US3070071A

US3070071A - Failure protected signal translating system

Info

Publication number: US3070071A
Application number: US14301A
Authority: US
Inventors: Arthur F Cooper
Original assignee: North American Aviation Corp
Current assignee: North American Aviation Corp
Priority date: 1960-03-11
Filing date: 1960-03-11
Publication date: 1962-12-25
Anticipated expiration: 1979-12-25

Description

Dec. 25, 1962 A. F. COOPER 3,0

- FAILURE PROTECTED SIGNAL TRANSLATING SYSTEM Filed March 11, 1960 4 Sheets-Sheet l i l I67 I72 I nee I76 n4 I69 I77 OUTPUT 5p 0 INPUT I I75 0 I760 I660 0 x T l l I I700 I720 FIG. 5

e4 26 INPUT 24 K KI 22 42 INVENTOR.

ARTHUR F. COOPER ATTORNEY Dec. 25, 1962 A. F. COOPER TINPUT 4 Sheets-Sheet 2 76 g a 72 B2 74 52 r/ 5 r PICK PICK PICK OFF VALVE =F0FFI r o ll- :4 e2 9O 40 IO FIG. 2

INVENTOR. ARTHUR F. COOPER Qua, 7 ll ATTORNEY Dec. 25, 1962 CQOPER 3,070,071

FAILURE PROTECTED SIGNAL TRANSLATING SYSTEM Filed March 11, 1960 4 Sheets-Sheet 3 2| I I g m 2 Z 11' I H II I 3 2| 2 Q 09 a 8 F6 Q g 7 K75 fi a a U q- N no a J E Q 5 rcu 8 Q 5 8 I33 I I FIG. 3

INPUT INVENTOR. ARTHUR F. COOPER ATTORNEY FAILURE PROTECTED SIGNAL TRANSLATING SYSTEM Filed March 11, 1960 4 Sheets-Sheet 4 OUTPUT INVENTOR. ARTHUR F. COOPER ATTORNEY United States Patent 3,070,071 FAELURE FRQTECTED SIGNAL TRANSLATING SYSTEM Arthur F. Cooper, Santa Ana, Calif., assignor to North American Aviation, Inc. Filed Mar. 11, 1960, Ser. No. 14,301 Claims. (Cl. 121--41) This invention relates to signal translating systems, and particularly concerns a multi-channel system which pro vides protection against partial failure.

The problem of apparatus reliability and longevity becomes of greater significance with increasing complexity of modern equipment. Closed loop or feed back systems are commonly utilized links in the over lengthening chain of components employed in electrical, mechanical, electro-hydraulic and electro-mechanical apparatus. A hydraulic servo system, for example, may include a hydraulic motor having an electrically-operated control valve and an error generating electrical system producing a driving error signal for operating the control valve in accordance with the difference between a commanded input and the motor output. In such a servo system, failure will occur most frequently in the electrical system or in the motor control system. Servo failure may be such as to provide either zero or maximum output for any given input. The latter type of failure, which may be termed hard over failure, is most troublesome since it prevents control by an alternate or manual system without totally disabling the failed system. Accordingly, it is an object of this invention to provide a system in which, while failure is a recognized possibility, the disadvantageous effects of failure are minimized and increased reliability is achieved.

The principles of the present invention are applicable to a wide variety of different types of signal translating systems whether either or both input and output signals are electrical, mechanical, fluid or the like.

In carrying out the principles of this invention in accordance with a preferred embodiment thereof, there is provided a signal translating system embodying a plurality of substantially similar feedback or closed loop channels, each of which includes an output or power device which is operable in response to a driving error signal. The driving error signal in each of one or more channels has combined therewith a portion of the driving error signal in one or more of the other channels. The outputs of the individual channels are summed. That is, they are arranged to provide a combined output drive from the system in accordance with the sum of the outputs of the individual channels.

In one specifically described embodiment, for example, the output devices of each channel are motors comprising valve-controlled fiuid motors each arranged with a feedback loop to provide a position follow up servo. All channels are responsive to a common command signal input or to mutually equivalent command signal inputs and all outputs are additively combined by suitable mechanical connections. Cross feed of all error signals (which drive the respective motors) is provided by combining with each error signal a portion of each of the other of the error signals.

With the above-described arrangement, using solely two channels, the combined mechanical output of the twochannel system will reflect very little or none of a hardover failure (a steady maximum output) of a single channel. The system merely becomes inoperable whereby a monitor system may be used to de-energize the failed channel. If three or more channels are utilized, upon failure of one channel the system will continue to provide linear operation without any monitor system.

An object of this invention therefore, is to increase the reliability of a signal translating system.

Still another object of the invention is to minimize the disadvantages of diiferent types of servo failure.

A further object of the invention is to provide a multichannel servo system wherein the output will not reflect the failure of any single channel.

Another object of the invention is to provide a multichannel servo system capable of operation after failure of one channel.

Still another object of this invention is the provision of a multi-channel servo system which will provide continuous operation in the presence of failure of one or more of its channels.

These and other objects of the invention will become apparent from the following description taken in connection with the accompanying drawings in which:

FIG. 1 is an illustration of a two-channel servo system embodying the principles of this invention;

FIG. 2 illustrates certain details of a three-channel system;

FIG. 3 illustrates a mechanical-hydraulic system;

FIG. 4 illustrates a modified mechanical-hydraulic system, and

FIG. 5 illustrates a closed loop rate system.

in the drawings, like reference characters refer to like parts.

The term servo or servo system as utilized herein designates :a conventional closed loop or feedback system having an output produced in accordance with an input and wherein a driving error signal representative of the difference between input and output is applied to drive a signal translating or amplifying device within the system.

The system of the invention incorporates for each of its servo channels :a substantially conventional servo system including a signal translating device such as an amplifier or a motor, whether electrical, hydraulic, or pneumatic, and a feedback loop providing a conventional closed loop servo system. In accordance with certain embodiments of the invention, two or more of such channels are interconnected by feeding the error signal in each to the input of the others. The mechanical outputs of the motors are combined so that while each motor may operate independently of the others (except for the cross feed) the system output comprises displacement of a driven member in accordance with the sum of the mechanical outputs of the several motors.

As illustrated in FIG. 1, a single conventional electrohydraulic servo channel includes :an electrical-to-mechanical converter such as a valve-operated fluid motor 10. Motor 10 comprises a cylinder 12 reciprocally mounting a piston 14 under control of an electrically-operated valve 16 which supplies fluid under pressure to one side or the other of the piston 14 in accordance with the polarity of an electrical valve control signal appearing on lead 18. The valve control signal on lead 18 is provided as the output of an amplifier 20 which receives the driving error signal appearing on lead 22 at the output of a summing network 24. As in conventional servo operation, the summing network 24 receives a command signal at input 26 and also an output position feed back signal on line 28 from a position pickofr 32. The pickofi 32 may comprise a potentiometer 33 having relatively

movable arms

34 and 35 fixed respectively to the cylinder 12 and piston 14 whereby the potentiometer pickofi provides an output voltage proportional to the relative displacement of piston and cylinder. The error signal output of the summing network 24 is, of course, the difference between the inputs thereto on

lines

26 and 28.

The above-described position follow-up servo channel comprises nothing that is novel in itself and is considered to be exemplary of but one of the many types of similar closed loop servo systems well known to those skilled in the art. The second chanel of the two-channel system illustrated in FIG. 1 is substantially identical to the first channel described above and comprises a valve-operated motor 40, cylinder 42, piston 44, valve 46, amplifier 56, summing network 54, and pickofi 52, all constructed and arranged as are the corresponding elements of the first servo channel.

In order to provide fail-safe operation of the system of FIG. 1, the two channels are interconnected by cross feeding the error signal in each to the input of the other and by combining the outputs of the individual motors. The error signal in the first channel which appears at the output of summing network 24 is fed through a network 60, having a gain K as a third input to the summing network 54 of the second servo channel. Similarly, the error signal appearing at the output of the summing network 54- of the second servo channel is fed through a network 61, having a gain K as a third input to the summing network 24 of the first servo channel. The gains K and K are preferably equal and less than unity, and must never be greater than unity.

A convenient, but exemplary arrangement, for adding the individual output displacements X and X of the

respective motors

10 and 40 is illustrated as simply comprising a rigid interconnection of the two

cylinders

12 and 42. With this arrangement, either piston such as the piston 14 of motor 10 is connected to a base or support member 62 while the other, piston 44 of motor 40, will be secured to a member (not shown) which is to be driven by the system. Thus, the total output dis placement X of the dual channel system relative to the support member 62 comprises the sum of output displacements X +X While the inputs to the two channels are illustrated as being derived from a common input terminal 64, it will be seen that individual equal or equivalent inputs may also be provided.

It may be readily appreciated that any number of channels can be combined in the manner described above in connection with FIG. 1 with the error in each channel being fed, all with equal gains, to every other channel which is used. Further, all of the individual outputs X X X may be added in such an arrangement to give the total output X Such an arrangement is illustrated for three channels in FIG. 2 which also illustrates details of the summing networks and cross feeding of error signals. As illustrated in FIG. 2, the first channel may comprise the previously described fluid motor 16 with its valve 16 and pickoflf 32, amplifier 20, and a resistive summing network. Similarly, the second channel will comprise, a described in FIG. 1, the fluid motor 40 with its valve 46 and pickoff 52, amplifier 50 and a similar resistive summing network. The third channel will be substantially identical to the other two channels comprising a fluid motor 65 having a control valve 66 and pickoif 67, together with an amplifier 68 and a similar resistive summing network. The error signals in the respective channels appear at the

junctions

70, 71, and 72 of the several resistors of the respective summing networks to comprise the input to the

respective amplifiers

29, 50, and 68. In the first channel (which operates motor the common input signal is fed to one input resistor 72 of the summing network while the feedback signal from pickoff 32 is fed to a second resistor 73 of the summing network. The error signal in the second channel [from junction 71 is fed to resistor 74 of the first channel summing network while the error signal in the third channel at junction 72 is fed to resistor 75 of the first channel summing network. Similarly, each of the second and third channels has the common input signal fed to summing

network resistors

76 and 77 respectively and its individual feedback signal from pickoifs 52 and 67 fed to summing network resistors 78 and 79 respectively. In the second channel, summing network resistors 80 and 81 receive the error signal from the first and third channels at junctions 7t and '72 respectively. In the third channel, the summing

network resistors

32 and 83 receive the error signals from the junctions and 71 of the first and second channels respectively.

When these error cross feed networks are pure resistors, pairs of resistors in parallel can be physically combined into one resistor, such as 83 and 81, 74 and 80, and 82. However, separate resistors are shown for ease of visual perception and understanding of the basic operations.

It will be seen that the error signal in each individual servo channel is summed with the error signal in each of the other servo channels with the same cross feed gain between any two channels since all of

resistors

74, 75, 80, 8t, 82 and 83 are of equal value. The cross feed resistors are so chosen that the cross feed gain is less than one-half for this three-channel system. That is, the magnitude of the cross-fed signal component appearing at any one summing network output is less than the magnitude of the error signal from which such component is derived. The gain may be more rigorously defined by:

where K =cross feed gain, nznumber of channels. The expression on the left side of the equation must always be positive.

it may be noted that gain (or resistor) matching in the error networks is not critical. Mismatching, alone, will not cause instability.

As in the two-channel system, a system involving three or more channels will have the outputs of the individual channels combined. Thus, as illustrated in FIG. 2, the outputs of motors it) and 40 are combined just as described in FIG. 1 while the output of motor 65 is combined with the other two motor outputs by having the cylinder thereof secured to the piston of motor 40. Thus, it will be seen that the motors are connected in tandem each motor having first and second parts, such as its cylinder and piston, mounted for relative motion in response to the error signal input which drives the motor valve. Each motor has one of its parts, such as the cylinder part of motors 1t) and 46, fixedly connected to one part of the next adjacent motor while having the other of its parts, such as the piston 14, mounted for motion relative to the other part such as piston 44 of the next adjacent motor 40.

It will be seen that in the embodiments described, as in others to be described hereinafter, the several channel outputs comprise mechanical motions that are mutually independent. For example, the output of motor 10 is the motion of cylinder part 12 relative to piston part 14. This motion is in no way affected (in the absence of cross feed of error signals) by the output of motor 40 which is the motion of cylinder part 42 relative to piston part 44. The several mutually independent outputmotions are combined as illustrated. With this arrangement, for example, one channel may be locked without preventing operation of another.

In normal operation (FIG. 1) the error signal in each channel cross feeds into the other channel and thus enhances and supports the individual outputs X and X of the respective channels. If one channel, such as that providing the output X fails and goes hardover due to such failure, its error becomes large and feeds into another channel such as the channel providing the output X In this'situation, the second channel operates in response to a total error signal having as one component thereof its own normal operation error signal and having as a second component thereof a portion of the error signal of the failed channel. In such a situation the output X of the operating channel becomes of opposite sense to the output X of the failed channel. Since the system output X is the summation of the output X and X the total output remains zero or at a relative small value depending upon the magnitude of the fractional cross feed gain. This cross feed gain, while being less than unity, may have any desired value such as, for example, 0.90-0.98. Preferably, the cross feed gain is as close to unity as possible. The system is unstable for a gain greater than unity. Due to inherent lack of absolute precision of the gain controlling components and to insure stability, the gain in this system is reduced.

In the above-described operation of a two-channel system with a hardover failure, the system output does not become maximum but does, in effect, provide a fail-safe condition. Nevertheless, the system is no longer operable. However, it will be readily appreciated that the failure of any single channel may be sensed by any suitable monitor system well known to those skilled in the art which may be utilized to shut off the power to the failed channel and recenter the latter to zero. Thus, the remaining channel could still be utilized to provide the total output.

The cross feed system lends itself to a very simple monitor system for a hardover failure (the most common type). A large pressure differential exists across the piston of the bad channel, while the good channels have zero pressure differential. Pressure switches with suitable time delay can positively sense the bad channel. Auxiliary devices can cut out the bad channel and allow the good channels to continue operation.

This monitor sensing capability is one of the appeals of using the cross feed system compared to other multichannel schemes, particularly for a dual channel scheme Where continued operation after a hardover type failure is desired.

It will be noted, however, that if more than two channels are used and combined as illustrated in FIG. 2, for example, the system will continue to operate linearly without any monitoring system even after failure of one channel.

A steady state operation under normal and failure conditions for both two and more than two (11) channel systems is shown in the table below in terms of expressions relating the individual outputs X and X and the total output X to the input R. Operation under two types of failures is described. The first type of failure is defined as a zero type wherein any one individual output such as X goes to zero and remains there for any input R. A second failure, the hardover type, is defined as a failure such that any one individual output such as X; goes to its maximum limit value and remains there for any input R.

channel system where n=2, the cross feed error gain K must be less than 1.

From the above table, it will be seen that zero failure of one channel of either a dual or n system will not prevent continued linear operation. Hardover failure in a two-channel system provides a fail-safe condition wherein the outputs X remains zero or a small value, depending upon the gain K With hardover failure of a single channel of an n channel system, only a small fraction of the failed channel output appears in the total system output X whereby the system will continue to operate linearly with its two or more non-failed channels. For example, with reference to FIG. 2 if hardover failure of the channel including the motor 65 occurs such that the piston of this motor is moved to the extreme left in the drawing, the cross feed from the failed channel to the other two channels will operate to cause the piston of motor $0 to move toward the right relative to its cylinder and the cylinder of motor 10 to move toward the right relative to its piston an amount sufficient to compensate for the failed position of the motor 65. The latter may thus remain in its failed limit position while motors 10 and continue to operate linearly in response to the system input.

In the triple channel system of FIG. 2, the desired output of the system comprises displacement of a member (not shown) which will be secured to the piston of motor 65 relative to a fixed support 62 which is secured to the piston of motor 10. It is to be understood that any suitable means may be provided to constrain relative motion of the several motors. In the disclosure of the relative motion of the individual motor parts such constraint is schematically illustrated in FIG. 2 as rollers 90 mounted between the several motors and the base or support 62. Additional rollers or slidable constraint may be provided if deemed necessary or desirable.

Illustrated in FIG. 3 is a mechanical-hydraulic system having a different type of summing of outputs, a mechanical signal input, and a mechanical linkage arrangement for the error network and cross-feed. There is provided a mechanically operated valve assembly 101 which is arranged to feed high pressure oil to or from actuator 104 through ports 102 and 103. The ram or piston 105 of the actuator moves left or right, depending on which side of the ram is supplied with high pressure. A second identical system of valve 101a and actuator 104a is provided for the dual system illustrated. Ram 105 is connected to a mechanical summer or walk- Normal Zero Failure Hardover Failure X /R Xj/R Dual system 28 S n Channel sys- 118 S tem.

In the above table,

It will be seen that for a multiple channel system of any number (n) of channels, the cross feed gain between any two channels (equal for all channels) must be such that (n1)K is less than unity. Therefore, in a dual ing beam 106 at pivot point 107, while the piston a is connected to walking beam 106 at pivot point 107a. The motions of the rams 105 and 105a are summed through the beam 106 to produce the net desired output in the form of motion of an output member 108 pivoted to the beam 106 at point 108a.

The valve 101 is operated by input motion left or right supplied through link 109 which is pivoted to a valve clapper arm 110. It is to be noted that this description of the operation of valve 101 is equally applicable to valve 101:: and its actuator 104a. Parts in the second channel of the dual servo system are designated by numbers corresponding to like parts of the first channel with the ad dition of the suffix a.

Clapper arm 110, pivoted in the body of valve 101, is operated by input link 109 to close either of orifices 111 or 112 formed in the valve body. High pressure from a pressure source, not shown, is supplied to chambers 113 and 114 and via

pressure dropping orifices

115 and 116,

flows to return chamber 11"] through one of orifices 111 and 112 which is not closed by the clapper arm. A valve spool 118 is mounted in the valve body and normally centered by springs 119 and 120. Upon closing of one of orifices 111 or 112, there is created a differential pressure across the valve spool in

chamber

121 and 122 provided between the respective ends of the spool and the valve body. This differential pressure moves the valve spool to the left, for example to a point where communica tion between pressure chamber 113 and conduit 102 is provided together with connection between chamber 117 and conduit 103. in such position the ram 105 is moved toward the right in the illustration as long as the pressure differential is maintained across the valve spool. Upon movement of the valve spool to the right by creation of an opposite sense pressure differential, conduit 103 is in connection with high pressure chamber 114, which conduit 102 connects with return chamber 117.

To summarize the operation of the valve actuator com" bination for one direction of motion, arm 109 moves to the right, causing clapper 110 to close orifice 111. and provide a larger pressure in chamber 121 than in chamber 122. Spool 118 therefore moves to the right provid ing high pressure from chamber 114 through conduit 103. A return path is now provided from conduit 102 to chamber 117. Ram 105 therefore moves to the left and continues to so move until input arm 109 is returned to neutral position by the error forming linkage to be subsequently described. When arm 109 is returned to neutral, the pressure in

chambers

121 and 122 is equalized and the springs return the spool 118 to neutral whereby the ram 105 stops moving.

The mechanical linkage illustrated between the two valves, provides cross feed of error to each valve input and also provides the position feedback which completes the closed loop of each servo system. The common input signal to each servo system is a mechanical signal in the form of motion to the left or to the right of an input link 123. Input motion is transmitted through link 123 to a lever 124 which pivots about point 125. Point 125 is held fast by connecting arm 126 which is rigidly attached to the ram 105. Pivotal motion of link 124 about point 125 is transmitted through a pivotal connection 127 to a link 128 pivoted at 129 to one end of a link 130. The input arm 109 is pivoted at 131 to an intermediate portion of link 130.

The input motion of beam 123 is applied to the clapper driving arm 109a of valve 101a via similar

linkage including links

124a, 128a, and 130a. Accordingly, upon motion of input beam beams 124 and 124a pivot about

points

125 and 125a to move

links

128 and 128a to the right. There is provided a link 133 rigidly affixed to link 128 and pivoted to link 130a at point 1320. There is provided a link 134 rigidly affixed to link 128a and pivoted to link 130 at point 132. Thus, upon motion of

beams

128 and 128a to the right, beams 130 and 130a move to the right without pivoting since both

links

133 and 134 move equally to the right. I

Upon movement of input arm 123 to the right, the above described linkage moves both

arms

109 and 109a to the right in equal amounts and operates the valve actuator system as previously described. Accordingly, both rams 105 and 105a move to the left equal amounts producing the desired output motion of arm 108 which is equal to one-half the sum of the motion of the two rams.

Position feed back results when motion of the ram such as 105 is transmitted by link 126 to point 125 moving this point to the left. This causes arm 124 to pivot about its pivotal connection to the input beam 123 to move arm 128 to the left and return arm 109 to neutral whereby the motion of the ram is stopped. Similarly, link 126a provides position feedback for the second channel.

The error signal in each servo system is, in this embodiment, the motion of the respective beams 128 and 12811.

123 to the right, for example, both The error motion of beam 128 is cross fed to the other channel by link 133 while the error motion of beam 128a is cross fed by means of link 134.

Under failure condition, it is assumed that, for example, a metal chip is lodged between the valve spool and the valve body such that actuator 104 is continuously pressurized to move ram hardover to the left where it remains regardless of input motion of beam 109. Motion of ram 105 to the left is transmitted by link 126 to move pivot point to the left thereby pivoting arm 124 about input beam 123. Motion of arm 128, the error, is to the left and via cross feed link 133 causes arm a to pivot to the left about pivot point 129a. This causes arm 109:! to move to the left and effect motion of ram 105a to the right. Ram 105a moves to the right substantially as far as ram 105 moves to the left and no net output motion of arm 108 results. Right motion of ram 105a is transmitted via 126a to link 124a which pivots about its connection to link 123. This moves link 128a to the right to bring link 109a to neutral position. Link 109 also returns to neutral by virtue of the cross feed motion through link 134 but under the asumed failure conditions, link 109 has no effect upon ram 105.

It is understood that little force is required to operate valve arms 110 and 110a while input beam has a substantial source impedance.

The cross feed gain is determined by the relative lengths of the two portions of

links

130 and 130a, respectively. Thus, to prevent re-generation and provide a gain of less than unity, the top section of the link between pivot points 129 and 131'is made slightly shorter than the bottom section of the link between pivot points 131 and 132. With the cross feed gain of less than unity, the net output would not be exactly zero under the described failure conditions.

The system of FIG. 3 is a direct corollary to the previously described electrical cross feed error system wherein the error of any one system is continually fed to the other system. Two separate actuators are utilized and the outputs added through a Walking beam arrangement which is an alternate way of mechanically combining outputs of hydraulic actuators.

Illustrated in FIG. 4 is a somewhat different embodiment which employs cross feed of error in a manner somewhat different than that shown in the other systems. Further, the operation of the system of FIG. 4 during failure differs in that the output motion can reflect directly through the valve motor combinations to the input.

The system of FIG. 4 includes a pair of hydraulic actuators and 140a connected to a fixed support structure 143. The actuator cylinders are connected at pivot points 144 and 144a to the respective ends of a summer or walking beam 145 having an output shaft 146 pivotally connected at 147 to an intermediate point thereof.

Valves

148 and 148a are rigidly affixed to the respective actuators to move therewith. The valves include valve spools 149 and 149a which are directly operated by mechanical motion of

input arms

150 and 150a respectively. The input arms are connected in turn to opposite ends of a beam 151 at pivot points 152 and 1520 respectively. The common input is provided by motion to the right or to the left of an input shaft 153 which is pivotally connected to the mid-point of beam 151 at point 154.

If input beam 153 is moved to the right, the motion is transmitted to the two servo channels and moves

arm

150 and 150a to the right. The valve spools 149 and 149a move the the right equally relative to the valve housings. High pressure oil which is provided at all times from a source (not shown) to

valve ports

155, 156, 155a and 156a is transmitted through

ports

155 and 155a to actuator chambers 157 and 157a. Also upon motion of the valve spools to the right,

valve chambers

158 and 158a are connected to the valve port's 159 and 1590 which are at all times connected to relatively low pressure or returns. Accordingly, the actuator cylinders move to the right (relative to common support 143) until the valve spools 9 are again returned to neutral (illustrated) wherein the

ports

155, 156, 159, 155a, 156a and 159a are blocked. The apparatus is arranged so that the force required to move the valve spools is substantially less than that required to move input beam 153. Therefore, upon motion of the actuator cylinders, the valve bodies which are fixed thereto are moved relative to the valve spools. This constitutes a position follow-up. That is, output motion of an individual actuator results in an input motion of its own valve spool in a sense and amount sufiicient to null the input to the individual servo channel. The spools now having returned to neutral, the actuators are fixed in position. The net motion of actuator cylinders 144i and 140a to the right is summed in beam 145 and produces the net output in arm 146.

Assuming a failure condition exists such as, for example, the lodging of a chip between a valve spool 149 and its housing 148, chamber 157 of actuator 140 is subjected to continuous high pressure while chamber 153 is subjected to continuous low pressure. Actuator 140 moves to the right until a limit position is reached. The chip of the assumed failure condition prevents closing of

ports

155 and 159 by the normal follow-up action. That is, rela tive motion of valve spool and body cannot occur because of the chip. Therefore, as the actuator cylinder 1.4% is moved to the right, the valve spool 149 together with its operating arm 150, is dragged to the right. Since no input exists in this assumed failure, the second channel of the dual system (that including the actuator 14%) is still at neutral. A torque is applied to beam 151 by virtue of the motion of arm 150 and the resistance to motion of the input arm 153. Therefore, beam 151 rotates about pivot point 154 (which has some threshold friction level) whereby the upper end of beam 151 moves to the right and the lower end of the beam moves to the left, both in equal amounts. Arm 150a and its spool 149a are moved to the left causing actuator 140a to move to the left an amount equal to the motion of actuator Mil to the right. The motion of the two actuators, one to the right and the other to the left, is summed in beam 145 whereby the net motion of the output arm 146 is zero in this condition of hard-over failure of the upper channel of the system.

The failure of actuator 140 is reflected in motion of arm 150 of its valve. This is effectively an error which is cross fed to the second system by means of rotation of beam 151 about pivot point 154 which creates a resultant motion of arm 150a of the second channel.

Illustrated in FIG. is an example of the application of the principles of the present invention to an electromechanical dual rate or integrating servo system (as distinguished from previously described position systems) 1f] the motor 169. Thus, the output X is the time integral of the input on lead 165. That is, for a constant input, the output X increases at a constant rate. Also, the rate of change of the output voltage is proportional to the input.

There is provided a second identical rate servo also driven from the common input 165 and providing an output X This second channel includes the error summing network 166a, amplifier 167a, motor 169a, tachometer generator 172a, gearing 171a, and pickoff potentiometer 175M, all illustrated and arranged as are the corresponding elements of the first channel of this dual rate system.

The outputs X and X of the two servo channels are electrically combined in summing network 174 to produce at output terminal 175 the total system output X The driving error signal of the several channels appears at summing points 176 and 176a of the respective summing networks. Each driving error signal is fed to the summing network of the other channel via the common resistor 177 which provides the desired cross feed gain.

Operation of the system under failure condition is similar to the electro-hydraulic position systems of FIGS. 1 and 2. Three types of failures are protected against in the system of FIG. 5.

(1) Either of the individual outputs X or X is zero and the pickoff potentiometer associated therewith does not move.

(2) Either X or X is constant but does not change with time.

(3) Either X or X runs away at a constant rate.

It is assumed that the tachometer generator is operative as long as its associated motor is operative. For the first menlioned failure, the non-failed system picks up the entire load and the output is approximately the same as it would be in the absence of failure, depending upon the value of the cross feed gain. It is noted that in this rate system, it is preferable that the cross feed gain be made equal to unity and that in this situation, the gain can be greater than unity without instability.

For the second type of failure, the non-failed system picks up the entire load, but now the output has a con stant (position) bias due to the constant output of the failed channel. Nevertheless, the output rate, the rate of change of X is still proportional to the input.

For the third type of failure wherein the failed system runs at a constant rate, the second system by virtue of the cross feeding of error (which in this case represents rate) is caused to run at a rate equal and opposite to the failure rate whereby the net output X is zero.

The following table is descriptive of the rate system of of FIG. 5.

employing electric cross feed of error for failure protection. The common input signal is fed from input terminal 165 to the summing and error forming network 166 of the first channel servo. The first channel servo includes ;put signal at lead 165. The output voltage of the potentiometer 170 is the individual channel output X which is proportional to the total angular shaft displacement of K-Gain of amplifier and motor combination.

K Gain of tachometer.

K Cross feed error gain.

S-Scale factor depending on gear-train and input and output and pot scaling.

xf-A constant rate of x output.

R-Input.

In the system of FIG. 5, there are no critical parameters. Parameters in each channel should however be of substantially equal value. The cross feed gain does not have to be less than unity as in the position system and shoud be as close to unity as possible.

There have been described a number of different mechanizations of a multi-channel servo system which is protected against failure without necessity of external dividual thereto,

monitoring apparatus. While the systems of FIGS. 1, 3, 4 and 5 have been described in connection with two channel systems, it will be readily appreciated that each of these systems can be mechanized utilizing three or more channels, combining all outputs and cross feeding all driving error signals substantially as described in the embodiment of FIG. 2. Each of the described systems employs an improved structure and method for operating a number of closed loop signal translating systems from a common input by virtue of combining with the error signal of one servo loop at least a portion of the error signal of another of the servo loops and further combining the outputs of the servo loops.

It will be seen that the above-described interconnection of conventional servo channels provides a highly reliable servo system wherein all of the components are utilized for normal operation. Failure of one channel of a dual channel system will have little effect on the system output while operation of a three or more channel system will be continuous despite failure of any one. Neither gains nor balance are critical providing only that the cross feeds be limited as described above.

Although the invention has been described and illustrated in detail, it is to be clearly understood that the sarre is by way of illustration and example only and is not to be taken by way of limitation, the spirit and scope of this invention being limited only by the terms of the appended claims.

1 claim:

1. A multi-channel servo system comprising first and second electro-mechanical converters each providing a mechanical output in response to an individual electrical error signal input, means for adding said converter outputs, first and second pickofif means for generating first and second pickoft signals in accordance with respective converter outputs, first and second summing means each individual to a respective converter for algebraically combining a common input control signal with respective pickoff signals to produce said individual error signals, and cross-feed means for modifying at least one of said error signals in accordance with the other.

2. A multi-channel servo system comprising first and second electro-mechanical converters each providing a mechanical output in response to an individual electrical error signal input, means for mechanically connecting "said converters to provide a combined system output in accordance with the sum of the individual converter outputs, first and second pickoif means for generating first and second pickoii signals in accordance with respective converter outputs, first and second summing means each individual to a respective converter for algebraically combining a common input control signal with respective pickoif signals to produce said individual error signals, and cross-feed means for adding to at least one of said error signals a portion of the other.

3. A multi-channel servo system comprising first and second electro-mechanical converters each providing a -mechanical motion output in response to an individual electrical error signal input, each converter comprising first and second parts mounted for relative motion, each converter having one of its parts fixed relative to one part of the other converter and having the other of its parts mounted for motion relative to the other part of the other converter whereby said mechanical outputs are summed, first and second pickoti means each individual to a respective converter for generating first and second pickoff signalsv respectively in accordance with relative motion of the parts of the converters infirst and second summing means each individual to a respective converter for algebraically combining a common input control signal with respective pickofi signals to produce said individual error signals, and cross-feed means for modifying at least one of said .error signals in accordance with the other.

4. A servo system comprisnig: a plurality of servo 1) (-l channels each individually responsive to a command signal and each producing an individual output as a function of said signal, each channel comprising an electrical to-mechanical converter for producing a mechanical output in response to an electrical error signal input thereto,- and means for producing said error signal in accordance with the difference between said command signal and the individual channel converter output; cross-feed means for combining with the error signal of each channel a portion of the error signal of each other channel; and means for combining said individual outputs.

5. A servo system comprising: a plurality of servo channels all responsive to a common command signal and each producing an individual mechanical output as a function of said signal; each channel comprising a valve controlled fluid motor having piston and cylinder parts relatively movable in response to an electrical error signal, pickotf means producing a pickoff signal in accordance with relative motion of said parts, and means for producing said error signal in accordance with the difference between said command and pickoit signals; cross-feed means for adding to the error signal of each channel a portion of the error signal of each other channel; said motors being arranged in a series, each motor having one of its parts fixed to one part of a next adjacent motor and having the other of its parts displaceable with respect to the other part of such next adjacent motor whereby relative motion of parts of each motor comprises said individual mechanical output and all such outputs are serially added.

6. A multichannel signal translation system comprising a plurality of closed loop signal translation channels each including an output device for producing an output in response to a driving error signal applied thereto, means for combining with the error signal in each of said channels at least a portion of the error signal in another of said channels, and means for combining outputs of said channels.

7. A rnulti-channel actuator system comprising first and second valve controlled fiuid motors, each comprising a movable cylinder, a piston mounted in said cylinder and connected to a support, a valve fixed to said cylinder for controlling fluid flow to and from said cylinder, said valve including a slidable valve core having a core driving link connected therewith; a summing link pivotally connected to both said cylinders, said summing link being mounted for translational motion to provide a combined motor output by translational motion of an intermediate portion thereof; an output link pivoted to said intermediate portion of said summing link; a driving link pivo-tally connected to the core driving link of said first and second motors; and an input pivoted to an intermediate portion of said driving 8. A multi-channel servo system comprising a plurality of motors connected in tandem, each motor comprising first and second parts mounted for relative motion in response to an error signal input thereto, each motor having one of its parts fixedly connected with one part of the next adjacent motor and having the other of its parts mounted for motion independent of the other part of said next adjacent motor, and means for combining with the error signal of each motor a portion of the error signal of each other motor.

9. The method of operating a number of closed loop signal translating systems from a common input, each said system being a closed loop and having a driving error signal representative of the difference between system input and output, comprising the steps of combining with the error signal of a first of said systems at least a portion of the error signal of a second of said systems, combining with the error signal of each of said systems at least a portion of the error signal of the other of said systems and combining the outputs of said first and second systems.

10. Multi-channel signal translating apparatus comprising a plurality of closed loop signal translating channels responsive to a common input signal, each channel comprising a signal translating device for providing an individual channel output in accordance with a driving error signal input rthereto, and a summing device for producing said driving error signal as the difference between said common input and said channel output; crossfeed means for combining the driving error signal of each of said channels with the driving error signal of the other of said channels; and means for combining the individual channel outputs of said first and second channels.

11. The apparatus of claim 10 wherein said signal translating device of each channel comprises an amplifier responsive to the driving error signal, an electric motor connected to be driven by said amplifier, and a tachometer generator having an input from said motor and having an output to said summing device.

12. The apparatus of claim 10 wherein said signal translating device of each channel comprises an amplifier responsive .to the driving error signal, and valve operated fluid motor connected to be driven by said amplifier.

13. The apparatus of claim 10 wherein said input and error signals comprise mechanical motions, said signal translating device of each channel comprising a fluid motor having a control valve and a valve operating link connected thereto, said summing device comprising a summing linkage having an input link for receiving an input motion, having a feedback connection with said motor, and having a driving error connection with 14 said valve operating link, said cross-feed means comprising first and second cross-feed links interconnecting the summing linkage of first and second channels.

14. A multi-channel servo system comprising a plurality of closed loop servo channels each transmitting an error signal to produce a mechanical output displacement in response to an input, means for combining with the error signal in each of said channels a portion of the error signal in another of said channels, and means for combining mechanical output displacements of said channels.

15. A multi-channel signal translation system comprising a first closed loop signal rtranslation channel having an output device for producing an output in re sponse to a first driving error signal applied thereto, a second closed loop signal channel having an output de vice for producing an output independent of said first named ouput in response to a second driving error signal applied thereto, said channels including mutually independent feedback means, means for effecting cross feed of said error signals between said closed loop channels, and means for combining outputs of said channels.

References Cited in the file of this patent UNITED STATES PATENTS 1,353,656 Heisler Sept. 21, 1920 2,561,654 Eller July 24, 1951 2,597,361 Mott May 20, 1952 2,763,990 Mercier Sept. 25, 1956 2,856,947 Hart Oct. 21, 1958