1. Motivation. The modular curves X 0 (N ) have been studied intensively since they provide a link between modular and elliptic functions. Among other things, they parameterise pairs of elliptic curves with a cyclic isogeny of degree N between them. Topics of interest include the search for models with few or no singularities or with small coefficients. For instance, the question which curves X 0 (N ) are hyperelliptic has been answered in [11] and complemented by concrete models in [12,9,16]. The factorisation pattern of these modular equations provides information on the rationality of isogenies and can thus be used to determine the L-function of elliptic curves over finite fields. In general, plane equations with nice properties are obtained by looking for two functions on X 0 (N ) with suitable pole orders and determining a polynomial relationship between them. To link the modular equations to the pairs of isogenous elliptic curves they parameterise, one needs to exhibit a relationship with the modular invariant j, which requires considerable work in each case (see [5,2]).It is thus convenient to fix one of the functions as j and to consider X 0 (N ) as a cover of X 0 (1), albeit this introduces further singularities. Equations for X 0 (N )/X 0 (1) with small coefficients have been exhibited in [7] and [8, Chapter 5]; some ideas presented there go back to Atkin (see [2]). In the context of point counting on elliptic curves, it is most efficient to use curves of prime level N .In this article we deal with an infinite family of modular curves X 0 (N ) where N is composite of the simplest form, i.e. a product of two primes p 1 and p 2 , which need not be distinct. Besides j, we also fix the second function generating the function field of X 0 (N ) as a certain product of η-functions. This is motivated by the observation that the singular values of these functions at ideals of suitable orders in imaginary-quadratic number fields lie in the corresponding Hilbert and ring class fields [13]. The modular polynomials relating these functions and j can thus be used to explicitly determine