The presence of mechanical loads at the limb socket interface is considered an initiating cause of tissue breakdown and ulceration. Historically, prosthetists have relied on past experience, patient feedback, and indirect indications of load to gauge socket fit. The clinical measurement of interface pressures has the potential to provide quantitative, objective information to help in the evaluation of prosthetic fit. Tests of the Tekscan F-Socket system have been performed over flat and curved surfaces, but not using axial load and the presence of a gel liner to more closely simulate a clinical setting. This study investigates how drift and cyclic drift errors of the F-Socket system are affected by surface curvature and the presence of a gel liner interface. For drift, a known constant axial load was applied by a servo-hydraulic testing machine for 20 minutes. Samples were taken at 1, 5, 10, 15, and 20 minutes. For cyclic drift, load was alternated in a known range at 0.5Hz for 10 minutes. Samples were taken at 1, 5, and 10 minutes. These tests were performed under four conditions: flat rigid, flat with liner, round rigid, and round with liner. It was found that drift error was significantly affected by the curvature of the model, but not by the presence of a gel liner. It was also found that cyclic drift was not affected by the curvature of the model, but was by the presence of a gel liner. Further study should be done with other curvatures, other liners, and other sensor properties such as hysteresis and accuracy.