US3196404A - Printer buffer load and read control means - Google Patents

Description

J y 1965 R. .1. FURLONG ETAL 3,196,404

PRINTER BUFFER LOAD AND READ CONTROL MEANS Filed June 26. 1961 2 Sheets-Sheet 1 F I 6.1 ,20

v COMPARE rg'cmcun REGISTER 9 STAGE YRING,

X DRIVERS w 11a 118 118 11a 11s I 35 14 35 as; 51} I 331- 340- 35m 310 310$ I 3131b 2141, 35b 611 i311 1o 3 1 7 HAMMER' 11111111 H 16 SELECTION] DRIVERS zz-R -a2 32- 1 I 181 I x R1NG, 5 STAGE EDP SUBSCAN GATING RING 3STAGE PULSE o GENERATOR COMPARE RC CONTROL CIRCUIT 1o L -1111 111 CONTROL 01110011 1050mm v PRINT 1 c1001 (286 281 282 INVENTURS 281 A BUFFER PRINT 011-3 x-s ROBERT g LOAD 1111115 0. CALVERT ATTORNEY United States Patent Ofiice 3,196,404 Patented July 20, 1965 3,196,404 PRINTER BUFFER LOAD AND READ CONTROL MEANS Robert Joseph Furlong and James Douglas Calvert,

Poughkeepsie, N.Y., assignors to International Business Machines Corporation, New York, N.Y.,

a corporation of New York Filed June 26, 1961, Ser. No. 119,588 6 Claims. (Cl. 340172.5)

This invention relates to high-speed printing apparatus, and particularly to improvements in the control means for line-at-a-time bufier storage for printer apparatus.

This invention can be used in printing machines which print on the fly; i.e., printing is done by engraved type imprints, which are in motion when the imprint is formed. In general, on-the-fly printers may comprise a print mechanism having an array of different type imprint members arranged in a predetermined sequence, means for moving the type members in the sequence relative to a print line having a plurality of print positions, control means for selecting type members to be struck at the moment of alignment of any selected type member at a required print position, and striking means impacting selected type members as they reach various print positions of a print line.

One form of such printing machine to which this invention is particularly related is described in a copending US. patent application for a High Speed Printer Apparatus, Serial No. 844,511, filed by F. M. Demer and E. I. Grenchus on October S, 1959, now Patent 2,993,437, and assigned to the same assignee as the present invention. Other pertinent US. patent applications assigned to the same assignee are an Improved Chain Printer, Serial No. 704,938, filed by F. M. Demer, R. H. Harrington, and A. T. Shalkey on December 24, 1957, now Patent 2,990,767; Chain Printer Timer, Serial No. 705,678, filed by E. R. Wooding on December 27, 1957, which is Patent 2,918,865; and Hammer Drivers, Serial No. 72,766, filed by A. M. Gindi on November 30, 1960.

In Patent 2,993,437 the type members of a chain printing mechanism are arranged in a predetermined manner and attached to an endless belt, forming a chain of type which is driven at a constant rate of travel along a print line. Print hammers located at the plurality of print positions along the print line are actuated to selectively strike the moving type members. A print medium (paper) located between the hammers and the type is driven by actuation of the hammers into contact with the type members; thereby imprinting a record.

Furthermore, in prior Patent 2,993,437 it was described how the spacing between print positions can be different from the spacing between engraved characters on a moving member; for example, the spacing between engraved type is 1.5 times the spacing between printing positions. This technique allows greater width for imprinted charactors, which produces a better quality printed page than can characters spaced with a ratio of one. It was described in said prior application how it was necessary to load a regenerative coordinate core torage buffer in a predetermined sequence, and to read the buffer with plural subscan sequences, wherein each subscan starts with a different character at the beginning of a line of print, and each subscan skip to every third print position and stored character position until every stored position representing one print line is interrogated and compared with the type characters existing at the respective print positions. The subscanning arrangement therefore read out buffer characters in correspondence with the mechanical alignment of the moving characters with print positions. The prior application also disclosed how three rings can beused to accomplish the loading and reading of the butter, wherein some of the rings are open.

The present invention is basically an improvement in the control means for ring circuits used to sequence characters being loaded and read from a butter storage device.

It is an object of this invention to provide an improved ring control means for a printer butter which is significantly simpler than prior ring controls for accomplishing buifer coordination with printing apparatus having a spacing ratio diiferent from one.

It i another object of this invention to obtain buifer storage control with buffer addressing rings which are all closed.

The invention controls buffer-addressing rings by providing a relatively simple switching means that controls the effective-cascaded operating order among the rings. One of the buffer-addressing rings has N number of sequenial outputs, wherein N is equal to the character-skipping increment within a buffer for compatibility with spacing ratios different from one. In such case, the N numbered ring is connected first in the cascaded order during one of the loading and reading operations; and for the other operation, the N numbered ring is switched to last position in the cascaded order of rings. The other rings do not change their relative positions in the cascaded order; and hence only the N numbered ring is efiectively switched.

The foregoing and other objects and advantages of the invention will be apparent from the following more particular description of a preferred embodiment of the invention, as illustrated in the accompanying drawings, in which:

FIGURE 1 illustrates in block form a general system of printer butter control using the present invention;

FIGURE 2 shows a detailed embodiment for a ring control and a pulse generator control found in FIGURE 1; and

FIGURES 3A and B illustrate dilterent effective-cascaded arrangements for butler-addressing rings obtainable by the ring control in FIGURE 2.

Considering the present invention in greater detail, reference is made to FIGURE 1 where there is illustrated a nondestructive buffer 14 which receives and temporarily store characters, so that they can be read out as required by mechanical limitations inherent with a printer that is printing the characters.

The printer apparatus and the reading of buffer 14 operate on a subscan basis, which principle of operation is explained thoroughly in prior US. patent application Serial Number 844,511, cited above. Briefly, subscan operation is required when the spacing among type members on a chain, for example, is diiferent from the spacing among print positions. This spacing ratio is sometimes called a pitch ratio; and when the ratio is different from one, only part of the type characters can be aligned with print positions at any one time. Since the buffercharacter positions correspond to print positions, only part of the buffer-character positions correspond to aligned print positions at any one time. Printing occurs at a print position only when an aligned type is equal to a butler character assigned to that print position. A character counter is used to designate the aligned type positions. A compare circuit determines when an equal state exists between an aligned type character and a buffer character assigned to an aligned print position; and the compare circuit causes printing by the type members alignable at any equal position during a subscan sequence. Printing during an equal condition can only occur at the moment of type alignment during a subscan, because the type members are continually moving. Many subscans are necessary to align every type member of an alphabetic array once with every print position. The number of required subscans depends on the spacing ratio between the type members and the print positions. For example, with a spacing ratio of 1.5 to one, a single subscan only finds every third print position aligned. It therefore takes three separate subscans starting with different type members to provide an alignment of one type member with each print position.

For example, assume that there .are 132 print positions in a line of print. Each print position is defined by an independent hammer operabl to strike type members alignable with it. During a first subscan, engraved type characters may be aligned with print positions 1, 4, 7, 10, 13, etc., 130. During the second subscan, the aligned positions are 2., 5, 8, 11, 14, 17, etc., 131. For the third subscan, the engraved characters are alignable with print positions 3, 6, 9, 12, 15, 18, etc., 132 to complete one scan of all print positions. The subscans continue in the same manner until every one of the type members has been aligned with every one of the print positions. Where there are 48 different type characters in each alphabetic array arranged successively along a moving member, it takes 48 print scans (equal to the number of characters in an alphabetic array) to complete a line of print. Since it takes three subs-cans to complete a print scan, it takes 144 subscans to present every character to every print position to complete a line of print.

in summary, the printing control required for a subscan print operation comprises: first means for presenting data from storage to identify characters which can be printed during each subscan, second means for identifying the type members alignable with print positions during the subscan, timing control means operable to synchronize the presentation of the respective stored data with the advance of the identified type members during the subscan, and means for comparing the identified aligned type members and presented stored characters for equality which causes hammer firing at the respective equal positions of the subscan.

In FIGURE 1, the means for identifying characters which can be printed takes the form of a storage device, such as a core storage matrix 14, having a plurality of storage positions corresponding in number at least to the number of print positions along a print line. The core storage matrix 14 comprises a multiple plane three dimensional array of bistable magnetic core elements capable of having their conditions of stability selectively switched in accordance with current applied to X, Y and inhibit windings inductively related to the core elements. Changes in magnetic states of the cores may be detected by sense windings provided on the core elements for that purpose. Means for applying current selectively to the windings takes the form of X drivers 15, Y drivers 16 and inhibit drivers 17 suitably connected to the X, Y and inhibit windings, respectively. The sense windings are connected to sense amplifiers 18 whose outputs are connected to a single-character register device 19. The latter may be any well-known register having a plurality of triggers or latches arranged to store a multiple bit electric signal and to provide the same through leads 20.

The storage device is regenerative, i.e., when the readout condition occurs, the information is put back into storage. For this purpose a feedback regenerative circuit is provided comprising leads 20 connecting the output of register device 19 to a set of AND gates 21, which in turn are connected by lines 22 to OR circuits 23 having output connections 24 to a second AND gate 25 connected by leads 26 to inhibit drivers 17. A second set of inputs 27 to OR circuits 23 ar provided from a tape transport, a computer, or other means for writing data into buffer 14. A sample pulse applied at suitable times to AND gates 21 will pass coded information to inhibit drivers 17 which put the information back into the storage location from which it was originally read. Various regenerative apparatus may be employed by persons skilled in the art for the purpose of the present invention.

Readout from core matrix 14 is effected by selective energization of the various X and Y windings threaded through the cores of the matrix. In accordance with the operation of the invention, the selective readout is performed on a subscan basis. Since as previously explained, the various type members are aligned only at spacedapart print positions, it is advantageous for the purpose of this invention to readout from storage by skipping storage positions in the sequence in which the type members become alignable at the respective separated print positions during each subscan.

The means for obtaining subscan readout operation comprises X ring 28, subscan gating ring GR, and Y ring 29. These rings, sometimes called ring counters or shifting rings, may be of conventional construction as disclosed in Arithmetic Operations of Digital Computers, by R. K. Richards (1955), pages 205208. Each of the rings 28, 29 and GR is a closed multiple stage ring comprised of a plurality of binary elements such as latches or triggers, the number of stages being selected according to the number of rows and columns of cores which are to be switched to effect readout for a complete line of print. In the particular embodiment illustrated herein, the core matrix comprises storage positions addressed by 15 X drivers and 9 Y drivers, respectively. Thus the Y ring 29 is a closed nine stage ring which can comprise nine binary trigger elements driven by pulses on an input lead 54. Trigger output leads 31 connect each stage to a corresponding Y driver. Each binary element forming the stages of Y ring 29 is of any suitable type operated to switch alternately ON and OFF conditions upon being successively pulsed. In addition each binary element is connected to the next adjacent binary element through gates so that the steps of the ring are switched ON and OFF sequentially in a given direction with re spect to the order of drivers.

Since each of rings X, Y and GR is a closed ring, the last stage of any ring is connected to its first stage to enable constant cycling of its outputs.

As already stated, subscan readout is effected in a skipped manner. For this reason, the X ring is a fivestage ring having the output of each stage connected by leads 32 to five groups of three AND switches 33, 33a, 33b, through 37, 37a, 37b. Each AND switch has an output lead connected to a corresponding X driver. The selection of which X drivers are to be energized during a sub-scan is the function of the subscan gating ring GR. Gating ring GR is a three stage closed ring stepped by pulses applied to its input 55. Gate GR has three output leads 40 from its three stages which respectively connect to the three horizontal groupings of five AND switches 33-37, 33a-37a, and 33b-37b.

With this ring arrangement, readout of every storage position is obtained at the end of three successive subscans (or subcycles), the particular subcycle positions;

being read out is controlled by subscan gating ring GR in conjunction with X ring 28 and Y ring 29. At the end of the third subcycle, every position will have been read out once in the following sequence:

Subscan 1, storage positions 1, 4, 7, 10 133 Subscan 2, storage positions 2, 5, 8, 11 134 Subscan 3, storage positions 3, 6, 9, 12 135 The process will be repeated until all 48 type members of an alphabetic array are presented to each print position, which takes 48 scans or 144 subscans.

It will be noted that since only 132 characters are in the print line, three of the buffer character positions are not loaded in the 135 character buffer. The unused positions are assumed to be positions 133, 134, and 135.

It is understood that as each character storage location of butter 14 is read out, the multiple bits of the character are being amplified by sense amplifiers 18 and stored in register device 19. The output from the register device 19 is also connected to a compare circuit (not shown). Thus, core readout is serial by character and parallel by bit.

A ring control RC is provided with outputs 53, 54 and 55 which are inputs to rings X, Y and GR, respectively.

A pulse generator control circuit 210 provides pulses on its output lead 50 in scanning groups of 135 during buffer loading and in subscan groups of clock pulses during the buffer reading operation. The output of circuit 210 is therefore input 50 of ring control RC. For loading the buffer, the addressing rings are synchronized with a clock in the external system supplying characters being loaded in the butter. Terminal 231 connects the external clock to lead 50.

FIGURE 2 illustrates detailed forms of ring control RC and pulse generator control 210. Pulse control 210 comprises a local oscillator 213 which provides a clockpulsed output for reading the buffer. A trigger 212 receives each subscan pulse obtained from a track on a magnetized drum (not shown) found with a chain printer of the type described in cited patent application No. 844,511. A subscan pulse is provided at each instantaneous alignment of type with print positions. The subscan pulses are connected to the SET input of trigger 212. Trigger 212 is RESET by a pulse on lead 282 from an AND gate 281, which provides the pulse at the simultaneous occurrence of the last pulses in cycles of the X ring and Y ring. During printing, gate 281 provides a pulse at the end of each subscan. An AND gate 232 has inputs connected to outputs of oscillator 213, trigger 212, and print signal lead 287. Hence gate 232 passes only that number of oscillator pulses required to generate a subscan, which in this case is 45 pulses, since a subscan is /3 of the buffer capacity of 135 characters. A clock advance signal on lead 287 is provided continuously during the entire period of printing a line. Ring control RC has three AND gates 236, 237 and 238, each having an input receiving pulses from lead 50. The outputs of AND gates 236, 237 and 238 are respectively provided to rings GR, X and Y as inputs on leads 55, 53 and 54.

The other inputs to AND gates 236, 237 and 238 control the reflective-cascaded driving order of rings GR, X and Y. This is accomplished in FIGURE 2 by using a pair of OR gates 242, and 243. OR gate 242 has its output provided as an enabling input to AND gate 236. Gate 242 has two inputs which respectively receive a buffer loading signal on lead 286, or an output from AND gate 281.

An output of OR gate 243 provides inputs to AND gates 237 and 238. OR gate 243 has inputs received from the 3rd output (GR3) of subscan ring GR, or a print signal on lead 287 which is activated only while the characters are being read out of the bufler for printing, and not while a butler loading signal on lead 286 is being provided.

The basic operation of ring control RC is to change the effective-cascaded driving sequence of rings GR, X and Y. They have the effective-cascaded order shown in FIGURE 3A during a buffer loading operation when lead 286 is activated. As shown in FIGURE 3A during a butter loading operation, ring GR receives all clock pulses from lead 50, ring X is stepped only once per cycle of ring GR by its last cycle pulse GR-3; and ring Y is stepped only once per cycle of X ring by its last cycle pulse X-5. This sequencing is obtained by ring control RC when the butter load line 286 is actuated to continuously enable AND gate 236 to pass all clock pulses from lead 50 to the input of ring GR. Then OR gate 243 provides last pulses GR-3 to inputs of AND gates 237 and 238. Therefore gate 237 steps the X ring once per GR cycle. AND gate 238 has a third input which receives the last pulse X-S. Hence, gate 238 is stepped only once per cycle of ring X.

During a print operation, the buffer is being read out; and the rings are reconnected in the effective-cascaded arrangement shown in FIGURE 3B. Now bufi'er load line 286 is deactuated, and print line 287 is activated.

5 Hence the output of OR gate 243 is continuously provided to gates 237 and 238; and gate 237 passes all clock pulses provided from line 50 as an input to the X ring on lead 53. AND gate also receives the last pulse X-5 of each cycle of the X ring; and hence the Y ring is stepped from lead 54 once per cycle of the X ring.

On the other hand, AND gate 281 provides an output pulse only at the simultaneous occurrence of the last outputs X-S and X-9 of the X and Y rings. The outputs of gate 281 is provided through OR gate 242 to AND gate 236, which therefore provides an output pulse on lead 55 to ring GR only once per cycle of the Y ring to obtain the arrangement of FIGURE 38.

The direction of cycling of the rings with respect to X and Y drivers 15 and 16 is a function of the direction of movement of the arrays of type members across the print positions. A ring having reversible operation is disclosed in US. Patent 2,831,150. If an opposite direction of type movement is provided, the cycling direction of each ring can be reversed in many ways, an obvious way being to reverse their connections to drivers X and Y. Then the last pulse for each ring during its reversed cycling is respectively substituted for last pulses X-5 and Y-9 to gate 281, and for last pulse GR-3 to gate 243 during forward cycling. Accordingly, if the 3 type movement is alternately reversed by a reciprocating drive, the ring directions and last pulsed outputs to gates 243 and 281 are switched in synchronism with the type movement reversals.

While the invention has been particularly shown and described with reference to a preferred embodiment thereof, it will be understood by those skilled in the art that the foregoing and other changes in form and details may be made therein without departing from the spirit and scope of the invention.

What is claimed is:

1. Butler addressing means to control buffer loading and reading operations in different predetermined manners, with one of said operations being in a given-order sequence, and the other operation involving sequentially skipping N number of buffer positions in relation to said given-order sequence, said buffer addressing means comprising, a plurality of closed rings effectively connected in cascaded order to a clock source, at least one of said rings having a number of stages equal to said skipping number N, said rings being connected in one order for controlling loading of said butter, and gate means for changing the cascaded order of said rings for reading said bufler.

o 2. Means as defined in claim 1 in which said rings control buffer operation for character printing by moving type, means cycling each of said rings in a specified direction for a given direction of movement of said type along a print line, and means for reversing said cycling directions upon reversal of movement by said -type.

3. Means as defined in claim 1 in which said N numbered ring is connected first in said cascaded order during one of said operations, and said N numbered ring is connected last in said cascaded order during the other of said operations.

4. Means as defined in claim 3 in which said N numbered ring is connected last in said cascaded order during the reading of said buffer.

5. Means as defined in claim 3 in which said plurality comprises first, second and third closed rings, with their effective-cascaded order being controlled by gating means, comprising first, second and third AND gates having outputs respectively providing inputs to said rings,

7 the first ring being said N numbered ring, means enabling said first AND gate either by a first signal continuously determining one of said bufier operations or a second signal occurring upon simultaneous outputs from the second and third rings, means enabling said second AND gate by either a third continuous signal indicating the occurrence of the other of said buffer operations or a fourth signal occurring at a predetermined output from said N numbered ring, and means enabling said third AND gate by said third or fourth signal, and means en- 8 abling said third AND gate with a fifth signal that is a predetermined output from said second ring.

6. Means as defined in claim 5 in which said first signal is provided during a buffer loading operation, and said third signal is provided during a buffer unloading operation to designate characters to be printed.

No references cited.

MALCOLM A. MORRISON, Primary Examiner.

Claims (1)

1. BUFFER ADDRESSING MEANS TO CONTROL BUFFER LOADING AND READING OPERATIONS IN DIFFERENT PREDETERMINED MANNERS, WITH ONE OF SAID OPERATIONS BEING IN A GIVEN-ORDER SEQUENCE, AND THE OTHER OPERATION INVOLVING SEQUENTIALLY SKIPPING N NUMBER OF BUFFER POSITIONS IN RELATION TO SAID GIVEN-ORDER SEQUENCE, SAID BUFFER ADDRESSING MEANS COMPRISING, A PLURALITY OF CLOSED RINGS EFFEXTIVELY CONNECTED IN CASCADED ORDER TO A CLOCK SOURCE, AT LEAST ONE OF SAID RINGS HAVING A NUMBER OF STAGES EQUAL TO SAID SKIPPING NUMBER N, SAID RINGS BEING CONNECTED IN ONE ORDER FOR CONTROLLING LOADING OF SAID BUFFER, AND GATE MEANS FOR CHANGING THE CASCADED ORDER OF SAID RINGS FOR READING SAID BUFFER.

Images