The MPIfR 19-channel bolometer array
In March 2000, a 19-channel bolometer array, developed at the MPIfR Bonn by
E. Kreysa and collaborators, was installed at the HHT. It consists of 19
individual broadband continuum receivers, and is sensitive around a wavelength
of 0.87mm (345GHz).
On this page we compile some information about this bolometer array, coming
partially from test observations carried out in April/May 2000.
|The 19 channels are arranged in two concentric hexagons
around the central channel. The figure to the right shows the array geometry
at an elevation of 0 degrees. The azimuth distance between two adjacent
channels is about 50", and the maximum beam separation about 200". The
channel numbers indicate the location on which the source will appear in
this channel when the central channel is centered on the source (to get the
actual position of the channels on the sky, you have to rotate the whole
picture by 180 degree). For increasing elevations, the image of the whole
array rotates counter-clockwise on the sky.
|The bandpass of a bolometer is usually not a flat profile.
The small figure to the right shows the sensitivity of the central channel
of the 19-channel array as a function of frequency. To get the response of
the instrument to any astronomical signal, you have to multiply this curve
with the atmospheric transmission.
|In the submillimeter range, most of the noise is due to the
atmosphere. With proper data reduction, this atmospheric noise can be
eliminated to a certain degree. The remaining noise level is about
|600 mJy (second)-1/2 per channel
Note that this value includes already the amount of observing time spent
on a reference position (blank sky) using the wobbling secondary. The pure
instrumental sensitivity is therefore a factor sqrt(2) better.
Even if this value is still preliminary, it can serve to estimate the
observing time needed.
- On-Off: This procedure should be used for point sources (flux
measurements, detection experiments, ...). The necessary observing time
can easily be calculated with the number above, e.g. to reach an rms noise
of 6 mJy, you should expect to observe for (600/6)2 seconds,
i.e. about 3 hours.
- On-The-Fly: Extended sources should be mapped. Consult the
Manual for necessary map sizes. With a scanning velocity of 8" per
second (which means about 40 minutes per 400" x 300" map) you can expect
an rms noise of 150 mJy/beam per coverage (i.e. after co-adding the 19
We again emphasize that these values include already the amount of time
spent off-source using the wobbler, and also the elimination of correlated
atmospheric noise during data reduction!
Remember that you have to add a certain overhead to the integration times
estimated. Pointing and focus measurements, as well as maps of planets and
skydip measurements for calibration purposes, may add up to 25 percent or
more of the total estimated integration time
|The reduction of data taken with the bolometer can
currently be done by two individual programs, MOPSI and NIC. MOPSI is
developed by Robert Zylka (MPIfR Bonn/ITA Heidelberg) and can read the
telescope raw data directly. The output format is NOD2. NIC is part of the
Gildas software package and is
mainly developed at IRAM Grenoble. It reads the Ascii-NMB data format, which
is produced on-site by the HHT2NIC program, and writes the output in GDF
format, which can be further processed by the GRAPHIC program.