onum NUM 1020/1040/1060 HIGH SPEED MACHINING 0101938956/1 06-97 en-938956/1Despite the care taken in te preparation of this documant, NUM cannot quarantee the accuracy of he information contains and cannot be held responsible for any errors therein, nor for any damage which might it rom the use or application of the document. ‘The physica technical and tunctonal characterises of the hardware and software products andthe services descrivedin this document are subject to modification and cannot under any circumstances be regarded as contractual. ‘Tho programming examples described inthis manual are intended for guidance only. They mus be specially adapted before they can be used in Programs with an industrial application, according tothe automated system used and te safety levels required © Copyright NUM 1997. ‘Allights reserved. Noppartof this manual may be copied or reproduced in any form orby ary means whatsoever, including photographic ormagnetic Processes. The transcription on an electronic machine ofall or pat ef the contents is forbidden. © Copyright NUM 1997 software NUM 1000 tne. ‘This sotwareis the property of NUM. Each memorized copy ofthis software sold conters upon the purchaser anon-exclsive loencestrictyimited to the use of the said copy. No copy or other form of duplication of tis product is authorized 2 n-938956/1Table of Contents T General Description TT 1.1 Document Presentation 1-3 1.2 General Information on the Tool Macros 1-3 1.3 Functional Description of Total Anticipation 1-4 14 List of Machine Parameters to Be Set 1-7 1.8 Total Anticipation Enabling Conditions 1-8 1.6 Order of Settings 1-9 2 Setting Procedure 2-1 2.1 Searching for the Optimum Gain on the Different Axes and Measuring the Corresponding Time Constant 2-3 22 Aligning All the Axes of the Group on the Longest Time Constant 2-3 23 Setting the Maximum Radial Error (P52) and Minimum Block Execution Time (P51) 2-5 24 Selecting the Filtering Time 2-5 25 Setting the Acceleration Anticipation 2-6 2.6 Setting the Maximum Permissible ‘Acceleration for the Axes (P32) and the Minimum Acceleration Time (P53) 2-7 2.7 Influence of the Reference Filtering Time on “Transients* 2-9 28 Setting the Approach Speed 2-11 29 Setting the "Anti-Pitch" Function 2-12 2.10 Setting the Peripheral Parameters Concerning Machining by Small Segments 2- 15 2.11 Examining the Path Error Trace 2-16 2.12 __ Displaying the Settings of the Coefficients 2-16 3 Description and Implentation of the Tool Macros 3-1 3.1. Macro G254 - Cycle for Aligning the Position Gains on the Desired Time Constant 3-3 32 Macro G255 - "Ball-Bar' Trace 3-6 3.3 Macro G253 - Setting the Approach ‘Speed - Examining the Position Loop Response 3-13 3.4 Macro G2S2 - Path Error Trace 3-18 3.5 Display Programme 3-25 4 Appendices at 4.1 Pitfalls to Be Avoided 4-3 42 Tool Mactos, Review of Call Possibilities 4-5 79389561 3— onum 4 en 9389561Record of Revisions DOCUMENT REVISIONS Date [Revision Description 01-96 O Document creation 06-07] 1 Conforming to software index K Manual upgrades: + Setting the speed transition time of sin? acceleration + Setting the minimum block execution time (P51) - Time constant per axis - Upgrading of macro G252 (window programming in machine dimensions) 0 93895641 5— onum 6 £0 938956/0Introduction Document History: Setting Procedures and Tool Macros Unreferenced (Prospective Studies), 12 December 94 Software version G: First issue. Unreferenced (Prospective Studies), 10 March 95 Software version G: Additional details. DES.DMP95012/A, 8 September 95: Inclusion of Changes in Version J Inclusion of the "maximum radial error” parameter (P52) on curves machined by small segments (TABCYLS): Until version H, P52 was used only for circles (G02/GO3)_On curves machined by small segments, there was no possibilty of speed control according to the curvature. This problem was solved by "special use of the "block preparation time" parameter (P51) In version J, the use of P52 is extended to curves by small segments and P51 recoversiitsnormal use. For High Speed Machining, P52 (maximum radial error) must be set to one ortwo micrometres. Setting of the parameters related to curves machined by small segments: ‘Two peripheral parameters were created to be able to distinguish between the segments forming a curve and normal segments. A segment is considered as part of a curve if itis sufficiently small and is at a small angle to the adjacent segments, 82004: Sagittal error: This parameter is used in conjunction with the angle formed by two consecutive segments to distinguish between "short" and “normal” segments. This parameter is expressed in micrometres. It must be set to 1.25 times the value of the “sagittal error" used by the CAD/CAM to create the current programme. In case of doubt, itis recommended to specify a high sagittal error of 200 to 400 um. £32003: Maximum connection angle: ‘This parameter specifies the maximum angle above which two consecutive segments areno longer considered as part of the same curve. Their connection is then analysed as an angle. This parameter is expressed in degrees. The usual settings are from around § degrees to 10 degrees. "On-line" setting of the reference filtering time and “anti-pitch” parameters Those settings do not modify the machine parameters in the EEPROM but only their image in the memory. = Thetwo"Anti-Pitch" compensation parameters (intial amplitude and compensation signal time constant) are visible and can be edited on the “Ball-Bar’ page (G255) ~The “teference filtering time parameter’ is visible and can be edited on the “Ball-Bar" page provided this page is called in "Static Trace" mode (P1: stop after each revolution): G255 F8000 G17 P41. This parameter canbe modified during a cycle stop but the modification is not taken into account until the next time the "CYCLE" button is pressed. Inclusion of the “path error trace” macro (G252) ‘The speed versus the time is displayed using a new menu called “Scope. en-9389560 7— onum 8 €-938986001 General Description 1.1_Document Presentation 1-3 T.2_General Information on the Tool Macros 1-3 1.3 Functional Description of Total Anticipation 1-4 1.4 List of Machine Parameters to Be Set 1-7 1.4.1 Parameters Concerning the Basic Control ‘System Already Existing in Version F 1-7 1.4.2 Parameters Concerning Total Anticipation and sin2 Acceleration 1-7 1.4.3 Anti-Pitch Compensation Parameters 1-7 1.44 Peripheral Parameters Concerning Machining by Small Segments 1-7 1.4.5 Synchronised or Duplicated Axis Dectaration in Po 1-7 1-8 T6 Order of Settings 1-9 1.6.1 Basic Control System Settings to Be Carried out with the Total Anticipation Option Inhibited 1-9 1.6.2 Total Anticipation Settings to Be Carried out with This Option Enabled 1-9 1.6.3 Setting The Default Values of the Parameters Conceming Machining by ‘Small Segments 1-9 1.6.4 Checking Machine Behaviour by “Dry Run* Machining of Qualification Parts 1-8 0-938956!1 1-1— onum 1-2 en. 938956)General Description 1.1. Document Presentation Inthe context of this document, the term "High Speed Machining” refers to the automatic control technique using total. speed anticipation and acceleration anticipation. This document gives an axis setting methodology which is valid for total anticipation. However, the first steps in the procedure also apply to normal feedback, .e. without anticipation. This methodology uses tool macros, developed to facilitate the settings. These tool macros are described in greater detail below. This document also covers the settings relative to the "ant-pitch compensation function’. From version J on, function no longer depends on the High Speed Machining function and is now included in the basic functions. This document includes the four following chapters: = Chapter 1, General Description, - Chapter 2, Setting Procedure, = Chapter 3, Description and Use of the Tool Macros, - Chapter 4, Appendices. 1.2. General Information on the Tool Macros All the macros can be executed in a part programme and in MDI mode. When axis coordinates are included in the call toa macro, the point defined by these coordinates in conjunction with the intial position determines an area of movement in which all the subsequent movements are carried out. ‘The ‘useful movements are all carried out with the feed rate override potentiometer locked on 100 percent (M49) However, before each operation, a complete cycle is carried out in dry run mode with potentiometer enabled (M48) to make sure there is no collision on the machine. For the tests to be significant, the macros must always be executed at the maximum work rate. This speed must be Programmed when the macro is called, ‘The macros provide a large number of functions. Not all the macros are needed for routine settings. The setting procedures described below specify the macros that are normally used. The others are provided as tools for adaitional diagnostics, The use of these macros requires certain precautions, which are described in Chapter 4, Section 4.1.1, “Tool Macro Implementation Conditions’. Itis absolutely essential to read this section before attempting to use the macros. en 9389561 1-31.3 Functional Description of Total Anticipat Total anticipation is characterised by four functionalities: = Total speed anticipation, + Acceleration anticipation, = Filtering of the "Position reference signal’, Damping of movements by a "sine square" spoed law. in Position settings Emergency stop potentiometer — a var ‘speed law Interpolator [ugSPE0d pe Remaining distance Total speed _. anticipation Acceleration anticipation X axis position rene Reference |_| Servo-drivesmotor+ gM fitering machine X axis position a Speed sett —X axis po Position gain peed setting Y axis position 1 reference Z axis position "> reference Functional block diagram of with total speed ant 1-4 (€n- 93895601General Description —eeeeeeeeeeeeeeeeeeEEeEeEeEeEeEeEeEeEeEeEeEeEeEeEeEeEeEEeeee ‘Speed anticipation ‘Total speed anticipation cancels the following error during a steady linear movement. This action is controlled only by machine parameter P56 (Position time constant). Without accompanying measurements, this action would cause: = Enlargement of circles by an extent related to the servo-drive response time, = tendency to overshoot the position when stopping. ‘Acceleration anticipation ‘Acceleration anticipation cancels the radial error on a circle travelled at constant speed. An increase in this parameter moves the real path toward the inside of the circle and a decrease toward the outside. This parameter also cancels the following error on a uniformly accelerated linear movement. It is set using tool macro G25S ("Ball-Bar" Trace). ‘Acceleration anticipation ‘Acceleration anticipation setting too low setting too high ZA Zero error ine Resulting "BALL-BAR" Resulting "BALL-BAR™ trace trace Position reference filterin, The “position reference filtering" function is necessary for machining curves by small segments (Taboyls). The main purposes tosmooth the anglesin the path as much as possible toallow the double derivative of the ‘position reference” signalo be calculated (.e. so thatitis dtferent from zero). The use of reference filtering moves the circular path toward the inside of the circle and therefore tends to “underset’ the action of acceleration anticipation. Underset the acceleration anticipation to compensate for this effect. To compensate for this effect, decrease the acceleration anticipation, en-938956/1 1-5— onum Sine square speed law With total anticipation, the “trapezoidal” speed law generally used would cause position overshoots, especially when stopping. The sine square speed law prevents the position overshoots by damping the movements. The sine square law effectively used in the current versions damps the movements better for large speed errors than for small errors. Itis planned to improve this law in forthcoming versions to optimise its behaviour. Sine square speed law Speed Slope: Ym Time 1-6 n-938956/1General Description 1.4 List of Machine Parameters to Be Set 1.4.1. Parameters Concerning the Basic Control System Already E: P21 (E980xx): Internal position gain per axis (KVAR). P56 (E4106): Position loop time constant per axis group (in ms). This is the ratio of the following error to the speed. P32 (E971x/E972xx): Maximum permissible acceleration (in mms"). There are two values per axis: work rate and traverse rate. se (E32002): Maximum permissible radial error on curves per group (in micrometres). P51 Minimum block execution time, 1.4.2 Parameters Concerning Total Anticipation and sin? Acceleration PSSINB..N15 (E3205): Position reference filtering time per axis group as a number of samples. P19/NO..Nt (E981XX): Acceleration anticipation coefficient per axis (in ms/1024) P33/NO..N31 (E973XX): Approach speed per axis (in mm/min). This parameter is used to adjust angle crossings. Itrepresents the maximum speed level that can be tolerated on an axis without causing any undesirable overshoots. P53 (E3206): Minimum time for speed transition with sin* acceleration (expressed in ms). 1.4.3 Anti-Pitch Compensation Parameters P19/N32..N63 (E982xx):_Anti-pitch signal amplitude at the initial time. P19/N64..N95 (E983xx): Anti-itch compensation signal time constant (in hundredths of ms). 1.44 Peripheral Parameters Concerning Machining by Small Segments 32004: Sagittal error. This parameter must be equal to 1.25 times the sagittal error parameter used by the CAD/CAM. It is expressed in micrometres. F32003: Maximum angle between consecutive segments of a curve. Itis expressed in degrees. 14.5 Synchronised or Duplicated Axis Declaration in P9 ‘The first four bts (nibble) of each word should contain the group number, as for normal axes, instead of FF as before. 9389561 1-7— onum 1.5 Total Anticipation Enabling Conditions ‘The Total Anticipation option must be enabled on three levels: ~ Option bit level: ‘The UGV (high speod machining) bit (bit 18) must be equal to 1. ~ Machine parameter level: The option is conventionally enabled by specifying a nonzero reference filtering time in machine parameter PSS/NB..N15. Each of the words of this parameter corresponds to an axis group. = Peripheral parameter level: E11012 must be set to 1 At initialisation, this parameter is forced to 1 or 0 according as the two above conditions are satisfied or not. This bit can of course be written by a programme. A\ CAUTION If total anticipation is enabled, the sin2 speed law must also be enabled. This law is enabled by programming, by setting peripheral parameter £11013 to 1. At initialisation, E11013 is initialised with the value of machine parameter P7 word 1 bit 5. 1-8 (609389561General Description 1.6 Order of Settings 1.6.1 Basic Control System Settings to Be Carried out with the Total Anti Option Inhibited + Search for the optimum gain (P21) and measure the corresponding time constant tobe carried out for each of axes. ‘The time constant corresponds to the ratio: (Following error)/Speed. + Align all the axes of a group on the longest time constant. 1.6.2 Total Anticipation Settings to Be Carried out with This Option Enabled ~ Set the maximum radial error (P52) and the block execution time (P51). = Select the approximate filtering time (PSS/NB..N15). - Set the acceleration anticipation coefficients for each axis (P19/NO..N31) using macro G255 ("Ball-Bar" trace). - Set the maximum acceleration per axis (P32). = Set the minimum speed transition time (P53) for software version K or above. = Check the behaviour on the transients. If necessary, readjust the filtering time and acceleration anticipation. This is done using the “Ball-Bar’ macro (G255 P1). ~ Set the approach speed. = Set the anti-pitch compensation, 1.6.3 Setting The Default Values of the Parameters Concerning Mac! Segments ‘These settings concern the peripheral parameters. The settings remain after a reset but can be reinitialised with the default values by programming in the reset subroutine (11000). 1g by Small 1.6.4 Checking Machine Behaviour by "Dry Run" Machining of Qualification Parts This check is carried out using the path error trace macro (G252). .0-938956/1 1-9— onum 1-10 0-938956/12 Setting Procedure 2.41 Searching for the Optimum Gain on the Different Axes and Measuring the Corresponding Time Constant 2-3 22° Aligning All the Axes of the Group on the Longest Time Constant 2-3 22.1 Digital Servo-Drives 2-3 2.2.2 — Analogue Servo-Drives 2-4 223 — Using Time Constants Adapted to Each Axis 2-5 2.3 Setting the Maximum Radial Error (P52) and Minimum Block Execution Time (P51) 2-5 2.4 Selecting the Filtering Time 2 25 Setting the Acceleration Anticipation 2.6 Setting the Maximum Permissible Acceleration for the Axes (P32) and the Minimum Acceleration Time (P53) 2-7 2.7_Influence of the Reference Filtering Time on “Transients” 2-9 2.8 Setting the Approach Speed 2-11 2.9 Setting the “Anti-Pitch” Function 2-12 2.9.1. Functional Description 2-12 2.9.2 Compensation Principle 2-12 293 Parameters to Be Set 2-13, 2.9.4 — Setting Technique 2-13 2.9.4.1 Setting in Version G or H 2-13 2.9.4.2 Setting in Version J (Using the “Ball-Bar* Trace Macro) 2-14 210 Setting the Peripheral Parameters Concerning Machining by Small Segments 2-15 240.1 Sagittal Error (E32004) 2-15 2.10.2 _ Maximum Connection Angle (E32003) 2-15 2.17 Examining the Path Error Trace 2-16 2.12 Displaying the Settings of the Coefficients 2-16 en-938956/1 2-1— onum 2-2 en 93895611Setting Procedure 2.1. Searching for the Optimum Gain on the Different Axes and Measuring the Corresponding Time Constant The time constant corresponds to the ratio of the following error to the speed without anticipation: (Following error) Speed ‘Two techniques are used for searching for the optimum gain: ~ Use of SETTOOL (see Operator Manual), = Use of macro G253 (Q3 call parameters). Macro "G253 Q3" traces the position response of an axis subjected to a speed level (constant speedi/zero speed transition) The position loop is kept closed. Example: GO X Y E11012=1 E1101: G253 X Y I Q3 F ‘The procedure for use of this macro and its syntax are described in detail in Chapter 3, "Description and Use of the Tool Macro: 2.2 Aligning All the Axes of the Group on the Longest Time Constant 2.21 Digital Servo-Drives With integrated digital servo-drives, equality of the gains (P21) ensures equality of the time constants (all other things being equal) It is suggested to set all the gain parameters (P21) of an axis group to the value found for the least performing axis. The axis time constant can then be accurately identified using macro G254. Considering the remark made above, itis sufficient to identify a single axis. The accurately identified value must be stored (in ms) in parameter P56 for the group considered Example: Go X G254 Fem ax feed rate> X A\_ CAUTION The value of the time constant obtained by the above operation must be equal to an exact number of milliseconds (within a few microseconds). If not, adjust the setting of P21. This precaution is required because parameter P56 expresses the time constant as a whole number of milliseconds. 60 93895601 2-32.2.2 Analogue Servo-Drives In the case of analogue servo-drives, the gain (P21) must be set individually for each axis to obtain the exact value of the time constant to be set in parameter P56 for the corresponding group. Itis necessary to set each individual axis because of the dispersion in the characteristics of the axes. A dispersion of the time constant would result in “ovalisation’ of the ""Ball-Bar™ trace according to the diagonals of the quadrants and therefore a deformation of the circle effectively machined, Such a deformation could not be compensated for by any other setting. A dispersion in the time constant causes a difference between the two extreme dimensions of the path that is equal to four times the product of the speed multiplied by the difference in time constants: ID1 -Dal=4*V+dT af th nthe path [Di - v2] = 4" vedr The relative accuracy required on the axis time constants is around 1/1000 ‘Suggested Procedure: ~ Set the group time constant parameter (P56) to the value measured for the least performing axis (longest time constant measured), ~ Align the gain of each axis (P21) on the time constant (PS8) using the alignment cycle (G254): G0 X 254 Femax_machining_speed> X R This cycle actually adjusts only the memory image of the gain (P21): E98Ox«. ‘The operator must therefore not forget to record the values obtained in the words of P21. Otherwise, these values will be lost the next time the systems reset. To facilitate read of the values set, programme %12000 displays the contents ofall the peripheral parameters E980xx on the Graphic Parameters page (Plane 1+2). (In versions before J, the display programme number was %8081) 2-4 0 93895641Setting Procedure es 223 Using Time Constants Adapted to Each Axis Starting with software version K, it is possible to set a different time constant for each axis. In this case, the time constants are declared in parameter P66, with one word per axis. When the contents of a word of P66is different from 0, its value is substituted for the value declared in P56 for the corresponding axis. This is theoretically possible with high speed machining, but has not yet been tested in practice. Ifthe same time constant is set on all the axes (conventional setting), it must be made sure that all the words of P66 are set to 0. 2.3. Setting the Maximum Radial Error (P52) and Minimum Block Execution Time (P51) With High Speed Machining (UGV), the maximum radial error parameter (P52/E32002) must absolutely be set, generally to a low value (one to a few micrometres). From version J on, this parameter is used not only for circles (as was the case in the previous versions) but also for curves programmed by small segments ("TABCYLS"). This Parameter limits the speed according to the curvature of the path so as never to exceed the maximum error specitiod, P52’is always enabled, whether or not High Speed Machining is used. However, for the same value of P52, the speed limiting is less penalising in UGV (E11012 = 1) than for normal machining (E11012 = 0) This behaviour is authorised by the gain in accuracy provided by High Speed Machining. This speed advantage justifies the use of UGV. Starting with software version K, itis recommended to set parameter P51 to 0 (minimum block execution time). This change is due to the fact that itis now possible to interpolate several blocks during the same sample. 2.4 Selecting the Filtering Time ‘The position reference signal fering time is set by machine parameter PSS/NB..N15 with one word for each axis group expressed as a number of samples. The setting of this time therefore depends on parameter P50 (Sampling period). ‘There is nomathematical rule for defining its value. The filtering times that seem best suited by experience are between 75 percent and 100 percent of the position time constant. It should be noted that the setting of this parameter affects the acceleration anticipation parameter. Too high a value of the filtering time can result in negative values for the acceleration anticipation coefficient. The filtering time parameter is available (version J) in parameter £32005. ‘The filtering time parameter also affects transient deformations in the path. For this reason, the first setting may need tobe trimmed, but this should not be carried out until the acceleration anticipation has been set. This aspectis detailed below. ZA\ CAUTION The filtering time cannot exceed the value of 13 sampling periods as the servo-system behaviour is unpredictable above this value. en-938956/1 2-52.5 Setting the Acceleration Anticipat Polar trace of the radial error: "Ball~Bar* Polar trace of the radial error (BALL-BAR") ‘Theoretical circle (entry of the settings) rd "__ Softkeys ‘The acceleration anticipation s set for each axis while observing the "Ball-Bar* trace on the screen obtained by circular interpolation of two axes. The trace is obtained by executing cycle G255 which animates the selected axis pair with a circular movement and by the trace on the graphic page ("Scale 142") of programme %9997.9. ifat least one of the two axes selected is not a main axis (X, Y or Z), circular interpolation is carried out by small line segments (Tabcy!) The use and syntax of macro G255 are described in detail in Chapter 3. As an illustration, the following calls are possible: E11012=1 £11013=1 GO Y Z (centre position), 6255 [G02] [G18] [F] [Reradius>] [P] (Circular interpolation by G02), or E11012=1 £11013=1 GO Acdimension1> B (centre position), G255 [G02] A B Qcchordal_errom> [F] [Reradius>] [P] _(Tabeyl. Optional arguments are enclosed in square brackets [ } 2-6 (€0-938956)1Setting Procedure a ibd Once interpolation has started, the programme can only be stopped by pressing CYHLD and RESET. Macro G255 does not force the total anticipation (E1 1012) or sine square (E11013) enable conditions, which must frst be forced if necessary by programming £11012 = 1 and £11013 = 1. ‘The programmed feed rate (F) must be as high as possible. Itis recommended to programme the maximum machining {eed rate for this value. The choice of a lower feed rate would lead to an insufficiently accurate setting. Ris the radius. Ifitis not specified, the macro uses the radius causing an error equal tothe resolution unit (micrometre) at the programmed feed rate. Itis recommended not to set parameter R. When the “Ball-Bar" trace display (%9997.9) is active on the "graphic" page, this page gives access to a keyboard dialogue to edit the acceleration anticipation coefficients. Increasing the coefficient associated with an axis contracts the trace in the direction of this axis and decreasing it expands the trace. When G255iscalled, parameter P specifies the type of trace. Itis recommended not to programme parameter*P* when setting the acceleration anticipation. If *P* is not programmed, the dynamic trace displayed immediately shows the effect of any changes made. 2.6 Setting the Maximum Permissible Acceleration for the Axes (P32) and the Minimum Acceleration Time (P53) REMARK —_Inversions G, H and J, the technique used to implement the “sine square" speed law is such that the maximum permissible acceleration is always reached at the knee of the curve. This means that the lower the amplitude of the speed transition, the higher the derivative of the acceleration. With total speed anticipation, there is therefore a risk of overshooting the position when stopping, especially if the transition is made at low speed. The higher the maximum acceleration declared for the axes, the greater this risk. These considerations could lead to restricting the range of acceleration on certain machines. As an illustration, the table below shows the overshoots in micrometres measured on a real machine according tothe speed difference and maximum acceleration declared: Overshoot when stopping (micrometres) Initial speed Maximum deciared acceleration (P32) (mmvmn) Tmis2 0.5 mvs2 2000 ° ° 1600 ° 800 "1 1 400 1 5 200 5 100 2 2 e0-938956/1 2-7— onum Starting with software version K, aparameter has been created to specity the minimum duration ofthe sin2 acceleration phase. This parameter, PS3, is expressed in miliseconds. It is generally set to a value equal to two or three times the position loop time constant (P56). This parameter limits the derivative of the acceleration. By definition, itaffects ‘only low amplitude speed changes. ‘pcos enor= av Sin acoceration Transition time according \\_Max accel. = P32 «Max accel. = 244 V/ P53 —__———» Pa Transition time With sin® acceleration, the derivative of the acceleration is equal to the maximum acceleration reached, divided by half the transition time, The derivative of the acceleration is a maximum when the speed error to be corrected is equal to DV, show in the figure below. In this case, the maximum acceleration is equal to P32 and the transition time is equal to P53, ‘The maximum value of the derivative of the acceleration is then equal to: Paz P53 Vmax = Software version K Large speed error: Small speed error: Max. acceleration Max. acceleration Software version J rsanical not reached ‘A Speed : Speed A Speed : 4 Speea | | ' ' | i T ' Time J | Time | Time | {Time } > ~ -_ 7 7 3P5S 8 ‘A Acceleration eeterten ' ‘Acceleration (ymax : P32) Acceleration 2-8 en-93895611Setting Procedure 2.7 Influence of the Reference Filtering Time on "Transients" For high speeds and low curve radii, a transient error on the theoretical path can occur during acceleration or deceleration This error can be illustrated using the “Ball—Bar’ trace with the “static trace" option (parameter P1). Example of call: 255 G18 F600 P1 The above macrocall forces the CNC to draw acircle whose radius is calculated sothat the radial error does not exceed the internal resolution (0.001 mm or 0.0001 degree). The path is travelled at the programmed feed rate (here F6000) When the circle is complete, the machine goes on wait on MO. It starts machining another circle when "Cycle" is pressed, For each circle, two path segments are shown on the screen (%8987.8). These two segments correspond {0 acceleration and stopping, Han deviation from the theoretical path occurs, it can generally be compensated for by adjusting the filtering time. ‘According as the deformation at startup is toward the outside or inside of the circle, a correction can be attempted by increasing or decreasing the filtering time respectively. (The path deformation during stopping is generally in the ‘opposite direction from the deformation during startup). ‘Acceleration ‘Acceleration. Braking Without reference filtering Reference filtering time too long Influence of the reference filtering time on transients (Starting and stopping on a circle) Ifa change is made to the “filtering time" machine parameter (PSS/NB..N15), the acceleration anticipation coefficients must then also be modified. The change in these coefficients is calculated as follows, based on the change of the filtering parameter: (nr - n2%2) tere c2-c1= — . + 1024 24 Tp C1: isthe old acceleration anticipation coefficient, C2: is the new acceleration anticipation coefficient, nt: is the old fitering time (as a number of samples), 12: is the new fitering time (as a number of samples), te : is the sampling period (in ms), Tp: isthe position loop time constant (in ms). en. 938956/1 2-9— onum Coefficient 1024 takes into account the fact that C1 and C2 are in ms/1024, i.e. approximately in microseconds. Itshout ted that st ring time ean fied using the "Ball {G255). In this case, parameters E961x - which are the images of the acceleration anticipation parameters for the ‘axes of the group - are automatically readjusted by the macro. Using the "Ball-Bar" trace macro (255) Start the “Ball-Bar* macro G25... Pt and call the page: PROCAM Graphic Parameters - programme 9997.9 - Scale 1+2 The filtering time is adjusted with the machine stopped by pressing a function key and entering a value from the keyboard. Static trace mode (MO after each revolution): ‘The operator can enter anew filtering time (Nfiltr) between 1 and 13. sampling periods. The value entered is only stored. !tis not applied until Cycle is pressed again to start a new revolution. The new behaviour of the axes can be observed directly on the screen. tis recommended to make the setting without setting parameter R (radius). The “Ball-Bar" trace has the following shape: ‘New Nir Stored: +5 PRESS “CYCLE” ! ¥pteht(X):610 ‘Tptent(x) +5036} ptenz(¥) 410 ‘TpIch2(Y) +5036] Kacel(X):+502, Kace2(¥):+502 Nours+5 Cale_R:+28.5 axe 1X 12: de) ace fn: est 13:Kacel_(4:Kace? 1$:Vpteh}_{6:Vpteb2_(7:Tptcht_{8:Tptch?_19:Nfiltr (0: No Change) Nii [1..13) = § I Parag | |tshould be noted that a change in the filtering time causes automatic readjustment of the image of the acceleration anticipation parameters (E981x2) for all the axes in the group by application of the equation described above. Too high a filtering time can result in negative acceleration anticipation coefficients. These values would be rejected with display of the following message: Kace would be negative “Nfilte ignores 2-10 en-938956/1Setting Procedure se E 2.8 Setting the Approach Speed With total anticipation, a speedievel applied to an axis always causes a position overshoot. The overshootis especially large if the speed level leads to stopping of the axis. The amplitude of the overshoot depends on the size of the level. ‘The “approach speed” machine parameter (P33) represents the maximum permissible speed level that can be used on an axis without excessive overshoot. This parameter is used to set the maximum authorised speed at angle crossings. In effect, when an angle is crossed at a constant speed V, crossing causes a speed level DVi on axis i equal to: DVi=V* (ICos(ait) - Cos(Ai2)!) Ait and Ai2 are the angles formed with axis i by the two line segments before and after the angle. Itis recommended to set this parameter using macro G253 with calling argument Q3. Macro G253 determines the value of the “approach speed! according to the specified overshoot. Values commonly used are around 200 mrrvmin for a tolerated overshoot of 5 micrometres. Z\ CAUTION In versions G and H, the approach speed machine parameter (P33) must be forced to 0 before using macro G253, Otherwise, operation would be disturbed by processing the angle crossing and the result would be meaningless. From software version J and above, this precaution is no longer necessary, since the macro temporarily inhibits processing of the angle crossing at the point of interest (use of 11015). 4 D A x ¢ B +—. Speed on Y | —J = ' Position on ¥ Time Crossing an angle with total anticipation, representation of the speed level and position overshoot (en-938956/1 2-112.9 Setting the "Anti-Pitch" Function 2.9.1 Functional Description “Pitch" is an error occurring on "Ball-Bar" type traces that results in spikes pointing outward from the circle just after changing quadrants. Real path with enlargement ot tie iain Secon) Display of the straight section on the polar trace of the radial error (BALL-BAR) These trace errors actually identify the presence of straight sections in the real path which occur after changing quadrants. These straight sections amplified by the "Ball-Bar* trace appear as spikes on the screen. These deformations are generally explained by the fact that when the direction of movement on an axis is reversed, ‘a discontinuity occurs in the inertia component due to drive friction. This component changes at that time from a finite value in one direction to a finite value in the other. The spikes are consequence of this discontinuity. Itis as though the servo-drive did not correct the effects of the discontinuity fast enough (play take-up or pitch on the axis?, servo-drive reaction time?..) REMARK Starting with version J, the values of these parameters are displayed on the "Ball- Bar" page and can be edited by dialogue. See below. 2.9.2 Compensation Principle ‘The principle used in version G to compensate for this error was to accelerate movement on the axis when reversal ‘occurred by “injecting a disturbance’ in the feed rate setting. This disturbance is called “compensation signal" below. ‘Although purely empirical, this technique considerably attenuates the pitch. It is however difficult or even impossible to.liminate pitch completely. In addition, this technique loses some of ts efficiency when too far fromthe setting range (feed rate and radius) 2-12 (€0-938956.Setting Procedure 2.9.3 Parameters to Be Set ‘The "compensation signal” level decreases exponentially with time. Its waveform is defined by two parameters: DVcm: Signal amplitude at the initia" time, ie. the time when the position reference changes sign. This ‘amplitude is set by machine parameter P19/N32..N63 Exponential time constant, set by machine parameter P19/N64..N95. DV. "Anti-pitch” compensation ‘Speed setting compensation signal waveform Time of movement reversal Vem (intial amplitude) isin micrometres (or ten-thousandths of a degree). The actual compensation signalis a speed. ‘This speeds generated by dividing DVom by the position time constant Tp. Physically, itis calculated in intemal units by multiplying DVem by the position gain, tt (compensation time constant) is in hundredths of milliseconds (0.01 ms) 2.9.4 Setting Technique Setting is carried out by trial and error, simultaneously adjusting both parameters (initial amplitude and time constant) relative to an axis. The result is checked by observing the appearance of the "Ball-Bar" trace after each modification. 20.4.1 Setting in Version G or H Inversion G or H, the only way to set these parameters is by Utility 5 and each new setting requires resetting the CNC and rehoming. Setting of each axis therefore requires repeating the trace, adjustment, CNC reset, homing cycle several times In versions G and H, the anti-pitch compensation setting procedure is therefore relatively long. 0 93895601 2-13— onum 294.2 Setting in Version J (Using the "Ball-Bar" Trace Macro) Starting with version J of the software, the values of these parameters are displayed on the Ball-Bar page and can be directly modified by dialogue using the following procedure: Start the Ball-Bar macro preferably in dynamic trace mode without parameter P or in semi-static mode (PO): G255 ... G255 ... [PO] and call the page: PROCAM, Graphic Parameters - programme 9997.9 - Scale 142. ‘The parameters are selected by the corresponding function key (F5 to F8) ‘The operator is prompted to enter a now value for the parameter selected. ‘Then he can see on the screen the new trace obtained by circular interpolation of two.axes together. The Ball-Bar trace has the following appearance: Kacet(x):+502 Vpteni(X) +10 Tpieb1(X):+5036| ‘Wptena(¥):+10 Tprcha(¥):+ 5034] Kace2(y) +502 / Cale_R-428.5 \ cea Y Nate axis 1x, (2: deb trace 11: edt 13:Kacel_f4:Kace? 1$:Vpteh1 {6:Vpteb2_f:Tpteh1_{8:Tpteh2_(9:Nfitr (O: No Change) Tpteht (unit: 0.01ms) = S000 5 6 7 im te, EEarage | 2-14 (€n-938996/1Setting Procedure ——— EE 2.10 Setting the Peripheral Parameters Concerning Mac! Segments Until version H, "small" segments were treated as “normal" line segments. Therefore, the angle crossing routine was always performed between two segments. Not only did that cause a loss of time, butit sometimes resulted in unjustified local decreases in speed along perfectly connected curves. This behaviour was due to the fact that angle crossing analysis affects the speed more adversely than curvature analysis, ing by Small In version J, the attempt was made to distinguish between the elementary segments of a curved path ("small* segments) and normal line Segments (‘normal segments) A segment is considered "small if the following relation is satisfied: Length < (8 * Sagital_errory/Angle. Length: ‘Segment length. Angle : ‘Angle formed with the previous segment. Sagittal_error: _ CAD/CAM parameter defining the accuracy of the path. This value is specified in parameter £32004. In adaition, in version J: + Curvature analysis is performed only for ‘srall" segments, + Angle crossing analysis is performed between two segments if one of the following conditions is satisfied: ‘Atleast one of the segments is not a “small" segment, The angle formed by the two segments is larger than the value specified by parameter £32003, Parameters £32004 (Sagittal error) and £32003 (Maximum connection angle) preserve their value after a reset. 32004 is initialised with 0.010 mm and £32003 with 0 degrees. These default values are probably not suitable for all applications. It is therefore recommended to set these parameters to the values required for the routine type of application on each site in the reset subroutine (%11000). 2.10.1 Sagittal Error (£32004) ‘This parameter must be equal to 1.25 times the sagittal error specified for he active programme in the CAD/CAM, The margin of 25 percent corresponds to the variation in length observed at constant curvature on the reall programmes produced by the CAD/CAM. In some cases, small segment programmes are not produced with a "constant sagittal error" (e.g. generation by macro). In this case, itis recommended to specify a high sagittal error value in parameter 32004 (200 um or 400 jum). This parameter is expressed in micrometres. 2.10.2 Maximum Connection Angle (E32003) Values of 5 to 15 degrees appear suitable. This angle must not be too small. Otherwise, angle crossing analysis could be activated along parts with a very pronounced curvature. It should be noted that the edges of fiat dihedrals (with an angle less than the setting of 32003) are crossed at a speed and may therefore be somewhat blunted. Example: A sagittal error of 0.010 mm on a circle with a radius of 3 mm gives segments forming an angle: a= ((8 * sagittal errory/R) * (1/2) = (8 * 0.010)/3) 4 (1/2) = 0.16 rad = 9.4 degrees. This parameter is in degrees. £n-938956/1 2-15— onum 2.11 Examining the Path Error Trace ‘When the settings are completed, itis advisable to examine the behaviour of the machine while machining one or more “qualification” parts. itis suggested to use the “Error trace" macro (G252) for this purpose. This macro displays the ‘graph of the position error along a selected segment of the path. Certain parameters, in particular the maximum radial error parameter (P52/E32002), may need to be adjusted, depending on the results. For instance, the path errors after angle crossings can generally be corrected by decreasing the “approach speed" parameter (P33/E973xx) Macro G262 also displays the graph of the tangential speed versus the time along the segment of path examined. This, allows any unjustified decelerations to be detected. Such decelerations can generally be corrected by increasing the sagittal error parameter (E32004) or the “maximum connection angle" parameter (E3203) 2.12 Displaying the Settings of the Coefficients When the coefficients are set by macro G255, only the memory image of the coefficients (E980xx and E981xx) are ‘modified but not the corresponding machine parameter (P19/NO..N31). ‘The machine parameters must therefore be set using Utility 5. In the eventof an accidental CNC reset, parameters E980xx and E981 x ] X < dimension > [P] Alignment of the position gain of an axis (KVAR) on the “position time constant" parameter (P56) G254 [ F < feed rate > ] X < dimension > RO [P] ‘Alignment of the position gain of an axis (KVAR) on a given value of the time constant (Tp in ms) G254 [ F < feed rate > ] X < dimension > R < Tp > [P] REMARKS _ X denotes an axis address: X, Y, Z, U, V, W, A, BorC. Only one axis address must be included in a G254 block. The back-and-forth travel is carried between the current position when the call to the cycle is made and the position indicated by the axis address, With address “R’, parameter E9800x (KVAR) is modified. The presence of optional address "P" allows this macro to be used even without the High Speed Machining option. G91 (incremental programming) is prohibited. 3.1.3.2 Movement Specification Movement can be programmed by G90 (absolute workpiece dimensions) or G52 (absolute machine dimensions to be Programmed on each call block). Programming by G52 is advantageous in a "warm-up" programme because it eliminates the problem of shifts, Programming by G91 (incremental) is prohibited since itcould cause a misunderstanding of the "modality* 3-4 0 93895601Description and Implentation of the Tool Macros Ht the programmed movementis too small the axis does not have the time to reach stabilised speed and the oycle then continues to oscillate indefinitely. However, the cycle is able to detect too small a movement, in which case It displays the value of the optimum movement while continuing to oscillate. The optimum movement displayed is the one which allows the gain to be evaluated by a single round trip (preceded by the “off load round trip systematically conducted with M48 at the beginning of each cycle). A movement two or three times smallest than the optimum movement can be programmed but will cause more oscillations. ‘With modulo type axes, it is recommended not to programme movements above 180 degrees. If the programmed movement exceeds 180 degrees, the first movement is carried out as programmed, but the consecutive movements are carried out on the complementary angular path (excursion < 180 degrees). 3.1.3.3 Programming of Additional Addresses Function °@52": Programming in machine dimensions. cee Normally, the G254 cycle forces the accurate feedback enable parameter (E1012) and the gradual acceleration enable parameter (£11013) to zero. The first of these parameters must always be forced to zero. Forcing the second to zero allows faster accelerations. These parameters are returned to their initial state at the end of the cycle. However, there are cases when access to these parameters is prohibited and results in a programming error. These cases are as follows: ‘accurate feedback’ option absent (bit 18 = 0), “accurate feedback" option not enabled by machine parameter (P55, NB =0). To avoid a programming errorin the above cases, address “P* must be programmed in block G254. The presence ofthis address inhibits access to peripheral parameters £11012 and E1013, 314 Display Specific to Cycle G254 [At the end of the cycle, the following messages are displayed on the “Axis page" and the "Info" page: Use for evaluation: IDEN Tp = <__> Uncer Use for adjustment: SET Tp =<__>Uncer=< >Kp=<__> Measured position loop time constant. During adjustment, Tp is the value obtained after adjustment. Uncer: Uncertainty. The time constant is measured separately for each direction of movement. Tp is the mean value and Uncer is the difference between the two measurements, Kp: ‘Open loop position gain (KVAR) resulting from the adjustment. 2n-938956/1 3-5— onum 3.2 Macro G255 - "Ball-Bar" Trace 3.241 Functionalities Macro G255 is used to display a "Ball-Bar" type trace on the screen in order to adjust the "acceleration anticipation” parameter (E981xx) while checking the effect of this setting on the screen. The trace is actually a polar diagram of the radial error obtained when the tool travels a circular path. The error is shown on an enlarged scale around a circle symbolising the theoretical path. This is the type of race supplied when using the "Ball~Bar" system to qualify servo- controls. Cycie G255 causes execution of a circular path. Depending on the parameters selected, the tool travels either a single revolution or an uninterrupted sequence of revolutions. This sequence can only be interrupted by CYHLD and RESET. The first case of operation (one revolution each cycle) |s called "Static Mode" and the second is called "Dynamic Mode" or *Semi-Static Mode" depending on the type of display. Part programme %9997.9 displays the trace associated with cycle G255 and supports the setting dialogue. The real- time trace is obtained by calling programme %9997.9 by the "Scale 1+2" function key on the "Graphic Parameters’ page. When the display is active: = the "acceleration anticipation" parameters, and + the “anti-pitch" compensation parameters (trom version J on) relative to the two axes involved in the circle being executed can be modified by dialogue. It should be noted that this dialogue does not modify the parameters stored in EEPROM but only their active image in the memory. Accordingly, the settings must be re-entered at the end of the operation using Utility 5. Starting with version J, the “filtering time” parameter of the current group can also be modified by macro G255, but this ‘modification can only be made in "Static Mode”. The modification can only be entered when the system is stopped between two revolutions ands not taken into account until the next cycle. When the filtering time is modified, macro G255 agjusts the acceleration anticipation coefficients of all the axes of the current group according to the new filtering value. In this case as well, only the memory images are modified. ‘Table summarising the parameters accessible by macro G255 (in version J) Name Machine Parameter Keys EParameter Mode = Acceleration anticipation (G)_ P19 Noo..N31 Fo andF4 — E981xx all = Anti-pitch compensation amplitude (J) P19 N32..N63 F5andF6 = E982xx all - QW) P19 N64..No5 F7andF8 = E983x all - Filtering time W) P55 NB..N15 Fo £32005 static 3-6 (e0-938956/1