Background of the Invention
As skid control systems for aircraft have become progressively faster and more sophisticated, they have evolved toward the use of digital computing techniques for solving certain required equations. In view of the many variables experienced in terms of surface condition, weather conditions, etc., and in view of the fast computing capabilities available, it becomes more and more important that the wheel or tire speed-responsive means have good resolution and fast response time. Prior art systems have relied upon wheel speed signal generators which are really wheel-driven alternators having a substantial number of poles so that an alternating current signal is generated whose instantaneous frequency is proportional to instantaneous speed of the wheel. Since wheel speed generators should be as small and light as possible, there is, of course, a practical limit as to the number of poles, hence cycles per wheel revolution, which can be built into such an alternator.
The braked wheel can, however, change its operating characteristics from an acceptable deceleration to an unacceptable skid in a relatively few degrees of rotation, which may mean that even with as many as 128 poles, the number of signal cycles available at the time of a skid to produce a meaningful input signal may be so few that the associated control system may not respond quickly. During this time the skidding may well either continue or get worse before any correction can take place. Thus the wheel speed sensor has come to be recognized as imposing a significant limiting factor preventing substantial improvement in the performance of skid controls.
Description of the Drawings
FIG. 1 is a schematic block diagram of a wheel-speed sensor according to the present invention shown in association with a skid control system;
FIG. 2 (comprised of a-d) is a graph showing a number of wave forms produced and used by the system of FIG. 1;
FIG. 3 is a flow chart indicating a manner in which the output of FIG. 1 may be used in a microprocessor-based skid control system.
Description of the Preferred Embodiment
FIG. 1 is a schematic block diagram of a wheel speed-sensing system incorporating my invention. Referring to FIG. 1, the aircraft wheel being controlled is shown at numeral 10. It is connected through a mechanical link to an alternator 12 having a large number of poles which produces an alternating current output whose frequency varies with rotational speed of said alternator. The output of the alternator 12 is supplied to a signal processor 14 where it is converted to a series of sensor pulses whose width varies with the frequency of said alternator. These sensor pulses are shown diagrammatically on a line 16 connected between processor 14 and the input of an AND gate 18. Gate 18 is also connected to receive the output of sample pulses from the microprocessor-based skid control system 19. The wheel-speed calculating system computes rotational velocity providing to a computer or microprocessor a plurality of numerical factors from which the microprocessor can compute a number representing wheel speed averaged over a sample period. This number or numerical representation of wheel (or tire) speed must be the virtual instantaneous value for correct skid control computation. This requires a sample value of wheel speed over a desired period. The lower limit of tire speed for which anti-skid is required determines the sample period of the sensor signal. This is given by:
(1) Sample Period=2/v.times.n
where v=lower limit of tire speed for which skid control is required in radians/sec.
and n=number of pulses per revolution of the speed sensor.
With programmed inputs as to v and n, any appropriate microprocessor can readily provide the desired sample pulse. Sample pulses from control system 19 are also connected at one input of a second AND gate 20 which also receive pulses from a high rate clock 22. Each of AND gates 18 and 20 thus provides outputs only during the period of a sample pulse, gate 18 providing sensor pulses only during the sample period and gate 20 providing high rate clock pulses at its output only during the sample period. These outputs are shown graphically in FIG. 2 wherein time is depicted horizontally and the sample period is represented by the two widely spaced vertical lines. Sample sensor pulses appear on the top line (Graph (a)) in their relative time relationship to the sample pulse. The sensor pulses supplied by AND gate 18 are received in a pair of edge pulse generators 24 and 26 with generator 24 responding to positive-going excursions of the sensor pulses to produce sharp marker pulses and generator 26 responding to negative-going excursions of the sensor pulses to also produce sharp marker pulses. Both sets of pulses from edge pulse generators 24 and 26 are supplied as input signals to an OR gate 28 which passes these marker pulses to a counter 30 which counts the number of marker pulses occurring during the sample pulse period, to a counter 32 where it is used to reset the counter to its initial state, and to a latch circuit 34 which changes state upon incidence of the first sensor pulse edge occurring during the sample period. With this change of state, the high rate clock pulses are gated via an AND gate 36 into a counter 38. At the end of the sample period, counter 38 has counted the number of clock pulses occurring from incidence of the first sensor pulse edge (the first marker pulse) within the sample period until the end of the sample period, as shown in graph (c) of FIG. 2.
Graph (b) of FIG. 2 shows the output pulses from OR gate 28 which occur within the sample period. The sample pulse is also supplied to AND gate 20 along with the clock pulses from clock 22 which transfers these clock pulses to counter 32 which counts the number of clock pulses occurring in the sample period. Since counter 32 is reset to zero each time the sensor pulse changes state within the sample period, the action of counter 32 is to provide a count of the number of clock pulses occurring between the last sensor pulse change of state (marker pulse) and the end of the sample time period. This is shown in FIG. 2, graph (d).
At the end of the sample period, all counters cease their counting and maintain their current state. The counters connect to a data bus in the microprocessor-based skid control system 19 where their counts are transferred at the end of each sample period. In the control system the computer (microprocessor) calculates tire velocity using the equation:
(2) Tire Velocity= ##STR1##
where K= ##EQU1##
It will be recognized that the factor "radians per sensor pulse" is built into the alternator 12 and the "high rate clock pulse" is the output of clock 22.
After transfer of data from counters to the microprocessor 19, the counters and latch are reset and a new sample period commences. During this sample period the computer works out the anti-skid control equations and provides an output that is used to control the anti-skid servo valve (or other output device). The computer then waits until the new sample period finishes and then repeats the calculation for a new number for controlling the servo valve.
FIG. 3 is a computer flow chart which shows the type of sequencing that takes place to make a calculation of tire speed. The numerical representation of value attained by each counter at the end of a real time period is first transferred sequentially over a common bus line into the computer. These numbers are next substituted into the above equation (2) which is then solved, giving the tire speed.
It will be recognized that the particular microprocessor-based skid control system used bears no relationship to the present invention other than as a means of utilizing the counts supplied by the counters 30, 32 and 38 and that these counter inputs provide a means whereby any appropriate microprocessor may solve the above equation for wheel or tire velocity. It will be apparent to those skilled in the art that modifications may be made within the scope of the present invention.