2. Observations

We present observations of SiO maser emission at 7 mm wavelength during the period July 1984 - May 1990, using the 13.7 m radiotelescope of the CAY, located at Yebes, Guadalajara, Spain. We started observing 14 objects in the v=1 transition in 1984. Ten of those were observed also in v=2 since mid 1985. In 1987 the program was extended to 16 objects, observed in both v=1 and v=2 lines. Finally, after some major works on the receiver in 1988, the program was extended up to 21 objects observed in the v=1 and v=2 masers. In principle, the observations were performed every about 3 weeks, each observational run consisting in two consecutive days of observations, the first day for the v=1 line and the second one for the v=2 line.

The main characteristics of the CAY-13.7 m radiotelescope, described by Barcia et al. (1985), remained unchanged during the observing period except for those concerning the 45 GHz receiver (and the aperture efficiency of the telescope). At the beginning of the monitoring, the receiver consisted of a Schottky diode balanced mixer and a FET amplifier at room temperature, yielding a single side band (SSB) system noise antenna temperature above the atmosphere ( $T^{\star}_{\rm a,sys}$) of about 900 K. In May 1987 the performance of the mixer was improved resulting in SSB $T^{\star}_{\rm a,sys}$ of $\sim$ 600 K. Finally, in December 1988, the receiver was replaced with a similar one but being cooled down to 20 K with a closed-cycle Helium refrigerator, with typical SSB $T^{\star}_{\rm a,sys}$ of $\sim$ 260 K (including the $\lambda$/4 plate, see point 3 in Sect. 3). All these changes also resulted in different polarization sensitivities of the system (see again point 3 in Sect. 3). The dependence of the antenna efficiency with elevation was tested, modeled and taken into account in the calibration procedure. We note that the telescope is enclosed in a radome, inside which the air is kept in circulation to maintain a uniform temperature. This prevents differential dilatation in the main reflector that could affect the efficiency of the telescope, as well as in the supporting structure which could introduce pointing errors.

The spectrometer used during the whole period was a filterbank of 256 channels 50 kHz wide, providing a spectral resolution of 0.35 km$\,$s$^{\rm -1}$ at 43 GHz, although for the data presented here we have degraded the resolution to 0.7 km$\,$s$^{\rm -1}$ to increase their signal to noise ratio. Most of the observations were done in frequency switching mode (by switching the IF by 6.4 MHz), which reduces the effective filterbank bandwidth to $\sim$ 44 km$\,$s$^{\rm -1}$. Only for three objects in our sample, VY CMa, VX Sgr, and Orion IRc2, this velocity coverage is not large enough because of their broad maser emission, resulting in difficulties for a proper baseline subtraction. For these three objects a position switching procedure was adopted, setting the "off'' position 10' away in azimuth from the star position. Note that at 43 GHz, the half-power beam width of the telescope is $\sim$ 2'.

For the calibration of the data we have followed the procedure explained in Barcia et al. (1985), that essentially uses as references an absorber at ambient temperature as hot load, and the blank sky as cold load. This method directly yields data calibrated in units of antenna temperature. This is the most usual output unit in radioastronomy, but it is not appropriate when dealing with point-like non-thermal emissions, as it is the case here, for which flux units are preferred. The conversion factor from the antenna temperature scale to flux units (Jy) has been determined from observations of the planets as standard calibrators.