1 Introduction

The eclipsing nature of the Herbig Be star TY$\,$CrA was discovered by Kardopolov et al. (1981) from photoelectric observations. Later photometry (e.g. Vitrichenko & Shevchenko 1995) and spectroscopy (Casey et al. 1993, hereinafter Paper I; Lagrange et al. 1993; Corporon et al. 1994; Casey et al. 1995, Paper II) have shown that the orbit is circular with a period of $P=2\hbox{$.\!\!^{\rm d}$}888779$. High S/N spectra show lines of both eclipsing components and in addition lines of a tertiary star in a longer period orbit (Corporon et al. 1994; Paper II; Corporon et al. 1996). An outstanding peculiarity of TY$\,$CrA is the very narrow spectral lines of the primary star, indicating a $v\sin i$ of less than 10 km/s (Paper I; Lagrange et al. 1993). Thus the primary appears virtually non-rotating despite its location in a young, short-period binary system with an already circularized orbit (see also discussion in Beust et al. 1997 and Casey et al. 1997, Paper III).

TY$\,$CrA is embedded in the densest part of the R CrA molecular cloud (Harju et al. 1993) in the reflection nebula NGC$\,$6726/7. It forms a pair with HD$\,$176$\,$386 (Sep. $1\hbox{$.\mkern-4mu^\prime$}1$), which is a visual binary (Sep. $4\hbox{$.\!\!^{\prime\prime}$}1$) and has occasionally been confused with TY$\,$CrA (e.g. Proust et al. 1981). The location, reflection nebula, and strong infrared excess of TY$\,$CrA are all evidence that it is still very young. After EKCep (Popper 1987), TY$\,$CrA is only the second double-lined eclipsing binary which includes a pre-main sequence star, and as such is of particular interest.

The cloud associated with TY$\,$CrA is very peculiar: its infrared emission (IRAS point source 18583-3657) peaks beyond $100\;{\rm \mu m}$, suggesting very cold dust as its origin (Friedemann et al. 1996). The densest clump of the CrA cloud is one of the coldest ever measured (an upper limit of 8$\,$K, according to Harju et al. 1993 and Haikala 1994), which should favour the formation of multiple stars (Durisen & Sterzik 1994). In fact, TY$\,$Cra is a triple system, located at a projected distance of 8500 AU from the visual binary HD$\,$176$\,$386, the companion of which is very red (private communication from Dr. H. Zinnecker from a $\rm 10\;\mu m$ exposure obtained with TIMMI and the ESO 3.6 m telescope). Practically all stars associated with the R CrA region are very young (Bibo et al. 1992) and we would expect that most of them are multiple.

In order to establish the precise properties and evolutionary stages of the stars in TY$\,$CrA, gain insight in the applicability of current theoretical models for PMS stars, and understand the variability and unusual photometric phenomena of TY$\,$CrA, a careful photometric analysis based on complete and accurate light curves is essential. The observational material for such studies is documented in the present paper, while a detailed analysis of the y light curve (including an ephemeris) and a discussion of the properties of the system are given in Paper III.

The light curves of TY$\,$CrA show a moderately deep primary eclipse and a very shallow secondary minimum, $0\hbox{$.\!\!^{\rm m}$}03$ deep in y and barely detectable in u. The unequal minima indicate different temperatures for the two components. Anticipating Paper III, the 3.16$\,M_{\odot}$, 12$\,$000$\,$K primary is located near the ZAMS (although curiously cooler than models predict) and the 1.64$\,M_{\odot}$,4$\,$900$\,$K secondary is at the base of its Hayashi track. A conspicuous "reflection effect'' is present, but no other strong proximity effects appear in the light curves. There is evidence of considerable intrinsic variability, on time scales ranging from days to years and not evidently correlated with orbital phase. We discuss the variability further below and suggest that it is primarily the result of variable obscuration.