The convergence and resolution of Poiseuille turbulent channel flows have received little attention in the last 35 years. Apart from the idea that the cell size must be close to the Kolmogorov scale, no reasons are given for a particular mesh selection. Typically, authors only compare their mesh size to ones in earlier publications. Often, a specific mesh size is chosen for numerical reasons (FFT performance) instead of physical or statistical ones. In this talk, we will analyze several simulations performed at a relatively small friction Reynolds number of 1000, studying the effect of the mesh size, the computational box, and the compiling time on the statistics of the flow.