Gears are important components in most power transmission mechanisms. Failures of gears can cause heavy losses in industry. Condition monitoring and fault diagnosis of gears is therefore important to improve safety and reliability of gearbox operations and reduce losses caused by gear failures. This research proposes a new diagnostic approach based on the statistical analysis of data. It investigates the use of Principal Components Analysis (PCA) to detect growing local faults in a two-stage industrial helical gearbox. In this research, the vibration signal is used to monitor fault conditions and a broken tooth is simulated as a local fault. Since the early detection of faults is a challenge, small fault conditions were tested first as well as severe fault conditions. In order to examine the ability of the PCA to detect fault conditions, first the PCA-based model was created for normal operating conditions. Any unexpected event such as a fault condition causes a significant deviation from the PCA model, which is obtained from the normal condition data of the gearbox. The Square Prediction Error (SPE) was calculated to detect the fault conditions. When the vibration signal from the gearbox is representative of normal operation, the value of the SPE shows very little fluctuation and remains under a certain threshold value. However, in the presence of the fault the SPE fluctuates considerably beyond the threshold value. It is shown that the PCA-based statistical approach cannot only be used to detect severe fault conditions, but that it also reveals small growing fault conditions at very early stage. The technique also provides information about the state of the fault such as the location of the fault as well as its severity.