Submillimetre instruments use heterodyne techniques to observe spectral lines at high resolution. Heterodyne receivers take a weak incoming signal and mix it with a reference frequency known as the local oscillator (LO). This results in a signal of lower (intermediate) frequency (IF) which can be amplified and more easily sampled. The current JCMT receivers have an IF of 4 GHz (RxA) or 5 GHz (HARP). This mixing produces two frequencies for each value of the LO () which are known as the upper sideband (USB) and lower sideband (LSB).
In a single-sideband receiver one of these is suppressed. Double-sideband receivers have the two frequencies superimposed.
Reducing the noise in the system is crucial to the efficiency of these receivers. The required integration time is directly proportional to the total noise in the system given by via the radiometer formula
where the terms are as follows:
The heterodyne instruments described below are those currently operational on the JCMT. For information on previous instruments see the JCMT web pages.
In the following is the main-beam efficiency and is the forward spillover and scattering efficiency. For the latest and historical calibrations visit ACSIS Beam Efficiencies.
Receiver A (RxA) is a single-element double-sideband receiver. It operates in the A-band between 211 and 276 GHz. In addition to stare observations, the most common observing mode for RxA is raster mapping.
The Heterodyne Receiver Array Programme (HARP) is a 16-receptor array arranged on a 44 grid . It operates in the B-band between 325 and 375 GHz. The receptors are adjacent on the focal plane but separated on the sky by 30 arcseconds (or approximately twice the beamwidth). The footprint of the full array is 2 arcminutes. Note that not all receptors are operational; historically between 12 and 15 receptors have been available at any given time. The tracking receptor is chosen to be the best performing of the four internal (22) receptors in the grid.
You can find the number of the tracking receptor and the number of working receptors in the FITS header
In addition to stare observations with the tracking receptor, the most common observing modes for HARP are jiggles and raster maps. These observing modes are described in Section 3.3 and an example image from each mode is shown in Figure 3.2.
Since 2006 September, the backend for heterodyne instruments has been the Auto-Correlation Spectrometer and Imaging System known simply as ACSIS (see Figure 3.1). Given the various combinations of ACSIS’s 32 down-converter modules and available IF outputs from the different instruments, a number of bandwidth modes and corresponding frequency resolutions are available. Listed below are the most common ones. Visit the JCMT web pages for a full list.
|Subbands||Bandwidth mode||Channel Spacing|
|1||250 MHz||0.0305 MHz (HARP/RxA/RxW)|
|1||1000 MHz||0.488 MHz (HARP/RxA/RxW)|
|1||1860 MHz (1600,1800)||0.977 MHz (HARP)|
|0.488 MHz (RxA/RxW)|
|1||440 MHz (400,420)||0.061 MHz (HARP)|
|0.0305 MHz (RxA/RxW)|
|2||any 250 MHz||0.061 MHz (HARP)|
|0.0305 MHz (RxA/RxW)|
|2||any 1000 MHz||0.977 MHz (HARP)|
|0.488 MHz (RxA/RxW)|
There are a number of special configurations available with ACSIS. These make use of multiple sub-bands to
cover lines at different frequencies. You can find the subsystem number for a given observation by the FITS
You will find each subsystem number corresponds to a different molecule or transition
(given by the FITS headers
TRANSITI). The most common pairings are
CO (3-2)/CO (3-2)
and DCN (5-4)/HCN (4-3) for HARP, and
SiO (6-5)/SO (4-3)
|Stare||A stare, or sample, observation is as simple as it sounds. When you open reduced cubes of stare observations in Gaia you will see a map consisting of one pixel for each receptor (just one in the case of RxA). The left-hand panel of Figure 3.2 shows a stare observation opened in Gaia. You can also view the central spectrum using linplot (see Appendix F) or Splat.|
|Jiggle||A jiggle map is a common HARP observing mode. They are designed to fill in the 30′′spacing
between the HARP receptors resulting in a 2′2
map. This is achieved by ’jiggling’ the secondary mirror in a 44
pattern to give a HARP4 or HARP5 map respectively. The HARP4 pattern has a 7.5
pixels while the HARP5 has 6′′pixels, slightly under- and over-sampling compared with Nyquist.
The central panel of Figure 3.2 shows a HARP5 jiggle pattern.
Each of these patterns can be pixel-centered, where the target co-ordinates fall on one of the central four pixels, or map-centered where the target co-ordinates lie at the centre of the map between the four central pixels. See Figure 3.3 for an illustration of the different jiggle patterns. Although designed for HARP this mode can be used with the single-element receivers.
|Raster||Raster mapping, or scan mapping, is designed for mapping large areas with maximum efficiency. The telescope continuously takes data while scanning across the source. When HARP is used for raster mapping the array is rotated 14.°04 to the direction of the scan in order to give a fully sampled map with 7.3 pixels. This will result in jagged edges to your reduced map which you may wish to trim. It is common when using HARP to repeat the map scanning along the other axis to give a ‘basket-weave’ pair of maps. Each one of the pair will appear as an individual observation. The right-hand panel of Figure 3.2 shows a HARP raster map.|
In position-switch (PSSW) mode, the whole telescope moves off source and on to the reference position. This allows a large offset to the reference position, useful in crowded environments such as the Galactic Plane. The main disadvantage is that it takes a longer time resulting in less-accurate sky subtraction or non-flat baselines.
|Beam-switch||In beam-switch (BSW) mode, the secondary mirror chops onto the reference position. This results in fast and accurate sky subtraction but it is limited to targets which have a blank reference position within 180′′(the maximum chop distance of the SMU).|
|Frequency-switch||In frequency-switch (FSW) mode the telescope does not move but the centre of the spectral band is shifted, typically by 8 or 16 Mhz, to a line-free region of the spectrum. It is very efficient with no off-source time and can be useful when no emission-free reference position can be found.|
Subsystems are normally arranged to cover different spectral lines and would be processed independently. However, there is also a mode in which the subsystems overlap in frequency, called hybrid. Such data are combined to yield a broader spectral coverage, say where there multiple lines in proximity. The number of subsystems is a power of two up to a maximum of four.