Building LOFAR-Sweden

While in the UK we've completed our LOFAR station and can retreat inside our homes when the icy weather comes, in Sweden, they have just started the construction of their station, at Onsala. The image below is a great snapshot taken today of the preparatory work underway amidst the snow and ice (credit John Conway).

Sweden will be the final one of the 'founding' members (NL, DE, France, UK, Sweden) of the International LOFAR Telescope to construct their station, and we expect full operation of the entire ~40-station array in 2011.

In the meantime, our station at Chilbolton has been taking some observations of the Sun, and should be connected to the Dutch supercomputer/correlator in the very near future, allowing the first high-resolution images utilising the British baseline. More posts coming soon !

Turbo-fans

Here's a rare peek inside one of the signal processing servers. Note the two forced-air heat-sinks on the motherboard. This computer is one of four servers that will be used to control the array of GPU-processors which, in turn, will perform the de-dispersion calculations (see Heavy Duty Computing).

And, yes... when they are running all of these machines make a LOT of noise!

Cable Calibration

In order to form the best possible result out of the LOFAR system, it is essential that we can calibrate for the signal travel-times. This means not only being able to know the direction from whence the signal came, but also how long it takes to travel through every cable and every circuit.

Each cable within the system is carefully manufactured and labelled. The serial number on the cable contains information about its performance and role. For example, the one in the photograph (CHY-0853425-1305090755) is an "C"able from the "H"igh-Band Array, "Y"-polarisation. It is "085" metres long and has a delay of "3425" tenths of a nanosecond. This means that, in laboratory conditions, it takes 342.5 nanoseconds for a signal to travel down this cable.

However, that is in the lab.

Although this is a pretty good measurement, to really get the best performance we need to know the signal delay in the field, now that the cable is in place. Things such as connectors, thermal variations and other deployment factors can contribute to subtle variations in the effective cable length. Calibrating these sorts of things is part of the vital commissioning work that is currently being carried out on the LOFAR Chilbolton station.

Shut the door!

Inside the RF-container lies all the electronics that does the signal processing. Because there is so much of it, and it is so close to the antenna array, it is crucial that it is shielded to prevent interfering signals from escaping from these electronics to feedback into the system.

Now that we've started taking routine calibration and test data, it is important to keep the RF-contain sealed against unwanted radio leaks.

The notice on the door reads: "INTERFERENCE HAZARD - The equipment in this RF-container can disrupt telescope observations. Do not open this door without first checking with the Operations Group."

Heavy Duty Computing

With the main antenna arrays now complete, we are now concentrating on getting the data processing equipment set up. One of the major components of this is the computer system that will perform the de-dispersion calculations. Dispersion is a physical phenomenon where the signals at different frequencies travel at different speeds through a medium. In our case, lower-frequency radio waves travel through the interstellar medium slower than higher-frequency radio waves.

In order to add signals together (for example to analyse the pulses from a pulsar), one must first apply a correction for the dispersion that occurs. Although it is possible to estimate what the dispersion will be, based on our knowledge of the Galaxy, ultimately one has to simply try lots of different dispersion measures.

That means a lot of calculations, which means we need some pretty high powered, specialist computing to do the work. The project to do this for LOFAR is being led by the University of Oxford on behalf of SEPnet.

Caption 1 :  Dispatch from the University of Oxford.

Caption 2 : Aris Karastergiou unloads another GPU-server from the delivery van.

LOFAR HAKU

The RF-Container at the Chilbolton LOFAR station is an adapted shipping container. Because ships and storage facilities can hold many tens of thousands of containers, it is important to be able to identify them. As a result, containers from a given manufacturer are marked with a unique 7-digit serial number (strictly speaking it is a 6-digit code plus a single-digit checksum). Additionally, the serial number is prefixed by a four-letter code to denote the company that built the container (these prefixes always end in the letter "U"). The code for the RF-container at Chilbolton is HAKU2961555.

HAKU is the prefix for containers manufactured by "Container Company Amsterdam BV".

However, HAKU has another meaning.

During the visit from the ionospheric physics group, Derek McKay-Bukowski described the LOFAR system, with its multiple beams, rapid direction-switching and de-dispersion back-end, as an "astronomical search engine". After the presentation, an e-mail was received from Thomas Ulich, asking if we knew what the "HAKU" on the side of the container meant. While we were aware of the container serial number, as explained above, what wasn't known was that "haku" is the Finnish word for "search". Thomas explained that "hakukone" means "search engine" in Finnish; "kone" is engine, and "haku" means "search, fetch, retrieval", but also "quest, trawl".

So our LOFAR container certainly has the correct name!

Ionospheric Visit

Because of the frequencies at which LOFAR observes, an understanding of the ionosphere is crucial to the interpretation of our results. On the other hand, LOFAR also offers a powerful tool for investigating this tenuous layer of the earth's atmosphere.

On 27-28 September 2010, a special meeting was organised by the UK solar-terrestrial physics community (MIST) to celebrate the life and work of the late Prof. Henry Rishbeth (1931-2010). The meeting took place in his home city of Southampton, and celebrated Henry's long commitment to the advancement of ionospheric science. And, on the afternoon of the 28th, a number of the attending physicists visited the LOFAR Chilbolton station. (Photograph by Th. Ulich.)

This post also appears on the STFC RAL website.