to learn and synthesize the observing techniques and requirements.  My informal training consisted of watching over the shoulder of a staff astronomer for a couple of days at the telescope, and then following remotely via a virtual network connection to the telescope during the morning shift preceding my afternoon shift.  I took note of important commands, and planned my strategy for the afternoon.

When the time came, so did the adrenaline, and to add to the excitement of observing, I was stunned when the first command I sent to the telescope by typing "go" triggered an alarm!  In fact, it was just a warning that the telescope was beginning to move, as intended, and the operations went ahead.

Over the next few days, I learned by experience the observing routine, the quirks of the system, and methods to make the job simpler.  The most valuable lessons which will help me in the future were the general observing routine, which might also interest you, so I'll describe it here.

In the control room at the telescope. Who knows what to do now?
Observing routine
An observing routine includes several general categories, including set-up, calibrations, and science targets.

Set-up
We need to tell the telescope where in the sky we want it to point, and what frequency to observe.  To test this, we point to an object we know well, for example a bright planet like Venus, and make a few observations called pointing and focus.  We run scripts that tell the telescope to scan across venus, and calculate the corrections in each direction we need to make to get the best signal.

Calibrations 
When we observe, we are essentially measuring the intensity of light that arrives to the telescope.  To tell anything physical about our target (in our case, the molecular gas of a star forming cloud), we need calibrations to measure all of the stray light that leaks into the signal between our detector and our target region XX light years away.

For the ambient light in the sky, we observe a point in the sky that we expect to be free of any light related to our target.  This is our "off position", and we will subtract this signal from our "science" target observations.

But wait, closer to home, we need to understand the signal from our electronics and the path the light takes within our telescope.  For this we measure the temperature of the system electronics.  So in reality our detection will be (target+sky+electronics) minus (sky) minus (electronics) equals just target.

If we don't have the set-up and calibrations done right, our science observations lose their physical meaning.  Sometimes we even spend more time on the previous than the latter.

Science
Ultimately, what excites us is the detection of radiation from the "science" target that we are studying.  We can study just one point in the sky and integrate for a really long time, or we can make a map by scanning over a larger region of the sky and observing each point for a shorter amount of time.

Etc.
APEX observes radiation with wavelengths called sub-mm, which are longer than visible wavelengths but shorter than radio.  A sub-mm telescope can operate 24 hours per day, observing different sources as they rise and set all day and all night.  We can only observe regions of the sky above the horizon but not directly over head, and also away from the sun or we will fry the telescope.  We need to make a detailed schedule of what is possible to observe and when.  As we observe, a source is either rising or setting, so we can only observe a given source for a limited time, in this case we generally stayed on a project for a few hours.

Ambient weather, such as wind and precipitation can damage the telescope, and at some point I noticed a flashing red box on the control screen notifying me that the wind gusts were exceeding 20 meters per second (45 mph).  Had this wind sustained, we would have stopped observing because this grad-student-turned-antenna-operator didn't want to be held responsible for the antenna blowing away.  The telescope is designed to sustain the extreme weather conditions of the Atacama, and in fact we had no problems.

In practice
The control system has been designed to relieve many of these concerns and simplify the observing procedure, at least in theory.  Scripts made by engineers or astronomers responsible for a particular project tell the telescope the sequence of calibrations and observations.  Blinking red lights on the control screen generally mean that something is not right, and a vigilant operator can catch a problem before it manifests.

Even so, the observer might think that the telescope is happily observing, when in fact for some reason a setting is not right.  During observations, it's good practice to take a quick look at data as they arrive and verify that the observations are what we expect.  Of course, we don't know the exact signal we expect, or we wouldn't need the observations in the first place, but we have a general idea of the frequency and intensity of radiation we hope to detect, and where in the sky to point the telescope. If we don't detect anything, we can do a thorough check of the system and perhaps re-calibrate.  By understanding the entire system and environment, from detector on the telescope to source in the sky, we can be more confident in the detections that we report as science.

Somewhat unrelated, but I think these are the coolest things at 5000 meters, the "penitentes" ice formations.