Patent US 5,222,022

Methods and apparatus, preferably carried out with the aid of a programmed digital computer for iteratively determining the actual speed of an engine and iteratively updating a fuel rate command signal so as to govern and correctively remove any errors from the set point speed. Jumps to a program subroutine are triggered by sensing the engine crankshaft to generate a predetermined number Q of pulses during the span of one revolution (where Q is an integer greater than 1), and each pass through the subroutine determines the actual engine speed as an average over the last whole revolution of the crankshaft. There are thus Q iterations per revolution but each interation signals the speed over one whole revolution. This results in wholly eliminating torsionally induced speed fluctuations, or the like, from the determined speed signal. Each pass through the program subroutine also determines the speed error and updates the fuel rate command signal as a PID function of such error. Because this iterated updating occurs, not at a regularly timed rate, but at intervals which vary inversely with speed, the instructions of the subroutine program can be made very short and simple while nevertheless causing effective automatic tuning adjustments of the integral gain, stability and compensation factors.