In semiconductor manufacturing, the circuit integrity of packaged BGA devices is tested by measuring electrical resistance using test sockets. Test sockets have been reported to often fail earlier than the expected life-time due to high contact resistance. This has been attributed to the formation of Sn oxide films on the Au coating layer of the probe pins loaded on the socket. Similar to contact failure, and known as "fretting", this process widely occurs between two conductive surfaces due to the continual rupture and accumulation of oxide films. However, the failure mechanism at the probe pin differs from fretting. In this study, the microstructural processes and formation mechanisms of Sn oxide films developed on the probe pin surface were investigated. Failure analysis was conducted mainly by FIB-FESEM observations, along with EDX, AES, and XRD analyses. Soft and fresh Sn was found to be transferred repeatedly from the solder bump to the Au surface of the probe pins; it was then instantly oxidized to SnO. The $SnO_2$ phase is a more stable natural oxide, but SnO has been proved to grow on Sn thin film at low temperature (< $150^{circ}C$). Further oxidation to $SnO_2$ is thought to be limited to 30%. The SnO film grew layer by layer up to 571 nm after testing of 50,500 cycles (1 nm/100 cycle). This resulted in the increase of contact resistance and thus of signal delay between the probe pin and the solder bump.