Contactless photoconductance measurements are commonly used to extract the emitter saturation current density (Joe) for crystalline silicon samples containing an emitter on the surface. We review the physics behind the analysis of Joe and compare the commonly used approximations with more generalised solutions using two-dimensional device simulations. We quantify errors present in such approximations for different test conditions involving varying illumination conditions and surface properties in samples with the same emitter on both sides. The simulated Joe obtained from the dark hole current from the emitter into the bulk is nearly the same as the simulated Joe determined by photoconductance measurements of the rear diffusion. The simulated Joe at the front emitter is equivalent to that at the rear emitter only when the sample is subject to a nearly constant and flat generation profile. For illumination conditions including visible light, the simulated Joe at the front emitter is smaller than the simulated Joe at the rear emitter. Both Joe at the rear emitter and from the dark hole current in the emitter remain nearly constant over a wide range of base doping densities. The approximations used for the determination of Joe from photoconductance measurements make Joe dependent on the excess minority carrier density. Lifetime measurements demonstrate that, even in high-quality silicon, Joe should be determined from the analytical solution as a function of excess minority carrier density including Shockley-Read-Hall recombination. Copyright © 2012 John Wiley & Sons, Ltd.