Field of the Invention

The present invention relates to a system for controlling draw on a moving web, and in particular, a system for maintaining a constant tension in a moving web to control the length of web moving past a cyclically operating apparatus in a printing press by controlling the angular position of an upstream roll relative to the angular position of the printing units.

BACKGROUND OF THE INVENTION In web fed printing press systems, a web of material, typically paper, is dispensed from a roll on an unwind stand and driven, in sequence, through a series of printing units, a dryer, a chill roll stand and various downstream processing apparatus such as a slitter, rotary cutter and folder. The printing units, which typically each include at least one set of asso¬ ciated plate and blanket cylinders, cooperate to imprint a multicolor ink image on the web. The dryer heats the

web to evaporate various solvents in the ink, and the chill roll cools the web to set the ink.

The various apparatus within the system are driven in synchronism, typically from a common shaft. A nominal ''press speed" is defined by the surface (circumferential) speed of the respective print unit cylinders. The surface (circumferential) speed of the various other units are expressed as a percentage of the press speed; some running slightly slower than and some running slightly faster than the print units. Such deviations in speed are small, typically on the order of a fraction of one percent. The tension of the web can be controlled by varying the relative speeds of the respective apparatus.

Uniformity of image repeat length can be critical to proper operation of the downstream processing apparatus. For example, a rotary cutter typically comprises a pair of cooperating cutting cylinders bearing one or more cutting blades. The cutting cylinders are rotated in synchronism with the printing units so that the blades intersect the moving web at predetermined points, e.g., between repeating images. It is necessary that the length of web moving past the cutting apparatus between cutting operations be substantially constant, so the cutting blades intersect the moving web on a repetitive basis in precise coordinated relationship with the repetition of the imprinted images on the web.

However, various conditions of the printing system tend to cause changes in the web substrate, and concomitant changes in position of images, and, thus, misregistry between the downstream apparatus and the images on the web. For example, changes in the tension of the web can affect the length of web between cuts. Variation in web tension can be caused by numerous factors, such as, for example variations in the tension at which the web was wound on the dispensing roll (e.g., paper is typically rolled at a higher tension at the

inside of the roll than at the outside) and variations in the moisture content (humidity) throughout the roll. Changes in tension are particularly apparent at splices between rolls. In addition, non-uniform shrinkage of the web during the ink drying process tends to cause variations in tension.

Misregistration problems tend to be particularly apparent where a plurality of dissimiliar webs, e.g., different types of paper, are fed to a folder to form composite signatures. The respective webs are often different types of paper, e.g., one coated and others uncoated, and at different tensions.

Cutoff control systems which adjust the effec¬ tive position of the cutting cylinders along the web path in accordance with the position of the signature on the web are available. An example of such a cutoff control system is described in U.S. Patent 4,736,446 issued to Reynolds, et al. on April 5, 1988.

Systems for controlling registry between respec¬ tive webs provided to a folder are also, in general, known. An example of such a system is described in U.S. Patent 3,556,510 issued January 19, 1971 to E.H. Treff. In that system, a slitter, disposed downstream of the chill roll, forms a plurality of ribbons from a web, and variable speed drive assemblies compensate for variations in the characteristics (e.g., modulus of elasticity) of the respective ribbons. The variable speed drives purport to vary the speed of the web rela¬ tive to the press speeds to maintain constant tension in the ribbon. The speed of the drives are controlled in accordance with optically sensed register control marks on the respective ribbons.

Web tension control systems in general, are also known. For example, U.S. Patent 3,561,654 issued on February 9, 1971 to Greisner describes a tension control system employing respective sets of rollers, one disposed upstream of the print units and the other

disposed downstream of the dryer (e.g., an infeed, and chill roll) cooperating to control the tension of the web. Both sets of rollers are driven from a main drive motor through a gearing mechanism with a finely variable transmission ratio. A floating roll disposed between the upsteam rollers and the printing unit generates a signal indicative of web tension, utilized to vary the ratio of transmission gearing.

The use of harmonic drives in tension control systems is also known. A harmonic drive typically includes a circular spline with internal spline teeth, a nonrigid cylindrical thin wall cup bearing external spline teeth (flexible spline) , disposed concentrically within the circular spline, and an elliptical ballbearing assembly (wave generator) disposed within the flexible spline. The flexible spline includes fewer spline teeth than the internal spline teeth of the circular spline and assumes the elliptical shape of the wave generator. The resultant spline pitch diameter at the major axis of the flexible spline matches that of the circular spline. The flexible spline teeth thus engage the circular spline teeth at respective points 180° apart. Each revolution of the wave generator produces a con¬ comitant revolution of the flex spline elliptical shape. This, in turn, causes a tooth to tooth rolling mesh at the two points of engagement between the flex spline and circular spline. The flex spline proper thus rotates in the opposite direction from the wave generator rota¬ tion by an amount equivalent to the difference in the number of teeth between the circular spline and flexible spline.

U.S. Patent 3,539,085 issued to Anderson, et. al, on November 10, 1970 describes an example of a system for controlling the web tension of a printing press employing a harmonic drive. In such system, a roller (e.g., a chill roll) which engages the web in cooperation with a nipping roller, is driven at selectively variable

speeds by a harmonic drive. The roller (e.g., chill roll) is fixed to the circular spline of the harmonic drive. The flexible spline is rotated at a constant predetermined speed, and in turn, drives the circular spline (and the roller) at a normally constant predeter¬ mined speed. The speed of the roller is varied by driving the wave generator at selectively variable speeds. The rotational speeds of the wave generator generates a signal which is fed back to a control panel from feedback tachometers connected to the roller drive. The control panel establishes a speed ratio to maintain a desired web tension on the web exiting from the roller.

Summary of the Invention The present inventors have found that by main¬ taining a constant tension on the webs after exiting the chill rollers, length variations can be controlled and lengthwise registration between respective webs maintained.

In accordance with one aspect of the present invention a printing system can be improved by sensing the tension of the web and controlling the angular position of the chill roll in accordance with tension, rather than merely controlling the angular velocity.

In accordance with another aspect of the present invention a system for controlling the angular position of a shaft may be improved employing a phase lock loop techniques. Such a shaft may be by way of nonlimiting example, a chill roll, or compensation shaft associated with the drive of the chill roll. More specifically, a signal indicative of the actual angular position of the shaft is generated by phase shifting a signal from a reference signal by an amount proportional to the angular position of the shaft. A signal of predetermined phase with respect to the reference signal is generated to indicate the desired position of the shaft. The deviation in phase of the phase shifted

signal from the desired position signal is then employed to develop a control signal to effect adjustment of the angular position of the shaft.

In accordance with still another aspect of the present invention, the signal indicative of the angular position of the shaft is generated by generating a magnetic field in accordance with the reference signal, and rotating a coil in the magnetic field in correspon¬ dence to the rotation of the shaft.

In accordance with still another aspect of the present invention, a control system is provided in which a single knob can be utilized to control a large number of parameters.

Brief Description of the Drawing A preferred exemplary embodiment of the present invention will hereinafter be described in conjunction with the appended drawing, wherein like numerals denote like elements and:

Figure 1 is a block schematic of a printing system employing a tension and draw control in accordance with the preferred embodiment of present invention;

Figure 1A is a schematic representation of the chill rolls and associated drive mechanism of Figure 1 (in reverse perspective from Figure 1) ;

Figure 2 is a block diagram of a suitable control circuit for the chill roll stand of Figure 1;

Figure 2A is a suitable microcontroller for the control circuit of Figure 2;

Figure 2B is a block schematic of a suitable communications interface for the control circuit of Figure 2;

Figure 2C is a suitable load cell interface for the control circuit of Figure 2;

Figure 2D is a suitable proximity detector interface for the control circuit of Figure 2;

Figure 2E is a suitable quadrature sine wave generator and amplifier for the control circuit of Figure 2;

Figure 2F is a suitable input filter and conditioner for the control circuit of Figure 2;

Figure 2G is a block schematic of a suitable programmable gate array implementation of a programmable divider and digital phase detector for the control circuit of Figure 2;

Figure 2H is a block schematic of a suitable loop filter, amplifier, summer and feed forward digital to analog converter for the control circuit of Figure 2;

Figure 3 is a block schematic diagram of a suitable operator interface for the system of Figure 1;

Figure 3A is a block schematic of suitable buffers and microprocessor for the operator interface of Figure 3;

Figure 3B is a block schematic of a suitable alpha numeric display for the operator interface of Figure 3;

Figure 3C is a suitable bar graph display for the operator interface of Figure 3;

Figure 4 is a suitable schematic flow chart illustration of a suitable menu hierarchy for the operator interface of Figure 3;

Detailed Description of a Preferred Exemplary Embodiment Referring now to Figure 1, in a printing press 10 employing tension and draw control in accordance with the present invention, a web 11 (typically paper) is dispensed from a take off (unwind) stand 12, and routed, in sequence, through an infeed unit 14 respective print units 16, 17 and 18, a dryer 22, a chill roll stand 24, and downstream processing apparatus such as a

folder 26. Print units 16-18 cooperate to imprint a multi-color ink image on web 11. Dryer 22 heats web 11 to evaporate various solvents in the ink and chill roll stand 24 cools the web to set the ink. Folder 26 slits, folds and cuts the web into predetermined lengths. If desired, other apparatus may be included in press 10, such as, for example, a silicone coater interposed between chill roll stand 24 and folder 26 and/or web guides disposed before infeed 14, before folder 26, or both.

As previously noted, print units 16-18 all operate at the same surface speed (also sometimes referred to as circumferential speed or linear web speed) typically expressed in feet per minute, inches per second, or meters per minute. By convention the surface speed of print units 16-18 is denominated "press speed' ^τ . The surface speed of takeoff stand 12, infeed 14, chill rolls 24 and folder 26 are typically expressed as a percentage deviation from press speed (referred to as draw) . The respective components of printing press 10 are driven from a common line shaft 28, cooperating with a conventional drive mechanism 30 (e.g., master and slave motors 30A and 3OB) . Print units 16-18 and folder 26 operate at fixed percentages (ratios) of the lineshaft rotational speed, typically driven from line¬ shaft 28 through conventional worm gears. Infeed 14 operates at variable speeds relative to press speed, driven from lineshaft 28 through variable speed gearing assemblies such as, for example, conventional harmonic drive 32. Chill roll stand 24 also operates at speeds variable relative to press speed, as will be explained.

Infeed 14 meters the web 11 dispensed from takeoff stand 12. The metering rolls of infeed 14 operate at a surface speed slightly greater than takeoff stand 12 but less than press speed. The differential in speed between infeed 14 and print unit 16 causes web 11 to enter print unit 16 under relatively high tension

(e.g., between 100 to 200 pounds for a 36" wide web). The tension of web 11 at successive print units 17 and 18 wilj. be slightly less than the tension at the preceding print unit; the moisture and ink deposited on web 11 by the preceding print unit tends to change the elasticity of the web, decreasing tension. The relative speed of infeed 14 is adjusted through drive ^ in a conventional manner to establish desired tensions through print units 16-18.

As previously noted, after imprinting, web 11 is routed through dryer 22 to evaporate the solvents in the ink. However, the heating process also removes moisture from web 11, and tends to cause web 11 to shrink in a non-uniform manner. Such shrinkage tends to cause misregistration in the operation of folder 26. Such misregistration is minimized by maintaining a constant tension of the web as it exits chill roll stand 24, controlling longitudinal shrinkage during the drying process. This is effected through chill roll stand 24 which cools the heated web to set the ink, and, in the preferred embodiment of the present invention, in one mode of operation serves the dual purpose of maintaining . a constant tension in the web to control longitudinal shrinkage of the web during the drying process.

Referring now to figures 1 and 1A (in reverse perspective from Figure 1) , chill roll stand 24 will be described. In practice, a press 10 may include plural tiers of apparatus, disposed one above the other, operating upon plural webs, as is well known in the art. Figure 1A depicts a chill roll stand 24 including two tiers of chill rolls for use in such a press. Each tier of chill roll stand 24 suitably comprises a plur¬ ality of hollow rollers 35-40, filled with a crolant liquid such as chilled water. A nipping roller 42 is selectively biased against chill roller 40 to ensure engagement between roller 40 and web 11. Web 11 passes over each of rolls 35-40 in turn, with the relative

disposition of rolls 35-40 (See Figure 1A) providing high surface contact with web 11 (on the order of 210 ^β - 220" wrap). As web 11 exits chill rolls 35-40, it is routed about respective idler rollers 44, mounted on load cells (e.g., Dover Flexo Electronics universal tension 45 transducers) for measuring the tension in web 11 exiting chill roll stand 24. Chill rolls 35-40 are driven by a controllable drive mechanism 34, cooperating with an operator interface 92, (not shown in Figure 1A) and a mechanism 94 for sensing press speed.

In the preferred embodiment, chill roll stand 74 operates in one of two modes. In a first mode of operation (constant draw) , the surface speed of chill rolls 35-40 are maintained at a constant percentage of press speed. In a second mode of operation (constant tension) the surface speed of the chill rolls relative to press speed is varied to maintain a constant preset tension on web.

Operator interface 92 provides signals indica¬ tive of the desired mode of operation and desired draw or tension values to drive mechanism 34, as well as a display of, for example, tension and draw speed to the operator. Operator interface 92 will be more fully described in conjunction with Figures 3-3C.

Press speed is suitably sensed by a conven¬ tional rotary encoder or proximity detector (PROX) 94 cooperating with, e.g., line shaft 28 or print units 16-18. In the preferred embodiment, a PROX 94 generates a pulse for every predetermined increment of angular travel by line shaft 28 (e.g., one pulse each 15 degrees; 24 pulses per revolution) .

Drive mechanism 34 suitably includes a conven¬ tional harmonic drive 58, a clutch 62, respective pullies 70, 74, and 78, a DC motor 80 (e.g., a 90 volt permanent magnet 3/4 Hp. motor), a resolver 84, and a control

circuit 90. A right angle box 64, and an air actuated brake mechanism 68 are also suitably included.

Harmonic drive 58 drives chill rolls 35-40 at a surface speed differing from (e.g., in excess of) press speed by a controlled percentage in accordance with drive signals provided by control circuit 90 to motor 80. Harmonic drive 58 includes a circular spline, flexible spline, and wave generator. The circular spline is suitably driven from line shaft 28 through a belt 56. The wave generator is driven by DC motor 80; the wave generator is connected to a compensation shaft 72 bearing pulley 74, which is coupled by a belt 76 to pulley 78 on the output shaft of motor 80. Chill rolls 35-40 are controllably driven from the flexible spline; the flexible spline is connected through an output shaft 60, clutch 62, and right angle box 64 to pullies 70 which are, in turn coupled by belts 71 to respective pullies (not shown) connected to chill rolls 35-40. Chill rolls 35-40 thus rotate at a rate proportional to, (depending upon pully and gear ratios) , the line shaft rate plus or minus the rotation rate, of compensa¬ tion shaft 72 and maintain an angular position correspon¬ ding to the angular position of compensation shaft 72. Resolver 84 generates a signal indicative of the instantaneous rotational (angular) position of compensation shaft 72 (and thus chill rolls 35-40) . Referring now to Figures 1 and 2, resolver 84 is suitably a brushless (e.g., transformer coupled) resolver, including respective stator coils 100 and 102 and a rotor coil 104. Rotor coil 104 is rotatably mounted in transformer relation (providing magnetic coupling) with stator coils 100 and 102. While a brush-type resolver can be utilized, it is particularly advantageous that resolver 84 be of the brushless type such as a Singer brushless resolver. Under many operating conditions, the rotor of resolver 84 rotates constantly. In such

circumstances, a brushless resolver is much less suscep¬ tible to wear than a brush-type resolver.

As will be explained, contrary to typical practice, the resolver stator coils are excited and an output taken from rotor coil 104. More specifically, stator coils 100 and 102 are excited with quadrature reference signals. The rotor of resolver 84 is coupled to compensation shaft 72 so that it rotates in coordina¬ tion with Shaft 72 through a suitable aligning coupling 82, so that rotor coil 104 rotates with shaft 72. Accordingly, an output signal is induced in rotor coil 104 having a phase difference from the reference signals which is proportional to the angular position of the rotor (and hence compensation shaft 72 and, ultimately chill rolls 35-40) . The induced signal thus manifests a difference in frequency (rate of change in phase) from the reference signal equal to the rate of change of angular position (angular velocity) of rotor coil 104 (compensation shaft 72) . The frequency of the induced signal is equal to the reference frequency, plus or minus (depending upon the direction of rotation of the rotor) one Hz for each rotation per second of rotor coil 104. Taking the reference signal to repre¬ sent press speed, the phase differential between induced and reference signals is indicative of the position of compensation shaft 72 (and ultimately chill rolls 35- 40) relative to the rotational position of print units 16-18, and the frequency differential indicative of relative surface speeds.

Control circuit 90 precisely controls the rate of change of angular position (angular speed) of motor 80, and thus the rotational position of chill rolls 35-40. Referring now to Figure 2, control circuit 90 suitably comprises: an input filter and signal condi¬ tioner 106; a digital phase detector 108, a 16 bit programmable divider (counter) 110; a suitable loop filter 112; an amplifier 114; a summer 116; a digital

analog converter (DAC) 120; suitable driver circuitry 122, a microcontroller 124; suitable quadrature sine wave generator and amplifier circuitry 125, suitable communications interface circuit (e.g., RS422 or RS232) 136 and suitable interface circuitry 138 and 140 to facilitate communication between microcontroller 124 and load cells 45 and press speed PROX 94, respectively.

As will be explained, microcontroller 124 receives control parameter data (e.g., draw rate, constant tension) from operator interface 92 (through communications interface 136) , tension data from load cells 45 (through load cell interface 138) , and press speed information from PROX 94 (through PROX interface 140) . Microcontroller 124 calculates actual press speed, and derives therefrom a feed forward value corresponding to the operational speed of motor 80 anticipated at that press speed and draw rate. In addition, micropro¬ cessor 124 also generates a draw command (from which a desired chill roll speed reference signal is derived by programmable divider 110) indicative of the desired chill roll speed. If in the constant draw mode, the draw command is generated in accorance with an operator input draw rate value (draw set point) . If in the constant tension mode, microcontroller 124 determines the deviation of the measured tension from an operator input desired tension value (tension set point) and determines the appropriate draw command to effect the necessary change in tension to eliminate the deviation. More specifically, the tension deviation is multiplied by a system reponse gain, and by the press speed, and the product integrated, to determine the desired draw rate. The system response gain value, typically in the range of 2 to 5, and preferably in the range of 3 to 4, controls the response time of the system to tension upsets.

Referring briefly to Figure 2A, microcontroller 124 suitably comprises an Intel 80188 microprocessor 124A, cooperating with a 16MHz system clock crystal, Yl, a suitable power up/reset circuit 12 B, address latch 124C, an electronically programmable read only memory (EPROM) 125A, a random access memory (RAM) 125B, and nonvolatile programmable memory (e.g., EEPROM) 125C. In addition, in the preferred embodiment, portions of microcontroller 124 are implemented in a Dual Universal Asynchronous Receiver Transmitter (DUART) chip 127 asso¬ ciated primarily with communications interface 136 (Figure 2B) and a programmable gate array 109 (Figure 2G) associated with phase detector 108 and divider 110.

Power up/reset circuit 124 is suitably in the form of a watchdog timer, comprising a conventional timer chip LM556. Upon powerup, or upon failure to receive an "Alive** signal (periodically generated during normal operation) , circuit 124B causes microprocessor 124A to reset, initiating an initialization sequence, as will be described. Microprocessor 124A communicates with various of the other components of the system through a multiplexed address and data bus AD, and address bus A. Communication between microprocessor 124A and various of the other components is effected through the intermediary of communications interface 136.

Referring briefly to Figure 2B, communications interface 136 comprises conventional DUART chip 127, such as a Signetics SCN2681, cooperating with a conven¬ tional output driver 127B, and a standard RS422 (and/or RS232) communications interface circuit 127C. Communi¬ cations interface 136 facilitates communications between microprocessor 124 and operator interface 92 (through RS422 interface 127C) , motor drive circuit 122 (enable and fault signals) , various other components of the

system, and peripheral devices. In the preferred embodi¬ ment various internal counters of DUART 127 are opera- tively employed as part of microcontroller 124.

Load cell interface 138 receives signals indicative of web tension from load cells 45, and develops tension data suitable for input to microcon¬ troller 124. Load cells 45, mounted on the ends of idler roller 42, generate analog signals indicative of web tension, typically on the order of 250 millivolts full scale. However, the response of load cells 45 tends to exhibit offsets that vary from unit to unit and application to application. (e.g., plus or minus 125 mV) . For example, the weight of roller 42 will cause an offset. In addition, the magnitude of the output signal from load cell 45 for a given load tends to vary from unit to unit. Accordingly, it is necessary that the system permit for calibration of load cell gain and offset to be tolerant of a wide range of load cell characteristics.

Referring to Figure 2D, a suitable load cell interface 138 comprises a differential amplifier 202, a variable gain/offset amplifier 204, and an analog to digital converter (ADC) 206. A 5 volt excitation signal is applied (J2) to load cells 45. Analog tension signals from load cells 45 connected* in a bridge configuration, are applied to differential amplifier 202 (J2, A, B) . Differential amplifier 202, removes the large common mode signal of the bridge connected load cells and provides a predetermined gain and modicum of filtering (e.g., a gain of 7.15 and a rolloff frequency of approxi¬ mately 10 Hz) . The tension signal is applied to variable gain/offset amplifier 204. Amplifier 204 provides for calibration of offsets and maximizes resolution by matching the range of analog signals to the range of ADC 206. The calibrated signal is applied to ADC 206, suitably an 8-bit analog to digital converter such as a

National ADC 0848, which generates a digital signal indicative of tension for communication to microprocessor 124A through bus AD.

The offset and gain of amplifier 204 are established by employing programmable electronic poten¬ tiometers 208A and 208B, such as XICOR EEPOT devices. Offset potentiometer 208A and gain potentiometer 208B provide ranges of resistances from zero to 10 killohms and and from zero to 50 killohms, respectively. The resistances of potentiometers 208A and 208B are pro¬ grammed through microcontroller 124. During the cali¬ bration process, initiated in response to actuation of a calibration switch SW1 physically located on the circuit board bearing control circuit 90 (Fig 2G) , potentiometers 208A and 208B are selectively enabled for change by respective signals (CSOFFSET/; CSGAIN/) commanded by microprocessor 124A (generated through DUART 127 of communications interface 136) . The resis¬ tance value is then varied by amount in accordance with increment signals (XINC/) , increased or decreased in accordance with an up/down control signal (XUD/) (pro¬ vided from gate array 109 upon command from micropro¬ cessor 124A) .

PROX interface 140 receives signals indicative of press speed from proximity detector 94 and converts those signals into an appropriate input (TMRINi) for microprocessor 124. Pulses from proximity detector 94 tend to be asymmetrical, noisy, and poorly defined. PROX interface 140 converts such signals into signals having sharp, definite transitions and precise levels. A suitable PROX interface 140 is shown in Figure 2D. In essence, pulses from PROX 94 are applied to suitable level shifting circuitry 140A, and AC coupled by capa¬ citor C2 to a suitable low pass filter 140. The filtered signal is applied to a suitable comparator 1 0C. Low pass filter 140B removes noise from the signal, and the input thereof is actively DC clamped to provide tolerance

to asymmetries in the PROX signals, and to provide a precise threshold level for comparator 140C. Comparator 140C, suitably comprising an LM556 chip, provides corres¬ ponding signals with clean transitions and well defined voltage levels. Active DC clamping is provided by feeding the output of comparator 140 through a transistor Ql to the input of filter 14OB. The PROX interface circuit of Figure 2D is particularly advantageous in that it is relatively immune to noise, ambiguities in DC input levels, and tolerant of input amplitude swings and asymmetries.

In accordance with one aspect of the present invention, the relative surface speed of chill rolls 35-40, and in particular, the angular position (phase) of chill rolls 35-40, relative to print units 16-18 is precisely controlled. Control of the angular position of the chill roll relative to the print unit cycle, rather than merely controlling relative angular velocity, provides a higher degree of accuracy; errors in velocity tend to cause cumulative position error. Further, speed control using tachometers provides poor regulation at low speeds, whereas phase control provides precise control at relative speeds approaching, and, indeed equal to, zero. In the preferred embodiment, this is accomplished employing feed forward and phase lock loop techniques.

Rough control of the speed of motor 80 is effected by calculating an anticipated operating speed for motor 80 from the press speed and desired draw rate, and applying a corresponding value as feed forward signal. As previously noted, microcontroller 124 calcu¬ lates the press speed from the PROX pulses provided through interface circuitry 140. The anticipated operating speed of motor 80 based upon the calculated press speed and desired draw rate is then determined and a corresponding digital signal applied to DAC 120. DAC 120 suitably comprises a National DAC1232 12-bit

DAC coupled to bus AD and a current output to voltage converting amplifier (Figure 2H) . DAC 120 converts the digital signal into a corresponding analog signal, which is applied as a feed forward signal to summer 116. Summer 116 algebraically combines the feed forward signal with a phase lock loop signal as will be explained. The resultant signal is applied to drive circuitry 122, and hence to motor 80.

Precision control of motor 80 is effected by phase lock loop techniques through cooperation of input filter 106, phase detector 108, loop filter 112, ampli¬ fier 114, and divider 110. In essence, the signal indicative of chill roll position (phase) generated by resolver 84, reflects chill roll position as deviation in phase from the reference signal applied to stator coils 100 and 102. The reference signal nominally represents the press cycle; the instantaneous phase of the reference signal represents the instantaneous posi¬ tion of the press units, derived from the 16 MHz micro¬ processor system clock by quadrature sine wave generator 125. A desired chill roll position (phase) signal is generated by divider 110, reflecting desired draw as a deviation in frequency (rate of change of phase) from the frequency of the reference signal. The deviation in phase of the resolver signal from the desired chill roll position signal, representing phase (position) error, is determined by phase detector 108. The phase error signal is suitably filtered and amplified by loop filter 112 and amplifier 114, and applied to summer 116 for application to DC drive 122 (in algebraic combination with the feed forward signal from DAC 120) . Drive circuitry 122 provides the control signals to motor 80, varying the rotation of the output shaft of motor 80 accordingly. Such variations are reflected in the rotation of stator coil 104, thus completing the phase lock loop.

As previously noted, stator coils 100 and 102 of resolver 84 are excited with respective quadrature signals at a predetermined frequency (e.g., 400 Hz). These quadrature signals are derived from the same source employed to generate the desired chill roll position (phase) signal output of square wave generator 110, and represent the instantaneous position (phase) of print units 15-18. More specifically, the 16 MHz signal is effectively divided in an internal timer (TIMER1) of microprocessor 124A, to provide a signal at a frequency (e.g., 400 Hz) nominally corresponding to press cycle. Quadrature sine wave generator and amplifier 125 converts the square wave press speed reference signal into corresponding quadrature sine waves for application to stator coils 101 and 102. Referring now to Figure 2E, quadrature sine wave generator and amplifier 125 suitably comprises a high pass filter 126, a clipped integrator 128, a second order low pass filter 130, a 90-degree phase shifter 132, and a two-channel power amplifier 134. High pass filter 126 removes DC components from the signal, which is then converted by clipped integrator 128 into a clipped triangle wave having a slope at zero crossing matching the slope of a sine wave of the prede¬ termined frequency (e.g., 400 Hz). Low pass filter 130 then removes the higher order harmonics from the clipped triangle wave to provide a corresponding sine wave. at the predetermined frequency. The sine wave is then applied to phase shifter 132, which generates quadrature signals for application to dual amplifier 134. The respective sine waves are applied to stator coils 100 and 102 of resolver 84.

As previously described, rotor coil 104 of resolver 84 rotates with compensation shaft 72 of harmonic drive 58. Accordingly, a sinusoidal signal is induced in coil 104, having a phase difference from the reference signal proportional to the angular position of shaft 72. The induced signal thus manifests a

frequency (rate of change in phase) equal to the frequency of the reference signal (400 Hz) , plus or minus one Hz for each rotation per second of shaft 72. Variations will add or subtract depending upon the direction of rotation. The induced sinusoidal signal is applied to input filter and conditioner 106.

Input filter and conditioner 106 converts the sinusoidal output of rotor coil 104 into a square wave of corresponding frequency and phase and of predetermined magnitude. A suitable input filter and signal condi¬ tioner 106 is depicted in Figure 2F. The induced sinusoidal signal is applied through an amplifier and band pass filter 106A (relatively wide band width, e.g. 200 Hz, centered at the reference frequency, e.g. 400 Hz) , to a phase lock loop (PLL) chip 106B including an internal voltage controlled oscillator, (e.g.. Motorola MC14046B) . PLL chip 106B generates a square wave (FIN) of appropriate magnitude at the frequency of, and phase locked to, the induced sinusoidal signal. The use of PLL chip 106B provides particularly advantageous noise immunity. The square wave signal (FIN) indicative of rotor coil output is applied to phase detector 108, and to microcontroller 124.

The square wave signal derived from the signal induced in rotor coil 104 is applied to digital phase detector 108 for comparison to a signal indicative of a desired compensation shaft 72 operation position (phase) . The desired position signal is provided by programable divider (counter) 110, in cooperation with microprocessor 124A. Programmable divider 110 is receptive of the 16 MHz system clock signal from microcontroller 124 and provides an output square wave divided down from the 16 MHz signal. The resultant output signal from program¬ mable divider 110 will be a square wave having a frequency equal to the reference frequency (F; e.g., 400Hz) plus or minus an offset frequency ( F) indicative of draw (desired rate of change of phase deviation) .

For example, for a zero draw (i.e.g, operation at press speed), the 16 MHz clock would be divided by 40,000 to generate a 400 Hz signal. The instantaneous phase of the square wave corresponds to the instantaneous desired chill roll position (phase) . The use of a relatively high clock frequency such as 16 MHz provides for rela¬ tively high resolution.

The press cycle square wave reference signal and desired position (phase) signal from programmable divider 110 are derived from a common source. Thus, the reference signal employed in resolver 84 (the quad¬ rature sine waves) and the desired position (phase) signal employed as an input to the phase detector 108 maintain a predetermined relation in frequency and phase. The phase error detected by phase detector 108 represents the phase deviation of compensation shaft 72 from the desired position indicated by the signal generated by programable divider 110.

Referring to Figure 2G, in the preferred embodiment, programmable divider 110 and digital phase detector 108 are implemented employing a programmable gate array 109, such as, for example, a XILINX XC2064 logic cell array. Gate array 109 is also employed to implement various I/O port functions, i.e., reading switches and generating XUD/ and XINC/ signals for control of load cell interface 138. The logic functions effected by, and internal connections of, programmable gate array 109 are determined by data stored in internal static memory cells. Data configuring programmable gate array 109 as a programmable 16-bit divider and phase comparator is developed in accordance with conven¬ tional techniques (using commercially available develop¬ ment systems, e.g., the XILINX XACT development system) and maintained in a table resident in EPROM 125A of microcontroller 124. The configuration data is loaded from EPROM 125A into the static memory cells during initialization of the system upon peer up. Gate array

109 receives a 16MHz signal from microcontroller 124, and generates phase error signal (PHO (the least signi¬ ficant bit) , PHI (most significant bit) ) indicative of the deviation. Programmable divider 110 and phase detector 108, may, of course, be implemented in other manners, such as, for example, employing discrete commer¬ cially available programmable counter and phase detector chips, or portions of commercially available programmable phase lock loop chips, or custom designed circuitry.

Referring now to Figures 2 and 2H, the phase difference signal (PHO, PHI) is employed in a control loop for motor 80. Phase error signal PHO, PHI is applied to loop filter 112. Loop filter 112 effectively removes the 400 Hz component from the phase detector output signal, providing an analog voltage proportional to the difference in phase between the desired position reference square wave and the square wave indicative of actual position (FIN) derived from the output of resolver rotor coil 104. Loop filter 112 is suitably a (SINX)/X low pass filter. The filtered phase difference signal is applied to amplifier 114. Amplifier 114 provides appropriate gain to the phase error signal. The phase error output of amplifier 114 is applied, together with the feedforward signal from DAC 120 to the summing junction of summer 116. The resultant combination (MVC) is applied to DC drive 122 which generates a commensurate drive signal for motor 80. Drive 122 is suitably a commercially available unit, such as Warner Electric, SECO DC drive. The motion of motor 80, and thus the signal generated by rotor coil 104 vary in accordance with the drive signals, thus closing the phase lock loop.

As previously noted, operator interface 92 provides communication between an operator and control circuit 90 of chill roll stand 24. It provides a mechanism for operator input of signals indicative of the desired mode of operation (constant draw or constant

tension) , desired draw or tension set point values, and various other parameters, and for display of indicia of measured tension and draw speed values and various other parameters.

In order to achieve operator acceptance in a pressroom environment, it is necessary that the control mechanism be relatively simple, preferably replicating various controls already familiar to the operators. Control of conventional chill rolls has been relatively simplistic; a calibrated potentiometer is provided to adjust draw rate, and respective switches are provided for selectively enabling a clutch and nip rolls. How¬ ever, it is also desirable that a number of parameters be adjustable from the control panel, and that display of various parameters be accessible. Operator interface 92 provides a control that is adjustable using one hand, leaving the other hand free to hold a signature. The single knob control replicates the conventional chill control in many respects, and thus should tend to engender a high degree of operator acceptance. At the same time, a high level of control is provided through use of a hierarchical menu system implemented in micro¬ controller 124 employing linking techniques, as will be explained. Referring now to Figure 3, operator interface 92 suitably comprises a rotary encoder switch 300, suitable buffers 302, a microprocessor 304, a suitable communications link 306 (e.g., an RS432 interface), an alphanumeric display 308, and a two-scale bar graph display 310.

Encoder 300 generates respective signals indicative of clockwise rotation, counter-clockwise rotation, and depression of a control knob. Encoder 300 is suitably a Grayhill Series 61 optically coupled rotary encoder switch. If desired a stiffer spring can be employed in commercially available encoders to provide more definitive tactile characteristics.

Microprocessor 304 arbitrates the operation of operator interface 92, communicating with control circuit 90 through RS432 interface 306. Referring to Figures 3 and 3A, microprocessor 304, suitably an Intel 8751 microprocessor, serves a relatively limited func¬ tion; the bulk of processing is performed in microcon¬ troller 124 of control circuit 90. Specifically, micro¬ processor 304 monitors the operation of encoder 300, and determines whether the encoder knob has been rotated clockwise or counter-clockwise or depressed, and initi¬ ates transmission of an ASCII U, D, or B, respectively, through RS432 interface 306 to control circuit 90. Microprocessor 304 receives ASCII characters from control circuit 90, and generates the appropriate signals to effect the display on alphanumeric display 308 or bar graph 310. If desired, a suitable watchdog timer 312 can be employed to reset microprocessor 304 in the event of inactivity extending beyond a predetermined period.

The displays of operator interface 92 are simple in design; the operator is not confronted with a formidable array of displays. In the normal running mode, a draw rate and measured tension are displayed on alphanumeric display 308, together with an arrowhead pointed at the draw or tension figure in accordance with the operational mode. At the same time, as will be explained, various other parameters can be set or displayed, through a hierarchical menu structure acessed by turning and depressing the knob of encoder 300.

Referring now to Figures 3 and 3B, alphanumeric display 308 suitably comprises three Hewlett Packard HDSP2112 eight-character, intelligent display chips cooperating in a conventional manner with microprocessor 304. As will be explained, the data for display is provided by microcontroller 124.

Bar graph 310 provides continuous indicia of the tension of web 11, suitably on one of two ranges (e.g., 0-100 pounds; 0-200 pounds). As will be explained.

the data for the bar graph is provided from microcon¬ troller 124. Tension is sampled at a relatively high rate (e.g., 100 Hz). Tension data is selectively filtered (digitally in software) to facilitate display at various response rates, and scaled to provide the desired range. More specifically, the measured tension data is, in effect, applied to an infinite impulse response (IIR) filter adjusted to a bandwidth of 0.5 Hz or 2 Hz in accordance with the desired display response (implemented in software) . Referring to Figure 3C, bar graph display 310 suitably comprises a commercially available 101 element bar graph display chip 320 (e.g., Hewlett Packard HDSP 8825) , cooperating with a commer¬ cially available cathode driver 322 (e.g., Sprague UCN 5821 8-bit Serial-In Latched Driver) and anode driver 324 (e.g., Sprague UCN 5810A 10-bit Serial/Parallel Source Driver. In addition, respective sets of LEDs 326 and 328 are provided to selectively back light alternative upper and lower range markings on the bar graph display. Range marker LED sets 326 and 328 are connected to respective outputs (UPSCA, LOSCA) of a cathode driver 330 which is in turn driven by micropro¬ cessor 304. Driver 330 also selectively provides actu¬ ation signals to respective LEDs 332 and 334 to provide visual indicia of nip roller and clutch engagement.

In accordance with one aspect of the present invention, a hierarchical menu structure is made acces¬ sible through a single knob. A suitable menu heirarchy is illustrated in Figure 4. By way of convention, right angled arrows in Figure 4 indicate a return to the run time display upon depression of knob of encoder 300 and straight arrows indicate a return to the previous display upon depression of the encoder knob.

During normal operation, alphanumeric display 308 provides a run time display, generally indicated as 402, consisting of a draw value 404, a tension value 406 and a mode indicator 408 (pointing to draw or tension

value in accordance with mode) . The menu hierarchy is entered by actuating encoder 300 (depressing or turning) . Upon an initial actuation, a main menu, generally indi¬ cated as 410 is displayed. Main menu 410 suitably includes "status", "operation", and "set up" menu items. Encoder 300 is rotated to poll the menu items, and depressed to select a given item. The selection of the "status" menu item provides for selective display of various parameters such as the speed of motor 80 (comp motor) , or press speed, or indicia of the operating condition of various system elements such as brake 68, or clutch 62 or the operational mode of nip roller 42.

The selection of the "operation" menu item permits adjustment of the operating parameters of the system, such as operational mode, draw rate set points, tension set points, nip control (selectively engaging nip roller 42 at certain press speeds) , and engaging and disengaging clutch 62.

Selection of the "set up" menu provides for setting the nip speed (press speed at which nip roller 42 is actuated) or selecting the bar graph response rate.

For example, to set a constant tension set point, main menu 410 would be accessed through actuation of encoder 300. The knob of encoder 300 would be rotated to poll the operation menu, and depressed to effect the selection. A second level menu 412 would then be dis¬ played, suitably including "draw," "tension", "nips" and "clutch" menu items. Encoder 300 would be rotated to poll the desired item, i.e., tension, and depressed to select tension mode operation. An "adjust tension" prompt would then appear on display 308. Encoder 300 would be rotated to adjust the tension value, and depressed when the desired value is displayed to effect the selection.

Similar procedures would be followed to select draw mode operation, and to set the draw rate set point, or to control clutch 62. If desired, however, a fail safe can be provided, inhibiting change in clutch status if the press is moving.

As will hereinafter be described, the hierar¬ chical menu structure is effected as a linked list in EPROM 125A. Each menu item has a descriptor associated therewith, including in sequential memory locations; an identifying header; links to the respective locations (e.g., menu descriptor) to be accessed in response to each of the three possible valid characters B, U and D; a pointer to the menu entry text, and a pointer to the starting address of any special instructions. Microcon¬ troller 124 maintains a pointer to the presently selected menu item. When a character from encoder 300 is communi¬ cated to microcontroller 124, the pointer is offset by an amount corresponding to the relative location of the link associated with that character, and the contents of the pointer replaced with the address of the menu description designated by the link, the text of the menu associated with the designated description is displayed and any special processing routines associated with the menu, such as incrementing the draw or tension set points, effected.

It will be understood that while various of the connectors/conductors/communication links are depicted as single lines, they are not so shown in a limiting sense and may comprise plural conductor/connec¬ tors as is understood in the art. Further, the above description is of a preferred exemplary embodiment of the present invention, and the invention is not limited to the specific forms shown.

For example, one aspect of the present inven¬ tion provides a particularly advantageous mechanism for detecting and/or controlling the angular position of any rotating shaft (including, but not limited to a

chill roll): a stationary coil (e.g., a resolver stator) is excited with a reference signal of predetermined frequency; a second coil (e.g., a resolver rotor coil), disposed in the vicinity of the first coil, is rotated in coordination with the shaft to induce a signal in the rotor coil; and the phase difference between the reference signal and induced signal is determined. The phase difference is indicative of the angular position of the shaft, and a control signal to the drive mechanism for the shaft can be derived.

Likewise, another aspect of the present inven¬ tion provides a particularly advantageous mechanism for controlling the circumferential speed of any rotating shaft. Controlling the angular position of the shaft through a phase lock loop as described above, and controlling the circumferential speed in accordance with the deviation in angular position, as opposed to merely employing a feedback loop based upon speed (e.g., using a tachometer) provides for circumferential speed control without accumulative error. A further aspect of the present invention provides a particularly advan¬ tageous mechanism for generating the signal indicative of the angular position of the shaft: generating a magnetic field in accordance with the reference signal, and rotating a coil in the magnetic field in correspon¬ dence to the rotation of the shaft.

Similarly, another aspect of the present inven¬ tion provides a particularly advantageous operator inter¬ face of general applicability through which a single knob can be utilized to control a large number of para¬ meters.

Further modifications may be made in the design and arrangement of the elements without departing from the scope of the invention as expressed in the appended claims. *