JAC Style guide: template
 Short Baseline Interferometry
The JCMT-CSO interferometry was split into two parts this semester, with one run in October 1995 and one
in January 1996. Although there were some patches of poor weather on both runs, we did have a reasonable
amount of time when interferometry was possible and in January there were some stretches (regrettably
rather short) of really good weather with the 230 GHz \ below 0.04 and the 'CFA seeing' at around 0.2
arcseconds. In October there were no significant technical problems and we were able to carry out a range
of programmes at 320 to 356 GHz. As on previous runs the most rewarding sources were bright protostars
which give strong fringes and often show a good deal of structure. On this occasion L1448-N and Orion-
IRC2 were the prize objects. Finding models to fit these data is proving an interesting challenge. We
detected several other ptorostars and found some more Ae-Be stars that have substantial emission from
small regions. In most of these cases the visibility did not obviously vary as the baseline projection varied
so we can only put limits on the sizes of discs or other structures present. We also observed some double-
protostar candidates and made new flux measurements of the compact object in the Galactic Centre Sag A*
(which is in too confused a region for single-dish measurements at these wavelengths).
In January the big push was to get the system working properly at 460 GHz for the first time. In addition to
the generally more challenging requirements - higher receiver noise, lower antenna efficiencies, much
higher atmospheric attenuation, and greater sensitivity to atmospheric and instrumental path length
fluctuations - there is an additional complication that the JCMT has a 4 GHz IF system at this frequency
whereas CSO's is at 1.5 GHz. We therefore had to introduce additional hardware into the system which
offsets the Gunn in the JCMT local oscillator system by 480 MHz. Since the Gunn frequency is then
multiplied by 5 this produces an offset of 2.4 GHz at the mixer, which is then removed in the IF system by
converting down from 3.9 GHz to 1.5 GHz. This has all to be accomplished without losing the phase
stability.
Figure 1a.
Figure 1b.
We checked out the system by setting up an artificial source at 460 GHz on the roof of the UKIRT control
room. We were pleased to see the fringes straight away, but unfortunately instead of one trace on the scope
there were three! This was eventually tracked down to two different 10 MHz reference oscillators fighting
each other in one of the synthesizers. With that resolved the fringe was clean but subject to nasty phase
glitches on timescales of only a few seconds. Another heavy bout of debugging finally traced this to a loose
connection on a power supply. We then turned to astronomical sources and were most gratified to see
strong fringes immediately from 3C273 (which was conveniently well above its usual 600 micron flux at
that time).
Our prime astronomical targets at this frequency were the nearby protostars HL Tau and L1551-IRS5,
which both show strong evidence at lower frequencies of having discs about 1 arcsecond across. Both were
detected on two nights over a fair range of hour angles. The first diagram shows the visibility of the fringes
as a function of projected baseline for L1551-IRS5 (Figure 1a), showing it is well resolved, and also for the
quasar 0528+134 (Figure 1b), which is of course unresolved. We hope to use these visibilities to refine our
models of the dust discs in these objects. We also found that we could see emission from the CO J = 4-3
line in the evolved star IRC+10216, which poses interesting questions given our 0.8 arcsecond fringe
spacing.
The other main technical project on the January run was to try out the prototype 183 GHz radiometer which
Martina Wiedner (an MRAO student) had built. The intention is to mount a pair of these on the two
telescopes and use them to measure the water vapour content along the paths from the source very
accurately so that we can correct the phase fluctuation on timescales down to a second or so. The first one
seemed to work well giving reasonable values for the total water and showing the fluctuations at about the
level expected.
Figure 2.
On the final night we were threatened by a large weather front moving in from the north-west. We spent
most of the time working at 346 GHz on NGC 2024 and L1641N but after these were set we tuned to the
H26I recombination line frequency (353 GHz) to look at the strong line in Eta Carina recently discovered
by Henry Matthews and colleagues. Since the Declination is below -59 degrees this was a challenge, but we
timed our attempt so that it was just coming above 10 degrees in elevation. We were amazed to measure a
system temperature of 1600K when we arrived at the source and the first 10 second integration showed
fringes and a strong line! We then tried to peak up the pointing, but the line seemed to be getting weaker
whichever way we went. Another check on the system temperature gave 4000K. The weather front had
arrived! In desperation we took a 100 second integration which is shown in Figure 2. (The lower line show
the amplitude as a function of frequency and the upper one the phase. Note that there is continuum flux as
well as the line.) Obviously the calibration is not very meaningful, but one can say that a lot of the emission
is coming from a very small region. A few minutes later the system temperature was above 10,000K, there
were clouds overhead and it was time to start taking the system apart and go home.
Richard Hills, MRAO.
Last Modification Date 1996/04/08 - Last Modification Author: Graeme Watt (gdw)
|
