Ink Density Control System

Abstract

Control of ink supply in a printing press in accordance with sensing of the density of ink being printed on imprint-receiving material, wherein sensing measurements are made and smoothed by considering previous printing press cycles and wherein the smoothed measurement is compared to a desired predetermined standard density and the ink feed is adjusted accordingly. Erratic density measurements are automatically identified and disregarded. Interaction between adjacent ink adjustment keys of the ink feed mechanism is automatically taken into account, and lift-off of keys from an ink fountain blade of the mechanism is prevented. Proportional, derivative, and integral control signals are produced and combined to provide a composite control signal for the ink feed. The invention may be implemented by analog or digital embodiments.

Description

CROSS REFERENCES TO RELATED APPLICATIONS

Two co-pending U.S. applications which involve ink density control systems and that are related to the present application are: application Ser. No. 73,319, of Jean R. Gaillochet, filed Sept. 18, 1970, and entitled "An Automatic Device for the Remote Adjustment of the Inking Blade of a Printing Machine" and a continuation application Ser. No. 182,538 now U.S. Pat. No. 3747524 of James N. Crum, filed Sept. 21, 1971, and entitled "Ink Fountain Key Control System."

BACKGROUND OF THE INVENTION

1. Field of the Invention

The invention relates primarily to a system for controlling the printing occurring in a printing press, by constantly monitoring or sensing the resultant print placed on print-receiving material, comparing it with a desired predetermined standard, and varying the feed of ink to said material in accordance with deviations from said standard.

2. Description of the Prior Art

Attempts to accomplish this result have become increasingly apparent in recent years. In an approach as shown and described in U.S. Pat. No. 3,353,484, issued to Koyak on Nov. 21, 1968, the thickness of ink on selected lateral portions of a printing press inker is monitored, and ink fountain keys aligned with the monitored section of the inker are adjusted correspondingly. Both the monitoring means and the adjusting means move laterally across the press inker and ink fountain and perform their monitoring and control function on a periodic basis.

Another form of apparatus for performing the general objectives of this invention is disclosed in U.S. Pat. No. 3,567,923, issued to Hutchison on Mar. 2, 1971. In this latter approach, the color density of ink on a printed web is sensed in a plurality of locations across the web and the speed of an ink fountain roll is increased or decreased according to an average of the several density measurements. Slight variations in density within a certain "dead band" of a reference signal are ignored and do not effect a control function. At selected periods controlled by a timer, sample measurements are made, and if the control signal is outside the predetermined "dead band," a control function occurs to increase or decrease the flow of ink across the entire width of the inkers.

U.S. Pat. No. 2,969,016 for Colour Printing issued to J.F. Crosfield et al. on Jan. 24, 1961, describes apparatus for measuring ink density by scanning printed patches, and suggests that corrections to the ink flow could be made automatically.

SUMMARY OF THE INVENTION

An inker control suitable for closed-loop or open-loop operation is provided with a plurality of ink density sensors located to monitor ink laid on print-receiving material across its width. Keys of an ink fountain are similarly located in line with the sensors and are individually or group responsive to the sensors to maintain ink feed from the fountain at a rate required to maintain the print density at the level of a predetermined standard with which signals from the sensors are compared.

Since various printing jobs are run at different press speeds, and since each printing job is normally run at a slower speed during a "make-ready" period as compared to a production speed, periodic control signals are related to press cycles or revolutions instead of to time. In addition, in order to establish each control signal, a plurality of density measurements are made, and these measurements are combined to provide an error signal which is indicative of many samples taken during different press cycles. Especially in lithographic printing, where the combined effects of ink and water may provide a large error in a single sample, this is important to avoid overcompensating. This system provides a "smoothing" action in the response of the ink feed to the cyclical measurement, and is relatively non-responsive to such occurrences as an ink-water imbalance due to a press trip-off for only a few press cycles.

Another aspect of the invention relates to control of neighboring or adjacent ink keys in a fashion which would prevent a tendency to destroy the effectiveness of the system resulting from certain of the keys moving out of contact relative to the ink fountain blade with which they cooperate.

DESCRIPTION OF THE DRAWING

Other aspects and features of the invention will become more apparent upon consideration of the following description taken in conjunction with the accompanying drawings wherein:

FIG. 1 shows one printing unit of a printing press having an ink fountain and an ink density control system;

FIG. 2 is a mechanical schematic view of an ink fountain showin individual control elements such as keys for controlling the lateral distribution of ink;

FIG. 3 shows a cylinder of the printing unit, several printed density test patches, and desitometer sensing heads for inspecting the test patches;

FIG. 4 shows an edge view of a cylinder of the printing unit and the placement of a densitometer head for reading optical densities from printed material;

FIG. 5 is a block diagram of an electronic control circuit of an analog embodiment of the present invention;

FIG. 6 is a graph of optical density of test patches as a function of the number of impression imprinted on material passing through the press, a first curve being shown for a prior art system and another for the present invention;

FIG. 7 is a graph of ink key position for a typical ink key as a function of the number of impressions imprinted on material passing through the press before and after a sudden change in optical density of printing is detected;

FIG. 8 is a block diagram of a portion of an electronic control circuit for an alternate form of analog embodiment of the present invention;

FIG. 8A shows an analog embodiment switching circuit for associating each densitometer channel with a printing unit in a multiple-unit press;

FIG. 8B is a table of possible pairings of densitometer channels with printing units in a multiple-unit press;

FIG. 9 is a block diagram of portions of the electronic control circuit for a digital computer embodiment of the invention showing especially the optical density sensing portions of the equipment;

FIG. 10 is a simplified block diagram of internal components of a digital computer employed in the digital embodiment of the invention;

FIg. 11 is a block diagram showing some details of the digital computer which relate to interfacing of the computer with external circuits;

FIG. 12 is a flow chart of "interrupt" controls of the digital computer;

FIG. 13 is a flow chart showing the utilization of press on-pressure signals and time-delay signals by the computer when obtaining density data from the density sensors;

FIG. 14 is a flow chart showing processing of sensed densitometer data in the digital computer embodiment of the invention;

FIG. 15 is a graph illustrating the operation of a density reading validity test subroutine, by which erratic density readings are identified and rejected in both the analog and digital embodiments of the invention;

FIG. 16 is a block diagram showing portions of a digital embodiment which are related to actuators for operating ink keys of the ink fountains of the printing unit; and

FIG. 17 is a flow chart showing steps in a process for preventing lift-off of keys from the fountain blade in presses in which a unitary fountain blade is employed.

ANALOG COMPUTER EMBODIMENT

Press Inking

Referring now to the drawings, which are only for illustrating the preferred embodiments and not for limiting the invention, FIGS. 1, 2, 3 and 4 illustrate the invention in conjunction with a conventional sheet-fed lithographic printing press. The press includes a plate cylinder 10, a blanket cylinder 12, an impression cylinder 14, and a transfer or delivery cylinder 16. The plate cylinder is inked by a conventional inker 18 comprising an ink fountain 20, an adjustable ducting mechanism 22 including a duct roll 24, and a plurality of ink transfer and vibrating rolls 26 and 28 located between the ducting mechanism 22 and the plate cylinder 10.

The ink fountain 20 includes a fountain roll 30 which rotates in the ink fountain to form an ink film on the roll 30. The duct roll 24 is reciprocated between a position in engagement with the fountain roll and a position in engagement with one of the vibrating rolls 28. While the duct roll 24 is in engagement with the fountain roll 30, the latter is rotated an angular amount determined by the setting of an adjustable mask 32 of a pawl and ratchet drive 34 for the fountain roll. The extent of rotation of the fountain roll 30 while in engagement with the duct roll determines, for a given film thickness on the fountain roll, the amount of ink transferred from the fountain roll to the duct roll and, in turn, the amount of ink transferred to the plate cylinder.

The ink fountain 20 includes in addition to fountain roll 30, a fountain blade 36 which extends for substantially the length of the fountain roll. The blade is flexible and is urged into engagement toward fountain roll 30 by means of a plurality of ink keys in the form of screws, such as key 38, shown in FIG. 1, by reversible motors 40, 42, 44 and 46 (FIG. 2) so as to control the flow of ink at various sections across the length of fountain roll 30. Although the preferred embodiment has 46 such ink keys and motors at each ink fountain, only four are shown in the drawing, FIG. 2, for simplicity. For a more complete description of inker 18 and ink fountain 20, reference is made to U.S. Pat. No. 3,185,088 to R.K. Norton, and assigned to the same assignee as the present invention.

Each ink fountain key 38 can be moved to different positions by means of an actuator motor such as motor 40 in order to admit more or less ink to the portion of the ink fountain which it controls. Changes in the position of the ik fountain keys 38 are not apparent at a printed sheet 50 immediately; in inker train delay occurs. Ink emitted to the ink fountain duct roll 24 must be transported around several inker rolls 26, 28 in succession to the printing plate 10, then to the blanket 12, and thence onto the printed sheet 50. Consequently, a delay of a number of impressions occurs before a change in ink fountain key setting can have any effect on the printed sheets 50.

Even at the end of the inker train delay time, the thickness of ink deposited on an ink test patch printed on each sheet 50 does not rise immediately to a steady state value corresponding to the new ink key settings when, for example, the ink flow is increased. Some of the recently admitted ink is wiped back along the inker roll train, so there is a further delay in the increase of printed image density following a step change in the setting of an ink key.

When no ink whatsoever is on the printed sheets 50, a small increase in amount of ink thereon makes a big difference in the printed density. On the other hand, when the density of ink being printed is already high, a similar small increase in ink film thickness makes very little difference. That is, the optical density of a test patch as a function of ink film thickness is, therefore, very nonlinear. The density nonlinearity becomes part of the transfer function of the principal loop of the ink control servomechanism.

Control System, General

In accordance with the present invention, a densitometer head 41 is adjustably positioned on a support bar 48 as shown in FIGS. 1 and 4 so as to monitor sheet material 50 carried by the impression cylinder 14. On multi-unit sheet-fed presses, the densitometer head 41 is preferably located at an impression cylinder 14 of the last color printing unit as illustrated herein. On web-fed presses the densitometer head 41 is preferably located after the dryer and the chill rolls, where the ink is dry.

As will be described in detail hereinafter, the densitometer head 41 includes a light source for transmitting light to sheet 50 to impinge simultaneously on at least one printed test patch surface area thereof and on an adjacent reference surface area, together with a pair of sensors for receiving light reflected from the two surface areas and providing output signals indicative of the amount of reflected light received. These signals are applied to a gated densitometer circuit 51 which determines the optical reflection density of the ink on the test patch surface area and provides output signals for application to a control computer 53 for controlling motors 40, 42, 44 and 46 to operate the keys 38 to control the positioning of the fountain blade 36 in dependence upon the measured ink density. Also, the gated densitometer circuit 51 and control computer 53 may provide signals to a suitable visual display M to indicate to the pressman the density of ink reproduction. The densitometer head 41 and circuit 51 are described in detail in the co-pending U.S. application of John M. Manring, Ser. No. 79,952, now U.S. Pat. No. 3,756,725, filed Oct. 12, 1970, and entitled "Measurement and Control of Ink Density."

The operation of the densitometer head 41, of gated densitometer circuit 51, and of control computer 53 are synchronized with the movement of the sheet member 50, as with a cam 52 provided with a lobe 54 for camming against a movable switch member 56 to contact electrically a stationary contact 58 so that an electrical signal, such as that taken from a DC voltage supply source B+, may be applied to gated densitometer circuit 51 and control computer 53.

Reference is now made to FIGS. 3 and 4 which are schematic illustrations of the impression cylinder 14 carrying the sheet 50 past the densitometer head 41. As shown in FIG. 3, a transversely arranged colored ink test patch 60 is provided on the trailing edge of the sheet 50. Immediately adjacent the test patch 60 is the associated reference surface area 68. This surface are is an uninked area on sheet 50, although it may be printed in advance to provide a reference level of ink density if desired. The test and reference areas are each of small size, such as three-eighths inch by one-half inch. The test patch 60 printed on the paper is ordinarily solid printing but may be a half-tone.

After being printed, the paper 50 bearing the test patch 60 travels through the press to the densitometer head 41. There the optical reflection density of the test patch 60 is measured and compared with the reference surface area 68 while the paper 50 is in motion. The gated densitometer circuit 51 is gated on for a short enough time so as to inspect only when the test patch is present at its field of view. A lamp 61, associated with and inside the densitometer head 41, is flashed to illuminate the test patch 60 and the reference area 68 at the time of measuring the density. The optical reflection density of the test patch 60 is ascertained by comparing light which is reflected from the test patch 60 with light reflected from the unprinted reference area 68, which is illuminated by the same flash of light from lamp 61. The ratio of the two reflected lights is used by the gated densitometer circuit 51 to determine the optical reflection density of the printed patch 60. The gated densitometer circuit 51 produces an analog output voltage 69 which is proportional to the logarithm to base 10 of that ratio, the lagarithm being called, in this art, the optical reflection density.

Included in the gated densitometer circuit 51, near its output, is a holding circuit which holds the results of each density reading until the next succeeding density reading is made. Thus, the gated densitometer circuit always provides an output signal 69 indicative of the most recently completed density reading.

In the analog circuit embodiment now being described, a separate complete system is shown for each individual longitudinal stream of test patches such as test patch 60, to simplify the description. The separate systems can communicate with each other, as will be illustrated below. If desired, the equipment can easily be arranged to share circuits among several streams of test patches such as patches 62, 64 and 66 (FIG. 3), and to provide for a plurality of colors as will be shown in the description of the digital embodiment below.

Filter and Reference Comparator

The density signal voltages 69 produced by the gated densitometer circuit 51 are conducted to the control computer 53, a block diagram of which is shown in FIG. 5. In the analog embodiment, control computer 53 is an analog circuit. At the input of control computer 53 is a filter 71 for smoothing of the measured density data. The filter 71 is a one-pole low-pass type having a cutoff frequency above which the density signals are greatly attenuated and below which the signals are not attenuated appreciably. The low-pass filter 71 prevents occasional erratic density readings 69 from excessively influencing the settings of the ink keys 38. Erratic readings may be caused by electrical noise and other factors. The filter 71 averages the density readings 69 appearing at its input, and produces at its output a voltage 73 which is influenced by previous readings as well as the most recent density readings 69. If the press were always to be run at a constant speed so that successive density readings were produced at a constant rate, the cutoff frequency of the data smoothing filter 71 could be constant. In the usual situation, however, in which variations in press speed occur, a different number of printing impressions and, therefore, a different number of density readings would occur within the time constant of the filter if the time constant were not adjustable. A low-pass filter with a controllable variable cutoff frequency is, therefore, employed, having a cutoff frequency that is proportional to the press speed, and therefore proportional to the number of density readings per unit time that are being produced. A tachometer 76 on the press provides a control signal to relays 78 which connect and disconnect shunt capacitors in the filter to adjust the cutoff frequency, as is known in the prior art.

An output voltage 73 from the low-pass filter 71 represents the filtered actual density; it is connected to one input of a comparator 75. There it is compared with a density reference signal 77' from a DC reference source 77 connected to a second input of the comparator 75, whose signal value is manually adjustable. The density reference signal 77' minus the output voltage 73 from the data smoothing filter 71, constitutes an error signal 79 which is sent out by the comparator 75. Error signal 79 represents a discrepancy between the desired density and the actual density. It serves as an input to the remainder of the system to control the settings of the ink keys 38 so as to correct the density and thereby to reduce the error signal 79 to a negligible amount.

The error signal 79 is amplified by an amplifier 81 whose gain is not constant, but instead depends upon the magnitude of the error signal itself. For high magnitudes of input error signal 79, irrespective of their sign, the gain of the amplifier 81 is less than its gain for lower values of the error signal. Consequently, a final output voltage from the amplifier 81 is a nonlinear function of its input voltage 79. This characteristic reduces and controls the amount of ink key overshoot that would otherwise result from the delays in the press. The sign of the output signal of amplifier 81 is responsive to the sign of signal 79. The nonlinear gain characteristic of the amplifier 81 can be manually adjusted by increasing the attenuation setting of an attenuator 80 which precedes the amplifier and at the same time decreasing the attenuation setting of another attenuator 82 which follows the amplifier, or vice versa. The proper attenuator settings for a job depend upon ink opacity, viscosity, and other factors.

Integrating and Differentiating Circuits of the Controller

An error signal voltage 83, which is present at the output of the attenuator 82, is effectively transmitted to three essentially parallel circuit channels in each of which that signal is treated differently. In one channel 92, the error signal 83 simply passes through, essentially unmodified, to a summing junction 94. In another of the channels 96 the signal is integrated with respect to press impressions, and connected to the same summing junction 94 as the essentially direct signal 92. The third parallel channel 100 has a differentiating circuit for anticipating future requirements for ink flow. An output voltage 102 of the differentiating circuit 100 is connected to the same summing junction 94 as are the other two channels 92 and 96.

Further details of the three parallel circuit channels are as follows:

Error signal 83 is connected to a first sample-and-hold module 104 for storing temporarily the most recent reading of the error signal. The first sample-and-hold module 104 accepts the most recent value of the error signal 83 which is presented at its input and holds that value available at its output until such time as a new value of signal 83 is made available and is accepted. Acceptance by module 104 occurs upon issuance of a pulse on a circuit 86 from a synchronizing circuit 84, as is conventional with sample-and-hold modules. The output of the first sample-and-hold module 104 connects to the summing junction 94 and serves as a proportional signal or direct essentially unmodified error signal components into that summing junction.

The output signal 83 of the potentiometer 82 is also connected to a sampling switch 85. Switch 85 closes and reopens once for each new reading 69 of density made for the inker keys being controlled. The functions of switch 85 can be performed by either a static or a mechanical switching device under indirect control of the synchronizer 52, which paces the synchronizing circuit 84.

The length of time during which switch 85 remains closed is always the same irrespective of its frequency of closing because of the manner of operation of circuit 84. While contacts 85 are closed, the voltage 83 derived from the output of the nonlinear amplifier 81 is applied to an input of an integrating amplifier 98. Small errors in density will therefore accumulate and cause a correction in ink key position to be made after a time.

A circuit 87 applies the output signal from the first sample-and-hold module 104 to a second sample-and-hold module 106 which stores the error signal reading of the immediately preceding press impression. At the time of occurrence of a pulse on a strobe circuit 88, which is shortly before the pulse on circuit 86, the second sample-and-hold module 106 accpets into its hold circuit the voltage that is standing on its input terminal at that time. This is the same voltage that was standing at the output terminal of the first sample-and-hold module 104 immediately prior to the most recent pulse mentioned above on circuit 86. Upon the pulse on circuit 88, the second sample-and-hold module 106 produces at its output terminal 108 a voltage equal to the previous press impression density error signal. Thus, upon each occurrence of pulse pair 88, 86, each sample-and-hold module 106, 104, respectively, produces at its output terminal a new value of voltage, the value appearing at the output of the second sample-and hold module 106 being the same value as that which was appearing previously at the output of the first sample-and-hold module 104. In this way, two voltage readings are made available at any time. One reading represents the most recently produced error signal 83; the other represents the error signal from the immediately preceding reading.

The output voltages of the two sample-and-hold modules 104, 106 are applied to a subtracting circuit 110, with such polarity that the previous reading's error signal 83 is subtracted from the most recent error signal 83 to produce a further signal representing the change which occurred between the most recent signal and the signal immediately preceding it. This change, which is of the nature of a derivative, is applied to an amplifier 112. The output voltage 102 of amplifier 112 therefore represents a rate of change of error signal 83 with respect to press impressions. Output signal 102 is applied to the summing junction 94 with the same polarity as were the proportional signal 92 and the integrated signal 96 described above.

Without the differentiating channel 100, the density of the printed test patch 60 would recover to its desired steady state value rather slowly following a sudden change in density caused by an external disturbance. This is shown in FIG. 6, as curve A, for a sudden decrease of density caused by something other than ink key settings. The number N.sub.A of printing pmpressions that must be made before the optical density has substantially returned to its initial and correct value is considerably reduced by temporarily opening the inker keys extra far, in anticipation of the setting delay N.sub.A, when the sudden change of density occurs. The exaggerated opening of the ink key settings then causes a compensatory excess of ink to flow at the beginning of the correction period, which reduces the required number of settling impressions from N.sub.A to N.sub.B, as shown in curve B of FIG. 6.

FIG. 7 shows a typical graph of ink key position versus impression count, which is carried out by the fountain keys 38 in order to compensate to the extent possible for the settling phenomenon. After an appropriate amount of extra ink flow has occurred through the unusually enlarged ink gap opening, the gap opening is reduced by circuit 100 to its steady state value so that the density actually printed on the train of test patches 60 will not overshoot its desired final value. As a result, the ink density patches 60 rises rapidly to a value close to its final value, and then tapers into its final value asymptotically at a somewhat earlier number of impressions N.sub.B than it would have without the compensatory ink flow. This compensatory ink key behavoir is accomplished by the differentiating channel 100 of the three-channel signal-processing circuit in the manner just described. To summarize, differentiating circuit 104, 106, 110, 112 produces a signal proportional to the first derivative of the error signal present at its input; the total signal that drives the ink keys therefore has a component which forces a rapid correction of density variations.

Actuator Drive Circuit

An output signal 114 from the summing junction 94 connects to an input 116a of a driver amplifier 116 for producing a signal 117 driving a duty cycle modulator 118. The duty cycle modulator 118 converts the signal 117 to a series to pulses or bursts of AC wave 120, to produce step changes in key positions. One group of the keys 38 are driven by key actuators 40, which are connected to modulator 118 in the simplified analog embodiment being described. Keys are selectively associated with particular densitometer heads 41 by patch connections 121 at the inputs to the actuators 40. (The modulator 118 can instead be connected to control the inker pawl and ratchet 34 is desired, or several densitometer heads can be multiplexed in an analog embodiment to control several groups of keys 38 independently, as described in the digital embodiment hereinbelow.)

A sample of the output signal 120 of the duty cycle modulator 118 is also rectified and smoothed by a diode and filter circuit 124 and applied to the input of an integrating amplifier 126, which serves as an accumulator. This accumulator 126 is resettable to zero by means of a momentary-acting relay 89 which is controlled by the synchronizing circuit 84. The accumulator 126 is reset to zero when a pair of contacts 89' or relay 89 close briefly immediately before the time of the strobe pulse on circuit 88. Contacts 89' close only long enough to short-circuit a capacitor 128 connected from input to output of the accumulator amplifier 126 so as to reset the accumulator amplifier's output signal 130 to zero; thereafter, the accumulator amplifier 126 accumulates a voltage at its output 130 which corresponds to whatever changes in position of the actuator occur during the current press impression interval, which is an error-correction interval. The output signal 130 from the accumulator 126 is connected as a feedback signal to a second input 116b of the driver amplifier 116 so as to subtract from the principal input signal 114 of the driver amplifier. During an error-correction interval, after the actuators 40 have been driven far enough for the signal 120 to build up a voltage 130 at the output of the accumulator 126 and at input 116b which is equal in magnitude but opposing the voltage 114 at the input 116a of the driver amplifier 116, the output signal 117 of the driver amplifier 116 becomes zero. The amount of ink flow correction that was required during the subject correction interval has then been accomplished and no further correction will be made until the next correction interval. Ordinarily, the full correction that is called for by the signal 114 will be completed before the next density error reading is obtained.

Alternative Actuator Drive Circuit

FIG. 8 shows an alternative embodiment of portions of the control computer 53 related to the actuators 140. This is an alternative to the portion of the aforedescribed circuits which follow the summing junction 94. The alternative actuator circuit of FIG. 8 differs from the first actuator circuit (FIG. 5) in that the alternative circuit has feedback from an actuator-driven potentiometer 136, while the actuator circuit uses, instead, an accumulator amplifier 126. Moreover, in the alternative circuit, duty cycle modulation is not employed. The alternative circuit of FIG. 8 has an inker ratchet-control feature also.

A signal 114 for the alternative actuator circuit is obtained from the summing junction 94, and is connected to a principal input 132a of an amplifier 132.

Another input 132b of amplifier 132 receives a positive feedback signal 134 indicative of the present position of a key actuator 122. Voltage 134 is derived from the potentiometer 136 whose movable arm is driven by actuator 122. The signals 114 and 134 are added in amplifier 132. Their sum represents a desired new position of the actuator, because it represents an actual present key position as indicated by signal 134 plus a desired change as indicated by signal 114. An output signal 138 of amplifier 132 connects to a third sample-and-hold module 140. A command pulse occurs on synchronizing circuit 89' shortly after occurrence of the pulse, mentioned above, on circuit 86. Thereupon, the third sample-and-hold module 140 accepts and provides at its output terminal a voltage 142 representative of a desired new position of the actuator 122, and holds it essentially throughout of the subject correction interval. The voltage 142 is maintained constant by the sample-and-hold module 140 even though a correction is being carried out by the actuator 122 during the present error-correction time interval.

The output signal 142 of the third sample-and-hold module 140 connects to a combining junction 144 where it is combined with the voltage 134 which represents the instantaneous position of the actuator 122. The voltage 134 is subtracted in the combining junction 144 from the desired position of the actuator 122, which is represented by the output signal 142, to produce a signal 146.

The signal 146 is connected through a relay contact D to an amplifier 148 and is an error signal, which at all times corresponds to the amount of correction remaining to be made by the reversible actuator 122 during the current error-correction interval. The actuator motor 122 receives a voltage output from amplifier 148 which drives the actuator 122 to change the position of corresponding ink keys 38, and also to move the transfer arm of the potentiometer 136, which affects the voltage 134 at the transfer arm. After key 38 and the transfer arm of the potentiometer 136 have been completely driven to a desired new position, the voltage 134 equals the voltage 142 held by the third sample-and-hold module 140; the error signal 146 is zero, and the actuator 122 does not operate any further during that correction interval.

Where one densitometer and control circuit must actuate a plurality of keys 38, each key has an individual respective actuator, all actuators of the same group are driven in common by amplifier 148, and only one of the keys 38 is selected to have its potentiometer 136 provide the signal 134 for its group of keys.

Ratchet Pullback, Analog

A ratchet pull-back circuit is provided to sense when any ink key 38 (or key group) has approached too closely to either limit of its possible range of adjustment. When any key has been adjusted to such a close position, the pawl and ratchet drive 34 for the ink fountain 20 is automatically re-adjusted. Ratchet re-adjustment changes the amount of ink provided, without a change in ink key positions, and therefore changes the density of all of the test patches. The equipment of FIG. 8 thereupon automatically responds by actuating the keys to a more central position where no ink key is near a limit. The ratchet change is accomplished by sensing the position of the arm of the feedback potentiometer 136 and comparing the position signals 134 produced by potentiometer 136 with key-limit reference voltages.

A high-low comparator 150 has a sensing input terminal 151 connected to the actuator position-indicating potentiometer 136. If the actuator position voltage 134 becomes too great or too small by comparison with high and low DC reference voltages 152, 153, which are put into the high-low comparator from a circuit 162a, the high-low comparator 150 produces an output signal which operates a relay 154. The high and low reference voltages 152, 153 are predetermined percentages of whatever voltage 162a exists over-all on the potentiometer 136. Contacts 154a of the relay 154 are shown in a de-energized position D of the relay, in which the output 146 of summing junction 144 is connected to the input of amplifier 148. This is the normal position of the relay 154 and is its position when the key 38 being controlled is not out of range in either direction. When the relay 154 is actuated by the high-low comparator 150 as a result of the key's traveling too far in either direction, the output signal 146 from the combining junction 144 is connected by means of a relay contact E to a summing resistor 155. The summing resistor 155 and other summing resistors of the same type from other key groups on the same printing press (color) unit are connected to an input of a summing amplifier 156. The summing amplifier 156 is used in common by all of the key groups for one color unit. An output of the summing amplifier 156 drives a bi-directional ratchet motor 158, which in turn moves the ratchet assembly 34, of which there is only one for each color unit. A ratchet position potentiometer 160 has its transfer arm controlled by the ratchet assembly 34 so as to produce a position signal 162. The ratchet position signal 162 is connected to one extreme terminal of every actuator position potentiometer 136. As a result, the output signal 162 of the ratchet position potentiometer 160 serves as a multiplying factor upon the position of the transfer arm of every potentiometer 136 and therefore the ratchet position signal 162 is one factor of the voltage signal 134 produced at the transfer arm of each actuator potentiometer 136. Each group of keys is represented by a potentiometer 136 and a signal 134.

The operation of the ratchet pullback circuit is as follows. When no key 38 is near a limit, the high-low comparator 150 outputs a zero signal, and the relay 154 is de-energized. The control loops behave routinely as described above. If, however, one key group's representative key 38 approaches too closely to a limit of its range of travel, the high-low comparator 150 puts out a signal to the relay 154, which energizes the relay, placing its contacts 154a in the position E. The amplifier 148 and the actuator 122 for the subject group of keys 39 thereafter receive a zero input signal and the actuator 122 does not move for the remainder of the correction time interval which is currently in progress. Instead, the error signal 146 from the combining junction 144 is connected through the relay contact E to the summing register 155 and hence to the summing amplifier 156. The amplifier 156 and the motor 158 operate the ratchet 34 to a new position to provide the remaining correction signal required through the circuit consisting of the potentiometer 160, its output signal 162, the potentiometer 136 and its output signal 134 for the subject group. The ratchet 34 operates during the current correction interval until such time as the feedback signal 134 is equal in magnitude to the signal 142 from the sample-and-hold module 140. At that time, the combining junction 144 puts out a zero signal 146 and the ratchet motor 158 stops.

While the ratchet 34 is being operated to its new position, the ratchet position signal 162 is changing; that signal 162 is applied not only as a reference for the comparator 150, and to the actuator position potentiometer 136 for the group of keys which has encountered a limit, but also to the corresponding actuator position potentiometer for other groups of keys through bus 157. Consequently, each of the other groups of keys experiences a change in its reference signal 134 within the same current correction interval. The other groups of keys have not caused their comparator relays (corresponding to relay 154) to operate, so the error signal 146 produced by the combining junction 144 of each of the unlimiting key groups passes through the normal position D of respective relay contacts 154a to amplifier 148 in each such group. Amplifier 148 in each such unlimiting group operates its respective actuator or actuators 122 until the feedback potentiometer 136 representing each group has changed to such a new position as to cause its error signal 146 to be zero. Actuators 122 for the unlimiting groups then have zero signals, and stop moving. This circuit permits unlimiting key groups to correct their actuator 122 positions in response to a change of ratchet 34 position without relying upon the principal feedback loop through the inker and the printed paper and the densitometers to perform the correction. The necessary changer are therefore made before the density readings are substantially affected, and are made independently of the deviations of the principal error signal 114. The term ratchet is used herein to represent any ink feed rate-control technique other than keys, which simultaneously affects an entire fountain, such as the ratchet itself, speed of the fountain roll, or duct roll dwell time.

In a similar manner, other changes in the printing process which are capable, in the absence of a change in the setting of the ink keys 38 of later affecting the density of ink deposited on the paper, can provide compensatory signals to change the key settings to prevent changes in ink density, without waiting for a density error to occur. Another example of such a change is a change in the water feed rate.

SEQUENCE OF OPERATION

A time sequence of operation of the analog system of FIG. 5 is as follows.

Usually the ink keys 38 have been set by the press operator to provide approximately correct ink feed for the job layout which is to be produced, when a printing unit first goes on impression. The inker rolls 26, 28 have ordinarily been pre-inked before a printing unit goes on impression. To illustrate the system's operation, a situation will be described in which the press goes on impression with some of the ink keys 38 initially too far closed and therefore with insufficient ink on the corresponding lateral portion of the inker 18. When the press goes on impression, auxiliary contacts 164 (FIG. 5) of an impression on-off solenoid, which is part of the press' electrical controls, start a delay device 166, which counts to a predetermined number of impressions and then puts out a signal 168 to enable the densitometer 51 and the modulators 118. This activates the controller.

In the example presently being described, insufficient ink is deposited on the paper at first, so the area of a test patch 60 has a semi-blank appearance not much different from the appearance of the neighboring blank reference area of the paper. After passing the blanket cylinder 12, the lightly-printed test patch travels a distance to a place where it passes under the densitometer head 41. The densitometer head 41 and the gated densitometer circuit 51 inspect the paper at the first test patch area 60 and find that not enough ink has been printed on it. The gated densitometer circuit therefore outputs a low voltage signal 69 corresponding to a low density reading.

The low signal voltage 69 from the gated densitometer 51 is conducted to the control computer 53, where it is filtered by the filter 71, which is storing zero voltage initially. The filter 71 averages the new low reading 69 with the previous zero initial condition, and outputs a low voltage 73 to the comparator 75. If desired, an additional impression-count or time-delay relay may be employed to permit a number of density readings to be accumulated before any ink keys are moved to new positions.

This low smoothed density reading is compared with the reference voltage 77 which has previously been adjusted to correspond to some non-zero desired value of density. Of course, a great error signal 79 results, which is applied to the non-linear amplifier 81.

The non-linear characteristic of the non-linear amplifier 81 has very little effect upon the servo operation unless there is a large error signal. For example, when there is almost no ink on the paper, a very large error occurs; because of the nonlinearity, signal 83 from amplifier 81 is not large in the same proportion. On a basis of error size alone, the nonlinear amplifier therefore acts as a signal compression circuit for large error signals to prevent over-shoot of the density correction.

A large signal 83 from the output of the nonlinear amplifier 81 is stored in the first sample-and-hold module 104. Because insufficient ink has been printed in this example, the proportional circuit channel 92 provides a large component of error signal to the summing junction 94. Also, a great rate-of-change-of-error signal 102 is created by the derivative circuit 100 and applied to the summing junction 94 because the second sample-and-hold module 106 stores zero error signal at the start. The integrating channel 96 provides only a moderate signal component. The three channel signals 92, 96, 102 are summed at the junction 94 and applied to the driver amplifier 116 of FIG. 5, whose other input signal 130 is zero because the accumulator 126 was recently reset to zero by relay contacts 89. The driver amplifier 116 puts out a large error signal to the duty cycle modulator 118, which starts to drive the ink keys 38 open rapidly by means of the actuators 40. Ink flows to the inker rolls 24, 26, 28.

More low density readings are made by the densitometer while the increased ink flow is being transported through the series of inker rolls 26, 28 to the paper 50, and the ink keys are driven open relatively far. When the increased ink flow reaches the paper 50 the optical density of the test patch 60 increases. After a time, the optical density becomes great enough that the signal 69 from the gated densitometer 51 is of such magnitude as to make the signal 79 become zero at the non-linear amplifier 81. Shortly thereafter, the key actuators 40 cease to receive any significant correction signal 120 from the duty cycle modulator 118. The control system is in equilibrium and is automatically controlling the optical density of the printed test patch 60 by controlling the ink keys 38.

Additional identical control systems are provided for other lateral portions of the fountain roll; the additional systems include densitometer heads 43, 45, and 47 of FIG. 3, more gated densitometer circuits for processing the signals, and actuator motors 42, 44, 46 (FIG. 2).

Open Loop Operation

The ink keys 38 can be controlled in open-loop fashion by a press operator instead of by the closed-loop method described above. In open-loop operation, which is simpler, the operator may manually adjust DC signals and apply them through a manual-or-automatic selection switch 170 from a point 172 in the circuit of FIG. 8, in place of the automatic signals 142. The same manual input provisions serve for pre-adjustment of keys before starting. A digital embodiment to be described below can also be operated either open-loop or closed-loop, with a press operator observing a display of density readings and making corresponding adjustments in the open-loop mode of operation by holding a switch depressed.

Associating Densitometer Channels with Printing Units & Displays

In a printing press having a plurality of printing units, each unit ordinarily prints a different color of ink. It is sometimes desirable to change the assignments of colors among the plurality of printing units so that, for example, yellow images may be printed by unit no. 1 on one job and by unit 2 on a different job.

It is convenient to associated a particular sensor channel in densitometer head 41 and a particular gated densitometer circuit of densitometer 51 always with the same color, regardless of the particular printing unit by which that color is printed. For example, the sensor channel A and the gated densitometer circuit channel A may always be associated with the color yellow. This is convenient because the location of the densitometer head 41 is usually fixed with respect to the press frame, the color filter used for each color is installed in a particular position of the densitometer head 41, and the test patch of that color is always printed in the same lateral position on the paper irrespective of which printing unit is employed to print that particular color. Also, a calibration adjustment peculiar to each color is made in the gated densitometer channels. Consequently, to avoid having to relocate the color filters and to recalibrate the densitometer channels, the electrical output of each measurement channel, consisting of a sensor and a channel of the gated densitometer circuit, is most conveniently associated always with the same printed color. When changes are made in the printing unit upon which the colors are to be printed, therefore, the outputs of the various gated densitometer channels, which remain with the same color, must be switched so that they control the different printing unit.

It is more convenient for the operator if each printing unit display be associated always with a particular printing unit rather than with a particular color. Therefore, when colors are interchanged among the printing units, each display, M, remains with the same printing unit rather than follow any particular color. This situation requires that the various gated densitometer channels be switchable at their outputs so as to operate different display units that are permanently associated with the printing units.

FIG. 8A shows a switching circuit for associating colors with printing units and displays. Three channels A, B and C are shown, each more or less permanently associated with a respective color A, B or C to be printed. Three printing units and displays M are shown, No. 1, No. 2 and No. 3. A six-position switch 174 is arranged so as to connect the outputs of the gated densitometer channels A, B and C in six different permutations to the three printing and display units No. 1, No. 2 and No. 3, FIG. 8B. In switch position 2, for example, the output of the dated densitometer circuit A is connected to control the printing unit and display No. 1, the output of gated densitometer circuit C is connected to control the printing unit and display No. 2, and the output of gated densitometer circuit B is connected to control the printing unit and display No. 3. Switch 174 has additional poles, omitted to simplify the drawing.

DIGITAL COMPUTER EMBODIMENT

A second embodiment of the invention utilizes, as part of the control computer 53 of FIG. 1, a digital computer instead of an analog computer. FIG. 1 applied to both the analog and digital embodiments. In the latter, the output voltage 69 of the gated densitometer circuit 51 passes through an analog-to-digital converter (which is an input portion of control computer 53), before being presented to the digital computer itself, which is also included in control computer 53.

FIG. 3 shows an arrangement for controlling only one printing unit. Where several colors are printed, as in the present embodiment, additional test patches, not shown, similar to patch 60, and additional reference areas similar to area 68 are provided. Additional sensors with color filters are incorporated in densitometer heads for measuring light reflected from the additional color patches, which should not be confused with patches 62, 70, etc., for other parts of the roll width. Only one lamp is provided in each densitometer head for serving all colors in common, but every printed surface area and reference area to be measured requires an individual sensor to receive reflected light from the area.

Control equipment for the digital embodiment is shown in FIG. 9 for a three-color press having eleven densitometer heads arranged laterally across the width of the press. The purpose of the densitometer heads, the gated densitometer circuits, and the analog-to-digital conversion equipment is to measure the optical reflection density of each test patch and to present the results to a digital computer 208 in the form of digital data.

Densitometer Multiplexing

Some components of the densitometer equipment are used in common for several measurement channels. They are time-shared by means of multiplexing equipment.

As shown in FIG. 9, eleven densitometer heads 180 to 190 are provided. Each densitometer head includes one flash lamp 180L to 190L.

In a three-color press, each flash lamp is positioned so as to illuminate three test patches of different colors and three unprinted reference areas near the test patches. (Instead, one unprinted reference area could serve three colors, if desired.) Each densitometer head 180 to 190 receives reflected light from the three test patches and from the three unprinted reference surface areas under it. In this way, each densitometer head obtains data regarding three colors. Four densitometer heads 41, 43, 45, 57, each having one pair of light-sensitive detectors for measuring the density of one color, were described above in connection with the analog embodiment of the invention; the digital embodiment is similar except that there are eleven densitometer heads 180 to 190 distributed across the width of the press and each densitometer head has three pairs of light-sensitive detectors to accommodate the three colors being printed. A synchronizing device 52' connects a lamp trigger signal to a trigger signal input 202L of a lamp multiplexer 191. Another set of input terminals for the lamp multiplexer 191 is connected to receive data on lines 194 from a ring counter 196 for selecting one lamp at a time. The multiplexer 191 has eleven outputs, each of which connects to and operates one of the eleven flash lamp units 180L to 190L.

For each printed color A, B, C, the eleven sensors which receive light reflected from test patches are all connected to a multiplexer 192A, 192B, 192C, respectively. Also connected to the multiplexers 192A, B, C, are digital data lines 194 from the ring counter 196, which is further described below, for selecting one of the eleven sensors. Only one output signal 193A, B, C is connected from each test signal multiplexer 192A, B, C, respectively, to each gated densitometer circuit 198A, B, C.

For each color A, B, C, the eleven sensors which receive light reflected from the unprinted reference areas are connected to another multiplexer 200A, B, C, respectively. Only one selected output signal 195A, B, C, is connected at a time from each multiplexer 200A, B, C, to a second input of the gated densitometer circuits 198A, B, C, respectively. Also connected to the gated densitometer circuits is a pulse signal 202 derived from the synchronizer 52' The synchronizing device 52' is of the type 52 described above in connection with the analog embodiment. The output pulses 202 on a bus from the synchronizer 52' are conducted to gate inputs 202A, B, C, of each gated densitometer circuit 198A, B, C, respectively, for gating purposes.

An output of each gated densitometer circuit 198A, B, C, is an analog voltage signal representing the density of whichever one of the eleven heads 180 to 190 was most recently sampled. Each analog density signal is connected to an analog-to-digital converter 204A, B, C, whose digital output lines are connected through switching gates 206A, B, C, respectively, to the digital computer 208.

Thus, identical sets of density measuring equipment are provided for each of the color units of the printing press, except that the flash lamps, synchronizer, ring counter, and digital computer and some miscellaneous components are used in common by all of the color units.

Forty-six ink keys 38 collectively span the width of the printing press, but only eleven densitometer heads 180 to 190 are provided to span the same width. Consequently, some of the densitometer heads 180 to 190 must serve to control more than one fountain key 38. One approach to distributing the fountain keys among the densitometer heads is to have nine of the densitometer heads each control four fountain keys and to have the other two densitometer heads each control five of the keys. It is more satisfactory, however, to group the keys 38 in accordance with the form density of the images to be printed, form density being the dependence of required ink flow upon the form of the images currently being printed by the press at various positions across the width of the press. At portions of the press where the form density changes rapidly (laterally across the width), each densitometer head 180 to 190 may control relatively fewer of the fountain keys 38.

Sequence of Operation of Density Measuring Equipment

The density measuring equipment operates as follows. Let the ring counter 196 be assumed to have a count of 1 standing at its output terminals, before a synchronizing pulse 202 occurs. The ring counter's output data 194, which is this count of 1, is connected to an input of the lamp multiplexer 191; it causes the trigger input 202 to the lamp multiplexer 191 to be connected internally in multiplexer 191 to lamp 180L and not to any of the other ten outputs of lamp multiplexer 191. Lamp 180L does not yet flash, however. When the press reaches a particular phase position, a cam 54' similar to cam 54, of the synchronizer 52' actuates a movable arm 56' similar to arm 56 and causes a voltage to be applied to the synchronizing line 202 which is connected to the trigger input of the lamp multiplexer 191 and to the gate terminals 202A, B, C, of the gated densitometer circuits 198A, B, C, respectively. The synchronizing device 52' produces a pulse 202 once for each impression time interval of the press. The leading edge of the synchronizing pulse 202 is timed to cause a selected lamp, in this example lamp 180L, to flash when the corresponding printed test patches are in a proper position for measurement, under the density head 180. The leading edge of the synchronizing pulse 202 triggers a flash unit of the lamp 180L, causing light to fall on three colored test patches and on three unprinted reference areas adjacent to them. None of the other printed test patches or reference areas are illuminated by their flash lamps during this particular measurement interval, that is, at the time of this printing impression.

Light is reflected from the three colored test patches into three sensors of densitometer head 180. Each test multiplexer 192A, B, C, connects only this one input (from head 180) of its eleven inputs through to its single data output 193A during the present measurement interval, the choice of input being under control of the ring counter 196 through data lines 194 connected from the ring counter's output to the test multiplexers 192A, B, C. While the ring counter's count is 1, each of the three test multiplexers 192A, B, C, connects a density signal received from density head 180 to its output and therefore to the test signal input 193A, B, C, of the gated densitometer circuits 198A, B, C, respectively, for the color involved.

Light is also reflected at the same time from the three unprinted reference areas into three reference sensors of density head 180. The three reference multiplexers 200A, B, C, connect the three reference signals to the respective three reference data inputs 195A, B, C, of the three gated densitometer circuits 198A, B, C. The multiplexers 200A, B, C, are under the control, (via lines 194) of the ring counter 196, which has selected the signals from density head 180 in the present measurement interval. Each of the three gated densitometer circuits 198A, B, C, receives a gating signal 202A, B, C, from the synchronizing device 52' which is the leading edge of the same synchronizing pulse that triggered the flash lamp 180L. The gated densitometer circuits 198A, B, C, thereupon accept both the test data and reference data into their circuits.

Each gated densitometer circuit 198A, B, C, produces an analog output signal which represents the optical reflection density of the test patch of its respective color. The three optical density signals, one for each color, are connected to the inputs of analog-to-digital converters (A/Ds) 204A, B, C, which convert them into binary data and apply them to output terminals of the three A/Ds.

The trailing edge of the aforementioned synchronizing pulse 202 sets a device flag flip-flop 210A, B, C, corresponding to each of the three A/D converters 204A, B, C, respectively, the pulse 202 having persisted long enough for the A/D converters to settle to steady output values. This digital information from the A/Ds is connectable to the computer 208 through the switching gates 206A, B, C. The data stand at the output of each of the gated densitometer circuits and of the A/Ds throughout one full impression time interval, because each gated densitometer 198A, B, C, includes a sample-and-hold module near its output which holds the analog density signal until another density reading has been obtained to replace it.

The computer 208 then reads the data sequentially from all three A/Ds 204A, B, C, within a single impression time interval. It does so by successively enabling only one at a time of the three sets of switching gates 206A, B, C, as will be described in more detail hereinbelow under the heading "Input Interfacing." The computer 208 also reads the status of the ring counter 196 on lines 207, after which the computer 208 outputs a pulse 212 to the ring counter 196 to increment it by one step. The ring counter 196 thereafter contains a 2.

A second measurement interval then begins, corresponding to the next impression interval of the press. The lamp multiplexer 191 internally reconnects its input trigger terminal 202L so as to be able subsequently to trigger lamp 181L. Each of the three test multiplexers 192A, B, C, is reconnected ss that it can put out a signal to be received from a second input of its eleven inputs. The reference multiplexers 200A, B, C, do the same. Density head 181 is now being employed. Upon a second occurrence of a synchronizing pulse 202 from the synchronizer 52', the operation described above is repeated, with three new values of density signals (one for each color) being accepted into the computer from the three A/Ds 204A, B, C.

Each of the eleven densitometer heads 180 to 190 is utilized in turn during eleven successive impressions or density reading intervals.

The ring counter 196 recycles to a count of 1 following a count of 11 so that it counts repeatedly from 1 to 11. For each impression of the printing press, the density measuring equipment measures as many optical reflection densities as there are color units, and puts the resulting digital data into the computer 208.

The full complement of eleven densitometer heads 180 to 190 need not always be used. The ring counter 196, which selects the heads, has a skip facility for selectively skipping heads, under control of the digital computer 208.

As is shown on FIG. 9, there is a separate analog-to-digital converter 204A, B, C, for each of the three gated densitometer circuits 198A, B, C. Each has a 12-bit output. The most significant bit position from each analog-to-digital converter contains the sign of the density. The other eleven bits of information state the density's magnitude. Negative density signs are made available for the printing of fluorescent inks, the light from whose test patches can be brighter than the light reflected from an unprinted reference test area.

Digital Computer, General

FIG. 10 is a simplified block diagram of the digital computer 208 including its console 214 and input and output devices, which include the A/D converters 204A, B, C. The computer 208 itself consists of an arithmetic unit 216, a control unit 218, and a memory unit 220.

The arithmetic unit 216 comprises a 12-stage accumulator 222 and a one-stage link 224 for storing a carry digit from the accumulator 222 during certain operations. Various electronic circuits, not shown, are arranged around the accumulator 222 for controlling it to perform elementary computations such as adding and shifting.

The control unit 218 comprises a program counter 225, which is a register containing a particular address of the memory unit 220 which it is desired to access next. In routine operation, the program counter 225 may be incremented one count at a time to retrieve successive instructions and data from the memory unit 220. The control unit 218 also has an instruction register 226 which usually contains a code identifying the instruction currently being executed. A major state generator 228 is provided to put the computer into an appropriate state for each computer timing cycle; typical states are fetch, execute, etc. Several computer timing cycles may be required to execute one instruction.

The principal component of the memory unit 220 is the core memory 230 itself, which stores 12-bit words at several thousand addresses. A memory address register 232 is also provided, to contain the address of the core memory 230 where data are currently being written or retrieved. A memory buffer register 234 is also included, for buffering data input and data output from the core memory 230.

The computer is programmed in accordance with flow charts such as those shown in FIGS. 12, 13, 14, and 17; programming from such flow charts is well within the state of the art.

A program is placed in the core memory 230 before starting the printing press. To execute the program, the program counter 225 starts with the first step of the program by addressing the program through the memory address register 232. The first 12 -bit word is then withdrawn from the core memory 230 into the memory buffer register 234 or the instruction register 226. If the word is an instruction, it is executed by the computer. The program counter 225 is then incremented one count to produce a new memory address, and the next instruction or data word from the program is withdrawn from the core memory 230. The entire program is executed by the computer in this sequential manner. The instructions themselves can cause the computer to jump to other addresses, more than one count away from the present address, when desired.

Input Interfacing at Computer, General

The computer 208 accepts density data at its input terminals from the three analog-to-digital converters 204A, B, C, via the data gates 206A, B, C, respectively, and from other devices. The status of the ring counter 196 must also be transmitted to the computer so that the computer can determine which of the density heads 180 to 190 is currently providing data. Other information which must go into the computer includes position feedback data for the ink key actuators, on-pressure signals from each of the printing units, positions of open-loop or closed-loop selector switches, status of the switch which selects a make-ready mode or a run mode of operation, etc.

A transfer of the data into the computer 208 is accomplished by passing the data through the accumulator 222 of the computer; this requires the computer to stop any background program which may have previously been in progress. This is called a programmed data transfer. Transfer of density data into the computer will be described in detail, and will serve as an example of the method of transfer of other data items also.

Components inside the computer 208 that are involved in transferring density data into the computer are shown in FIG. 11. Twelve input terminals 238 of the computer 208, which are provided for accepting the input data, are connected in parallel to all three of the gates 206A, B, C. Input data lines 240 from the data gates 206A, B, C, conduct signals through the terminals 238 to the accumulator register 222 of the computer. The same lines 240 are connected in parallel also to other external devices which may at some time have to provide information by this method. The input bus 240 is time-shared by the external devices; only one external data-producing device applies data to the information bus 240 at any one time.

Sequence of Inputting Density Data to Computer

A time sequence of operation for transmitting density data into the computer 208 starts with the production of digital data by the analog-to-digital converters 204A, B, C, FIG. 9. After these A/D converters have had time to make the data available, device flag flip-flops 210A, B, or C, corresponding respectively to the A/D converters, are set by the trailing edge of the synchronizing pulse 202. A device flag signal 236A, B, or C is transmitted through an OR gate 237 to a program-interrupt facility 254 of the computer 208 over a program-interrupt request line 256, FIG. 11. When the computer 208 receives the interrupt flag signal, it finishes the instruction which it was in the process of executing, then transfers data relating to the background program which was previously in progress, if any, into storage 230, as shown in a flow chart of FIG. 12. By a step 246, the computer 208 transfers into storage the number in the program counter 225 to save the address where the background program was interrupted. The contents of the accumulator 222 and the 224 are also saved in storage 230 in a step 247. The computer next takes a jump by putting a particular address in the program counter 225, which directs the computer to call out from the core memory 230 an "interrupt" sub-routine starting with a step 249. The interrupt sub-routine is carried out by having the program counter 225 call out one instruction at a time from the memory 230 starting at that particular address. A 12-bit instruction word for an input-output transfer instruction corresponding to analog-to-digital converter 204A, wbich is for color A, is called out first from the core memory 230 into the instruction register 226. The instruction register 226 examines the word and recognizes the first three bits as being a code calling for an input or an output data transfer (FIG. 11). Bits numbered 3 to 8 of the words, which include a device select code 242, are supplied to some output terminals 244 of the computer for device interrogation. Three device decoders 260A, B, C, corresponding to the A/Ds 204A, B, C, respectively, are connected in parallel to the terminals 244. In the present example, decoder 260A for the analog-to-digital converter 204A, color A, is identified by the current device select code 242, and decoder 260A recognizes its unique code 242 standing on the terminals 244. Decoder 260A responds by putting a level on an AND gate 253A, which sets a flip-flop 251A, whose output goes through an OR gate 255 to a bus 250. The OR function of gate 255 may alternatively be performed inside the computer as a wired OR. The signal on bus 250 tells the computer 208 that the A/D converter 204A is not to be skipped because it is the device, or at least one of the devices, that has called for an interruption. The skip bus 250 is connected internally in the computer 208 to a skip flip-flop facility 248 so as to permit a skip signal to set that skip flip-flop, and the skip flip-flop is connected in the computer so as to enable incrementing of the program counter 225.

Upon finding that the first densitometer, for color A, is the one whose data should be accepted, the computer 208 goes to a densitometer interrupt service subroutine, point 3, FIG. 13. The computer accepts the offered density reading, step 251, from the A/D 204A of color unit A. The density data are connected from the A/D 204A to the twelve lines of the input information bus 240 through the data gates 206A, which are enabled by the latching flip-flop 251A. The density information goes into the accumulator 222 of the computer 208, and then into the memory 230. The computer clears the first densitometer's interrupt flag 236A by resetting the flag flip-flop 210A and flip-flop 251A, in accordance with FIG. 13, and then goes through certain checks. These include determining whether or not the color unit in question is on pressure, step 253, whether or not a control change by the operator is in progress, step 255, whether the system is in manual or in closed-loop control mode, and whether or not a required number of printing impressions have been made since going on pressure, step 259. A delay counter provides a time delay following the press' going on pressure, after the expiration of which the density control program is entered. The results of any of these checks can cause the computer to return to the main program (point 7) without making any control adjustments. When external contacts, such as, for example, on-pressure indicating contacts and make-ready mode indicating contacts, change their positions from open to closed or vice versa, the change is communicated to the computer 208 and put in memory 230. Positions of the contacts can thereafter be ascertained by interrogating the memory.

If a control adjustment is to be made, the computer goes, via a point 5 of the flow chart of FIG. 13, to the density control program of FIG. 14, which will be described in more detail hereinbelow, and it performs a density computation for the particular received data.

After the A/D 204A for color unit A has been serviced by the computer 208, the program returns to the interrupt subroutine by way of the "interrupt return" terminal (point 7) of FIGS. 14 and 12, and restores the data that was saved from the background program in the accumulator 222 and the link 224. The ability of the computer to accept interruptions is again enabled, and the computer returns to the background program, if any, by putting the previously stored contents of the program counter 225 back into the program counter

If an interrupt request is still present, as it will be in the present example because the A/Ds 204B and 204C for color units B and C are also calling for interruption, the interrupt program of FIG. 12 starts all over again. This time, when the A/D 204A for color unit A is tested by the computer 208 to see whether or not it was the one calling for the interruption, it is found not to be the one. The A/D 204A puts a skip signal on the skip bus 250 when the device select code 242 corresponding to the A/D 204A is applied to the device select bus 244. The computer skips the A/D 204A.

The computer then extracts from its memory 230 the word that is stored in the next sequential memory address. That word is another input or output transfer instruction and it calls for the A/D 204B, by means of its six identifying bits 242, i.e., the device select code, in step 261 of FIG. 12. A/D 204B responds with a signal on the common skip bus 250, to tell the computer 208 that it is the one (or one of several) which is presently requesting to interrupt the computer. The computer accepts the density data from the A/D 204B in the same manner as it did for the A/D 204A, and again returns to the background program after servicing the A/D 204B. The fact that color unit A was on pressure and its delay had been fulfilled does not guarantee that color unit B was also ready.

The background program does not as yet get started again, because in the three-color press being described, A/D 204C is also requesting to interrupt the computer in step 263. A/D 204C is then serviced by the computer 208 in the same manner as the others, after which the computer returns to its background program.

The status of the ring counter 196 is transmitted to the computer 208 to tell the computer which of the eleven densitometer heads 180 to 190 is currently providing density data. The ring counter 196 communicates data to the computer by the same method described above for A/D converters, so details for this are omitted from the drawings.

Processing a Density Reading

The flow chart of FIG. 14 shows the steps in processing a density reading in the computer 208 after a densitometer has interrupted the computer. Starting at the point 5, the computer already has in storage 230 the density reading from a densitometer's A/D converter 204A, B, or C, as described in detail above. The computer then goes through the routine described below, shown in FIGS. 14 and 17, and finally returns via point 7 to FIG. 12.

Reading Validity Check

The digital computer 208 examines data received from the selected densitometer A/D 204A, B, or C, and rejects any reading whose value is so unexpected so as to suggest that the reading is probably incorrect. A reading validity computation is employed for this purpose. Such erratic readings are identified as being any reading which does not exceed a predetermined fixed minimum value 266 (FIG. 15), or any reading which is below a floating minimum value 268, or any reading which is above a floating maximum value 270. The predetermined value 266 differs from the floating minimum value 268 in that the predetermined minimum value is semi-permanently preadjusted, while the floating minimum value may change its value in response to current conditions of operation, when the press is in a make-ready mode of operation. Thus, the reading validity computation accepts as being valid, only data which falls between certain limits, and disregards data outside of those limits. When the press is in a make-ready mode of operation, a center value 272 of the valid data range is determined by simultaneously considering a number of recent data readings. After the system has been put into a run mode of operation, the center value 272 of a valid range 273 is determined by density reference voltage 274 which is predetermined for the job. Changeover to the fixed density reference voltage 274 occurs when the operator switches from a make-ready mode to a run mode of operation. If the current reading is invalid, thp computer returns immediately to its main program. To summarize the reading validity checks, the first step is to ascertain whether the computer is in a make-ready mode or in a run mode, by reading its own memory, step 265. This determines the reference to be used as the center of the valid range. Validity is then examined in step 267, and if the reading is invalid, the computer returns to its main program via the point 7.

Data Smoothing

If the density reading is valid, it is used in a data averaging sub-routine to compute a running smoothed value of the density readings. This step 269 is for the purpose of data smoothing or filtering. It is done in a manner which minimizes the number of storage locations 230 in the computer 208 which must be provided for storing density readings. The previous smoothed value A, computed for the immediately preceding density measurement interval, is subtracted from the most recent density reading D to obtain a figure representative of the departure of the new reading from the previous smoothed value A. This departure is divided by a number K in order to reduce its impact, and the resulting quotient is added to the previous smoothed value A.

Thus, in the digital embodiment, the low pass filtering for data smoothing is accomplished by a computation sub-routine which operates in accordance with the equation A' = A + (D-A)/K. The new smoothed density reading A' is equal to the smoothed density reading A for the immediately preceding measurement level, plus an incremental change which is represented by the second term, namely, (D-A)/K. The numerator of the second term is the departure of the new density reading D from the previously smoothed value A. The denominator K of the second term is a selectable number which controls the amount of smoothing. The subroutine therefore discounts previous values at a predetermined rate.

The effect of this equation is to permit the most recent density reading D to affect the smoothed value A' only by a fractional amount which is 1/K of the deviation of that new density reading D from the previously smoothed value A. The smoothing process has a stabilizing influence on the inking system, giving it better behavior than it would have if each new density reading were taken as the only density reading then available for determining the settings of the ink keys 38. This new data smoothing system differs from a running average of the K most recent readings, in that this new system gives some weight not only to the K most recent readings, but also to earlier readings obtained before them.

The newest smoothed value A' is a weighted average of A and D in which the old smoothed value A is given (K-1) times as much weight as the new density reading D. That A' is a weighted average of A and D is better seen when A' is expressed in the following form, which is equivalent to the foregoing equation:

A' = (K-1/K) A + (1/K) D

Square Root of Error Signal

The output of the smoothing computation is multiplied by a gain constant, step 271, and a predetermined density reference value is then subtracted from that product, step 273. The resulting difference is multiplied by a second constant to produce a proportional error signal 280. A variable multiplying factor can be substituted for the second constant if desired, with its value dependent upon the present setting of the key involved, to take account of non-linearities in the system. The non-linear functional relationship is stored in the computer, and can be dependent upon actual ink key setting if desired.

Next, the error signal 280 is operated on by the computer 208 to extract the square root 282 of its magnitude. A subroutine in computer memory performs the square root computation by a conventional numerical method. The square root 282 of the error signal 280 is less responsive than the original signal 280 to system demands for very great corrections, which may be caused merely by system delays. The sign of the error signal is not altered by the square root subroutine, which operates only upon the magnitude.

Integration and Differentiation of Error Signal

If no error integral component of control signal is desired in the control loop, the computer next directs the resulting error signal 282 to a ratchet pullback computation, to be described in more detail below. If, instead, the computer has been directed by means of an external switch to produce an integral component 284 of the error signal in addition to a "proportional" component 282 which is always present, the computer reads and resets an elapsed impression counter 288. If one or more readings were previously discarded because of being invalid, the integration over impressions is thus maintained correct. It then multiplies the result by an integral gain constant. The product 290 is multiplied by the proportional error 282, and added to the previous value of integral error in a step 291. The new value 284 of integral error is then added to the proportional error 282 to produce a total error signal 292. A derivative term may be added to the error signal also, when desired, by a derivative subroutine, not shown. The effects upon system behavior of a derivative component of error signal were described in detail above in connection with the analog embodiment, and are the same here.

If integral control has been selected, the signal 292 is employed for the remaining steps; if integral control has not been selected, signal 282 is, instead, employed. In either case, a determination is made in step 293, FIG. 14, as to whether or not the density reading which is being processed came from the last (eleventh) density head. If it did not, a computed prospective change in actuator setting is merely stored for later use and more density readings are obtained. If it did, the actuators of one printing unit are all then moved simultaneously to their prospective new positions, in accordance with the error signal 282 or 292, as modified, perhaps, by some ratchet pullback and anti-lift-off computations that are now to be described.

Ratchet Pullback, Digital

The output data 282 or 292 resulting from the foregoing computations is further processed by a "ratchet pullback" subroutine 294, similar to that which was described in connection with the analog embodiment above. Briefly, when one of the keys 38 is prospectively directed to a position which is undesirably close to either one of its limits of travel, the ink fountain ratchet 34, which independently controls the amount of ink flowing at all of the keys 38 of one color unit, is operated to a new position. This changes the ink flow at all of the ink key positions in such a direction that the particular key which was becoming too close to one of its limits need not go any closer, and may nevertheless fulfill its ink demand. All of the other keys 38 are then moved automatically to compensate for the change in ratchet 34 position in order to recover their original ink flow rates. In the digital embodiment, prospective positions of the actuators and the ratchet motor 297 are computed, but the actuators and ratchet motor are not moved immediately. A key liftoff prevention routine, about to be described, is computed before anything is moved. Routine details of the ratchet pullback concept are described in the analog embodiment and not in the digital embodiment.

Key Liftoff Prevention

Following the ratchet pullback routine shown as step 294 in the flow chart of FIG. 14, the information flow of the digital embodiment goes to a key liftoff sub-routine starting at a point 4. The key liftoff subroutine, which is shown in the flow chart of FIG. 17, will first be described broadly without reference to the flow chart, then in more detail with the flow chart. The key liftoff subroutine is especially desirable in the invention if a single continuous duct blade instead of a segmented duct blade is used in the ink fountain. It corrects the following problem. If a desired position of a first key were a much more open position than that of an adjacent key, the adjacent key would prevent a single continuous duct blade from following the first key outward as it opens. The first key would then lift off of the duct blade, and for positions of the first key farther open than its liftoff position, the first key would have no further control of the blade, the blade being restrained by the adjacent key.

Typical experimental data obtained with a particular duct blade show that control by each individual key (or key group) can be maintained for several thousandths of an inch beyond the position of adjacent keys (or key groups), but that further movement of the key has no effect. If a required rate of ink flow has not yet been fulfilled by the time the key being adjusted lifts off of the duct blade, additional fountain opening must be accomplished by moving the adjacent keys or gey groups farther open. A similar situation occurs when one key is turned very far inward with respect to adjacent keys. The inward-going key would lift the duct blade off of the stationary adjacent keys, absent some preventive measures. The purpose of the ink key liftoff subroutine is to prevent ink fountain key liftoff from the fountain blade, while at the same time fulfilling the commanded ink flow requirements, to produce the desired optical density of printing. When ink flow requirements call for an ink key position which would be expected to result in key liftoff, the adjacent key positions are adjusted to prevent key liftoff so that all key groups remain in contact with the ink fountain blade and are capable of control without lost motion.

The principle of operation of the key liftoff sub-routine involves the following steps. First, the desired changes in key positions, which are ordinarily indicated by optical density readings on the printed material, are computed for all of the keys or key groups. Prospective positions of keys are then computed by adding the desired changes to the actually existing key positions. Before any of the keys are moved to the prospective positions, however, the prospective position for each key (or key group) is compared with the prospective positions of adjacent keys to ascertain whether or not liftoff would occur if the prospective positions were to be taken by all of the keys. If any liftoffs are predicted, the keys are not driven to the prospective positions. Instead, the prospective position information is modified so that the maximum difference of offset between positions of acjacent keys does not exceed a difference .DELTA. , below which liftoff is certain to be prevented. In comparing the prospective position of each key with that of adjacent keys, it is convenient to start with the key whose prospective position or gap is the greatest of any of the keys, and proceed to consider all of the keys in a sequence of decreasing magnitude of key position. The subroutine could equally well have called for starting instead with the key having the smallest prospective position and proceeding to treat the keys in ascending order of prospective position.

Each key is offset from an adjacent key by some amount; if that amount exceeds the maximum permissible amount .DELTA. beyond which liftoff may occur, an additional adjustment is required for an adjacent key. An amount X by which the offset between a key and an adjacent key exceeds the permissible offset .DELTA. is computed by the following formula: x= (P.sub.K + C.sub.K) - (P.sub.A - C.sub.a + .DELTA. ). In this equation, X is a tentative computed excess of offset between a key and an adjacent key. The maximum allowable key offset distance is .DELTA. . The current actual position of any particular key is designated as P, followed by a subscript. The key under examination is indicated by the subscript K. An adjacent key is indicated by the subscript A. A prospective change in key position is C, followed by a subscript. The primed quantities to be introduced below designate final values, while unprimed quantities represent original values. If X is negative or zero, the offset is not excessive and no adjustment of the adjacent key is required. If X is positive, the offset is so great that liftoff is likely to occur if the keys are moved to the indicated positions; the adjacent key must therefore be moved enough to bring the final offset down to the maximum permissible offset value .DELTA. . Thus, if X is positive, liftoff is predicted, and new prospective positions and corrections must be calculated as follows: P'.sub.A = P.sub.K - .DELTA. ; C'.sub.A = P'.sub.A - P.sub.A ; C'.sub.K = C.sub.K (unchanged); P'.sub.K = P.sub.K +C.sub.K (unchanged).

Actual present positions of the keys are sensed by individual feedback potentiometers 295, each of which is mechanically connected to the actuator of one key, FIG. 16. All of the potentiometers 295 are energized at their extreme ends by a power supply and, by means of a movable contact, each sends an analog signal indicative of its key's position to a position feedback multiplexer 297. Under control of the digital computer 208, the position feedback multiplexer 297 selects the position data of one of the keys at a time, and sends this one analog signal to an analog-to-digital converter 299. The analog-to-digital converter 299 produces a digital output signal and transmits it to the digital computer 208. By the same interfacing means which was described above in detail, the digital computer accepts data into its memory 230 from the key position feedback potentiometers, details being omitted this time.

A patch panel 301 is provided for specifying to the digital computer 208 the grouping of the ink keys, a group of one or more keys being controlled by each density sensing head. The patch panel 301 is a manual data input device by which a set-up man can specify which of the ink keys are to be controlled by each of the density sensing heads.

In the flow chart for performing the key liftoff subroutine, shown in FIG. 17, the steps of the process are numbered. In general, the first six steps compute prospective positions and identify the key currently to be examined. The seventh step performs the computation of the quantity X in the formula above, with respect to an adjacent key in an adjacent group. If liftoff is predicted by the seventh step, an adjacent key correction is made in the eighth step, and the new information is substituted for old values in storage 230. Details of the steps of the key liftoff subroutine are as follows: In the first step, the key actuator, (for example, one of the actuators in the set 300A) which corresponds to the left-hand end of a first group of keys, is identified. In the second step, the position of that actuator is determined by the computer by reading the actuators's feed-back potentiometer 295. In the third step, a prospective position of the subject actuator is calculated by algebraically adding a calculated correction to the present actual key position. The correction can be based upon either density readings or manual commands. The prospective key position datum which is thus computed is stored in a temporary table of the computer memory. In the fourth step, the first three steps are repeated for the actuator which is at the right-hand end of that same group. In the fifth step, prospective positions for the left-hand end and the right-hand end of each of the other groups of fountain keys are computed, and stored in a temporary table in the computer memory. In the sixth step, the particular key, of those listed in the tamporary table, which has the greatest prospective position (i.e., ink gap), is identified. It will be the first one which will be studied to determine whether or not it will lift off. In the eighth step, the quantity X of the foregoing equation is computed for that key. If X is positive, indicating that liftoff would occur, the correction for the adjacent key, in a different group from the subject key, is changed to a value C'.sub.A such that its new position would be P'.sub.A as shown in the equations. As a result, this new prospective position P'.sub.A of the adjacent key is different from the position of the subject key which is being studied by an amount exactly equal to .DELTA. , so that liftoff should not occur. The new values replace the old values in all the stored data regarding prospective positions. If, instead, the computation of the seventh step shows that liftoff will not occur because the offset is not excessive, no adjustment need be made in the adjacent key's prospective correction value as previously computed. In the ninth step, the subject key which was just taken care of is deleted from the temporary table so that the second greatest key position of the original array is now the greatest key position still remaining in the temporary table. A check is made in the tenth step as to whether or not all of the keys have as yet been studied, for this particular printing unit; if they have not, the subroutine returns to the sixth step and operates on the highest remaining key position in the temporary table. If, upon the tenth step, all of the actuators for the printing unit are found to have been done, the computer goes to the eleventh step and transfers the data describing the final prospective changes through an external output decoder and multiplexer 296 to a corresponding data register of the printing unit, for example, data register 298A, FIG. 16. Also in the eleventh step, the computer issues a command to a gate flip-flop 303 which enables pulses from an AC source 305 to actuate all of the keys 38 to take up their final prospective positions as computed above by computer 208. All of the keys which are to be moved are moved simultaneously by their actuators. More details of the operation of the actuator circuits of FIG. 16, that is, the last step of the key liftoff routine, are given below. The end of the key liftoff subroutine is indicated by a point 6 which returns the control of the computer to the main program as shown on FIG. 14.

Actuator Circuits

The last step of the key lift-off routine involves the following details. The computer 208 sends signals to the external output decoder and multiplexer 296, FIG. 16. The output decoder and multiplexer 296 multiplexes the computers's output to data registers such as register 298A, for the various actuators such as actuators 300A, preparatory to later driving those actuators simultaneously to their new positions.

Each of the actuators 300A, B, C, FIG. 16, includes groups of bi-directional synchronous motors 302A, B, C, one motor being associated with each key. The motors are driven by short bursts 304A, B, C, of AC voltage from the AC source 305, transmitted through logic gates 306A, B, C, for the actuators 300A, B, C, respectively. In the sequence of operation, information intended for the actuators 300A, B, C, is multiplexed out of the computer 208 to the groups of data storage registers 298A, B, C. The direction of each motor is controlled by a direction flip-flop stage or sign bit of each register. Each register, such as register 298A, accommodates eleven words. All 33 storage registers 298A, B, C, (three for each of the eleven densitometer heads 180 to 190) are first preset by the computer 208. Then all of the actuators which are to move during an adjustment interval are moved simultaneously, to minimize transverse slipping of the ink keys 38 on the inker blade 36.

The ink keys 38 are moved in discrete steps. As a numerical example of one typical embodiment, each step of a key 38 moves the key 0.000125 lineal inch. Four steps are accomplished per second when a maximum rate of change of key position is called for. The key then travels at an average speed of 0.0005 inch per second. Typically, one impression per second may be made on a sheet-fed press, in which case the keys can travel 0.0005 lineal inch per impression. If a density reading is made by a particular one of the eleven density heads 180 to 190 only upon each eleventh impression, a corresponding key has time to travel 0.0055 inch for each density reading interval of that particular key 38. Full scale travel of the keys is approximately 0.03 inch, as an example.

The burst signals 304A, B, C, of AC voltage, in addition to driving the actuators, are rectified and filtered by rectifier-filters 308A, B, C, to produce groups of pulses 310A, B, C, which count-down the data registers 298A, B, C. The data registers 298A, B, C are pre-settable down-counters which stop at zero reading and thereupon produce a pulse. When the count in a data register 298A, B, C becomes zero, the pulse thereupon produced disables the gates 306A, B, C by means of an OR gate 313 whose output resets the gate flip flop 303. This turns off the AC signals 304A, B, C to the corresponding actuators 300A, B, C, and stops the actuators.

End of a Density Data Interrupt

After the key liftoff routine of FIG. 17 has been completed, the computer is at a point 6 of FIG. 14. It then returns to the main program by going first to the terminal 7 of FIG. 12, where it performs a few interrupt return steps 136, 318 and 320. Step 316 involves putting back into the accumulator 222 and into the link 224 the data from the background program which had been stored in the memory 230 when the background program was interrupted. The capability of the computer to be interrupted is then re-enabled in the step 318, and the computer returns to the interrupted background program in the step 320.

If another densitometer channel, such as the channel for color B or for color C, then has some data ready, (and it should have), the computer then goes through another entire interrupt routine, starting again at the top of FIG. 12, before accomplishing anything on the background program.

When the density readings obtained from the printed sheet 50 become equal to the reference density settings called for by the operator and stored in the computer, the error signal 292 is zero and the actuators 300A, B, c do not operate until the next occurance of an error signal.

Prospective changes in key positions are computed and stored in the computer for all of the keys or key groups of a printing unit before any of the keys of the unit are adjusted. Then all keys of the unit are adjusted simultaneously upon a command generated by the computer.

Other Controls

The actuators can be controlled manually also, in an open-loop mode of operation, by the operator, who may observe density readings and intervene at the digital computer 208 by interjecting control signals 312, FIG. 16. To adjust an ink key, the operator holds a spring-return manual switch down; the ink key moves as long as the switch is held down.

All of the features of the digital embodiment can be provided also in the analog embodiment and vice versa although not all of them are shown in both embodiments.

It is necessary to associate the various sensors of the densitometer heads 41 and the channels of the gated densitometer measuring circuits 51 with the various printing units, in the digital embodiment of the invention, for the same reasons which were described in connection with the analog embodiment, FIG. 8A and FIG. 8B. In the digital embodiment, a single six-position selector switch 174', FIG. 9, is provided for setting by the operator. Each of the six positions of switch 174' correspond to a different permutation of connections between densitometer channels and printing units, as shown in FIG. 8B. The position of switch 174' is read by the digital computer in the course of executing its program. On the basis of the switch's position, a correct set of printing unit and display pointers is selected and stored for program use. The selection information may be entered into the computer from a keyboard or other data input device instead of from the manual selector switch 174' if desired.

The ink keys 38 can be present to provide approximately correct ink feed for each job in accordance with data entered either manually by the operator or automatically from a data storage and handling device, such as a reader for punched paper tape. Both of these methods are indicated in FIG. 16 by the block "various inputs from operator." The data thus entered are stored in the computer's memory 230 until a command occurs to call it forth to be utilized for control.

Data stored in the computer 208 and data available to the computer can be read out by various methods. For example, a successful set of key settings for a particular job can be read out and stored on punched paper tape for future use in presetting the keys.

A single computer can be employed for a plurality of printing presses if desired.

Claims

What is claimed is:

1. In a printing press, a method of controlling the feeding of ink from an ink fountain to a printing member for printing images during successive impression cycles onto a material comprising the steps of: cyclically measuring the density of ink deposited on the material and establishing first signals representing the measurements, establishing second and third signals representing maximum and minimum acceptable values of density, respectively, comparing the first signals with said second and third signals, blocking those of said first signals that are greater than or smaller than said second and third signals, respectively, and passing the acceptable first signals whose values are between the values of said second and third signals, periodically producing a control signal in dependence upon said acceptable first signals, and varying the feeding of ink from the ink fountain in response to deviation of said control signal from a value corresponding to a specified ink density to reduce said deviation.

2. A method of controlling the feeding of ink as defined in claim 1 wherein said second and third signals representing maximum and minimum acceptable values are established by establishing at least one tolerance signal indicating predetermined tolerance margins from an effective running average of previous acceptable first signals, establishing a running density signal indicating an effective averaging of previous acceptable values of said first signals, and combining said tolerance and running density signals to establish the acceptable values for maximum and minimum measurements.

3. A method of controlling the feeding of ink as defined in claim 1 and wherein establishing said signals representing the minimum acceptable values comprises establishing a predetermined first minimum value signal, establishing a running density signal from an effective running averaging of previous acceptable first signals, automatically subtracting a tolerance signal from said running density signal to produce a second minimum value signal, and employing the greater of said first and second minimum value signals to serve as said third signal for comparison with said first signals.

4. The method of controlling the feeding of ink as defined in claim 1 and further comprising the steps of: establishing an error signal that is representative of said deviation, producing a primary signal that is responsive to only a present value of said error signal, producing a secondary signal that is a function of a derivative of said error signal, said derivative being taken with respect to a quantity that varies monotonically with time, producing a tertiary signal that is a function of an integral of said error signal with respect to said quantity, combining said primary, secondary, and tertiary signals to produce a composite control signal, and varying the feeding of the ink from the fountain in response to said composite control signal.

5. A method of controlling the feeding of ink as defined in claim 4 and wherein the quantity that varies monotonically with time is the number of impression cycles of the printing press.

6. A method of controlling the feeding of ink as defined in claim 1 and wherein said varying of the feeding of ink comprises the steps of establishing a first error signal indicative of said deviation, modifying said first error signal to provide a second error signal which varies disproportionately in accordance with a nonlinear compression function of said first error signal, the feeding of ink being varied as a function of said second error signal, whereby a varying of said feeding of ink is less than a proportionate amount as said deviation increases.

7. A method of controlling the feeding of ink as defined in claim 6 and wherein said step of modifying said first error signal in accordance with a nonlinear compression function comprises the step of modifying to provide a second error signal which varies in accordance with the square root of said first error signal when said first error signal is positive, and in accordance with the negative of the square root of the magnitude of said first error signal when said first error signal is negative.

8. In a printing press, a method of controlling the feeding of ink from an ink fountain to a printing member for printing images during successive impression cycles onto a material comprising the steps of: establishing an error signal that is representative of a desired change in the rate of feeding of ink from the fountain, producing a first signal that is responsive to only a present value of said error signal, producing a second signal that is a function of a derivative of said error signal, said derivative being taken with respect to a quantity that varies monotonically with time, producing a third signal that is a function of an integral of said error signal with respect to said quantity, combining said first, second and third signals to produce a composite control signal, and varying the feeding of the ink from the fountain in response to said composite control signal.

9. A method of controlling the feeding of ink as defined in claim 8 and wherein the quantity that varies monotonically with time is the number of impression cycles of the printing press.

10. A method of controlling the feeding of ink as defined in claim 8 and wherein said step of producing said second signal comprises the steps of storing one value of said error signal as produced during one cycle of the press, establishing a second value of said error signal as produced during a later cycle of the press, an inanimately subtracting said one value of error signal from the later-produced value of error signal to produce said second signal.

11. A method of controlling the feeding of ink as defined in claim 8 and wherein said step of establishing an error signal comprises the steps of cyclically measuring the density of ink deposited on the material, producing density signals representing said measurements, establishing at least one reference signal corresponding to a desired density of ink, and inanimately subtracting one of said density and reference signals from the other one to establish said error signal.

12. A method of controlling the feeding of ink as defined in claim 8 and wherein establishing an error signal comprises the steps of: cyclically measuring the density of ink deposited on the material and establishing a density signal dependent upon the measurements, producing a first control signal indicative of deviation of said density signal from a specified ink density value, modifying said first control signal to provide a second control signal which varies disproportionately in accordance with a nonlinear compression function of said first control signal, said error signal being varied as a function of said second control signal, whereby a varying of said error signal is less than proportionate amount as said deviation increases.

13. A method of controlling the feeding of ink as defined in claim 12 and wherein said step of modifying said first control signal in accordance with a nonlinear compression function comprises the step of modifying to provide a second control signal which varies in accordance with the square root of said first control signal when said first control signal is positive, and in accordance with the negative of the square root of the magnitude of said first control signal when said first control signal is negative.

14. In a printing press, a method of controlling the feeding of ink from an ink fountain to a printing member for printing images during successive impression cycles onto a material, comprising the steps of: cyclically measuring the density of ink deposited on the material and producing a principal control signal in dependence upon the measurements, varying the feeding of ink from the ink fountain by adjusting a setting of at least one ink feed control element of said fountain in response to deviation of said principal control signal from a preestablished ink density value, detecting the occurrence of a change in the printing process that would be capable of later affecting the density of the ink deposited if no compensating change were made in the setting of said ink feed control element, producing an additional signal indicative of said occurrence before said occurrence has substantially altered said principal control signal, and additionally adjusting said ink feed control element in response to said additional signal and independently of the value of said principal control signal to effect compensation for said occurrence.

15. A method of controlling the feeding of ink as defined in claim 14 and wherein the step of cyclically measuring comprises a step of measuring in synchronism with press impression cycles.

16. In a printing press, a method of controlling the feeding of ink from an ink fountain to a printing member for printing images during successive impression cycles onto a material, comprising the steps of: cyclically measuring the density of ink deposited on the material and establishing a first signal dependent upon the measurements, producing a first control signal indicative of deviation of said first signal from a specified ink density value, and varying the feeding of ink from the ink fountain in response to said first control signal, wherein said step of varying the feeding of ink comprises the step of modifying said first control signal to provide a second control signal which varies disproportionately in accordance with a nonlinear compression function of said first control signal, the feeding of ink being varied as a function of said second control signal, whereby a varying of said feeding of ink is less than a proportionate amount as said deviation increases.

17. A method of controlling the feeding of ink as defined in claim 16 and wherein said nonlinear compression function is adjustable as to degree of nonlinearity of said second control signal with respect to said first control signal.

18. A method of controlling the feeding of ink as defined in claim 16 and wherein said step of modifying said first control signal in accordance with a nonlinear compression function comprises the step of modifying to provide a second control signal which varies in accordance with the square root of said first control signal when said first control signal is positive, and in accordance with the negative of the square root of the magnitude of said first control signal when said first control signal is negative.

19. A method of controlling the feeding of ink as defined in claim 16 and wherein the varying of the feeding of ink is done in accordance with a composite signal comprising at least two of the following three components of signal: (a) said second control signal, (b) a second component of signal dependent upon an integral of said first component, and (c) a third component of signal dependent upon a derivative of said first component, said integral and said derivative being with respect to occurrences that are successive in time.

20. In a printing press, a method of controlling the feeding of ink from an ink fountain to a printing member for printing images during successive impression cycles onto a material, comprising the steps of: cyclically measuring the density of ink deposited on the material and establishing a first signal dependent upon the measurements, producing a control signal indicative of deviation of said first signal from a specified ink density value, varying the feeding of ink from the ink fountain in dependence upon said control signal, providing a further signal during each cycle indicative of whether or not said press is currently cyclically printing onto said print-receiving material, and blocking at least one of said steps listed preceding said step of providing a further signal to prevent said varying of the feeding of ink when said further signal indicates that the press is not currently printing.

21. A method of controlling the feeding of ink as defined in claim 20 and comprising the additional step, following the step of providing a further signal, of providing an additional signal indicative of whether or not a predetermined time delay has expired since the press started to cyclically print onto said print-receiving material, and wherein said step of blocking comprises blocking at least one of the steps listed preceding said step of providing a further signal of said additional signal indicates that said time delay has not expired.

22. A method of controlling the feeding of ink as defined in claim 21 and wherein said predetermined time delay corresponds to a predetermined number of printing impressions.

23. In a printing press, a method of controlling the feeding of ink from an ink fountain to a printing member for printing images during successive impression cycles onto a material, comprising the steps of: providing a plurality of ink feed control elements laterally spaced along the ink fountain and adjustable by signals to greater and smaller positions for controlling the flow of ink, establishing a plurality of signals the value of each of which represents a tentative position of a group of at least one of said ink feed control elements, inanimately computing the difference between a pair of said tentative positions corresponding to a pair of adjacent laterallyspaced groups of said control elements, comprising said difference with a predetermined maximum allowable difference, modifying the value of at least one of said pair of signals toward the value of the other signal of the pair to reduce said difference to said maximum allowable difference if said difference initially exceeds said maximum allowable difference, substituting the modified value of signal for its pre-modification value to represent a modified tentative position of the corresponding group of said control elements, repeating the foregoing steps of computing, modifying, and substituting with pairs of said groups of control elements until all pairs have been modified so as not to exceed said maximum allowable difference, and adjusting said groups of control elements to positions in accordance with the modified values of said signals.

24. A method of controlling the feeding of ink as defined in claim 23 and wherein the first pair of tentative positions whose difference is computed corresponds to an extreme group of control elements whose tentative position is at least as great as the tentative position of any other group of control elements, the other group of the first pair being laterally adjacent to said extreme group on a first side of said extreme group, and where the second pair whose difference is computed corresponds to said extreme group and to an adjacent group on a second side of said extreme group, and wherein subsequently treated pairs are treated in descending order of extremity of position similarly until all adjacent pairs of groups have been treated, and wherein said step of modifying the value of at least one of said signals of said pair comprises the step of modifying the signal of said pair corresponding to the control group having the less extreme tentative position.

25. A method of controlling the feeding of ink as defined in claim 23 and wherein the pair of laterallyspaced control groups whose difference is computed first consists of an extreme group whose tentative position is at least as small as the tentative position of any other group, the other group of the first pair being laterally adjacent to said extreme group on a first side of said extreme group and where the pair whose difference is computed second consists of said extreme group and an adjacent group on a second side of said extreme group, and wherein subsequently treated pairs are treated in ascending order of extremity of position similarly until all adjacent pairs of groups have been treated, and wherein said step of modifying the value of at least one of said signals of said pair comprises the step of modifying the signal of said pair corresponding to the group having the greater tentative position.

26. In a printing press which has a printing member printing images onto a material during successive impression cycles, an ink control apparatus comprising a source of ink, an inker for feeding ink at a controllable rate from the source of ink onto the printing member including first means having at least one adjustable ink control device for varying said rate of feeding ink, means for successively measuring printed densities of the images on the material and successively producing signals in dependence thereon for adjustment of said first means, means for establishing limiting values between which said signals are more reliable indicators of density for ink control purposes than are signals outside of said limiting values, and circuit means receiving said signals and said limiting values and comparing said signals with said limiting values for rendering ineffectual for ink control those of said signals that are outside of said limiting values.

27. An ink control apparatus as defined in claim 26 and wherein said circuit means further comprises means for producing a reference signal, and feedback means adapted and arranged for automatically adjusting said first means in dependence upon said signals and said reference signal.

28. In a printing press which has a printing member printing images onto a material during successive impression cycles, an ink control apparatus comprising a source of ink, an inker for feeding ink at a controllable rate from the source of ink onto the printing member including means comprising at least one adjustable able ink control device for varying said rate of feeding ink, first means for periodically measuring printed densities of the images on the material and periodically producing measurement signals in dependence thereon, data memory means for storing indicia of density of print determined as a function of the measurements of said first means, second means for receiving said measurement signals and said indicia from said data memory means and periodically combining them to produce a new indicia and for replacing the previous indicia with said new indicia in said data memory means upon production of each new indicia, and circuit means for successively receiving said new indicia for use in adjusting said ink control device in dependence thereon, said second means comprising means for discounting said previous indicia at a predetermined rate when combining them with said measurement signals.

29. In a printing press which has a printing member printing images onto a material during successive impression cycles, an ink control apparatus comprising a source of ink, an inker for feeding ink at a controllable rate from the source of ink onto the printing member including means comprising at least one adjustable ink control device for varying said rate of feeding ink, first means for periodically measuring printed densities of the images on the material and periodically producing measurement signals in dependence thereon, data memory means for a storing a running value indicating density of print determined as a function of the measurements of said first means, second means for receiving said measurement signals and said running value from said data memory means and periodically producing a new running value and for replacing the previous running value with said new running value in said data memory means upon production of each new running value, and circuit means for successively receiving said new running value for use in adjusting said ink control device in dependence thereon.

30. An ink control apparatus as defined in claim 29 and wherein said circuit means further comprises means for producing a reference signal, and feedback means adapted and arranged for automatically adjusting said ink control device in dependence upon said new running value and said reference signal.

31. An ink control apparatus as defined in claim 29 and wherein said second means comprises means for combining each measurement signal with the running value stored in said data memory means in a weighted average in which the stored running value has (K-1) times as much weight as does said measurement signal, K being a predetermined number greater than 1.

32. An ink control apparatus as defined in claim 31 and wherein said means for combining comprises filter means having a nominal cutoff frequency for attenuating signals which are above said frequency much more than signals which are below it, said filter means comprising a one-pole, electrical analog low-pass filter.

33. An ink control apparatus as defined in claim 32 and wherein said filter means further comprises means for sensing frequency of impressions made by said printing press, and means for altering said nominal cutoff frequency to maintain said cutoff frequency proportional to said frequency of impressions.

34. An ink control apparatus as defined in claim 31 wherein digital computer means provides said data memory means and said second means for computing said weighted average.

35. An ink control apparatus as defined in claim 34 and wherein said digital computer means further comprises means for storing the latest-received signal, D, of said measurement signals and said number K; means for subtracting said running value from said signal D to produce a difference; means for dividing said difference by said number K to produce a quotient; and means for adding said quotient to said running value stored in said data memory means to produce said new running value.

36. An ink control apparatus as defined in claim 29 and wherein said circuit means for adjusting said ink control device comprises a source of reference signals, a comparator connected to receive and compare said new running value and said reference signals and to produce an error signal proportioned to a difference between them, and sub-circuit means receiving said error signal and connected with said adjustable ink control device for adjusting said device as a compression function of said error signal, said sub-circuit means comprising means receiving said error signal and providing nonlinearly therefrom a second signal for adjusting said ink control device, whereby a varying of said rate of feeding ink is less than a proportionate amount as said error signal increases in magnitude.

37. An ink control apparatus as defined in claim 36 and wherein the magnitude of adjustment of said ink control device produced by said sub-circuit means is porportioned to the square root of the magnitude of said error signal, and a directional sign of said adjustment tracks the sign of said error signal.

38. In a printing press which has a printing member printing images onto a material during successive impression cycles, an ink control apparatus comprising a source of ink, an inker for feeding ink at a controllable rate from the source of ink onto the printing member including means comprising at least one adjustable ink control device for varying said rate of feeding ink, means for successively measuring printed densities of the images on the material and successively producing measurement signals in dependence thereon, a source of reference signals, a comparator connected to receive and compare said measurement signals and said reference signals and to produce an error signal proportioned to a difference between them, and circuit means receiving said error signals and connected with said adjustable ink control device for adjusting said device as a compression function of said error signal, said circuit means comprising means receiving said error signal and providing non-linearly therefrom a second signal for adjusting said ink control device, whereby a varying of said rate of feeding ink is less than a proportionate amount as said error signal increases in magnitude.

39. An ink control apparatus as defined in claim 38 and wherein the magnitude of adjustment of said ink control device produced by said circuit means is proportioned to the square root of the magnitude of said error signal, and the directional sign of said adjustment tracks the sign of said error signal.

40. In a printing press which has a printing member printing images onto a material during successive impression cycles, an ink control apparatus comprising a source of ink, an inker including a fountain for adjustably feeding ink from the source of ink onto the printing member including means comprising a plurality of ink control devices laterally spaced at locations along said fountain and adjustable to different settings for varying rates of feeding ink respectively at said locations, means for producing a plurality of signals corresponding to proposed settings for said ink control devices, means for computing data for each subject ink control device representing af inal setting which provides the greater ink flow of the following two settings: said proposed setting, or, a setting which is less than the final setting of an adjacent ink control device by a predetermined amount, memory means for storing said data until said data has been computed for all of said plurality of ink control devices, and means for reading said data from said memory means and simultaneously adjusting all of said ink control devices to said final settings in accordance with said data automatically when all of said data have been computed.

41. An ink control apparatus as defined in claim 40 and wherein said means for producing a plurality of signals corresponding to proposed settings comprises a plurality of means connected respectively with said ink control devices for sensing actual settings thereof and producing indicia respectively of said actual settings, and means for algebraically adding proposed increments and decrements to said indicia to produce said plurality of signals corresponding to proposed settings.

42. An ink control apparatus as defined in claim 41 and wherein said means for simultaneously adjusting all of said ink control devices to said final settings comprises actuator means drivably connected with said ink control devices for adjustably driving said ink control devices by electrical drive signals, and said means for computing comprises computer means receiving said indicia of said actual settings and receiving said signals corresponding to proposed settings for providing said electrical drive signals, said computer means comprising arithmetic means for adding said predetermined amount to said proposed setting of said subject ink control device to produce a sum and subtracting said sum from the prospective setting of said adjacent ink control device to produce a difference, means for ascertaining whether or not said difference has a positive sign, means for enabling said electrical drive signal of said subject ink control device only if said difference has a positive sign, and means for establishing a value of said electrical drive signal of said subject ink control device proportioned to drive said subject ink control device to a final actual setting which is less than the final actual setting of said adjacent ink control device by said predetermined amount.

43. In a printing press which has a printing member printing images onto a material during successive impression cycles, an ink control apparatus comprising a source of ink, an inker including a fountain for adjustably feeding ink from the source of ink onto the printing member including means comprising a plurality of ink control devices laterally spaced at locations along said fountain and adjustable to different settings for varying rates of feeding ink respectively at said locations, means for producing a plurality of signals corresponding to proposed settings for said ink control devices, means for computing data for each subject ink control device representing a final setting which provides the smaller ink flow of the following two settings: said proposed setting, or, a setting which is greater than the final setting of an adjacent ink control device by a predetermined amount, memory means for storing said data until said data has been computed for all of said plurality of ink control devices, and means for reading said data from said memory means and simultaneously adjusting all of said ink control devices to said final settings in accordance with said data.

44. An ink control apparatus as defined in claim 43 wherein each of said ink control devices comprises a group of one or more contiguous ink control elements.

45. In a printing press which has a printing member printing image onto a material during successive impression cycles, an ink control apparatus comprising a source of ink, an ink fountain having at least one ink feed control element for feeding ink to the printing member from said source, means for producing a first control signal, means for varying the feeding of ink from the ink fountain by adjusting a setting of at least one ink feed control element of said fountain in response to said first control signal, means for detecting the occurrence of a change in the printing process that would be capable of later affecting the density of the ink deposited on the material if no compensating change were made in the setting of said ink feed control element, means for producing a second signal indicative of said occurrence before said occurrence has substantially altered said ink density, and means for additionally adjusting said ink feed control element in response to said second signal and independently of the value of said first control signal to effect compensation for said occurrence.

46. An ink control apparatus as defined in claim 45 and wherein said fountain includes means comprising a plurality of groups of at least one ink feed control element laterally spaced at locations along said fountain and adjustable within limits to different settings for varying the rates of feeding ink at said respective locations, and wherein said means for producing a first control signal comprises means for producing a command signal for establishing a proposed rate of ink feed for at least one of said ink control elements, means including computer means for detecting a condition wherein a proposed setting corresponding to said proposed rate of ink feed would exceed one of said limits and would leave a deficiency portion in said proposed rate, said ink control apparatus further including master control means for altering rates of feeding ink simultaneously across the entire fountain upon receiving a master signal, and means for producing a master signal to adjust said master control means to bring said element as adjusted by said first control signal within said limits, and wherein said means for producing a second signal comprises means responsive to the occurrence of the adjustment of said master control means, and wherein said means for additionally adjusting comprises means for adjusting the others of said elements to counteract their change in ink feed due to the change in said master control means.

47. An ink control apparatus as defined in claim 46 and wherein said means for producing a command signal and said means for detecting a condition and said means for producing a master signal, comprise digital computer means.

48. In a printing press which has a printing member printing images onto a material during successive impression cycles, an ink control apparatus comprising a source of ink, an inker for feeding ink at a controllable rate from the source of ink onto the printing member including means having at least one adjustable ink feed control element for varying said rate of feeding ink, means for establishing an error signal that is representative of a desired change in the rate of feeding ink from said ink feed control elements, means for storing data defining a functional relationship between values of said error signal and amounts of change of setting of said feed control element which would substantially effect said desired change, means for using said error signal to inanimately compute as a function of at least said error signal and said stored relationship an amount representing the change of setting of said feed control element which would substantially effect said desired change in the rate of feeding ink, means for sensing the current position of said feed control elements and producing a position-indicating signal accordingly, means for algebraically adding the computed amount to said position-indicating signal to produce a sum signal representing a prospective new position for the feed control element, and means for operating said feed control element to said prospective new position, said means for establishing an error signal comprising means for cyclically measuring the density of ink deposited on the material and establishing a first signal dependent upon the measurements, and means for producing said error signal in response to deviation of said first signal from a specified ink density value.

49. In a printing press which has a printing member printing images onto a material during successive impression cycles, an ink control apparatus comprising a source of ink, an ink fountain for feeding ink to the printing member from said source, means for cyclically measuring the density of ink deposited on the material and establishing a first signal dependent upon the measurements, means for producing a control signal indicative of deviation of said first signal from a specified ink density value, means for varying the feeding of ink from the ink fountain in dependence upon said control signal, means for providing a further signal during each cycle indicative of whether or not said press is currently cyclically printing onto said print-receiving material, and means for disabling at least one of the aforesaid elements to prevent said varying of the feeding of ink when said further signal indicates that the press is not currently printing.

50. In a printing press which has a printing member printing images onto a material during successive cycles, an ink control apparatus comprising a source of ink, an inker including a fountain for adjustably feeding ink from the source of ink onto the printing member including means comprising a plurality of ink control devices laterally spaced at locations along said fountain and adjustable by means of signals to different settings for varying rates of feeding ink respectively at said locations, means for producing a plurality of data corresponding to proposed changes of settings for those of said ink control devices whose settings it is currently desired to change, memory means for storing said data until said data are available for all of said plurality of ink control devices, and means for reading said data from said memory means and simultaneously adjusting all of said ink control devices whose settings it is currently desired to change to proposed settings in accordance with said data automatically when all of said data are available in said memory means.

51. In a printing press which has a plurality of printing units each having a printing member printing images onto a material during successive impression cycles and an ink control apparatus comprising a source of ink for each printing unit and an inker for each printing unit for feeding ink at a controllable rate from the source of ink onto the printing member including first means having at least one adjustable ink control device for varying said rate of feeding ink, a plurality of sensor means each to be selectably associated with a respective one of said printing units for successively measuring printed densities of the images placed on the material by the associated printing unit and successively producing signals in dependence thereon for adjustment of the respective first means for the printing unit, and switching means having a switch status for every permutation in which said sensor means can be associated with the various printing units and selectively actuatable to any one of said statuses for selecting one such permutation.

52. An ink control apparatus as defined in claim 51 and further comprising a plurality of display means each one of which remains associated with the same respective printing unit irrespective of the status of said switching means.

53. An ink control apparatus as defined in claim 51 and wherein said switching means comprises a digital computer including memory means for storing data descriptive of all of said permutations, said computer including means responsive to input data for selecting one of said permutations.