United States Patent 3743055

An elevator control system that utilizes a digital signal representative of the car position to provide the necessary information inputs to the selective logic system and motor control so that a smooth acceleration to maximum velocity and deceleration to stop at a selected floor. A drum selector is utilized to produce the digital number corresponding to elevator position. The motion input to this drum selector is linked to actual elevator motion. The system includes electronic provision for simulating the advance of the car position an amount dependent on it's velocity and locks out all floors behind that advanced position in order to allow adequate time for deceleration. The velocity command is compared to actual velocity and the power control section called on to make up any difference. Differences persisting for a period of time produce increasing commands in response, while damping means prevent overshoot.

Hoelscher, William R. (La Mesa, CA)
Vildibill, Alvin J. (El Cajon, CA)
Application Number:
Publication Date:
Filing Date:
Primary Class:
Other Classes:
International Classes:
B66B1/16; B66B1/24; B66B1/30; B66B1/34; (IPC1-7): B66B1/16
Field of Search:
187/29 318
View Patent Images:

Primary Examiner:
Gilheany, Bernard A.
Assistant Examiner:
Duncanson Jr., W. E.
1. An electronic system for controlling elevator motion comprising:

2. An electronic system for controlling elevator motion comprising:

3. The elevator system of claim 2 wherein said sensors comprise photodiodes sensitive to the infrared portion of the spectrum.

4. An electronic system for controlling elevator motion comprising:

5. The electronic system of claim 4 wherein:

6. The electronic system of claim 5 wherein:

7. The electronic system of claim 4 wherein:

8. An electronic system for controlling elevator motion comprising:

9. The system of claim 8 wherein:

10. The system of claim 9 wherein:


An increasing number of structures are being built whose height classes them in the high rise category. The utility of any high rise building is importantly affected by the efficiency of the elevator system there installed. The increasing demand for elevator efficiency in such structures has drawn attention to the conflicting needs of elevator speed and passenger comfort.

In conventional high rise elevator systems, various means have been utilized in an attempt to meet the demands of speed and comfort by producing a signal corresponding to the elevator position and disabling the selector logic so that a command to stop the elevator at a particular floor, will take place only if sufficient distance for comfortable deceleration to that floor remains.

A typical high rise elevator system utilizes a scale model elevator car moved by gearing or other means in a relationship to the movement of the actual elevator car. Due to space and other limitations, the "shaftway" for this model elevator must be of a limited height, for example, 10 feet for a main shaftway of 300 feet. The model's shaft contains numerous microswitches or other means of sensing the car's passage and therefore is capable of providing signals indicative of the position of the main elevator. The point at which the car triggers the microswitches may be advanced ahead of the car's position, in response to a signal corresponding to the car's velocity, by an electromechanical device. Thus these systems are capable of producing a signal indicative of the advanced position of the car, so that the selector logic will disable the floor selection switches on all floors behind the car's advanced position. When a floor in advance of the car's advanced position is selected, the selector logic in these systems, will command the car to decelerate. This deceleration is frequently at a fixed rate, and means are provided, in association with the selected floor to obtain some degree of local control over the point where zero velocity and floor position correspond. Systems of this type must utilize position information that is inherently inaccurate, since the initial positioning of the microswitches and their subsequent wear will be magnified by a factor, in the example, of 30 to 1. The linear position of the car is therefore not precisely known, and the selector logic must compensate for this inaccuracy by inactivating floor selector switches in excess of the car's ability to stop. Further, these systems produce a crude deceleration characteristic that is not easily tailored to reduce passenger sensation and discomfort upon deceleration or acceleration. Finally, the system inaccuracies require another transition that may produce deceleration changes as the car approaches the immediate proximity of the floor selected. The transition to this local control and the final deceleration to reach zero velocity at the level position of the car produce additional jerks that may be felt by the passengers.

In addition to the objectionable characteristics of these prior art systems in operation, there are other disadvantages that relate to the maintenance of such systems. The primary maintenance deficiency of such systems is that they must incorporate numerous electromechanical devices, resulting in mechanical wear, and necessitating adjustment and replacement with the attendant problem of locating the particular part with deficiencies.

Thus it would be desirable if a motion control system for elevators could be produced that would provide a motion control system to overcome the disadvantages of prior art devices.


An exemplary embodiment of the invention incorporates a drum selector for producing a digital signal in response to an input related to the car position. The linear motion of the car is transmitted, by means attached to the car, directly to the drum selector and the circumference of the drum therefore moves at very close to the same linear velocity, and for the same distance as the car itself. The drum selector produces a digital signal comprising a discrete binary number for each of a large number of possible car positions. This binary number is produced directly from the rotation produced by the drum and is summed with a similar binary number corresponding to the advance of the car position necessary to allow for deceleration. The actual car position signal, is utilized in conjunction with inputs from the logic circuitry to determine the distance remaining between the next selected stopping floor and the elevator's present position. This distance inputs a deceleration signal matrix that produces a discrete velocity signal for a predetermined number of points along the distance between initial deceleration and stop. Deceleration signals produced by the deceleration matrix are converted to a voltage analog of velocity, and filtered, to produce the desired velocity curve. The curve is inputed to an error determining means.

The error means functions to determine the amount of velocity error between the velocity commands produced in the digital selector means and the actual velocity, as determined by a tachometer on the motor output or other suitable measure of velocity. If a difference exists between the commanded velocity and the actual velocity, a signal is produced and inputed to the power control means to vary the power delivered to the motor-generator set accordingly. The motor-generator set has a relatively large time constant, as with all large rotating machinery, and therefore would tend to develop a residual velocity error that would be eliminated slowly, if at all. Thus, there is provided a residual error elimination system that includes a means for integrating the residual error over time. In this manner a residual error that exists for a period of time will have an increasing effect and call for increasing corrections to the power control systems. Overshoot of the system is prevented by an additional feature of the error means utilizing a signal that increases and decreases in advance of the actual velocity of the motor-generator set, to anticipate the velocity thereof, and damp the error signal accordingly. In the instant embodiment this signal is extracted from the armature loop voltage which by its inherent nature leads the shaft velocity in time.

The current delivered to the generator field and therefore the mechanical power output of the motor-generator set, is controlled by a pulse width modulated oscillator. The oscillator frequency is selected to be sufficiently high to ensure that no power surges will be felt and the "on" time of the generator field is varied by modulating the pulse width at the oscillator frequency. If the modulator, for example, is producing a negatively oriented pulsed signal, a driver associated with the negative supply will be producing a signal to the power transistor controlling the output of that supply. A set of gates ensures that, during the operation of a particular driver and power transistor, the opposite sense driver and transistor will be gated off. Dual clamps are provided to by-pass excessive voltage surges precipitated by the high inductance of the generator field, so that voltage in excess of a predetermined value is clamped to ground, and destruction of the power transistors is thereby avoided.

The modulator input is a function of the difference between the voltage output of the error means and a voltage that is proportional to generator field current. Therefore the power control means becomes a current source for the generator field, reducing the time lag inherent with an inductance driven by a voltage source.

It is therefore an object of this invention to provide a new and improved electronic system for controlling elevators.

It is another object of the invention to provide a new and improved electronic system for controlling elevators, that produces acceleration and deceleration of the elevator car without inducing significant sensations of jerking or other evidence of changing acceleration.

It is another object of the invention to provide a new and improved electronic elevator control system that is capable of a high system reliability.

It is another object of the invention to provide a new and improved elevator control system that closely follows velocity commands.

It is another object of the invention to provide a new and improved electronic elevator control system with reduced reliance on electromechanical components.

It is another object of the invention to provide a new and improved electronic elevator control system, that accelerates to the optimum speed for producing minimum transit times, without inducing objectionable sensations to the passengers.

It is another object of the invention to provide a new and improved electronic elevator control system that is capable of more precise control over the point at which it reaches zero velocity in stopping at a floor.

Other objects and many attendant advantages of the invention will become apparent from a reading of the following detailed description together with the drawings, wherein like reference numerals refer to like parts throughout and in which:

FIG. 1 is a diagram of the basic system.

FIG. 2 is a block diagram of the digital selector circuit.

FIG. 3 is a block diagram of the error loop and power control circuitry.

FIG. 4 is a side elevation view of the drum selector mechanism.

FIG. 5 is an enlarged face view of one of the coating discs of the drum selector.

FIG. 6 is a sectional view taken on line 6--6 of FIG. 5.

Referring now to the drawings, particularly FIG. 1, there is illustrated the electronic motion control system of the invention. The digital selector 10 is provided with a digital binary number from the drum selector 12 through a line bundle 14. The rotation of the drum selector is directly related to car movement because the drum is driven from a cable 16 which is wound around drum 18. One end of the cable 16 is affixed to the elevator car 20 and the other end to the counter-weight 22 through compensation spring 24. As the car moves vertically under the influence of the drive motor, transferring force through cable 26, the cable 16 is forced to move through the same linear translation and therefore the circumference of the drum 18 moves the same linear distance. The structure of the drum selector for producing a binary number, will be described more fully hereinafter. The digital selector 10 receives commands and information from the logic 28 through connection bundle 30 and provides the logic 12 with information as to car position and velocity through cable bundle 32. Also inputing the logic are car buttons 34 through lines 36, and the floor buttons illustrated at 38, through line 40.

The output of the digital selector is an analog of the desired velocity of the car, and this analog signal is inputed to the servo loop compensation system or error means 42 through connection 44. An analog of the actual velocity is developed by the tachometer 46 and is inputed to the servo loop compensation system by line 48. The servo loop compensation system compares the inputed velocities and determines the action of the drive control that will be necessary to maintain the actual velocity in very close conformity to the velocity commands from the digital selector. The commands from the servo loop compensation system to the drive control 53 are carried by line 52. The drive control causes increases or decreases in voltage to the generator field and through line 55 thereby controls power in the drive motor 19.

Referring now to FIGS. 4 through 6, the structure of the drum selector is illustrated. The drum 18 is illustrated as being driven by several turns of light aircraft cable 16 around its circumference. A drive shaft 57 connects the drum to a plurality of discs 54, 56 and 58. The discs are substantially identical and an exemplary disc 58 is illustrated in greater detail in FIG. 5. Each disc is divided into 16 radial segments and 4 concentric tracks. The track for generating the least significant binary digit comprises a plurality of openings 62, alternating with blanks of equal angular extent. The next digital positions are generated by the innermost track, utilizing a plurality of openings 64 alternating with equally spaced blank areas, followed by the next digital position, slots 66 positioned between the two previously described tracks. Finally, the most significant digit is created by an reduced diameter portion 59, of the disc 58 over one-half of its circumference. The arrangement of the tracks in the foregoing manner is advantageous in that it permits much thinner materials to be utilized, while retaining the same rigidity and ability to hold dimensional tolerances. This is due in part to the reduction in concentration of the lines of weakness, and in part, to the placement of the most significant digit the maximum distance from the center, so that a cut out is not necessary to achieve this digital position.

The system employs a plurality of light emitting diodes and corresponding solid state photocells located in close proximity to the disc along the tracks and aligned with the slots or apertures. In this manner, the rotation of the disc will cause the slots to expose the photocells to the illumination of the light emitting diodes producing alternating on and off indications from the photodiode. This information is readily convertible and may be stored in a register in accordance with the well known practice of binary signals.

The FIG. 5 illustration shows a plurality of photodiodes associated with each track. The outer track utilizes, for example, photodiode 60 to indicate the leading edge of the bit that is represented by the solid area immediately adjacent, photodiode 61 to indicate the trailing edge. In this manner a conflict between adjacent tracks is avoided, since the leading and trailing edge signals may be utilized to disable the following digit until the preceeding digit has been switched. The technique avoids transients in the car position register. A similar arrangement of photodiodes is utilized for the remaining tracks and the least significant track utilizes photodiodes 70 and 72 for the leading edge and trailing edge signals respectively. The adjacent track employs photodiodes 74 and 78 and the innermost track, diodes 80 and 82.

The disc 54 differs from the other discs, in that a single photodiode 63 is provided at the reference position for the least significant track since this position is required as a reference for all following digits.

Referring again to FIG. 4, the array of photodiodes is illustrated at 82 and the corresponding array of light emitting diodes at 84. The succeeding discs 56 and 58 have similar arrays of photodiodes 86 and 88 and light emitting diodes 90 and 92.

The mechanical interconnection between the discs 54 and 56 is through a gear box 102 that reduces the shaft speed of shaft 57 by a ratio of 16 to 1. Similarly, the mechanical interconnection between disc 56 and 58 is through a gear box 104 that reduces the shaft speed from the rotation of the disc 56 to the disc 58 by a ratio of 16 to 1. In this manner, each disc produces four binary digital positions for a total of 12 binary digital positions for the entire drum selector.

The photodiodes utilized with the drum selector may be selected to have their sensitivity in the infrared range of the light spectrum, and would be matched with light sources, such as the light emitting diodes illustrated that have light characteristics in the same wave length. In this configuration the system is relatively insensitive to ambient light conditions, reducing the chance of error signals. Additional error protection is provided by maintaining close dimensional tolerances on the disc, so that the photodiodes and light emitting diodes may be placed as closely as 0.150 inches from the disc surface. As a result the likelihood of stray light entering the range of sensitivity of the photodiodes is reduced. The close spacing is made possible, in part, by the ability to utilize relatively thin discs as the result of the structural strength provided by the arrangement of the adjacent tracks. In some environments it may be desirable to provide even further rigidity to the discs by skewing adjacent tracks. That is, it may be desirable to displace each track around the circumference so that there is no alignment of the leading edges of the slots. When this technique is employed, it will obviously be necessary to skew the photocell and light emitting diodes pairs an equal amount.

Referring now to FIG. 2, the drum selector 12 is illustrated as inputing the binary number generated through the rotation of the drive motor to the car position register 104 through cable bundle 14. The register 104 then contains a binary number, which corresponds to a specific position of the elevator car within the range of resolution of the system. The resolution of the system is dependent upon the number of discs, the angular spacing of the digits, the number of tracks, and the gear ratio between adjacent discs. For the system illustrated, it is possible to have a total of 4,096 positions. Thus for a 300 foot building, the selector of the preferred embodiment is capable of providing a resolution of plus or minus less than 1 inch.

The digital car position is utilized directly in the slow down system of the digital selector as will be described more fully hereinafter and is summed with an advance number in adder 103 through cable bundle 106. The number in the advance register 108 is determined by a signal provided through gate 110 and line 112. This signal is determined by a voltage control oscillator 114 through line 116. The frequency of the voltage controlled oscillator is determined by input through line 118 from the digital to analog converter 120. The operation of the digital to analog converter will be described more fully hereinafter, it being sufficient for the present purpose to indicate that the signal through line 118 represents a voltage analog of the desired velocity. The system functions in such a precise manner that the desired velocity and actual velocity are substantially equal and therefore the frequency of the voltage controlled oscillator will be an analog of the actual car velocity. The effect of the voltage controlled oscillator is to cause the register 108 to accumulate an increasing number as long as gate 110 remains open.

If, for example, in a particular system the equivalent distance of 4 floors is necessary for the elevator to decelerate, the advanced register will accumulate a number which will equal in binary code, that distance representing the four floors, at the same instant that the car reaches its maximum velocity. At this point, line 120 will gate off gate 110, to prevent further accumulation of the advance in the advance register 108, since the maximum advance necessary for deceleration is already stored therein. This advance is inputed through lines 122 to the adder 103, and summed with the car position, to produce an advance position binary number at its output on line 124.

The advance position in binary code is converted to floor position by floor decoder gating matrix 126. This matrix contains information corresponding to the binary position of each floor and therefore is tailored to the particular building to accommodate variable floor spacing. As the advance position of the elevator reaches each new floor this information is transferred in the form of a pulse to the floor indicator storage register 128 through lines 130. The floor indicator storage register 128 disables the appropriate selector relays at all those floors which the advanced elevator position has passed, through lines 132.

When the call answered portion of the logic determines from the advance position, traffic load, and other information programmed into it that a call received is to be answered, then gate 134 is opened by the coincidence of a signal on line 136, and a clock pulse on line 140 from clock 138. The gate 134, thorugh line 142, commands the floor position encoder 144 to deliver the binary number corresponding to the selected floor position for landing, to the storage register 160, by the 12 line bundle 158. The computer slow down logic gates on gate 162 at the instant that the advance position reaches the floor selected for landing. Thus the car has the appropriate distance remaining to decelerate according to its predetermined deceleration schedule. Deceleration is accomplished by forcing the car to follow a predetermined velocity curve. This velocity curve is based on the desired velocity for each of a discrete number of distances from the landing floor. The distance remaining to landing is generated by comparing the binary number in storage register 160, representing the landing floor position in binary increments, to the present car position, which is produced in car position register 104. The current car position is inputed to the subtractor 180 when allowed to pass gate 162 through line bundles 182 and 184. The binary number representing the distance to go is inputed to deceleration matrix 190 through line bundle 192. In the preferred embodiment 24 steps to the velocity curvature are utilized and they are inputed to shift register 194 by 24 line bundle 196. Thus the shift register comprises 24 flip-flops.

There are 12 diodes corresponding to the 12 decimal positions of the binary number for each output line to shift register 194. Thus, in the instant embodiment there would be 248 diodes necessary to input shift register 194 with a pulse, as the binary number representing the distance to go, corresponds with the number coded with diodes to the particular output line. Successive pulses on lines 196 reset corresponding flip-flops in the shift register and the voltage associated with the flip-flop change appears at the output bundle 200, when the clock 202 reaches an enabling time. The clock releases the pulses through the use of a gate 204. This operation is clocked as are the other operations in the digital selector for the purpose of reducing the possibility inadvertent actuation due to noise and other interference and also for the purpose of regularizing the output signal from the digital to analog converter. The signal is regularized to the extent that each step occurs at a precisely timed interval.

The digital to analog converter 120, sums the outputs of the flip-flops which appear on lines 200 producing a step function similar to that illustrated at 208, with the exception that the actual function, in the instant embodiment, would comprise the 24 steps to the maximum voltage.

The output of the converter is conveyed by line 210 to a rate filter 212 that performs the function of smoothing the output of the digital selector as is illustrated by the voltage versus time curve at 214. Additional input to the rate filter is provided through line 216 so that the level of filtration may be varied at selected voltage levels.

When the car has reached a floor, its velocity will have decayed to zero and the voltage analog will be at its minimum. If the car is then commanded by the logic to acceleration, the low frequency function of clock 202 is activated through line 220. The frequency of this low frequency clock output will be determined by the acceleration characteristics desired for the system. Each time the clock pulses, it will set one flip-flop in the shift register 194 through lines 224 and 226. Since the converter is summing the outputs of the flip-flops through lines 200, an increasing voltage will appear at the output 210 of the converter and will result in the increasing step function illustrated at 208.

Should the elevator have insufficient distance to accelerate to maximum velocity prior to the beginning of the deceleration cycle, the deceleration logic will discontinue the pulses initiated by the low frequency clock function and the distance to go information will again be delivered to the deceleration matrix as previously described. Since less than the maximum velocity was reached it is not necessary to begin deceleration until that portion of the curve, that corresponds to a lesser distance to go to the floor for landing is reached. When the binary number corresponding to this distance is inputed by line 192 to the deceleration matrix it will reset the last shift register flip-flop having been set during the acceleration phase, and call for a reduction in velocity causing deceleration of the car to that new velocity.

Thus, the digital selector produces an analog output that corresponds to a desirable velocity for the car at that particular time. The output is on a line 229 and is inputed to the drive control system through an input buffer 230 as is illustrated in FIG. 3. The input buffer isolates the two systems and provides a proportional output at line 232 to the greatest signal selector 234. The selector functions to select the larger of the two signals inputed to it, the second signal being inputed from a portion of the floor leveling system and being pertinent only at relative close spacing to the particular floor at which the elevator is landing. This floor leveling system will be described more fully hereinafter.

The output 236 of the selector will be the same polarity for both up and down velocities and therefore an input from the logic system it made to develop the proper negative or positive sense for the velocity analog voltage. Gate 240 or 241 is gated open by the logic during the appropriate phase of the elevator's operation, if the elevator is operating in a downward moving mode, then negative one multiplier 238 reverses its sense to correspond to the actual velocity, voltage analog. The appropriate signal is delivered by line 242 to junction 244. Here the voltage analog for the actual velocity of the car is subtracted to produce a difference output on line 246. In the instant embodiment the voltage analog of elevator velocity is developed by tachometer 46 driven by the motor shaft of motor-generator 19. The voltage is delivered to the junction by line 250. This voltage appears at amplifier K1 illustrated at 252 and the amplifier's output is delivered to summing junction 254.

Were the output of amplifier 252 used solely to control the motor in producing a velocity responsive to the velocity commands, a very poor response would result. This is due to the time contants and mechanical lags involved and such a system could be expected to produce varying results depending on the elevator load, temperature, and other variables. To prevent this time lag in the system's ability to follow the velocity commands, a second order servo 256 is incorporated in parallel with the proportional amplifier. The servo 256 integrates the error input with respect to time and the time is reset through lines 257 at the initiation of each acceleration or deceleration cycle. Since the system starts with zero error, both the actual and command velocity being zero, there will be no initial output from servo 256. However, if the actual velocity lags or leads the command velocity an error signal will be inputed to the servo. At the same time the proportional amplifier will produce a signal that ultimately commands power changes to correct the error. If the commands resulting from the porportional amplifier are not sufficient to correct the error, and a residual error remains over a sufficient duration, then the output of servo 256 will increase and will increase at an increasing rate. Summing junction 254 combines the outputs, and the power controls means will therefore be called upon to follow the velocity commands with great precision. Summing junction 264 combines the output of the junction 254 and a signal from the floor leveling system to be described more fully hereinafter.

The output of junction 264 is modified by difference junction 266 through input from the motor-generator armature circuit. The loop voltage in the motor-generator armature circuit has the characteristic of leading the actual velocity of the system and therefore subtracting this voltage from the value present at junction 266 by line 268 has the effect of damping the commands to the power control means so that power surges are avoided. The final output of the servo loop compensation system or error means through line 270 is amplified by an appropriate amplifier 272 for use in the power control means.

Returning to the input to summing junction 264, it is illustrated that this junction sums the signals resulting from the error between actual and command velocity, and the signal from floor leveling system for the elevator. The technique for developing the output signal created in the floor leveling system does not form a part of the invention herein but is included for completeness since it inputs the power control means in a manner similar to that of the remainder of the system. The output of the floor leveling sensor 274 may be taken to be a voltage analog of the distance from the level floor position that the car is at the moment. This system is designed for utilization within the immediate proximity of the floor and is extremely accurate, in that the means for determining the distance displacement is located physically in relationship to the floor, and because the voltage developed by the system is not dependent on the discrete binary number but rather is a voltage analog of actual displacement. The output of the sensor on line 276 is gated off by a gate 278 and remains off until the car is within a predetermined distance to the floor. The distance to the floor may be determined by a vane and light beam system or other means of measuring the car's displacement from floor center line. After the predetermined distance is reached, the gate allows the signal from the floor level sensor to pass through line 280 and is delivered to the squaring amplifier 282 by line 284. The distance value at which gate 278 is opened would normally be selected so that the output of the amplifier 282 is less than the expected value of the input on line 232 from the digital selector. Thus, the greatest signal selector will continue to pass the digital selector generator signal until the car is in close proximity to the floor center line. For example, in the instant embodiment the greatest signal selector will normally begin to pass the signal on line 286 only after the car is within approximately 4 inches of the floor. At these short ranges the voltage function generated by utilizing the displacement information from the drum selector may be insufficiently accurate for the purposes of local control.

When the squaring function generated by input relating to the distance from center line is greater than the velocity signal it is passed by the greatest signal selector and processed in the same manner previously described to cause power adjustments as necessary to produce a reducing velocity command. The floor level sensor output is also utilized to hold the car level with the floor. The parts of this system pertinent to its function with respect to the system of the instant invention include a proportional amplifier 296 that receives the input delivered on line 298 from the floor level sensor 274. This signal is passed by gate 300 when it is turned on through a signal on line 302 from the logic, indicating that the car is at rest on a floor. The same signal is utilized to reset the integrating amplifier 304. A third amplifier whose output is summed prior to gate 300 is differentiating amplifier 306.

The leveling system functions to detect any movement away from the center line of the floor and correct by the addition or reduction of power through the power control means. When the system is first activated and in balance the output would normally be zero, but, for example, if an additional passenger were to enter the elevator, some movement of the elevator could be anticipated due to cable stretch. Differentiating amplifier 306 provides the first response to this movement since it is sensitive to the derivative of position (velocity) and creates an immediate restoring signal, by commanding the power control means through summing junction 264 and proportional amplifier 272. Any position error generated by the additional weight will also result in a signal being amplified by proportional amplifier 296 to retain the restoring effort after the differential amplifier output dies out. Finally, the integrating amplifier 304 is effective to reduce and eliminate residual errors that are present over a sufficient time duration. Since the amplifier integrates with respect to time, the value increases at an increasing rate the longer the displacement error is present and will increase the power commands to the power control means until the elevator is restored to its level position.

The signal output from the servo loop compensation system is delivered to a junction 312 is in excess of the voltage across resistor 315 representing the current in the generator field, a positive error signal will be amplified by proportional amplifier 322 and will result in an increase in voltage to the generator field to rapidly increase current as will be described more fully hereinafter. Similarly, if the generator field current is in excess of the voltage from the servo loop compensation system, then a negative voltage will be amplified by proportional amplifier 322 and a reduction in the voltage to the generator field will result.

The proportional amplifier delivers the amplified positive or negative voltage to a pulse width modulator 324 by line 326. An oscillator 328 also inputs the pulse width modulator through line 330. The oscillator may desirably produce a relatively square wave output of, for example, 1 kilohertz. The effect of the pulse width modulator is to vary the duration of the on pulses in the output of the modulator. The modulator output is positive or negative in accordance with the sense of the signal input. Positive and negative drivers 342 and 344 are utilized to increase the power and isolate the modulator from the solid state switches. The drivers are responsive only to signals of the proper sense, however, as a means of preventing damage to the power switch that would result if both power switches were on simultaneously, the outputs of the drivers are utilized to gate off the opposing driver 342, gating off driver 344 through line 346 and gate 348. Driver 344 gates off driver 342 during its cycle of operation, through line 350 and gate 352. Solid state switches 354 and 356 are utilized to switch on the generator field supply with the appropriate sense, when driven by their corresponding drivers. As additional protection against simultaneous operation of the switches each switch operates a gate to prevent the operation of the opposing switch. In this manner switch 356 gates off gate 358 through line 360 and switch 354 turns off gate 362 through line 364.

The selected switch for the current mode of operation will deliver generator field voltage at the supply voltage which in the instant embodiment is 80 volts. However, the modulation of the square wave from oscillator 328 will determine the time duration of that 80 volt pulse which has the effect of a lesser voltage, in a well known manner. When switch 354 turns on it activates a clamp 372 through line 374. The clamp is necessary to deliver to ground a portion of the high reverse voltage transient which results from the removal of the switch voltage to the generator field. The voltage transient is high energy due to the high inductance of the generator field, which may be on the order of 5 henries. The clamp 372 is preselected to deliver all voltage transients above a predetermined level to ground. In this manner the voltage may be pre-set to correspond to the level above which damage to switch 354 would occur. A similar clamp 382 operates in the same manner with respect to switch 256 and is controlled by switch 356 through line 384. Thus the switches 354 and 356 deliver voltage of the proper sense to the generator field and therefore rapidly develop the current flow in the generator motor armature circuit necessary to produce the power changes commanded.


In an elevator installation incorporating the motion control system of the invention, the operation characteristics may be tailored to the particular installation. For example, the exact spacing between floors will be incorporated in the floor decoder gating matrix so that deceleration to the floor will be as smooth between floors spaced by 20 feet, such as with a building lobby as floors spaced by 10 feet in the upper stories. Additional tailoring is possible by a selection of a particular deceleration characteristic desirable for the user's requirements. This characteristic is entered into the system by changing the deceleration matrix as desired. Finally the system is adaptive in that it compensates for variables such as variable motor-generator operating characteristics, differing load, and variable cable stretch.

A typical operation might commence with the elevator being located in a resting position at the first floor. If a call is received from one of the upper floors, the logic will command the car to leave its position and answer the call. After door closure, acceleration will commence, and will produce a constantly increasing velocity to the maximum selected whereupon the velocity will remain constant. During this operation the logic will be inputed with the advanced position of the car. The advanced position corresponds to the first floor at which the car can stop due to its present velocity. Thus, when the advance position corresponds to the selected floor, the logic will indicate that the call is answered and at the same time initiate deceleration. Any floor selector, which at the instant of actuation is behind the car's advanced position will be inactivated and therefore will not produce a deceleration since the car is incapable of stopping at that floor without uncomfortable deceleration. The deceleration sequence proceeds while the logic stores the floors being subsequently selected for future answering. The basic information provided to the deceleration system is generated by a comparison of the car position in binary digits to the binary number corresponding to the selected floor's height. As the car approaches the floor, the car position will approach the floor position and therefore the difference between the binary number for the floor position and that for the car position will represent a decreasing number. This number is inputed to the deceleration matrix that includes a discrete number of deceleration steps, each step being identified with a particular binary number. When the binary number for a particular step is reached a pulse will be affected to reset a flip-flop in the shift register. Since there is one flip-flop for each input the summation of the flip-flops output will result in a step function and this step function is a voltage analog of the preprogrammed desired velocity for that point. Given the proper velocity for any particular point, the remainder of the system functions to ensure that the actual velocity closely follows the desired velocity. This is basicly accomplished by comparing the actual and commanded velocity and monitoring the system with respect to time to eliminate residual errors. The deceleration will continue until the car is in close proximity to the selected floor. At a predetermined point, approximately 4 inches in the instant system, local control will take presedence over the control provided through the digital number from the drum selector. The local control is analog in nature and therefore provides a smooth continuous signal rather than a step function to produce velocity commands that decrease until the car reaches zero velocity at the floor level. While the car is at the floor during the loading and unloading of passengers, for example, the system holds the car position through the use of motor power rather than applying a braking system. This results in a much smoother take off because the motor has already taken up the load as required. In another mode of operation when traversing between floors that are insufficiently spaced to allow the car to reach its maximum velocity the system nevertheless produces a smooth acceleration and deceleration without jerking because the acceleration is terminated and the deceleration initiated at corresponding points on the respective curves.