Detecting Certain Power Line Faults That Disrupt Sensitive Electronic Loads

AC (alternating current) power is generally distributed in a voltage waveform that closely approximates a sinusoidal function with respect to time. Sensitive electronic loads, including computers and medical equipment, may be disrupted by substantial deviations from this sine wave approximation. A floating-window algorithm is described which performs a triggering function that detects changes in the voltage waveform which may disrupt sensitive electronic loads. This triggering function may be used to capture digitally sampled power line disturbances so their cause may be identified and corrected. (The' algorithm may have other applications where deviation from a repetitive waveform must be detected, such as electrocardiograms.) IN GENERAL, power supplies for electronic loads are designed to accept AC power within a certain range of RMS voltages and a certain range of frequencies. It is generally presumed by the power supply designer that the voltage will be delivered to the power supply in an approximation of a sine wave with respect to time. However, in practical applications, substantial variations from a pure sine wave may be encountered. These variations may include reduced peak voltage, sub-cycle dropouts, and other variations in the waveform. In some practical applications, these variations from an ideal sine wave may be constant and can be observed with an oscilloscope or a spectrum analyzer. But in most applications, these variations are introduced by changing loads or switches on the distribution system and are thus transitory in nature and can only be observed if they are captured and held for later examination. These transitory faults in AC power can take on increased significance in servicing electronic equipment in the field, as they must be distinguished from other causes of intermittent failures that may occur with similar or identical symptoms. Traditionally, power line sine waves have been examined using oscilloscopes or their digital sampkg equivalents. This sampling technology is well understood. However, because power disturbances axe relatively rare events, sampling generates vast quantities of data points, most of which describe points which are not interesting, i.e. data points which do not deviate enough from the ideal waveform to disrupt the operation of an electronic load. Consequently, a triggering algorithm is required which allows selective recording of data points which show significant flaws. (Typically, a triggering algorithm is applied to a FIFO (First In First Out) register of data points to allow capture and display of pre-trigger data.)