Magnetic viscosity of rocks associated with magnetic relaxation of ultrafine ferrimagnetic mineral grains (superparamagnetism) is employed in magnetic grain size measurements. Magnetic viscosity is most often estimated from dual frequency measurements of magnetic susceptibility. The measured susceptibility values bear uncertainty that comprises two components: an instrument error and a drift. The instrument error refers to the accuracy of the measurement system and shows how precise the data are in ideal operation conditions. This error affects especially the low susceptibilities of weak samples, which thus should be measured on a high sensitivity range. Drift is due to external factors, such as changes in the temperature of sensors and/or samples, as well as in the orientation of the samples relative to the sensor, vibration, electromagnetic noise, etc. Drift, more critical for measurements on strong samples, is manageable by the operator. To reduce drift, every effort should be made to maintain suitable ‘quiet’ operation conditions.