UNIT-1

1. FUNCTIONS OF MCU IN NC

PN RAO 215218

2. RELATIONSHIP BETWEEN RESOLUTION ACCURACYAND REPEATABILITYResolution Accuracy, Repeatability, Compliance

ResolutionThe resolution of a robot is a feature determined by the design of the control unit and is mainlydependent on the position feedback sensor. It is important to distinguish the programming resolutionfrom the control resolution. The programming resolution is the smallest allowable position increment inrobot programs and is referred to as the basic resolution unit (BRU). For IRB2000 ABB robot it isapproximately 0,125 mm on linear axis. The control resolution is the smallest change in position thatthe feedback device can sense. For example, assume that an optical encoder which emits 1000 pulsesper revolution of the shaft is directly attached to a rotary axis. This encoder will emit one pulse for eachof 0, 36 of angular displacement of the shaft. The unit 0, 36 is the control resolution of this axis ofmotion. Angular increments smaller than 0, 36 cannot be detected. Best performance is obtained whenprogramming resolution is equal to control resolution. In this case both resolutions can be replacedwith one term: the system resolution.

AccuracyAccuracy refers to a robot's ability to position its wrist end at a desired target point within the workvolume, and it is defined in terms of spatial resolution. At first accuracy depends on robot technologyand how closely the control increments can be defined for each of the joint motions, excluding for themoment the mechanical inaccuracy which include the robot manufacture quality. Initially we defineaccuracy as onehalf of the control resolution (Fig. 1.1), considering the worst case where the targetpoint is directly between two control points. A more realistic considerations include mechanicalinaccuracies with a statistical distribution (Fig. 1.2), in that case accuracy is defined as onehalf of thespatial resolution.

Figure 1.1: Diagram of accuracy in two dimensions frame, without mechanical inaccuracyconsideration [Gro86].

Figure 1.2: Diagram of accuracy and spatial resolution in which mechanical inaccuracies are representedby a statistical distribution [Gro86].

The term accuracy in robotics is often confused with the terms resolution and repeatability. The finalaccuracy of a robotic system depends on its mechanical inaccuracies, the computer control algorithms,and the system resolution. The mechanical inaccuracies are caused mainly by backlash in themanipulators joints and bending of the links. The backlash exists in gear mechanisms, in lead screws,and in actuators of hydraulic drives. The minimization of the link bindings is the main designrequirement for the link, as any deflection of the link due to the load at the robot's end causespositional errors. A higher rigidity of the links, however, should not be achieved by a substantialincrease in their mass. A larger mass causes an increase in the time response of the arm. Controlalgorithms might cause position errors due to roundoff errors in the computer. Computer roundofferrors might be significant if a robot controller uses scaled integer representation of Cartesian andangular coordinates. If the computer uses floating point representation, then the roundoff errors willprobably be insignificant. Different definitions of robot accuracy exists, for example in Robotics forEngineers [Kor85], system inaccuracy due to resolution is considered to be 1/2 BRU (Basic ResolutionUnit). The reason is that displacements smaller than 1 BRU can be neither programmed nor measuredand, on the average, they count for 1/2 BRU. A realistic accuracy system is determined in the followingrelationship:

Robot accuracy = (BRU + mechanical accuracy)/2Important: The definition currently used is the ISO 9283 definition (ISO / TC 184 / SC 2 / WG 2), itconsiders errors in 3 dimensions.

Figure 1.3: Errors affecting the robot structure [Naw93]

Accuracy is affected also by external factors. As torque moment is becoming more important on thewrist point with the fully extended configuration, accuracy decrease within the work volume, from thecloser position of the arm to its base, till positions out of work volume range. The term error map isused to characterize the level of accuracy possessed by the robot as a function of location in the workvolume. Accuracy is improved if the motion cycle is restricted to a limited work range, while itdecreases if the load being carried by the robot becomes important Position accuracy depends on theposition in the workspace envelope. For this reason, it is difficult to do robot offline programmingwithout using sensors. The sources of position error that affect accuracy can be grouped into fourcategories: 1) digitization error, 2) calibration error, 3) deterministic kinematic error, 4) stochastickinematic error [Ram88]. As the assembly robots axes are often vertically, accuracy is slightly better because it is less affected byexternal factors and mechanical inaccuracies; but this is not enough to perform some parts matingapplications as peginhole without problems.

RepeatabilityRepeatability is a statistical term associated with accuracy, it describes how a point is repeated. If arobot joint is instructed to move by the same angle from a certain point a number of times, all withequal environmental conditions, it will be found that the resultant motions lead to differingdisplacements (Fig. 1.4). Although a target is always missed by a large margin, if the same error isrepeated, then we say that the repeatability is high and the accuracy is poor. Repeatability does notdescribe the error with respect to absolute coordinates. System repeatability is the positional deviationfrom the average of displacements. For example, +0, 2 mm indicates that any point might be as muchas 0, 2 mm beyond or short of the center of the repeatability pattern. Most robot manufacturers provide a numerical value for the repeatability rather than the accuracy oftheir robots. The reason is that the accuracy depends upon the particular load that the gripper carries. Aheavier weight causes larger deflections of the robot links and larger load on the joints, which degradethe accuracy, while the repeatability value, however, and is almost independent of the gripper load. The repeatability of robots will usually be better than the accuracy, it is normally measured inhundredths of an inch. Repeatability definition which is currently used is the ISO 9283 definition (ISO /TC 184 / SC 2 / WG 2).

Figure 1.4: Example of representation of resolution, accuracy, and repeatability of a robot arm [Kor85].Compliance

The compliance arising out of the slip between object and fingers and from the elasticity of the skin,endows a human operator with extraordinary dexterity for precision assembly. The same complianceenables robot manipulators to perform a variety of manipulation tasks which require fine motion skills(Fig. 2.1). Compliance refers to the displacement of the wrist end in response to a force or a torque exertedagainst it. A high compliance means that the wrist is displaced a large amount by a relatively small force.Compliance is important because it may reduce the robot precision of movement under load as in thecase of robot pressing a tool against a work part, the reaction force of the part may cause deflection ofthe manipulator.

Figure 2.1: A typical example of parts mating operation. The robot arm compliance allowed thepeg to be directed by the whole wall.

SCARA robot, which stands for Selectively Compliant Assembly Robot Arm, is particularly designedfor assembly tasks. The robot shoulder has elbow joints which rotate around the vertical axes (Fig. 2.2).The configuration provides substantial rigidity for the robot in the vertical direction while allowing forcompliance in the horizontal plane which enhance accuracy by a compensation of residual errors (Fig.2.3). The robot arm has four degrees of freedom with four links and joints, it confers a sufficient work

volume for requested assembly applications, and makes it cheaper than full freedom degrees robots.SCARA robots are produced by almost all robots manufacturers and still the more used in assemblycells.

Figure 2.2: A four axes IBM SCARA robot.Figure 2.3: Importance of compliance in Peginhale

applications.

3.

4. DIFFERENTIATE B/W MANUAL AND APTPROGRAMMING

PNR 424

5

UNIT21. BENEFITS OF AC IN MACHINING

BENEFITS

A number of potential benefits accrue to the user of anAC machine tool. AC has been successfully applied insuch machining processes as drilling, milling, tapping,grinding, boring. Following are some of the benefitsgained from AC.

. Increased production rates. Productivity improvementwas the motivation force behind the development of ACmachining online adjustments to allow for variations inwork geometry, material and tool wear provide themachine with the capability to achieve the highest metalremoval rates that are consistent with existing cuttingconditions.

. Increased tool life. In addition to higher productionrates, AC will generally provide a more efficient anduniform use of the cutter throughout its tool life.

. Greater part protection. Instead of setting cutter forceconstraint limit on type basis of maximum allowablecutter and spindle deflection, the force limit can beestablished on the basis of work size tolerance.

. Less operator intervention. The advent of ACmachining has transferred control over the process even

further out of the hands of the machine operator andinto the hands of management via part programmer

2. CONCEPT OF ATC IN MACHINING CENTRES

3. ADAPTIVE CONTROL WITH CONSTRAINT

ADAPTIVE Control, WITH CONSTRAINTS

We have seen that considerable research anddevelopment are required before ACO systems becomepractical for industrial use. Actually all the AC systems forrough turning and milling used in production today are ofthe ACC type and seldom involve control of more thanone machining variable [3031). Unlike ACO, ACC systemsdo not utilize a performance index and are based on

maximizing a machining variable (e.g., feed rate) subjectto process and machine constraints (e.g., al lowablecutting force on the tool, or maximum power of themachine). The objective of most ACC types of systems isto in crease the MRR during rough cutting operations.This is achieved by maximizing one or more machiningvariables within a prescribed region bounded by processand system constraints [3, 3234]. One useful approach,for example, is to maximize the machining feedrate whilemaintaining a constant load on the cutter, despitevariations in width and depth of cut [3538]. This isillustrated in Fig. 3 for a slab milling operation. In anormal CNC system, the feedrate is programmed toaccommodate the largest width and depth in a particularcut, and this small feed rate is maintained along theentire cut. As a result the machining rate is reduced. Bycontrast, with the ACC system, the maximum allowableload (e.g., cutting force) on the cutter is programmed. Asa result, when the width or depth of cut are small thefeedrate is high; when either the width or depth of cut(or both) are increased, the feedrate is automaticallyreduced, and consequently the allowable load on thecutter is not exceeded. The result is, the average feedwith ACC is much larger than its programmed

counterpart. Likewise, when the tool moves through airgaps in the work piece, the feedrate reaches itsmaximum allowable value. Operating at the maximumallowable feedrate, maximizes the MRR (See Eq. 1). TheACC system guarantees maximum productivity whileminimizing the probability of cutting tool breakage.Commercial AC systems are available for end milling. Anend milling cutter may break in bending or in torsion or atooth may break away. An appropriate ACC system musttherefore continuously check the radial cutting force andthe cutting torque on the cutter, and vary the feedrateso as to keep both these variables below the permissiblelimit [39]. The most commonly used constraints in ACCsystems are the cutting force, the machining power, andthe cutting torque [27. 40]. The operating parametersare usually the feedrate velocity V (in millimetres perminute or inches per minute) and the spindle speed N (inrevolutions per minute); both can be easily manipulatedunder computer control the machining feed f is definedby the ratio

f V/pN (4)

Where p is the number of teeth in the cutter in themilling operation; in turning and drilling p 1 is substituted

in Eq. (4). The main cutting force F is proportional to thedepth of cut a and the feed:

F = K,aj" (5)

Where K, is the specific cutting force, and u is aparameter in the range 0.6 < u < 1. Both K, and u dependon the work piece and tool material. The cutting torqueT is proportional to the force and the work piecediameter in turning and the tool diameter in milling anddrilling. The machining power P is proportional to thetorque and the spindle speeds. A typical computerizedACC system, which applies the concepts introduced inthis section, is described for turning on a CNC lathe witha constant cutting force constraint [36]. The ACC systemshown in Fig. 4, is basically a feedback loop where thefeed adapts itself to the actual cutting force and variesaccording to changes in work conditions as cuttingproceeds. The CNC computer executes the originalcontrol pro gram and an additional AC routine, which islinked to the feedrate routine contained in the controlprogram. The AC loop functions in a sampleddata mode.The actual main cut ting force F is sampled every Tseconds (typically T = 0.1 s), then converted to a digitalsignal F, and sent to the computer. The actual force

representation Fe is immediately com pared in thecomputer with a predetermined allowable referenceforce Fr. The difference between Fr and F,. Which is theforce error E (E = Fr Fe) is used as the input to the ACcontroller. The latter sends a correction signal to thefeedrate routine, which, in turn, produces the feedratecommand nal. A positive error increases the feedrateand consequently increases the actual force, therebydecreasing the error E, and vice versa. In order toeliminate completely the force error, the controlleroutput command U should be proportional to the timeintegral of the force error W. The simplest structure forsuch an integral strategy can be written

W(i) = Wei 1) + TE(i) (6)

and then the command signal U from the controller is

(7)

Equations (6) and (7) can be combined to give a moreefficient form for programming:

(8)

Where K,., the controller gain, is proportional to thesampling period T, I.e., Kc TK;. The resulting computerfeedrate command VI to the servo loops at the ith

sampling (or to the interpolator in a multiaxial mode) isgiven by

(9)

Where KI is a constant associated with the feedrateroutine. The initial feedrate value K IUO may bepreselected so as to avoid tool breakage when the toolinitially impacts the work piece at the start of thecutting process. As long as there is an error, thecommand U varies the machine feedrate in a direction tocorrect this error. At the steady state, however, the errorin the force is zero, causing the condition U(i) U(i I),which means that the feedrate command is constant,maintaining the actual force equal to the required one.This integral strategy has been implemented on a highpower CNC lathe [41]. A typical result for Kc 0.5 is shownin Fig. 5. The feed before engagement was selected as0.5 mm/r. At the start of cutting, the feed isautomatically reduced to approximately 0.25 mm/r. Thedepth of cut is in creased by increments of 2 mm, andeach time, after a small transient, the force reaches thepreselected reference value of Fr = 1500 N, and thecorresponding feed is decreased. The selection of thegain K, is critical to the operation of the AC system. It is

known from control theory that the lower the gain K" thegreater the tendency for stability. Although a small gaincauses a sluggish response, the steadystate error alwaysbecomes zero

4. DNC ADANTAGES

DNC, CNC and Adaptive control

The significant achievement done in both batch and jobshop manufacturing caused the further development ofNC systems. These new enhancements and extensions ofNC technology, include:

1. Direct numerical control

2. Computer numerical control

3. Adaptive control

The Direct numerical control system (DNC) sets thedirect connection between the computer and the NCmachines. What is actually a DNC system? A DNC systemis consisted of a large master computer used to directthe operation of a number of separate NC machines.

Due to the advances of computer technology and thereduction in the cost of the computers it became

economical to install a microcomputer on the NCmachine in order to control each machine tool.

Adaptive control marked as the last achievement in theNC machines is a system which is able to measure one ormore process variables. Such process variables arecutting force, temperature, power etc. With theknowledge of this process variables the system is able toregulate feed and/or speed in order to compensate forundesirable changes in the process variables. The finalscope of this control system is to optimize the machiningprocess.

3.3 Direct numerical control (DNC)Direct numerical control is defined as a manufacturing system in which a number of machinesare controlled by a computer through direct connection in real time. The tape reader is omittedfrom the system. The information or the part program is being transferred directly to the machinetool through communication lines from the main computer. This is done in real time and thecommunication is two way, both the computer and the machine tool can send information to eachother. The computer sends information to the machine tool upon request of the latter and whenthis occurs the request for instructions must be satisfied almost instantaneously. The computerstores the information in a bulk memory and can control more than 100 machine tools. Thegeneral configuration of a DNC system is illustrated in figure 3.1. In addition in cases where the computational capability of the main computer is not enough tosatisfy the needs of the vast number of the machine tools additional computers are used linked tothe main server to satisfy group of machine tools. These smaller computers are called satellitecomputers. The configuration of this system is illustrated in figure 3.2.

FIGURE 3.1 General configuration of a direct numerical control (DNC) system

FIGURE 3.2 DNC in hierarchical configuration using satellite computers

3.3.1 Advantages of DNCThe advantages of DNC systems are laid below:1. Time sharing the control of more than one machine by the computer2. There is greater computational capability for operations such as circular interpolation,tapering etc. 3. There is no need for the computer to be located near the machine because the computer cannow control the machine tools remotely, while situated in a computer type environment. 4. The tape and tape reader are eliminated and replaced by communication cables making themachine tool reliable. 5. The hardwired controller unit is eliminated 6. Programs stored as cutter location data can be post processed for whatever suitable machine isassigned to process the job.

The only disadvantage of the DNC is if the computers breaks down, but the computer is soreliable that this is unlikely to occur.

5.PNR380