|
|||
| Joint Astronomy Centre |
|
|
|
| Show document only |
IntroductionSCUBA's flux sensitivity, measured by the Flux Calibration Factor (FCF, see the Calibration Cookbook) and image characteristics both depend on the chop throw used for the observation. During every observation made with SCUBA, the secondary mirror unit (SMU) chops between the "on" position and an "off" position at a constant rate of 7.8125 Hz. Since this is a real physical movement, there is some portion of every cycle during which the SMU is at neither position but is moving between the two. This leads to a smearing out of the image along the chop axis. The effect is strongest with the largest chop throws since the constant rate means that less time per cycle is spent in the "on" and "off" positions. This leads to lower flux in the "on" position, and therefore higher FCFs. This can be seen in the sequence of images below, which shows jiggle-maps of CRL618 with three different chop throws:
(Note about blanking times: When performing beamswitching heterodyne observations with the JCMT, a "blanking time", dependent on the chop throw, is applied to the signal. This effectively cuts out the signal from the periods where the SMU is known to be travelling between "on" and "off" positions. Plans were made to do the same for SCUBA observations, however when starting to probe this issue we discovered that the blanking time was in fact never activated for SCUBA. As a result the entire signal, incuding travelling time, is included in a SCUBA image.) Hence we strongly recommend that you make calibration observations using the same chop parameters as your science observations! However, there are times when this is not possible and so we have attempted to evaluate the relationships between FCFs at different chop throws. At 850 microns this turns out to be a relatively simple task; at 450 microns the data is not so friendly! Also below is a brief discussion of the image shape characteristics at different chop throws.
850 micron FCF relationsThe FCF relations at 850 microns are relatively easy to pick out. We reduced all Uranus jiggle-map observations in the archive that were a) made in good observing conditions, b) made between 8pm and 4am, to avoid thermal effects on the dish, c) made during times when we know SCUBA's performance was not atypical, using the monthly FCF summaries, d) non-polarimetric observations. The table below summarises these results (note that the aperture FCFs are calculated on the basis of our standard 40" aperture):
Some notes on these results:
The trends shown in the above table are plotted below, along with lines of best fit. For the Beam FCF dependence, a quadratic fit to all data points is given, whereas for the Aperture FCF dependence, the fit was constrained only for data points ≥ 60" (see point 2 above). The error bars shown in these plots are reduced by the factor of √N from the values in the table above.
450 micron FCF relationsWhen we tried to repeat the same exercise with 450 micron data, we found that, even taking the steps described above to minimise effects such as dish temperature and known SCUBA problems, we were unable to find smooth trends in the averaged data. 450 micron observations with SCUBA are extremely sensitive to small changes in focus and opacity as well as to the night-to-night state of SCUBA, and so these sources of noise in the datasets drown out the trend due to changing chop throw. To minimise the night-to-night changes, we instead chose to only make intra-night comparisons to produce typical ratios of FCFs, rather than worry about the absolute values. Unfortunately, nights when more than three different chop throws are used are rare. However, by combining all the useful data for Uranus and CRL618 (Mars was not used since it can become too large) we were able to get some consistency. These results are summarised in the table below. Note that the ratios are always expressed in the order shown i.e. FCF(smaller chop throw)/FCF (larger chop throw).
As is obvious from the above table, there is a wide variety in the number of data points available for each chop throw ratio. The most represented chop throws are 30", 44" and 68", and these often come together, since these are the three chop throws used for SHADES observations. Other areas of "chop throw space" are markedly underrepresented including, unfortunately, > 120", which is the region where the largest changes are expected. For the five most represented chop throws in the above table, we have produced plots to show the trend of their ratios. In the plots below, the constant chop throw is always regarded as the numerator; e.g. most of the ratios involving 120" have been inverted for inclusion in the 120" plot. The aperture ratios are shown by a dotted circle whilst the beam ratios are shown by a cross. Only those data points with errors ≤ 0.1 have been included.
Some notes on these results:
Image characteristicsThe change in FCF with chop throw reflects a change in the image shape, since the actual signal is conserved across the whole image. We have evaluated this effect using the KAAPA command PSF, which fits a Sersic profile, with D = A exp (-0.5 * [r/σ]^ γ), where r is the radial distance from the fitted centre of the source, σ is the profile width and γ is the radial fall-off parameter. If γ = 2, the fitted profile is a perfect Gaussian. The results of running PSF on the above data sets are shown below.
Notes:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
