U.S. patent number 3,913,719 [Application Number 05/552,283] was granted by the patent office on 1975-10-21 for alternate memory control for dot matrix late news device.
This patent grant is currently assigned to The Mead Corporation. Invention is credited to James A. Frey.
United States Patent |
3,913,719 |
Frey |
October 21, 1975 |
Alternate memory control for dot matrix late news device
Abstract
A printing system combines the economical features of a long run
conventional printing press with the versatility of a programmable
digital printer. The conventional press prints a first field of
information, which requires no changing throughout a relatively
long run, and the digital printer prints a second field of
information which requires relatively frequent modification. A
disclosed example is a newspaper wherein a conventional press
prints a page having a blank column, and this column is imaged with
late news or special interest material by coordinated control of a
dot matrix printer, which is preferably an ink jet array printer.
Graphics and text material for printing by the ink jet array
printer are stored in a memory and read out to the charge rings of
the ink jet printer in synchronism with the movement of the paper
and at the proper time for printing the column which has been left
blank by the conventional press rolls. In order to enable changes
in printed copy without shutting down the press there are provided
two separate memories for controlling the ink jet printer. Each
memory is loaded with a code which is read out at the end of the
data intended for use by the digital printer, and one memory is
loaded with new information while the other memory is in operation.
A change in printed copy may then be effected by merely throwing a
switch which shifts from one memory to another, but the actual
shift does not occur until after the operating memory has read out
the above mentioned code. Thus there is prevented production of any
newspaper pages having split copy.
Inventors: |
Frey; James A. (Kettering,
OH) |
Assignee: |
The Mead Corporation (Dayton,
OH)
|
Family
ID: |
27030402 |
Appl.
No.: |
05/552,283 |
Filed: |
February 24, 1975 |
Related U.S. Patent Documents
|
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
Issue Date |
|
|
435030 |
Jan 21, 1974 |
|
|
|
|
Current U.S.
Class: |
358/1.16;
101/374; 400/29; 346/2; 347/4; 400/61; 101/91; 101/490;
358/1.4 |
Current CPC
Class: |
B41J
5/44 (20130101); B41F 13/46 (20130101); B41J
2/505 (20130101) |
Current International
Class: |
B41J
5/44 (20060101); B41J 2/505 (20060101); B41F
13/00 (20060101); B41F 13/46 (20060101); B41J
005/30 (); G01D 015/18 () |
Field of
Search: |
;197/1R,19 ;346/140,75,1
;101/374,91,DIG.13,426 ;354/5,7 |
References Cited
[Referenced By]
U.S. Patent Documents
Primary Examiner: Burr; Edgar S.
Assistant Examiner: Suter; R. E.
Attorney, Agent or Firm: Biebel, French & Bugg
Parent Case Text
CROSS REFERENCE TO RELATED APPLICATION
This application is a continuation of Ser. No. 435,030, filed Jan.
21, 1974, now abandoned.
Claims
What is claimed is:
1. The method of continually producing and revising printed
material comprising the steps of:
a. printing a fixed field of information on a print receiving
member with a printing press,
b. transporting said print receiving member past a dot matrix
printing device positioned for printing a revisable field of
information upon said print receiving member in an area reserved
for copy subject to revision,
c. storing in memory a first set of digital data corresponding to a
first version of said revisable information,
d. gating said first set of digital data out of memory in
synchronism with the movement of said print receiving member past
said dot matrix printing device to control said dot matrix printing
device and cause printing of said first version of said revisable
field of information in said reserved area,
e. repeating aforesaid printing transfer and information gating
steps for production of a multiplicity of printed copies of said
first version of said revisable field of information in
registration with said fixed field of information,
f. during production of aforesaid multiplicity of copies preparing
a second set of digital data corresponding to a revised version of
said revisable field of information,
g. loading said second set of digital data into said memory,
h. discontinuing control of said dot matrix printing device by said
first set of digital data and continuing aforesaid transfer for
production of a further multiplicity of copies of said fixed field
of information,
i. gating said second set of digital data repeatedly out of memory
to control said dot matrix printing device and cause printing of a
multiplicity of copies of said revised version of said revisable
field of information in registration with said further multiplicity
of copies of said fixed field of information, and
j. timing the discontinuance of dot matrix printing control by said
first set of digital data and commencement of dot matrix printing
control by said second set of digital data such that the first copy
of said revised version of said revisable field of information is
printed in registration with the first copy of said fixed field of
information following the copy thereof which is printed in
registration with the last copy of said first version of said
revisable field of information.
2. A method according to claim 1 wherein said print receiving
member is a continuous web.
3. A method according to claim 1 wherein said first set of digital
data is read into a first memory area and said second set of
digital data is read into a second memory area, the loading of said
second memory area being done simultaneously with unloading of said
first set of digital data from said first memory area.
4. The method of continually producing and revising printed
material comprising the steps of:
a. transporting a continuous print receiving web at constant speed
through a zone in printing relation with a pair of printing
rollers,
b. transferring an image from said printing rollers to said web on
a repeated basis during passage of the web through said zone and
thereby printing a multiplicity of copies of a fixed field of
information upon said web,
c. storing in memory a first set of digital data corresponding to a
first version of a revisable field of information,
d. gating said first set of digital data repeatedly out of memory
in synchronism with the movement of said web,
e. using said first set of digital data as it is gated out of
memory, to print multiple copies of said revisable field of
information in dot matrix form on said web in registration with
associated copies of said fixed field of information,
f. during aforesaid gating out from memory of said first set of
digital data preparing a second set of digital data corresponding
to a revised version of said revisable field of information,
g. loading said second set of digital data into said memory while
continuing to transport said web through said printing zone for
printing of said fixed field of information,
h. commencing repeated gating of said second set of digital data
out of memory in synchronism with said web movement,
i. discontinuing aforesaid printing use of said first set of
digital data,
j. using said second set of digital data as it is gated out of
memory to print multiple copies of said revised version of said
revisable field of information in dot matrix form in registration
with associated copies of said fixed field of information, and
k. timing the discontinuance of printing use of said first set of
digital data and commencement of printing use of said second set of
digital data such that the first copy of said revised version of
said revisable field of information is printed in registration with
the first copy of said fixed field of information following the
copy thereof which is printed in registration with the last copy of
said first version of said revisable field of information.
5. A method according to claim 4 wherein said first and second sets
of digital data are loaded into separate memory sections with one
memory section being unloaded while the other memory section is
being loaded.
6. A method according to claim 5 including the steps of
incorporating into said first set of digital data an "end of text"
code and using said code for controlling the timing of aforesaid
timing step.
7. A method according to claim 5 wherein said memory sections are
loaded on an alternate basis with sets of digital data
corresponding with further revised versions of said revisable field
of information to cause printing of multiple copies of each of said
revised versions in dot matrix form, all of such copies being
printed in registration with associated copies of said fixed field
of information.
Description
BACKGROUND OF THE INVENTION
This invention relates to printing systems for application to
printing jobs wherein a high degree of flexibility is required. An
example is the newspaper publishing industry which has been under
ever-increasing pressure to meet the diversified reading interests
of an increasingly sophisticated reading public. In today's world
the average reader is continually exposed to communication media
which bring him vast amounts of information of both general and
special interest. The sources of this information include
television, radio, telephone, weekly news magazines, newspapers and
professional journals.
Metropolitan newspapers have attempted to be all things to all
people by carrying special interest columns as well as a wide
variety of advertising and local and world wide news. However, the
news is of little interest to most readers unless it includes
accounts of the latest events being carried on radio and
television. Consequently the larger newspapers have been forced
into a multiple edition mode of operation wherein five or six
different editions may be printed each day for distribution in the
central city. Each of these editions is similar in general content,
but each is updated to cover news not available at the time when
earlier editions went to press.
In addition to the numerous editions required for up-to-the-minute
coverage of the news, many large metropolitan newspapers publish
several different editions for suburban sale. Again each of these
editions is similar in general content, but one or more pages or
page portions are made up to carry news articles and advertising of
special interest in the suburb wherein distribution is made. These
suburban editions have been necessitated by competition from small
suburban newspapers which are well situated to satisfy the suburban
appetite for news of local interest.
In spite of the above mentioned multiple editions large newspapers
have been having difficulty meeting competition from other media,
and many have failed. An important contributing factor to such
failure is the great expense of preparing a newspaper edition which
in turn prevents profitable operation unless a great number of
copies of each edition are sold. It is readily apparent, however,
that editions which are to be up to the minute or which are highly
specialized in their content cannot be printed in large numbers.
Accordingly newspaper failures have continued, and a great deal of
public interest has focussed on increasingly specialized news media
such as for example individually ordered facsimile newspapers of
the type as described generally in Regunberg U.S. Pat. No.
3,479,451.
In order to decrease in some measure the great cost of preparing a
newspaper, a number of improvements have been proposed. For
instance it has been proposed to place the composition of newspaper
pages under the control of a computer. Typical examples of such
computer controlled composition systems are shown in Hadley U.S.
Pat. No. 3,593,305 and in Kolb U.S. Pat. No. 3,626,824. In addition
to such improvements in the composition process, there have been
many improvements in type setting, including phototypesetting
wherein type is set on photographic film under the control of a
computer and without any actual handling of type. The final result
of such improvements may be a series of photograhic negatives each
corresponding to a complete newspaper page. Such negatives are used
for making printing plates which may be of several types depending
upon the type of printing press employed. In any event a series of
new plates must be made for each new edition of the newspaper, and
printing of each new edition must be preceded by stopping of the
presses, changing of the plates, accomplishing the necessary press
makeready, and thereafter starting up the presses again. It will be
appreciated that this is a very tedious and expensive process and
one which greatly limits the number of special editions which
economically may be printed.
SUMMARY OF THE INVENTION
This invention enables improved printing of fixed and variable
fields of information as for publication of multiple editions of
newspapers and the like. In accordance with the invention the
printing is accomplished without press shut-down by providing dot
matrix printing apparatus which may be reprogrammed on line and
which prints within a column left blank by a cooperatively printing
conventional press.
The dot matrix printer is preferably an ink jet array printer, and
this printer may be controlled by one or the other of two memories.
For printing of multiple edition newspapers each memory may be
loaded with both graphics and text information, and either memory
may control the ink jet printer while the other memory is being
loaded with new information. The graphics information is loaded in
the memories in dot matrix form, but text information is stored in
the form of character codes. At the end of the character codes
there is a special code termed an "end of text" code which is used
in connection with memory selection. There is a switch which is
used for switching from one memory to the other, and this switch is
enabled with each occurrence of the end of text code. This prevents
the dot matrix printer from changing copy in the middle of a
newspaper page.
In order to synchronize the operation of the dot matrix printer
with the movement of the paper being printed, there is provided a
tachometer which produces pulses at a rate which depends upon the
speed of paper movement. There is an output register which holds
printing information for one row of matrix cells extending across
the column being printed by the dot matrix printer, and this
register is loaded with new printing control information each time
the tachometer generates a new pulse. By this means the printed dot
matrix information is maintained in mutual register, and the
vertical extent of the dot matrix column is invariant with changes
in paper speed. In order to cause registration of the dot matrix
column within the space left blank by the conventional press, the
conventional press prints a special "begin print" code which is
read by a sensor. The sensor is placed to observe the paper being
supplied to the dot matrix printer, and when the sensor sees the
begin print code, then it activates a timer which, after an
appropriate time delay, causes dot matrix printing to begin.
It is therefore seen that it is an object of this invention to
provide a printing press which can make long continuous runs, but
which may change portions of the printed copy during the middle of
a run.
It is another object of this invention to combine an ink jet
printer with a conventional printing press so as to provide a
printing system having the combined capabilities of both
devices.
Still another object of this invention is to provide a system for
cost effective printing of multiple editions of a newspaper.
Other objects and advantages of the invention will be apparent from
the following description, the accompanying drawings and the
appended claims.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is an illustration of the front page of a newspaper
comprising conventional printed columns and a special column
printed in accordance with this invention;
FIG. 2 is a pictorial illustration of a printing system employing
this invention;
FIG. 3 is a schematic cross sectional drawing of an ink jet print
head;
FIGS. 4A and 4B are a schematic illustration of logic for
controlling an ink jet printer in accordance with this
invention;
FIG. 5 illustrates a diode matrix;
FIG. 6 is an enlarged illustration of a portion of a newspaper
column printed with characters on a dot matrix basis.
DESCRIPTION OF THE PREFERRED EMBODIMENTS
As an example of copy which may be printed in accordance with this
invention, there is illustrated in FIG. 1 a specially printed
newspaper 10. The front page of this newspaper comprises a first
field of information which may be the usual headline together with
columns 11 which are printed by a conventional press and a second
field of information comprising two areas 12 and 13 which are
printed by a dot matrix printer. Area 12 includes pictorial
information, while area 13 may comprise text. Area 13 may be
positioned below area 12 and together the two areas may extend the
full length of a page or any part thereof. The two areas preferably
have the same width, and this width depends upon the size and
number of the characters to be printed therein.
FIG. 2 presents a schematic diagram of a complete printing system
in accordance with this invention. As illustrated therein a data
input unit 14 supplies coded printing data to a printing control
unit 18. Input unit 14 preferably includes a monitor 15, a keyboard
16 and a scanning unit 17. Printing control unit 18 preferably
comprises operator controls, a memory, character generating
circuits, printing control logic, and input and output buffers.
Preferably the storage capacity of the memory within control unit
18 is twice the size required for the copy being printed, so that
the memory may be loaded with revised copy without disturbing
system operation. Output signals from control unit 18 are fed to a
dot matrix printer 19 which is preferably an ink jet print head,
and the ink jet print head is connected to an ink supply 20 and a
power supply 22. Power supply 22 provides a DC voltage for
generation of drop deflecting electrostatic fields and a regulated
AC voltage for drop stimulation all as described below.
A web 25 is carried between a pair of print rolls 24, over a pair
of guide rolls 26, and thence toward other apparatus which cuts and
folds the printed web in the usual manner for assembly of
newspaper. For printing of 60,000 newspapers per hour, the speed of
web may be about 2,000 ft. per. min. During its passage between the
two rolls 26, web 25 passes under an optical sensor 23, thence
under the ink jet print head 19, and thereafter under a dryer 21.
Print rolls 24 may be letterpress or other conventional rolls, and
they print the major portion of the newspaper while leaving a blank
nonprinted area 27 for printing by the ink jet print head 19. The
top rail 24 prints a "begin print" code at the beginning of each
new section 27, and this code is sensed by the optical sensor
23.
Optical sensor 23 may be connected by activate a timer which in
turn initiates ink jet printing at the proper time. For this
purpose the begin print code may be any convenient code which may
be read by conventional automatic readers. For instance the code
may be a width coded bar code, and the sensor may be a device of
the general type described in Eckert et al. U.S. Pat. No.
3,716,699. Alternatively the timer may be triggered by a digital
encoder attached to one of the print rolls 24.
Data input unit 14 generates both pictorial and text information
for storage in the system memory, and in general the two different
types of information are differently formatted. Thus pictorial
information may be presented on a film transparency or other
graphic form for scanning by scanning unit 17, while text
information is entered by keyboard 16. The output from the scanning
unit is a digital representation of black and white spots as seen
by a scanning photosensor, and this digital information is stsored
within the system memory in binary form; that is, each observed
white resolution cell is represented by a binary "zero" and each
observed black resolution cell is represented by a binary "one".
Accordingly a typical pictorial area comprising 10 mil resolution
cells and having length and width dimensions of 4.80 inches is
represented by the binary states of 230,400 storage locations
within the system memory. Obviously the text material may be
treated as pictorial information and scanned into memory also. For
16 inches of pictorially generated text material 768,000 storage
locations would be required, so that something in the neighborhood
of 1,000,000 bits of storage would be required for scanner
controlled printing of both of areas 12 and 13.
Preferably data unit 14 scans into memory only the pictorial
information for printed area 12 and employs keyboard 16 for
controlling the storage of text for printed area 13. This reduces
memory capacity requirements considerably, because a character
matrix which may typically comprise 16 rows of 12 cells each may be
represented by a code which conveniently may be a single 8 bit
word. Thus the above mentioned 16 inches of text (100 lines of 40
characters each) may be represented by 4,000 character codes or a
total of 32,000 information bits. These information bits are loaded
into the system memory after the bits representing the pictorial
dots and are read out serially as hereinafter described. Ideally
the system memory is connected for controlling the display 15 as
well as printing unit 19, so that an operator can check the copy to
be printed.
As described above, data unit 14 is a rather conventional apparatus
which is connected to load either of two memories; each memory
having a graphics section and a character code section. There is a
special requirement, however, in that keyboard 16 requires a
special key to cause generation of an end of text code. Keyboard 16
will also have the usual keys for various letters, numbers, and
symbols as well as a spacing key. The spacing key (not illustrated)
causes storage of a code which is interpreted by the printer as a
blank space. Usage of this key is important, because each line must
be completely filled with letters or blank spaces. Depending upon
the desires of the operator, extra blank spaces may all appear at
the end of each line, or may be interposed between words within the
lines to produce line justification. Alternatively, line
justification as well as proportional character spacing may be
performed automatically by using any of many known phototypesetting
systems in place of data unit 14, but in such a case the output
from the phototypesetter must be stored in graphic form rather than
as character codes. In still another embodiment the character
information may be stored in the form of character codes with
proportional spacing being achieved upon readout in accordance with
the teachings of Taylor U.S. Pat. No. 3,174,427.
Print head 19 may be constructed generally as described in Mathis
U.S. Pat. No. 3,701,998 and may appear as shown in the simplified
cross section of FIG. 3. Thus there is an ink supply 30 maintained
within a manifold 29 by an orifice plate 31. Orifice plate 31 is
provided with a series of orifices 32 which may be arranged in two
rows. For printing an area such as area 13 of FIG. 1, which
comprises lines of 40 character matrices each 12 cells wide,
orifice plate 31 will have a total of 480 orifices or two rows of
240 orifices each. The two rows of orifices are staggered for full
printing coverage as described in detail in Taylor et al. U.S. Pat.
No. 3,560,641. Typically the distance between rows may be about 60
mils and the distance between orifices within each row may be about
20 mils. The orifice diameter may be about 1.8 mils.
Ink is supplied to manifold 29 under pressure so that 480 streams
of ink flow continuously through orifices 32. A stimulator (not
shown) vibrates orifice plate 31 to cause each of the streams of
ink to break up into uniformly sized and regularly spaced drops.
These drops pass through a series of charge rings 34 in a charge
ring plate 33 enroute to the web 25. For operation at a press speed
of about 2,000 ft. per. min, the vibration frequency of the
stimulator may be about 120 kHz.
Charge ring plate 33 is positioned relative to orifice plate 31
such that break up of the ink streams into drops occurs within the
charge rings 34. Charge ring plate 33 is made of insulative
material and charge rings 34 are coated with a conductive material.
Each of the charge rings is connected to an electrical lead 73
(FIG. 4B) which in turn is connected to receive output signals from
the system memory. Electrical signals are placed on leads 73 for
selective charging of the drops 38.
There is a pair of catchers 36 which are made of electrically
conductive material, and a conductive deflection strip 35 is
stretched therebetween. Catchers 36 and deflection strip 35 are
connected to sources of electrical potential to set up electrical
deflection fields between the deflection ribbon and each of the
catchers. All of drops 38 which carry an electrical charge are
deflected by these fields and caught by one or the other of
catchers 36. Drops so caught are injested into one of chambers 37
for evacuation by a vacuum source (not shown). Drops which are not
charged pass undeflected through the deflection fields and deposit
on web 25.
FIG. 6 illustrates the character matrices which are printed within
the printed area 13. Each matrix 39 comprises a series of cells 40
which may be 10 mils square. Each matrix 39 has 16 rows of
character cells, and there are 12 cells in each row for a total of
192 cells per matrix. There are 40 such matrices across the width
of section 13, so that as each of the rows of cells passes under
print head 19 there are presented 480 cells for printing (or
non-printing) by drops from the 480 ink jets. FIG. 6 illustrates
how various letters may be printed by depositing drops in the
various cells. The dots 76 each may represent the printed result of
about 3 deposited drops and may have diameters of about 14 mils so
that overlapping coverage may be achieved. The printed characters
occupy only a portion of the area within each of matrices 39 so
that spacing may be provided between characters and between rows of
characters. Thus the characters are limited to a 9 .times. 12
submatrix within the overall 12 .times. 16 matrix.
FIGS. 4A and 4B present a simplified schematic diagram of the
control logic for printer 19. Within the control system as
illustrated in FIG. 4A there are two memory sections 41a and 41b,
each of which stores sufficient data for controlling the printing
of a complete news column comprising pictorial information and text
information such as the information contained within areas 12 and
13 of FIG. 1. Memories 41a and 41b are divided into graphics
sections 42a and 42b and character code sections 43a and 43b. It
will be understood that sections 42a and 42b additionally comprise
input and output buffers as well as logic to initiate output from
character code sections 43a and 43b after the graphics sections 42a
and 42b have been unloaded.
Initiation of output from memories 41a and 41b is under the control
of a print control flip-flop 55 which is set by an output pulse
from a timer 44. Timer 44 is connected to receive an output from
optical sensor 23 whenever the sensor recognizes the above
mentioned begin print code and this initiates a timing cycle which
will result in an appropriately timed pulse to begin unloading of
memory sections 42a and 42b. The length of the timing cycle will
depend upon the movement speed of web 25 and the separation
distance between sensor 23 and print head 19. For a web speed of
2,000 ft. per. min. and a separation distance of about 6 inches,
the required delay time is about 15 milliseconds. While a number of
suitable timers are commercially available, it is desirable to
provide a variable time delay which is automatically adjusted in
accordance with changes in web velocity. Thus there is provided a
tachometer 51 which generates pulses in synchronism with the
rotation of one of the guide rolls 26. Typically tachometer 51 will
generate about 100 pulses for each inch of web movement, so that at
a web speed of 2,000 ft. per. min., tachometer 51 will generate
pulses at a rate of 40 kHz. These pulses from tachometer 51 are
applied to an AND gate 52 and to timer 44 which preferably
comprises a digital counter. For an arrangement as described above,
timer 44 will count 600 pulses from tachometer 51 after activation
by sensor 23. Upon reaching the 600 count, timer 44 sets flip-flop
55 which in turn enables AND gate 52. Thereafter pulses from
tachometer 55 are supplied via AND gate 52 to AND gates 45a and 45b
for selective application to the two memories 41a and 41b.
Flip-flop 55 is reset by a pulse on line 96 from an AND gate 50
when the operative memory 41a has been completely unloaded.
Since there are provided two separate memories for operation of
print head 19, there is a switch 46 which may be positioned by an
operator to control memory selection. Switch 46 is connected as
illustrated to the set and reset terminals of a flip-flop 48. When
switch 46 is in position B then flip-flop 48 enables an AND gate
45b which is connected to the unload control terminal of memory
41b. When AND gate 45b is enabled by flip-flop 48 then pulses from
tachometer 51 are supplied to the graphics section 42b of memory
41b, and graphics information is unloaded therefrom in 12-bit
bytes. It will be appreciated, however, that a memory which handles
data in 12-bit bytes must be custom built. If it is desired to
employ a memory comprising commercially available building blocks,
then the graphics data may be handled in 16-bit bytes, with the
last 4 bits of each byte being blank. The 12-bit bytes of unloaded
graphics information then would represent the first 12 bits of 16
bit words, while the 4 bits of blank data would represent wasted
storage space. Thus for such an alternate arrangement each of
graphics sections 42a and 42b would necessarily have to be
increased in size by 33 percent.
Continuing with the unloading of graphics section 42b, each new
pulse from tachometer 51 causes unloading of 40 bytes of
information, and these bytes are unloaded at a frequency which
depends upon the frequency of an internal clock (not illustrated)
within the memory. The clock frequency for the example described
herein must be at least 1.64 mHz, and preferably the clock operates
at a frequency of 2 or 3 times that rate. Pulses from the clock
also appear on a line 104 for purposes hereinafter discussed.
When switch 46 is placed in position A, then flip-flop 48 disables
AND gate 45b and enables another AND gate 45a. This then causes
pulses from tachometer 51 to be supplied via AND gates 52 and 45a
to graphic section 42a of memory 41a. Under these circumstances
graphic section 42b produces no output, and graphic section 42a
produces 40 12-bit bytes of output data for each pulse from
tachometer 51.
Memory section 42a has 12 output lines 83a, and memory section 42b
has 12 output lines 83b. Output lines 83a and 83b are connected to
a set of twelve OR gates 82 as illustrated in FIG. 4a, and the
output terminals of OR gates 82 are connected to a series of lines
56. Lines 83a and 83b carry zero or LO signals except when data is
being unloaded from their associated memory section 42a or 42b.
Therefore when flip-flop 48 is set to select one of graphics memory
sections 42a or 42b, the information gated out therefrom appears on
lines 56 for printing head control as hereinafter discussed.
Flip-flop 48 is connected to AND gate 50 via a line 49. AND gate 50
is activated whenever an end of text pulse is read out of one of
memory sections 43a or 43b, and this activation of AND gate 50
provides a clocking signal for flip-flop 48. Setting and resetting
of flip-flop 48 therefore does not occur immediately upon throwing
of switch 46 but awaits the end of a printing cycle. Accordingly,
lines 56 continue to receive information from the same memory, and
the printer is prevented from printing a newspaper page having
split copy.
Memory sections 42a and 42b contain counters which keep track of
the amount of data unloaded therefrom. When the operative one of
memory sections 42a 42b has been unloaded as above described, then
an output pulse from the unload counter (which may be connected to
one of lines 87a or 87b) initiates unloading from the associated
character code memory 43a or 43b. Thereafter the graphics memory
42a or 42b which had previously been operative provides a LO or
zero output to OR gates 82.
Once one of character code memories 43a or 43b has been activated
it begins reading out character codes in serial form on one of
lines 108a or 108b. The bits comprising these codes are read out in
strings of 320 bits representing codes for 40 characters. Reading
out of these bit strings is initiated by pulses on line 99 from a
counter 80, with a new string being read out for each counter
output pulse. Again the memory read out occurs at the basic memory
clock frequency which may be at the rate of about 5 mHz.
The bits which are read out onto lines 108a and 108b are applied
respectively to AND gates 58a and 58b which are selectively enabled
by flip-flop 48 in accordance with the position of switch 46. The
outputs from AND gates 48a and 48b then are applied to an OR gate
59 which is connected to load a 320 bit register 60 via line 97.
Register 60 therefore is loaded with 320 new data bits each time an
output pulse is generated by counter 80, and the bits which are so
loaded are gated out from that one of memory sections 43a or 43b
which has been selected by the positioning of switch 46.
Register 60 is connected to supply character data in 8-bit bytes to
a series of lines 62, which are connected to AND gate 50 and also
to a series of other AND gates 63. Unloading of information from
register 60 is under the control of memory clock pulses gated out
of AND gate 81 onto line 98. These clock pulses originate within
control unit 18 and may occur at a frequency of about 5 mHz as
mentioned above. The clock pulses are applied via line 104 to an
AND gate 94, which is enabled by a flip-flop 78, and from AND gate
94 the clock pulses are supplied to a counter 79 and to AND gate 81
as well as along a line 95 to the load control terminal of printing
control register 67. AND gate 81 is enabled by flip-flop 91 as
hereinafter described.
As mentioned above a selected one of memory sections 43a or 43b
causes production of 320 character code pulses on line 97 each time
a pulse appears on line 99. These 320 data bits are loaded into
register 60 under the gating control of pulses on line 105 from AND
gate 89. AND gate 89 is enabled by a flip-flop 90, and upon being
enabled, AND gate 89 transmits clock pulses via line 105 to the
load control terminal of register 60. A counter 88 counts the clock
pulses gated out of AND gate 89, and when the count reaches 320,
counter 88 resets flip-flop 90. Resetting of flip-flop 90 prevents
transmission of any more clock pulses to the load control terminal
of register 60 until after flip-flop 90 has again been set. As
shown by FIG. 4A, flip-flop 90 is set by an output pulse from
counter 80. The pulse which sets flip-flop 90 is the same pulse
which initiates unloading of the 320 bits from memory sections 43a
and 43b. This same pulse is also applied via line 84 to the store
control terminal of register 60 to cause the 320 bits of data
(representing character codes for 40 characters) to be shifted into
storage for subsequent output reading. It will be seen that this
output reading goes on simultaneously with loading of the next 320
bits into register 60.
After 320 bits of character code information have been loaded into
register 60 and have been stored, then output reading begins under
the control of clock pulses appearing on line 98. For each pulse on
line 98 an 8-bit byte of data is unloaded or read onto lines 62.
Forty pulses appear on line 98 to cause forty bytes of data to be
unloaded. The pulses which are provided by AND gate 81 to line 98
are counted by counter 79. When forty such pulses have been
counted, then counter 79 resets flip-flop 78, thereby disabling AND
gate 94 and preventing any further data unloading from register 60.
However, the set terminal of flip-flop 78 is connected to receive
tachometer pulses along line 109 from AND gate 52. Therefore, upon
occurrence of the next tachometer pulse, AND gate 94 is again
enabled and forty more clock pulses are applied to the unload
terminal of register 60. This causes register 60 to read out onto
lines 62 the same forty data bytes which had been read out during
the previous unloading cycle.
Each time counter 79 resets flip-flop 78 a count is added into
counter 80. Counter 80 is a row counter which counts the cell rows
being printed within the character matrices 39. When counter 80
reaches a count of sixteen, it initiates the loading of a new
string of character codes into register 60 and also causes shifting
into memory of the character codes loaded into the register during
the previous load cycle. This is all under the control of
tachometer pulses gated out from AND gate 52, so that every 16th
tachometer pulse causes 320 bits of character code data to be
loaded serially into register 60. Aftere sixteen more tachometer
pulses, these 320 bits of character code information are
transferred into storage within register 60. Once the data bits are
in storage they are unloaded in 8-bit bytes in cyclical fashion,
with each new tachometer pulse causing unloading of all of the 40
bytes in storage. All bytes of data stored within register 60 are
unloaded 16 times, before an output pulse from counter 80 causes
320 new data bits to be loaded into storage.
In order to limit cycling of register 60 to that period of time
when print head 19 is printing a text section of newspaper 10,
flip-flop 91 has its set terminal connected for activation by an
output from an OR gate 92. OR gate 92 is connected to lines 87a and
87b, which as above mentioned carry an activating pulse upon
completion of read out from one of the graphic memory sections 42a
or 42b. Thus when graphics read out has been completed, flip-flop
91 is set thereby enabling AND gate 81 to supply unload pulses on
line 98 to register 60. When printing of text information has been
completed, and AND gate 50 supplies the above mentioned pulse for
resetting of flip-flop 55, this same pulse resets flip-flop 91.
The character codes unloaded from shift register 60 are applied to
AND gates 63 and also to AND gate 50. AND gate 50 decodes the end
of text code as discussed above, and AND gates 63 decode the
character codes. There are as many of AND gates 63 as there are
different character codes. When using an 8-bit character code as
illustrated, there may be as many as 256 different codes. If one of
these codes is an end of text code then there may be as many as 255
character codes and 255 AND gates 63.
Each AND gate 63 is connected to enable a set of 12 associated AND
gates 64. There are as many sets of AND gates 64 as there are AND
gates 63, and each set of AND gates 64 is connected to 12 output
lines 65 from an associated character matrix 66. Each set of AND
gates 64 is connected as illustrated in FIG. 4B, for loading into
printing control register 67. Printing control register 67 has a
set of twelve input lines 110 for loading of printing information
in 12-bit bytes, and these bytes may come either via lines 56 from
OR gates 82 (pictorial information) or from selected ones of
character matrices 66 (character information). Selective enabling
of the sets of AND gates 64 by decoded outputs from AND gates 63
determines which of character matrices 66 will load register
67.
Each character matrix 66 has 16 input lines 68, and these lines 68
are activated in sequence by row select shift register 69. Shifting
of row select shift register 69 occurs at the tachometer pulse rate
under the control of output pulses from counter 79. These pulses
are applied via line 106 to AND gate 100. The set output from
flip-flop 91 which enables AND gate 81 also follows line 102 to
enable AND gate 100.
Printing control register 67 has 480 storage locations at its input
side and 480 storage locations as its output side, so that loading
and unloading can continue simultaneously. The register is so
constructed that the input locations are loaded in 12-bit bytes,
and the output locations are loaded by parallel shifting of the
data in all 480 input locations. The contents of the output
locations are read continuously so long as a HI signal appears at
the UNLOAD terminal.
The LOAD terminal of printing control register 67 is connected via
line 95 to the output terminal of AND gate 94, so that register 67
receives 40 load control clock pulses for each tachometer pulse
gated out of AND gate 52. This causes loading of 40 bytes of
information from OR gates 82 during graphics printing or 40 bytes
of information from AND gates 64 during text printing. Accordingly
during printing of text information, register 67 receives character
matrix bits for a corresponding matrix row of each of 40 different
characters before row select shift register 69 causes loading into
register 67 of character matrix bits for the next row of the same
40 characters. After 40 12-bit bytes have been loaded into register
67, an output pulse from counter 79 appears on line 106 which is
connected to the STORE terminal of the register. This then causes
parallel shifting of the forty bytes into the output side of the
register for parallel reading onto lines 71. Parallel shifting
within register 67 is carried out at the tachometer pulse rate, so
that registers 67 and 69 are shifted in synchronism with the
loading, storing, and unloading of register 60.
Now referring to FIG. 5 it will be seen that diode matrices 66 each
comprise a series of diode connections 74 between input lines 68
and output lines 65. Such diode matrices are commonly used for
character generation and need not be described in detail herein.
Such matrices may be procured with the characters being "hard
wired" therein or with inactive connections which may be
selectively activated as desired. It readily may be seen that
activation of any one of lines 68 by row select shift register 69
causes each of diode matrices 66 to generate a 12 bit code
corresponding to the information within one row of a corresponding
character matrix. All of diode matrices 66 are tied together, so
that they generate simultaneous output codes all corresponding to
the same row of their associated character patterns. Thus when the
first of lines 68 is activated all of diode matrices 66 generate
output codes corresponding to the first row of their associated
character patterns, and the output from register 60 determines
which one of those 12 bit codes will be loaded into register 67 at
any point in time. Forty such 12 bit codes are loaded into register
67 for printing control, and thereafter row select shift register
69 activates the next of its output lines 68. Register 60 then
cycles through its forty stored character codes a second time to
cause storage in register 67 of 40 12-bit codes corresponding to
the second row in each of the above mentioned 40 characters. This
process repeats until all 16 of lines 68 have been activated in
sequence and the output register of shift register 60 has been
unloaded 16 times. By this time the input register of shift
register 60 has been loaded with another 320 bits of information,
and the system is ready to begin printing a new line of forty
characters.
Referring again to FIG. 4B it will be seen that output lines 71
from printing control register 67 are connected to charge rings 34
via inverting amplifiers 72. Amplifiers 72 invert the output from
register 67, because a printed cell is represented as a one within
the register 67, and to accomplish printing by a drop 38 its
associated charge ring 34 must be uncharged at the instant of drop
formation.
As further shown in FIG. 4 the rows of charge rings 34 are
separated by a distance d, and this distance typically may be about
60 mils. If it is assumed that the web 25 moves 2,000 ft. per.
min., then about 150 microseconds will elapse during web movement
from a point under the upper illustrated row of charge rings to a
point below the lower illustrated row of charge rings. This means
that the charging control signals supplied to the lower illustrated
row of charging rings 34 must be impressed with a delay of 150
microseconds in order to produce registration of the printed dots
made by the drop streams controlled by the two rows of charge
rings. Thus there are provided a series of delay circuits 77 which
are connected to those of output lines 71 which service the lower
illustrated row of charge rings. These delay circuits 77 preferably
are simple 6 stage flip-flop chains; each stage of which provides a
24 microsecond delay at the nominal shifting rate of 40 kHz. Pulses
for time delay circuits 77 are provided by line 103 which is
connected to the output of counter 79.
It will be readily apparent that the delay produced by time delay
circuits is self-adjusting in accordance with variations in the web
speed. This can be understood by noting that tachometer 51 causes
counter 79 to generate 100 pulses per inch of web movement, so that
there are 6 pulses generated on lines 106 and 103 when the web
moves 60 mils. Delay circuits 77 always produce a delay of 6 pulse
periods, regardless of what the pulse frequency may be, and this is
exactly the required delay time when the distance d is 60 mils. If
d happens to be 70 or 80 mils rather than 60 mils, then the time
delay may be correspondingly adjusted by adding 1 or 2 flip-flop
stages to time delay circuits 77.
As described above there are 40 12-bit bytes of data loaded into
register 67 for every tachometer pulse gated out of AND gate 52.
However, the charging signals applied to charge rings 34 are
clamped to a constant level and are switched to a new level only
after the above 40 bytes have been loaded and shifted into storage.
The new charging levels then are held until after the next 40 bytes
have been loaded and have replaced the previous 40 bytes in
storage. For this reason there is provided a flip-flop 93 which
controls the readout from register 68 via a line 107. Flip-flop 67
is reset by pulses from counter 79, which as above mentioned, are
transmitted along lines 101 and 106 to the storage control terminal
of register 67. This briefly interrupts readout from register 67,
but the readout again resumes upon occurrence of the next clock
pulse on line 104, which it is seen is connected to the reset
terminal of flip-flop 93.
it will now be understood that the printed dots 76 of FIG. 6 do not
represent the printed result of single non-caught drops 38, but
rather are each the printing result of approximately 3 non-charged
and non-caught drops which are generated during the interval
between the setting of flip-flop 93 and the next resetting thereof.
At a web speed of 2,000 ft. per. min. this time interval is about
25 microseconds, and this represents about 3 drop periods at the
above mentioned stimulation frequency of 120 kHz. At a lower
stimulation frequency in the order of 40 kHz, the dots 76 would
each be printed by a single drop 38, but this would require very
precise control of the phase of drop formation at all 480 jets. At
the present time such phase control is quite difficult to
achieve.
It should be clear that the dots 76 need not be printed by an ink
jet printer but could be printed by other dot matrix printing
apparatus. Thus the electrical signals applied to lines 71 could be
used for selective activation of an array of solenoid driven wires,
in which case an inked ribbon would be positioned between the wires
and the moving web of paper. As another alternative dot matrix
printing could be accomplished by sequentially sampling lines 71
and using the sampled data for amplitude modulation of a scanning
laser beam.
While the methods and forms of apparatus herein described
constitute preferred embodiments of the invention, it is to be
understood that the invention is not limited to these precise
methods and forms of apparatus, and that changes may be made
therein without departing from the scope of the invention.
* * * * *