6 First observations

We describe now the observations done in March 1995 during a three and a half week campaign at the Millimetre and Infrared Testa Grigia Observatory (MITO) that gives an 8 arcmin beam with the Diabolo photometer. Here we present some results that were acquired with a sawtooth modulation of the secondary at 1.9 Hz (which provided a constant elevation scan across a source, of typically 26.4 arcmin width), combined with a slow drift of the elevation offset relative to the source (by a total of 40 arcmin, with steps of 4 arcmin i.e. half beam width every 10 s). The acquisition frequency of 61 Hz is twice the AC modulation readout frequency. It is synchronous with the wobbling secondary frequency of 1.9 Hz, giving 32 measurement points per period.

To our knowledge, these data are the first ever to be acquired on the sky in a total power mode using unpaired bolometers. It anticipates and proves the feasability of the total power readout mode that is planned for next submillimetre ESA missions ( PLANCK and FIRST).

6.1 The MITO telescope

The MITO telescope has been specifically designed for submillimetre continuum observations at the arcminute scale up to the degree scale, and as such is a unique facility in Europe. The telescope (De Petris et al. 1996), which was designed in parallel with OLIMPO (Osservatorio nel Lontano Infrarosso Montato su Pallone Orientabile) ex TIR (Telescopio InfraRosso) one, is a classical Cassegrain-type 2.6 m dish with a wobbling secondary mirror designed with very low levels of vibration (Mainella et al. 1996). The MITO facility is situated on a dry cold site at an altitude of 3500 m close to Cervinia-Breuil in Italy, very near the Swiss border and the Gornergrat infrared and millimetre observatory TIRGO (Telescopio InfraRosso del GOrnergrat). During our observations, we routinely had outside temperatures of -20 Celsius (most of the nights) and good weather for about one third of the time, making this site excellent for (sub)millimetre high angular resolution astronomy (the opacity is less than a tenth at zenith in the whole millimetre range).

6.2 Data reduction

The data were acquired with two independent acquisition systems, the first one based on the development of the PRONAOS one and the second one custom made to allow for the new readout technique. The following data analysis is based on the last system.

After deglitching, a raw map is made with the values of the signal for given azimuth and elevation offsets. The azimuth offset is deduced from the position in a given period of the secondary while the elevation offset is (or should be) a sawtooth function of time. The registration of the instrument data with the telescope pointing information is done with an absolute time line which happened to be inaccurate after ten minutes of observations. Therefore, we can only show here the data which are post-synchronised with the help of the occurence of a strong source detected in the raw data. The data present a strong systematic effect which is quite reproducible and function of the azimuth offset angle. This is easily removed from the maps by computing the mean effect (over elevation offset angles) after the source has been masked. This effect is most likely due to the instrument "seeing'' the asymmetrical back of the secondary during its sawtooth motion. This can and will be reduced by adding a secondary mirror baffle as described in Gervasi et al. (1998).

Another phenomenon is the slow drift of the detectors during time which is removed with a running constant elevation average after the source is masked. Each map is then rotated by the parallactic angle and coadded to the others to make a final map in astronomical coordinates. The maps of planet Mars as obtained in the two Diabolo channels, are shown in Figs. 9 and 10. It corresponds to the average of 9 individual maps of $28 \times$ 40 arcmin, and a total integration time of 1050 s. The beams are quite similar at both wavelengths and coaligned within a precision of a tenth of a beam. The beam FWHM is of 7.5 arcmin. The integrated beam efficiency is the same as that of an 8 arcmin FWHM Gaussian beam. The signal expected from the planet after dilution in the beam is equivalent to a 110 mK blackbody. The Orion BN-KL nebula is detected in the raw data and the final maps are given in Figs. 11 and 12. It is calibrated with Mars signal, but no correction for differential extinction was applied. As Mars was at a larger elevation at the time of the observations, the fluxes of Orion which are found as $860\pm 48 \rm\,Jy$ and $330\pm 40\rm\,Jy$ at 1.2 and 2.1 mm should really be considered as a lower limit (especially at 1.2 mm). The Orion spectrum, which is dominated by dust emission in the infrared and submillimetre domains, clearly behaves differently at the 2.1 millimetre wavelength, because the flux scales as the frequency to the power 2 between 1.2 and 2.1 mm rather than of 3 to 4 for dust submillimetre emission. Free-free emission from the compact central HII region is most likely at the origin of the 2.1 mm excess. The atmospheric noise is evident in all the data that were taken. Sensitivities were deduced from blank sky maps as 5, 8 and 7 $\rm\, mK_{RJ}~s^{1/2}$ at respectively 1.2, 2.1 mm and 2.1 mm after atmospheric noise decorrelation.