1998A&A...331..328K

The X-ray light curve of CAL 83 as measured with the ROSAT HRI covering ∼1(1)/(2) years (from July 95 to Dec 96) is derived. These observations show the pre-off state (persistent state), the April 96 off-state and the post off-state which returns after some decline to its previous persistent level. The intensity variations are discussed in the model of variations of the temperature of the envelope of the nuclear burning white dwarf. In order to explain the observed X-ray off-state a drop in temperature from 4.5x10⁵K to at least 3.2x10⁵K (i.e. by 40%) is required. This might have been achieved - assuming an accretion rate close to {dot}(M)_RG of a ∼1M_☉ white dwarf - if the radius of the envelope of the white dwarf increased by a factor of ∼2.0 possibly due to an increased mass accretion episode onto the white dwarf. CAL 83 resembles then the supposed wide binary supersoft source (and symbiotic star) AG Dra where a radius expansion by a factor of ∼2 and a temperature drop by ∼35% has been required in order to model the observed X-ray fall in 1994. In both systems the accretion rate is supposed to be close to {dot}(M)_RG and to pass episodically above this value and to cause the X-ray fall. The main differences between the two systems are the nature of the donor star, the mass of the white dwarf, and the size of the binary orbit. It is not known whether CAL 83 is similar to the supersoft X-ray transient RX J0513.9-6951 as no recurrent X-ray off-states have been found for CAL 83. The extreme short duration of <100 days of the X-ray off-state would require in case of an unstably nuclear burning white dwarf (a recurrent supersoft source) a white dwarf mass of M_WD>1.2M_☉.