The thermoelectric inhomogeneity of wires is one of the main components of the measurement uncertainty when using thermocouples. During calibration, it is therefore important to determine how much the inhomogeneities affect the measurement result. Thermoelectric inhomogeneity is normally assessed by gradual, or step-wise, insertion of a thermocouple into a furnace or liquid bath. With this type of equipment, the length that can be scanned is typically limited to about half a meter. To assess thermoelectric inhomogeneity over greater lengths, it is necessary to adopt a different technique. Therefore, an apparatus with a short, movable heating zone has been set up and evaluated. The apparatus produces a short, well-defined heating zone that is moved along the thermocouple while both the measuring and reference junctions are kept at 0 °C. Heating is done by means of a hot-air fan that produces a temperature-controlled heating zone up to 700 °C. Two directionally controllable cold-air fans, one on each side of the heating zone, make it possible to vary the slopes of both temperature gradients of the heating zone. Two temperature gradients influence the thermocouple when performing measurements of inhomogeneity with this setup. The results are, therefore, not directly comparable to the results from measurements taken in a bath or furnace, where only one temperature gradient is present. The resulting curve obtained with the two gradients is approximately equivalent to the derivative of the curve obtained with one gradient. It is possible to convert the two-gradient curve to a single-gradient curve by numerical integration, as shown in this work. Comparisons with inhomogeneity measurements obtained using salt baths show good agreement with the calculated results.