Do note, however, that converging a system with semiconducting electrodes might be harder.
Also, people often try "the simplest thing imaginable", which might be a perfect 1D chain of metal atoms, for instance. This has a very simple analytic solution (transmission T(E)=the number of bands crossing energy E), but for various reasons it's not a very simple system to compute using ATK, esp. at finite bias. Essentially, the ATK algorithms are designed to work best when there is a clearly defined place in the structure for the applied bias voltage to drop, i.e. there should be some resistance.
This is particularly noticeable for 3D systems, if one tries to compute the transmission spectrum of bulk gold, for instance. In fact, "conductance" in the Landauer picture (which is what ATK calculates) is not a well-defined quantity for a macroscopic bulk conductor, but rather it applies to some "quantum point contact", which should be understood in a broad sense but still clearly is related to microscopic (or nanoscopic) geometrical features.
