Testing stellar evolution models with detached eclipsing binaries

Stellar evolution codes, as all other numerical tools, need to be verified. One of the standard stellar objects that allow stringent tests of stellar evolution theory and models, are detached eclipsing binaries. We have used 19 such objects to test our stellar evolution code, in order to see whether standard methods and assumptions suffice to reproduce the observed global properties. In this paper we concentrate on three effects that contain a specific uncertainty: atomic diffusion as used for standard solar model calculations, overshooting from convective regions, and a simple model for the effect of stellar spots on stellar radius, which is one of the possible solutions for the radius problem of M dwarfs. We find that in general old systems need diffusion to allow for, or at least improve, an acceptable fit, and that systems with convective cores indeed need overshooting. Only one system (AI Phe) requires the absence of it for a successful fit. To match stellar radii for very low-mass stars, the spot model proved to be an effective approach, but depending on model details, requires a high percentage of the surface being covered by spots. We briefly discuss improvements needed to further reduce the freedom in modelling and to allow an even more restrictive test by using these objects.