Electrochemical micromachining (ECM) of p-type Si substrates is accomplished in HF-based solutions by applying nanosecond potential pulses between the substrate and a tungsten tool electrode. With sufficiently high potential pulses, the silicon potential locally reaches the electropolishing regime and microstructures may be machined. ECM precision is investigated as a function of pulse height, pulse duration, solution composition, and silicon doping level. Results show that micrometer precision may be obtained with highly doped substrates and that experimental data can be explained within a simple model, taking the charging time of the interface capacitance into account. In highly doped p-Si, well-defined microstructures can be realized without application of a mask on the surface. In addition, the isotropy of the process allows fabrication of structures not constrained by the crystal direction. In the case of low-doped material, ECM is only possible for very short pulses (<3 ns).