All information and opinions presented in this paper are the authors. Design News online did no editing or confirmation of the information provided.

Introduction

In the world of fluid power, advancements in the areas of electrical controls have opened many new frontiers for the application of hydraulic and pneumatic components. As certain industrial requirements dwindle, new areas of interest are being expanded with the use of electro-hydraulics.In the Eighties, the major interest for closed loop controls were in the areas of robotics and defense systems. Hydraulic servo controlled missiles were developed along with numerous "pick and place," welding and painting robots. The expensive and difficult to maintain robots of the '80s have been replaced with simple single axis movements. The function is still very much the same but the result is achieved from many individual systems. This allows problems and malfunctions to be isolated and replaced with minimal effort, expense and down time.

In the Nineties, the new and exciting challenges have been in the areas of entertainment. Hollywood has discovered Fluid Power. Movies such as: Free Willie, Jurassic Park, True Lies and Congo have all used fluid power in their films. Fluid power gives the film maker the ability to manipulate and implement ideas that could not be previously be handled. These ideas include, natural disasters, animals, and other futuristic ideas. Earth quakes, whales, alligators, and hippopotamuses have been simulated with closed loop controls.

The entertainment market is not limited to only motion pictures. The trend for the studios is to invest in a project, shoot their film and then move the investment to a theme park for further revenues. Interactive rides and attractions are popping up in local malls across the nation all of which use hydraulic closed loop servo controls.

The success of the system is solely dependent on each of the components in the loop performing it's given function along with their ability to work together. This article briefly covers definitions, sizing and implementation of a hydraulic closed loop system.

Definition of Control Variables

We can control virtually anything that can be measured. The most common "loop" in hydraulics is position. It is advantageous for a machine designer to have infinite control of the starting and stopping point of a cylinder throughout its stroke. This type of control loop is referred to as "position loop". The major concern of the actuator is its ability to reach and maintain a commanded (go to) position.

A positional loop consists of a controller, servo valve, actuator and a device to measure the actuator's position. This is by far the most popular and requested control variable. It is also the basis of the second type of loop-- Velocity (position over time).

Along with positioning a cylinder, applications are requiring the ability to control acceleration, running velocity and deceleration. All of these elements fall into the "velocity loop" category. Combining two loops (position and velocity) is not an easy task. Most positional controllers will give the option of "ramps," but this must not be confused with controlling the actual cylinder velocity.

A velocity loop is very similar to a positional loop with the exception that the device measuring the actuator's movement is different. Velocity is determined by displacement over time. In most analog systems, the feedback device is scaled to deliver an analog voltage at a predetermined velocity (speed). More sophisticated controls will use a digital feedback device and a time constant to determine velocity. This feature is beneficial because a position feedback may be derived as well.

Since force is defined as restriction to movement, we could use position to determine the force output but the potential for mathematical error would be too high. Strain gauges are a preferred measurement device, but what do we measure? The use of a strain gauge in a pressure transducer will work if we multiply the applied pressure on the piston end, multiplied by the piston area, minus the resistive pressure on the rod end, times the rod end area. Here again the potential for mathematical and scaling errors is high. For this reason the most common means for measuring force output is with a load cell mounted between the cylinder and the specimen.

A force loop is similar to a positional or velocity loop with the exception that a feedback device measuring the force generated by the cylinder is used.

Typically force loops are applied in the tensile testing area and engineers select a high end controller to interface with the system. Such controllers use a positional feedback device (LVDT) and a force feed back device (load cell) for closed loop control. Typically the cylinder selected has equal piston area (double rod) and one end of the cylinder (blind end) will house an LVDT for the positional feedback.

Definition of Control

The command sequence of an open loop system flows in series from one component to the next. The reason it is referred to as an open loop is that there is no feedback to the command device to indicate the position of the load. Open loop system command signals are typically generated from PLCs, battery packs or joystick type potentiometers This voltage signal is then directed to a proportional control valve. Most manufacturers offer an external excitation voltage for the driving current and use the voltage control as an input for the spool position. The control valve then directs metered flow into and out of the cylinder based on the command signal.

This has superior benefits over conventional directional control valves. A proportional control valve will out perform any soft shift solenoid in controlling a load's acceleration and deceleration. Traditional multi-speed applications requiring a number of different solenoids with flow controls can be replaced by one proportional control valve. The key is to select the right valve for the each job.

An open loop system has significant differences when compared with a closed loop system. The command signal, usually a voltage, will control the proportional valve's spool position. If a PLC generates a 5 VDC command signal, the proportional valve will deliver 50% of its rated flow at its rated pressure drop. For this reason we can use a proportional valve as an infinitely adjustable flow/directional control valve.

For purpose of clarification we will refer to proportional valves as valves with overlapped spool conditions and servo valves as valves with axis null cut (zero overlap).

The most common application in hydraulics is referred to as "bang-bang" control. We unleash flow to a cylinder without control of maximum speed, force or velocity. Then at a predetermined time or by tripping a proximity switch, we disengage the control valve, and the actuator abruptly stops. In the beginning, these systems actually went "bang-bang." The starting and stopping of the actuators was abrupt and produced shock throughout the system. This type of control utilized on/off control of directional control valves. Since shock fatigues components, structures and fluid carrying lines, it was desirable to reduce this shock to a minimum. System engineers started by replacing the conventional directional valves with servo valves. Since the servo valve is a directional control valve with electrical control of flow metering capabilities it was an obvious solution for eliminating abrupt starting and stopping of the actuator.

Without a feedback device a standard axis cut (zero overlap) servo valve can not be used because of cylinder drifting problems. To solve this problem a servo valve with 5-10% overlapped spool area should be selected.

Even though a servo valve is used in this type of control, it is still referred to as an "open loop." Open loop can be defined as a control means which does not have a continuous feedback of performance. Typically today this type of valve is referred to as a proportional control valve.

The improvements in the proportional control valve area, has made them preferable to a servo valve (zero overlap) in an open loop system. A servo valve will allow the actuator to drift when we are at a null state. Using a proportional valve with 10% overlap on the spool and sleeve assembly, at null ("0" command), the actuator may be stopped as a directional control valve restrains movement, while offering flow control of the fluid through an electronically adjustable orifice.

Closed Loop Control

Components in a closed loop system are very similar to that found in an open loop system. The differences are mainly in the areas of performance and control theory. Open loop circuitry is often overlooked due to the accuracy and confidence associated with closed loop control. In systems with tight budgets and loose accuracy requirements, open loop systems may have a solution for you, but some external means must be incorporated to determine if the actuator is at its desired position.

The closed loop system works in a circular format. This network of information is continuously informing the appropriate component of what to do and what the outcome is. This control method brings the designer benefits that could not normally be reached with human movements. How many people can pick up a part, move constantly at a speed of 4"/sec. and place it exactly 12" away, plus or minus 0.001 of an inch? This type of control is easily achieved with this method.

An analog closed loop will provide all the intercommunications in an analog format. Analog signals typically have two categories: voltage and current. Voltage loops typically range in the following categories: 0 to 10, -10 to +10, 0 to 5, and -5 to +5. 0 to 10 VDC is the most common. Since current is more immune to electrical noise, current loops are becoming increasing popular. Commonly used current signals are 4 to 20 MA and 0 to 20 MA.

Analog differs from digital in the ability to measurement between two points. With a digital system only the end points are defined . With an analog system there are an infinite number of possibilities between these two points. Based upon this, one would naturally think that analog circuitry is the way to go, but one must look at linearity and repeatability as well as resolution in determining which type of control to use. Most digital devices will offer the
ability to measure below 0.001".

Digital devices use a digital (high/low) pulse as a measurement tool. The trick with digital circuitry is to get a device with the right number of "counts" and accuracy beyond what you wish to achieve. They exist today in the form of feedback devices and controllers. The only thing missing is a control valve with serial type communications and response to match that of its analog cousins. There are digital interface valves, however, they still close the inner loop with an analog system. Before the turn of the century we will start to see full digital closed loop control systems, utilizing frequency controlled poppet valves that will vary the output flow by the frequency rate we turn them on and off. Currently these valves are in the experimental stage.

Building Blocks of a Servo System

This section will provide an overview of the entire closed loop circuitry with basic operational and selection criteria.

Electronic Controls

The controller is the brain of a closed loop system. The simplest electronics available are the summing/ power amplifiers. These Printed Circuit Boards (PCB) typically operate on either straight DC voltages or on an AC power source and provide the power for the rest of the components in the loop.

A command signal (usually a + voltage) is compared to a feedback signal (usually a - voltage). This is true with all analog loops. If the feedback is anything other than a DC voltage, the signals are conditioned on the PCB to give a DC voltage to the summing junction at the correct polarity. The key is polarities, since a + 5 VDC summed with - 5 VDC equal 0 VDC , closed loop control is easily obtainable. Some manufacturers offer an inverter on the feedback input so that common polarities may be used.

The difference between the command voltage in relationship to the feedback voltage or "error signal" at the summing junction is amplified based upon the gain setting of the amplifier.

Gain is defined as the difference between an input and an output. All components in the loop have gain. The Servo Controller accepts an input of voltage and delivers an output of current milli-amps (MA). The Servo Valve accepts an input of current (MA) and delivers an output of flow (cubic inches per second ). The cylinder accepts an input of flow (cubic inches per second) and delivers an output of movement (inches). The feedback device accepts an input of movement (inches) and delivers an output of voltage.

Amplifier gain can be explained as a multiplier. An amplifier with 10 to 1 g ain will amplify an input of 2 VDC to 20 VDC. This is significant in closed loop control. For example, a cylinder with a 24" stroke and a feedback of 0 to 10 VDC, the cylinder gain is 0.42 V/inch. To position this device 0.001" we will need to see variations of 0.0004 VDC. If we did not have a gain (multiplier) we would not move enough current to the servo valve to correct this error signal. With an amplification of 100 to 1 we will have 0.4 VDC going into the power amplifier to generate enough current to move the servo valve torque motor, which in turn will move the actuator.

This error signal that is sent to a current driving amplifier will saturates either a (+) or (-) transistor which delivers the power to the servo valve torque motor. Based on the gain setting, the power amplifier will saturate the servo valve coil with the maximum current available from the supply. T he output (current) from the amplifier will drive the servo valve force motor to its full position thus allowing the actuator to travel at it s designed maximum velocity. When the error falls below the gain setting, the valve will balance pressures across the unequal area of the piston and the cylinder will stop. Any over travel or under travel will generate an error signal, thus making the amplifier responded accordingly. If any external disturbances alter the position of the actuator, the feedback device will measure the true position of the actuator. The summing amplifier will measure the difference between the desired position and the actual position and the process will repeat.

Typically servo amplifiers have extremely high frequency response . This allows the amplifier to send an alternating (+) and (-) current to the servo valve at a very high frequency. The servo valve also has a relatively high response characteristic . This allows the valve to oscillate at a higher frequency than the actuator can respond to. Since a servo valve is nothing more than an electronic flow control, it will be metering the oil in and out of the actuator to obtain the desired position. However, if the gain setting on the amplifier is too low, the error signal sent to the power amplifier will not allow the servo valve to correct any deviation in position.

If the amplifier's gain setting is too high, the system will oscillate. Most servo amplifiers will use a +/- 15 VDC power supply for the servo valve with a maximum current output. Since the servo valve has a flow rating at a maximum current level, the amplifier selected should not exceed the coil rating. When the error signal reaches this maximum rating, the servo valve torque motor is saturated and the valve will deliver the maximum flow to the actuator.

With this type of system the gain setting controls the point at which the valve is driven to it s full position. If the amplifier saturates at a level the cylinder can not position to, it will overshoot the position and the amplifier will notice the large (+) error. It will drive the system in the reverse direction causing an additional (-) error and the process repeats. This is referred to as oscillation.

Servo amplifiers offer two (2) types of amplifiers: Proportional and Integral . This is also known as "P.I.D. " control. The "D" stands for derivative and in simple analog loops derivative is not normally used. The amplifier selected primarily is based on the type of loop we are closing.

Most servo controllers use operational amplifiers. An operational amplifier is a high gain amplifier with feedback. Without feedback the amplifier would be very unstable due to the high gain. This feedback in positional loops is generally a resistive device. This type of arrangement is referred to as "proportional" control and is commonly used in positional loops.

By replacing the resistive feedback with a capacitor an integrating amplifier is developed. When integrating amplifiers are used the gain of the amplifier is altered to provide milli-amps/second/volt. This type of gain allows integrating amplifiers to be used in velocity loops .

Programmable Controllers

Most PLCs (programmable logic controllers) today have the capability of accepting and delivering analog output. They are available in both current (4 to 20 MA) and voltage (0 to 10).

The PLC will "move" a voltage level when tripped by an input. The "move" is similar to changing the voltage level by activating a selector switch. The only way it will change its output is to "move" another voltage level from its register.

This is still referred to as an "analog loop" because the control means is primarily analog type signals. The command signal, whether generated from a potentiometer or from an analog output module (PLC), is fed into the amplifier in an analog format.

The PLC does not have any feedback to determine whether the actuator has reached its desired position . However, using a PLC to generate the command signal and monitoring the feedback is a highly economical form of positioning. With the addition of an analog to digital converter monitoring the feedback device, we can synchronize events after the cylinder has reached it s commanded position. Figure 2.2 illustrates the implementation of a PLC as the command signal generator.

Where accuracy requirements are looser and processor time is available the PLC may be used to close the loop by utilizing analog inputs and outputs. This is common with temperature controllers and other crude closed loops. However it is not always practical in the fluid power industry. If we remove the servo amplifier and drive the valve with the PLC, we will have to monitor the analog signal, convert it to digital, compare the digital signals and then convert the signal back to an analog signal to drive the valve . This process usually requires too much time to be effectively used in "closed loop" control.

With the increased demands for tighter controls, PLC manufacturers are offering the addition of stand alone motion controllers. These controllers are typically installed in an expansion slot on the PLC back-plane. The controllers are generally minicomputers configured to "close" a loop digitally. These controllers typically require a digital feedback device so that the analog to digital process is eliminated. These controllers typically offer either a current output (+/-100 MA) or a voltage output (+/-
0-10 VDC). This allows easy interface to most existing servo valves.

Motion controllers offer many features that are not obtainable from standard analog loops. Since a motion controller accepts a digital feedback device (counts) the output may be configured to give a velocity and position feedback simultaneously . With the controller delivering the analog output to the servo valve through a digital to analog converter, the programmer may select a desired velocity he wishes the actuator to travel to its commanded position. This feature also allows control of acceleration and deceleration that may be significant to the application.

With such power, manufacturers will offer the ability to control multiple axis (actuators) simultaneously .

Since the loop is closed digitally, the controller will offer a menu to select the proper gain settings. Motion controllers typically offer "P.I.D." and "feed-forward " (velocity) gain settings. Proportional gain or "P" will control the "stiffness" of the system. Integral gain or "I" will control the "steady state error " or dynamic control. Differential gain or "D" will provide "dampening" or static control. Feed-forward will allow velocity, acceleration and deceleration control. These types of gains may also be viewed as "filters" that allow wider control.

Servo Valves

The servo valve is the heart of the closed loop system. This valve will deliver the necessary flow to the actuators. There are several different designs available, but they all provide the same output. These valves actually are precise flow control valves with electronic actuation .

Where manufactures differ in design is how the spool is positioned. All the valves available today may be grouped into four separate categories.

Double Flapper. The double flapper design is the most common servo valve available. This type of design uses hydraulic pressure to shift the spool assembly. System pressure is diverted through two fixed identical orifices to reduce pressure for pilot use. This pressure is applied to each side of the valves spool. In the pilot stage there are two variable orifices (nozzles) that bleed pressure to the tank. This pressure is sprayed on a "Flapper": hence the name "Double Flapper."

The torque motor consists of a permanent magnetic field with an armature assembly and a pair of coils. The armature assembly is connected to the flapper and feedback device. All servo valves have feedback. The flapper design usually incorporates a mechanical spring as the feedback device with a close fit to the spool assembly. Without this feedback, the spool would be actuated fully in one direction and then the other.

In a nulled state (no input signal) the torque motor will center the flapper between the nozzles and the pilot pressure will balance the spool in the mid-position. When the torque motor coil receives an input, the coil assembly will alter the ma gnetic field and cause the flapper to pivot closer to one nozzle and further from the other . Thiswill cause a pressure imbalance on the spool area causing the valve to divert flow to the work ports.

This pilot pressure applied to the spool area will work against the feedback spring limiting the distance the spool travels . The direction the spool travels may easily be changed by changing the polarity of the input signal.

This design allows frequency response up to 300 hertz , solid pressure stability and excellent metering capability. The torque motor assembly requires extremely low power and unlimited null cuts and flow gains are available. However, it is highly sensitive to contamination.

Servo Solenoid/Voice Coil. The servo solenoid and voice coil design use a linear electrical force to position the spool.

The servo solenoid valve uses the improved technology of proportional solenoids. Its construction consists of either a single or double proportional solenoid , a spool and sleeve assembly, a LVDT for feedback and an optional integrated voltage to current converter.

This single stage design uses an electrical feedback to monitor the spool position. Because the device positioning the spool is a solenoid working against a spring, higher current is required. Most manufacturers will offer a voltage to current converter as an integral part of the valve for easy control application. Typically the servo solenoid valve will accept a bi-polar 10 VDC signal for control. Its simplicity has some unique features over the more traditional servo valves.

The valve basically operates similar to standard NFPA directional valves. The body and construction, although very similar to a directional valve, are manufactured differently . As in all servo valves, the spool and sleeve assembly are machined as a matched set. This is required to achieve the linearity in flow metering. The valve's spool is positioned fully in one direction. This position usually referred to as the abort position, is available in numerous different formats. Most common is the "closed center" format. This will block all the ports with enough overlap on the annular area to prevent leakage and may be used for load holding. The "onboard" electronics requires an excitation signal, usually (+) 24 VDC, which will position the spool against the spring until the LVDT (feedback) is in the desired position very similar to how a closed loop system work . When a voltage signal is applied to the "onboard" electronics, the spool is positioned proportionately . A simple change in polarity will reverse the valve's function.

Another advantage this design has is an electrical feedback device, which allows the valve's null to be factory pre-set. The mounting arrangement is usually to NFPA standards that eliminates the need for special mounting sub-plates and manifolds. Since this valve does not have a pilot stage and does not have any orifices it is still susceptible to fluid contamination. The spool and sleeve arrangement is ma nufactured with extremely tight tolerances and contamination will deteriorate and potentially prohibit the spool from movement. However, the design does eliminate the potential for a "hard over" state. "Hard over" is defined as loss of control resulting in hydraulic energy released to the actuator at the valves maximum capacity.

A voice coil or force motor design incorporates a magnetic torque motor assembly similar to the double flapper. The spool of the valve is mechanically attached to the armature. A change in current will cause the armature to shift which will position the spool accordingly . A mechanical feedback device is usually used when larger flow and shifting forces is required.

Rotary ElectricMotor.There are two different types of "electric motor" servo valves, digital and analog.

The stepper motor design is as close to a true digital servo valve as the present technology permits. Its construction consists of a rotary DC motor, an eccentric cam arrangement , a spool and sleeve assembly.

The analog motor uses a limited angle brush-less DC motor instead of a stepper motor. This design also incorporates a rotary feedback device and is offered with an optional integrated controller.

When the controller receives an input usually (+/- 0 to 10 VDC ) it indexes the rotary motor to the desired position. T he cam arrangement changes the rotary motion to a linear motion that positions the spool accordingly . The controller receives the feedback from the motor closing the position loop. The spool and sleeve arrangement is similar to all other servo valves and provides a variable metering orifice.

This arrangement eliminates the potential for a "hard over" state but as with the servo solenoid valve it too, is susceptible to contamination.

Jet Pipe Servo Valve. This type of design uses hydraulic pressure to shift the spool assembly similar to the double flapper. System pressure is diverted through a jet pipe orifice. This pressure is then sprayed on a target area. The target area is bevel shaped with an orifice on each side . The flow through the orifice is then diverted to the appropriate spool area.

The torque motor consists of a permanent magnetic field with an armature assembly and a pair of coils. The armature assembly is connected to the jet pipe and feedback device (mechanical) similar to the double flapper.

When pressure is applied to the valve the fluid is sprayed on to the target area, pressure is maintained on the annular area of the spool and valve is nulled. The torque motor operates in the same fashion as the double flapper, but instead of moving the flapper the jet pipe is positioned on the target area. All other mechanics are the same as the double flapper design.

Servo Actuators

The actuator provides the muscle in a servo system. Actuators are available in two forms: Linear and Rotary, which are also referred to as cylinders and motors respectively. There are three key things to look for when selecting an actuator: leakage, f riction and feedback implementation . Leakage is normally not a problem , but frictional characteristics are extremely important. Most manufacturers of servo grade cylinders and motors will offer various feedback devices. It is good practice to have the feedback device as an integral part of the actuator for protection from the environment and accurate measurement of the prime mover.

Most hydraulic servo motors are either of piston or roller vane design. These two designs will offer superior frictional characteristics to other motor designs and will have smooth transitions during low speed operation .

A good servo cylinder will have extremely low frictional characteristics as well. Seal material selection is imperative for optimum performance. This is not only limited to the piston. Since most servo cylinders are equipped with feedback devices the smaller bore sized cylinders require oversized rods that attribute to higher friction.

A conventional cylinder with steel piston rings and a bronze bushing can induce contamination into the system thus causing the servo valve to fail prematurely . For this reason, when selecting an actuator close attention must be paid to the seal material. Rod bearing area coated with a Teflon(R) material will provide very low coefficients of friction while eliminating the metal to metal contact of the bearing area. Similarly, the use of Teflon as your primary piston seal will further reduce friction and the introduction of foreign material into the hydraulic system.

Feedback Devices

The feedback device may be referred to as the "senses" of a servo system. The feedback device informs the controller of what is happening in the system. Just as our eyes, nose and feeling tell our brain about the outside world.

This device is extremely important in determining how accurate the system will be. Obviously the system can only position to a point equal to or above what the feedback device can measure.

Positional Feedback.The simplest form of position feedback is the Linear Resistive Transducer (LRT). This is nothing more than a linear potentiometer. In the early days of positioning, these pots were wire wound devices that would not tolerate temperature changes. Currently, w ith the addition of conductive plastics, LRT's offer acceptable temperature tolerance.

The basic construction of a LRT is illustrated in figure 4.1(on prior page). The potentiometer is mounted stationary in the cylinder cap (blind end) and the wiper arrangement is mounted to the piston of the actuator. The collection strip is zero resistance material used to divert the divided voltage back toward the electrical connector.

Voltage is applied across the resistive element and a simple voltage divider is developed through the wiper arrangement . As the cylinder is actuated, the wiper moves down the resistive strip thus changing the resistance seen between the resistive element and the wiper. In theory, when the cylinder is fully retracted, the wiper will be positioned on the common side of the resistive element in respect to ground. As the cylinder is moved through its full working stroke, the wiper will travel across the resistive element varying the resistance seen at the wiper in relationship to the power supply common.

Another device used for position feedback is the Linear Variable Differential Transformer (LVDT ) . The LVDT consists of three coils and a high permeability core. The primary coil will be excited with a high frequency AC signal. The two secondary coils are wired in series-opposition. When the core is positioned exactly in the center of the two secondary coils, the demodulated output is zero. Since the LVDT is primarily an AC device, the demodulator will convert the AC signals to a DC voltage for use in a servo system. As the core is positioned either side of center the LVDT will produce an output of amplitude and phase shift from 0 to 180 degrees. Due to the nature of the LVDT theyare primarily installed in short stroke double rod cylinders, with the blind end housing the LVDT. The demodulator will typically produce +/- DC output with the nulled position in mid-stroke.

One of the most popular transducers is the magnetostrictive type. It is also known as Magnetostrictive Linear Displacement Transducer ( MLDT). This device works on the principle of measuring the time between the excitation pulse applied to the transducer and the receipt of the return pulse (echo) generated by a rare earth magnet positioned around the transducer.

This transducer is a pure digital device with a wide array of digital and analog outputs available. The nature of this transducer provides very high resolution with stroke lengths up to 300 inches.

Cylinder modifications for adding an MLDT consists of gun drilling the piston rod, attaching a rare earth magnet to the piston and modifying the cylinder cap to accept the transducer. Integrated electronic modules allow for all the various signal conditioning required. Extremely high resolution , superb linearity and minimal hysteresis make the MLDT a very popular feedback device.

As the magnet moves with the piston, the distance that the return pulse travels increases. By calibrating each transducer with a time constant, the transducer generates a signal proportional to magnet position.

Another type of digital feedback device available is commonly referred to as an optical encoder. There are two basic types : incremental and absolute.

An incremental encoder consists of many small, precisely spaced lines, on a glass disc. On one side of the disc there is a light emitting diode ( LED ) and on the opposite side is a photo transistor receiver . The disc is rotated and the light beam is broken by the lines on the disc. The signal produced is primarily a digital pulse of high and low signals, also known as counts. The encoders are selected by the number of counts per revolution .

An absolute encoder functions in the same fashion, but with the addition of a process to the lines on the disc to give an angular feedback to generate an absolute position of the disc. Absolute encoders do not need to be "zeroed" and remember their position when powered down.

Encoders and rotary potentiometers work effectively in providing positional feedback for hydraulic motors. Most servo motor manufacturers will offer a through shaft arrangement to mount theses' devices opposite the work shaft. Synchs are also used in positioning a motor. The synch works on the same principle as a LVDT, only in a rotary fashion. This device is preferred over the potentiometer because of the infinite turning capability and it offers 360 degrees of positioning.

Implementing encoders into cylinders is not an easy task. Since the cylinder is a linear device and the encoder requires a rotary movement , a linear to rotary conversion must be made. This is typically done with a ball screw arrangement and because of all the idiosyncrasies of ball screws, most cylinder manufacturers are discouraged from implementing them into their products.

Velocity Feedback. Since the MLDT works on the principle of time between sending a pulse and receiving a pulse, this device is easily altered to provide a velocity output.

Tachometers are commonly used in providing velocity feedback for a rotary motion. They generate a DC voltage proportional to the speed of the shaft. One of the features with this feedback device is the lack of any excitation voltage to generate an output.

Acceleration transducers use a fixed mass in conjunction with a device that measures force. Based upon Newton's Second Law, (f orce equals mass times acceleration ), as the accelerated force is applied to the sensing device (usually a strain gauge or piezoelectric material) the resistance changes in the sensing element, which will generate an output proportional to acceleration.

Force Transducer. All force transducers operate in a similar fashion, the only difference is where the force is measured . Force transducers typically use a strain gauge for the sensing device. A strain gauge is a resistive device that changes resistance as force is applied. Strain gauges are extremely sensitive, and do not have any moving parts. For this reason they are considered excellent instruments.

Piezoelectric crystals are another way to measure force. The crystal will produce an electrical signal proportional to the amount of force exerted on it. However, these values vary with temperature that limits their performance and usefulness.

A pressure transducer will measure the hydraulic pressure in fluid carrying lines or vessel. A differential pressure transducer will provide an output proportional to the difference of two pressures. This is typically applied to each side of the actuator. Load cells are the most accurate form of determining the force generated. The strain gauge will be bonded to a device that will be mounted between the actuator and the specimen it is applying the force to.

Hydraulic Power Supply

The hydraulic power supply is the fuel in for the system. It provides the necessary power to do the work, conditions and stores the fluid. Careful selection of individual components will result in long term dependable service.

The power unit is sized based on a flow and pressure requirement of the actuator. Once these parameters are determined electric motor and pump combination can be selected.

The reservoir should contain not less than three times the output of the pump. This will allow enough dwell time in the reservoir to let suspended air molecules dissipate from the fluid. The reservoir, if sized properly, will also dissipate a p ortion of the heat generated by the system.

The nature of the servo system will determine the type of pump selected . Most positional systems spend a lot of time around "null" and do not require much flow. The flow demand is considerably larger during transition. For this reason, a pressure compensated, variable volume pump should be selected so the entire pump volume is not directed across the system relief valve during low demands. Pump compensators will induce a lag time in the system. To reduce this lag time and for shock suppression, a small accumulator may be installed in the pressure line.

Filtration is extremely important in a hydraulic power unit regardless of its application. It is false economy too undersize or eliminate a filter from the system. All of the components in the system, pumps, safety controls, and valves use metal to metal seals. Therefore the hydraulic system itself is responsible for generating contamination that must be filtered out to avoid breakdown and premature replacement of parts. Once a system is contaminated, it is extremely difficult to clean it to it s original state.

With all the components contributing contaminates into the system, it is a good practice to install a non-bypassing pressure filter at the servo valve. The filter should be sized to handle the amount of fluid the valve requires along with a degree of filtration below what the manufacturer recommends.

It is also imperative to provide a return line filter of the same value located at or in the reservoir. Dependable servo systems require periodical preventive maintenance schedules to check the fluid, replace the filter elements and even send the servo valves back to the manufacturer for calibration.

Heat removal is also important. A servo system by design is not highly efficient and does generate a lot heat. This heat must be removed. This is commonly done with a heat exchanger. There are two different media used to transfer the heat to: water and air. Both of which work well if sized correctly.

Reference Material

James E. Johnson, 1973 and 1977, 13th International Symposium on Industrial Robots

Conference Proceedings Robotics International of SME, 1983

Hydraulic control is generally used..., Sales Brochure, Catalog No. 801 482, Moog Inc.