Method and apparatus for baseline drift correction for a sensor

Abstract

A method for continuously correcting for the baseline drift of a sensor which measures a cyclically varying parameter. The output signal of the sensor is sampled rapidly over an integer number of cycles of the measured parameter. The sampled values of the output signal are stored as a data set in a memory circuit for subsequent analysis by a computing circuit to determine the actual baseline level of the sensor, which drifts at a rate relatively slowly in comparison to the time interval during which the data set was collected. In one embodiment of the invention, baseline level is calculated as the average of a fixed percentage of the points in the data set which have the lowest values found in the data set. In another embodiment, a data mask with multiple selection segments spaced at time intervals equal to the period of the cyclically varying measured parameter is used to determine those points from the data set which are averaged to calculate the baseline level. The data mask is placed at multiple positions on the data set, and the lowest calculated baseline level is chosen as best representing the actual baseline level of the sensor. This method is particularly useful for continually determining the baseline level of a grain mass flow rate sensor which is used in a harvester. The baseline level of this type of sensor drifts significantly due to varying thermal, mechanical and operating conditions. Correction for this drift by continually recalculating the baseline level allows accurate measurement of grain mass flow rate at all times.