Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
i'S~
B_ckqr_und _f the Invention
The presen~ invention xelates to raster scanniny
systems such as dot ma-trix printers, televisions, CRT
monitors, and the like.
Many devices use a so-called "raster" scanning
system. These include dot matrix prin~ers and cathode
ray tube (CRT) displays such as used in television sets
and computer monitors. In any raster device, a dot-
producing entity moves across the display area repeatedly
on a line-by-line basis. In a printer/ it is a printhead
moving across the paper. With the CRT, it is an electron
beam moving across the phosphors of the CRT screen.
There are two dot matrix principles used in dot
matrix print-ers today. One uses a matrix which is
approximately equal to the dot size. The second uses a
dot size that is up to ten times larger than the matrix.
As used herein, "Matrix" means the spacing between the
possible dot positions or the number of possible dot
positions per inch horizontally and vertically.
All quality printing by impact dot matrix printers
uses a dot size considerably larger than the matrix.
While in the draft mode, the dot size and the matrix are
usually approximately the same.
Quality printing is generally performed by using a
seven to nine pin printhead and using multiple passes of
the printhead while moving the printhead or paper a
fraction of a dot between passes (Reference: R.C.
Sanders Patent No. ~,159,882). Another method currently
used is to use eighteen to twenty-four pins arranged in
two or three staggered rows. At the same time, the
matrix used for firing pins during the horizontal sweep
of the carriage is a fraction of a dot width.
Most thermal and laser printers use a dot size equal
to the matrix, while ink jet printers have been designed
both ways. Most, if not all, CRT monitors use a dot size
equal to the matrix size. The new improved CRT monitors
will use dot sizes that are equal to the matrix size as
~ 7 ,~
~7~
well as dot sizes that are much larger than the matrix
size.
~ . ~
This invention is based on using a dot size much
larger than the ma-trix size regardless of the type of
device. To achieve scanning by a dot matrix printing
system accoxding to the present invention, the printing
means is moved relative to the recording means with a
scanning motion having a major direction and means are
provided for imposing on the scanning motion a cyclical
variation in the direction of the motion transverse to
its major direction. In one preferred form of the
invention (intra dot scanning), the cyclical variation
completes at least one cycle during a scan equal to the
dot size. The cyclical scan has a duration equal to or
longer than the dot refire time and an integral number of
dots can be created at predetermined points along the
cyclical path. In order to minimize confusion, the
application of this principle to various types o~
~0 printers is described separately in the following
detailed description. The use of the present invention
with CRT monitors is then addressed. Where dimensions
and performance capabilities are given, they are by way
of example only and reflect tested modifications of
actual devices performed by applicant herein.
In accordance with one aspect of the present
invention, there is provided a raster scanning system
comprising: (a) means Eor repeatedly sweeping an image
producing means across a display medium to produce a
series of sweeps for the creation of visual images in
said series of sweeps; (b) deflection means for
cyclically deflecting the image producing means in a
direction substantially normal to the sweep direction in
a repetitive series of like oscillations throughout each
image producing sweep across the display medium; (c)
means for controlling the image producing means to vary
the brightness of pixels created; (d) means for
supplying a first image information signal synchronized
with said repetitive series of oscillations to con~rol
the controlling means to produce vlsual image information
on said display medium at one set of desired
S corresponding locations on each of desired ones of the
oscillations during a the sweep; and (e) means for
supplying at least a second image information siynal
synchronized with the repetitive series of oscillations
to control the controlling means to produce visual image
information on said display medium at another set of
desired corresponding locations on each of desired ones
of said oscillations during the said sweep.
In accordance with another aspect of the present
invention, there is provided a cathode ray tube display
such as a computer monitor, television, or the like
including a raster scanning system having a source of an
int~nsity controllable electron beam and a main means for
deflecting the beam to create a series of horizontal
sweeps of the beam from the top to the bottom of a
2~ display screen for selectively creating a series of pixel
dots in a series of parallel rows comprising: (a)
deflection means for deflecting the dot creating means in
a direction substantially normal to the sweep direction;
(b) drive means operably connectad to the deflection
means for cyclically deflecting the dot-producing means
in a repetitive pattern throughout each dot-producing
sweep across the display medium, the driving means
comprising a signal producing means for applying a
repetitive electrical signal to the deflection means to
produce a peak to peak amplitude of the pattern which is
equal to the spacing between horizontal sweeps; and (c)
means for providing an image information signal
synchronized with the repetitive pattern to control said
intensity controllable electron beam to generate two
parallel image lines for each sweep.
'L~7'5r~
-- 3a -
Descri~L~ L~ ~LZ ~lls
Figure 1 is an end view of one type of dot matrix
S printhead that can be used with the present invention.
Figure 2 is a side view of the printhead o-f Figure 1
showing the means for oscillating the printhead
transversely to the major scanning direction.
Figure 3 is an enlarged schematic diagram of the
path of the print pins across the paper for the printhead
oE Figures 1 and 2.
Figure 4 is an enlarged view of typical type font
like characters producsd by the printhead of Figures 1
and 2 when moved through the path of Figure 3 according
lS to the technique of the present invention.
Figure 5 is a side view of a vertical scanning
mechanism for one and two pin prin-ters employing the
~,7
pres~nt invent;on~
Figure 6 ;s an enlarged schemat;c diagram of the
path of ~he electron beam scan in a CRT for a 1~'~
computer ~onitor when operated according to the present
5 invent;on.
Figures 7 and 8 are front and top v;ews~
respectively, of a magnetic CRT adapted to operate
according to the present ;nvention.
Figure 9 is an enlarged schematic diagram of the
path of the electcon beam scan in a CRT according to the
presen~ invention when using offset deflection to obtain
pseudo sawtooth scanning~
Figures 10 and 11 are front and side vie~s,
respectively, of an electrostatic CRT adapted to operate
according to the present invention.
Figure 12 is an enlarged schematic diagram of the
path of the electron beam scan in a CRT according to the
present invention when using a clippéd sine wave with
electrostat;c deflection to obtain pseudo square ~ave
Scanning
Figure 13 is an enlarged schematic diagram of the
path of the electron beam scan in a CRT according to the
present invention during hor;zontal scan ~ith fractional
scan shown.
F;gure 14 is a diagram of a group of adjacent pixels
in a CRT.
Figures 15 and 16 are block diagrams of circuits
according to the present invention for interpolating
between horizontal scans in a CRT~
F;gure 17 is a block diayram of another embodiment
of the invention~
~ I
Serial Impact Dot Matri~ Pr;nter Using
A twelve pin printhead 10 ;s used in th;s first
:
: - , .
.
~7~
~ 5
example arranged in two ro~s of six p;ns 12 as sho~n in
F;gure 1. The vert;cal spacing of the p;n centers is
1/~8" irl each row and d~e~ to the staggeringD the
e~fective spacing between the p;n centers is
appro~imate~y 1/96". The pin d;ameter ;s1/72" and the
dot diameter made on the paper is sixteen mils. If the
p;n diameter is reduced or increased, it is important to
scale the above dimensions as well as to increase or
decrease the number of pins in the printhead.
The ends of the pins are moved ;n sinusoidal fashion
five mils as the carriage (not shown) is moved
hori20ntally across the paper. Referring to F;gure 2,
this notion is created by driver 14 operatin~ on the
pr;nthead 10~ Since the printhead 10 is supported on a
frame 18 by spr;ng means 1b, the printhead ~;ll
oscillate at a freqlJency controlled by the drive and the
spring mount~ One cycie is made every time the carriage
is moved 1/96" horizontally~ The perm;ss;ble firing
points (;.e. the matrix) are approx;mately ten points --
equally div;ded ;n time for every complete vert~cle sine
wave cycle. This gives a matrix size of 9601;nch
horizontally and about 300/inch vertically. It should
be obvious to one skilled in the art that the intra-dot
scanning cycle could be divided into a lesser or greater
number of permissible firing points, say six to t~enty
po;nts. It is ;mportant that there be an integral
number of fir;ng points per intra-dot scanning cycle~
The motion of the end of the pins for a twelYe p;n
pr;nthead ;s shown ;n Figure 3. A typical dot as
produced by the pins 12 is indicated as 20 while the
sinusoidal path traversed by each of the pins 12 as the
carriage moves across the paper is indicated as 22~ The
carria~e speed ;s adjusted so that the refire time is
equal ~o 3t4 of an intra-dot scan cycle for quali~y
printing~ If the refire time is 400 microseconds~ the
3~ 7 t~ ~
~ ~ .
carriage speed would be 18.75 inches/second~
Figure 4 is an enlarged illustrat;on of a word
printed with various Letters as produced w;th the Figure
3 matrix. One can see by the layout o-f the dots 20 that
5 the qual;ty is at least as good as four pass, nine pin
output. Using a 250Q Hz printhead, the speed of the
intra-dot scan printer (twelve pins) would be 187.5 cps
tdraft speed 500 cps~ for the four pass, nine pin
printer. The speed advantage is obvious. If these same
letters were made ~ith a 2sno Hz eighteen pin printhead
with two passes~ the speed would be 120 cps ~draft speed
SOOcps)~ A 1875 Hz twenty-four pin printhead would give
the same speed and almost the same quality.
The advantage of the intra-dot scanning system of
the present invention is that it gives the same
performance as a twenty-four pin printhead ~ith, of
course, far less parts and cost. It gives superior
performance to both the nine pin and eighteen pin
printhead used in a multipass printing system and does
not have the stringent horizontal position indicator
requ;rements. (For additional information see recent
R.S. Sanders Patent No. 4~533,269.~ The intra-dot
scanning system need not be sinuso;dal, although that or
a triangular scan is near optimum and, therefore,
preferred ;n most cases. To obtain a square wave, one
preferred ~ethod comprises having the paper stationary
during the hor;zontal scan and a stationary printhead
during ~he vertical scanO Obviously, scan waves could
; be used that lie bet~een the trianglar wave and the sine
~ave and between the sine ~ave and the square waveO
The vertical scanning ~otion is given to the end o~
; the p1ns by moving the printhead 10 supporting the p;ns
12 plus or minus five mils at a frequ~ncy of ~875 Hz.
The mounting of the printhead 10 is made resonant at or
near 1875 Hz to min;mize the driv;ng power. This
.~ .
.
.
. . :
- .
J'5~
frequency is der;ved by subdividing the matr1x pulses
that deter~ine the matrix points a~d i5 phase adjusted
so that one of the matrix pulses is synchronized with
the top o~ the sign wave~ It is also ;mportant that the
separation of the two rows of p;ns 12 is an exact
multiple of the 1/96" horizontal motion during a single
vertical scan.
Alternately, the entîre print rnechan;sm, generally
ind;cated as 2~, can be moved so as to g;ve the desired
motion to the end of the pins 12 either by rotating ths
~echanism 24 or mov;ng it vertically as an alternate to
moving the printhead 10.
While dot ~atr;x printing using print w;res has been
described above, the ;nvention is equally applicable to
ink jet printers o~ the type described in "Printout"~
Vol. VIII~ No. 3, March, 1984. In this case the jet
nozzle assem~ly is osc;llated to achieve the transverse
scanning motion superimposed on the major scann;ng
mot;onO
Scaling ~;th Dot Size
If the dot size were reduced to twelve mils, the
vert;cal scan would be plus or minus 3.75 mils and the
horizontal mot;on for a vert;cal scan would be 7.5
~ miLs. Th;s would increase the number of p;ns to
;~ 25 ~ixteen. If the refire time was then 333 micro seconds~
the carriage speed would be 17"1second. The letter
quality (LQ~ speed wouLd be 170 cps. The design
cons;deration being used ;s ~o keep the dot overlap in a
45 line at 1.8 of a dot diameter~ Vertical and
horizontal lines would have a dot overlap of 3.8 of a
dot diameter. Lines at intermediate angles or parts of
curves would have dot overlaps that range between 3.8 of
a dot and 1/8 of a dot with the fast majority at 3/8 dot
d~ameter overlap~ If the dot is reduced to eight ~;ls
or belo~, it is not necessary to overlap the dotsr
,
Between e;ght and twelve mil dots, the amount of overlay
needs to be determined emp;rically.
Effect~v ~
The effective matrix size of the example of Figure 3
is 960/inch horizontally and somewhat variable in the
vert;cal direction~ but worst case is 300/;nch
vertically. Th;s is somewhat better vertically than a
twenty-four pin head ~216/inch vertically) and
equivalent to four pass~ nine pin printing (28R/inch
verticallY)-
It is re~atively easy to increase the matrix
definition ~ith no reduction in printing speed and only
minor penalties in electronic cost by doubling the
number of fir;ng points. This gives a matri~ ;ze of
lS 1920/jnch hor;zontally and 600/;nch verticaily. The
only limitation to increasin~ the definit;on with
synchronized intra-dot scanning or multipass printing ~s
that the dot si~e ;s the mintmum ~ine width.
There i5 an ;mproved variat;on of the invont;on 1
deflection of the scan is ;ncreased to plus or minus 10
mils. The matrix then can be increased to t~enty points
per cycle~ This ~ives a matrix size of 1920/;nch
horizontaliy and 480t;nch vertically. This provides a
hybrid inter-dot and intra-dot scan. Us;ng this
technique, the possibLe matrix points are a lot more
uniform~ When printing letter quality~ the carriage
speed can be increased to 25"/second because of the
increased interleav;ng. The ref;re time of 400
microseconds is one complete vert;cal scan. ~his
increases the Letter qual;ty speed to 260 cps for ten
~¦ pitch and 325 cps for proport;onal Times Roman.
When the carriage ;s speeded up to 41.6 ;ps with
pLus or m;nus 10 mils d~flection, a matrix size that is
960J;nch hor;Yontally and 430t;nch vert;cally can be
~ '
' : '
: ' ' '
: .
.: .
rl5~ L~
_ 9 _
achieved. This is a pure inter~dot scan in both
directions~ The only limitation is that there is a
reduced nu~ber of vertical pos;tions (i~e., two) that
can print horizontal l;nes~ One must therefore reduce
the point size of near letter quality (NLQ) or ;ncrease
the number of pins by one or two pins.
When the speed is increased up to 83~2 ips, a matrix
size of ~80Jinch hori~ontally and vertically ;s
achieved. ~8 dpi draft characters ~hich uould be
printed at 832 tten p;tch~ have a quality which is about
the same as conventional 48 dpi draft characters. The
only limita~ion is the reduced number of vert;cal
positions that can print horizontal lines~ One must
either reduce the po;nt s;ze of draft letters or
;ncrease the number of p;ns.
In summary, using a triangular scan with plus or
minus 10 m;ls, letter quality is obtained at 260 cps
(ten pitch) and 325 cps ~proportional~. Th;s is a four
to one improvement over conventional four pass
printing. Near letter quality is obtained at 416 cps
~ten pitch) and 520 cps (proportional), a three to one
improvement. A 48 dp; draft at 830 cps ~ten pitch)
gives a 1~3 to 1 improvement~ This also m;nimizes the
registrat;on problems of multipass pr;nting and is,
therefore, a s;gnificant improvement for that reason
alone.
As can be seen9 the complete scanning cycle is
longer than the dot d;ameter (;nter-dot scanning)~
However~ this embodiment shares with the intra~dot
scanning system the common point of novelty of the
present inventionr i.e~ a transverse scan imposed on the
major scanning motion. When using inter-dot scann;n~, ¦
it is necessary to have the scanning cycle complete
within about three or four dot diameters; otherwiseO the
ability to achieve des;red quality of prlnting ~ill be
~ ~>7~
- lO -
lost. Preferably, the inter-dot scanning is complete
w;thin one and one-half to t~o cyclesa
The few minor disadvantages of this embodime.nt are
as follo~s: The printhead ;s tw~lve p;ns rather than
S nine pins. Additionally, one actuator is needed for
vertical scanning~ Thus~ there is an added cost of four
pins and the associated pin driversa Due to the higher
pin tip velocity~ there is also slightly higher dot
position uncertainty and dot elongation~
There is an extension of the in~er-dot scanning as
described above. ~y changing the angle of the
transverse pin motion to slightly off ver~;cal~ the
generally triangular path over the paper prev;ously
described can be changed to a sa~t~oth path. This ~orks
for e;ther direct;on of carriage travel~ although there
: ~ould be minor changes ;n the front depend;ng on ~he
direction of carriage travel. Just ~hich ;s more
desirabLe ;n a given pr;nter des;gn depends on the
detailed printer specif;cation~ ~a
Figure 5 sho~s the preferred vertical scann;ng
mechanism, generaLly indicated as 2~, to be used on one
and two pin printers, and the l;ke. A T-shaped member
28 has the pr;nthead 10 mounted to the crossmember 30 of
the "T". Also mounted thereto are two spring-b;ased
: 25 soleno;ds 32 used to dr;ve the two print pins 12. The
other end of the member 28 ;s mounted to p;vot pin 34
~ith torsion spring 36 pass;n~ through the pin 3~ to
: bias the ~ember 28 to a neutraL pos;tion from ~h;ch it
Gan be oscillated in e;ther direction ~i~e~ up and down)
by the drive solenoid 38~ The ~hole scanning mechanism
26 1s osciLlated by the dr;ve colenoid 38 to oscillate
the printhead 10 to moue the pins 12. As mentioned
above~ to create a sa~toothed scanning path, the p;vot
: pin 34 is positioned slightly off of horizontal to make
the plane of oscillation of the pr;nthead 10 slightly
,
,:''' '. ~" ' ' ,
~5;~ ~
off vertical. Tes~s have shown that with one pin, one
can ach;eve speeds of 44 cps dra~tp fourteen cps La and
that a two pin pr;nthead can ach;eve print speeds of 104
cps draft, 28 cps LQ ~ten pitch~ and 35 cps L~
~proportional).
A four pin pr;nthead is really a hybrid cross
between the in-tra-dot scann;ng and the extra dot
scanning of the one and two pin printers. Here, one can
achieve 156 cps draft, 52 cps L~ (ten p;tch)o and ~5 cps
LQ (proportional). The mechanism for vertical scann;ng
could be as shown and described above.
Also poss;ble is a s;% pin printer w;th pLus or
m;nus twenty mil vertical scanning. The print speed
ach;evable ~ith this approach is 234 cps draft~ 78 cps
LQ ~ten pitch) and 97 cps (proportional). The ssanning
again is a some~hat different hybrid from the four pin
scanning.
Serial Thermal Printer Using
Synchronized Intra-Dot Scanning
A serial thermal printer would be implemented ;n a
very similar fash;on to the serial impact dot matri~ ¦
printer as described above. Several examples of such
printers th~t can be modified in accordance w;th the
present invention are shown on page 3 of the Kyocera
brochure "~hin Film Thermal Pr;ntheads",
CAT/1T8504FTK/2192E, and page 4 of "Printou~", Vol. VI~ i
No~ 12, Dec., 1983 as well as "Printout", Vol. IX~ No.
9, Sept. 1985. The design of the printhead 10 of Figure
1 can be modified for use ~i~h a thermal printer ~ith
twelve printing elements in place of the p;ns 12 being
mounted ;n t~o rows of six ~ith the element centers
twenty mils apart and staggered so that the ef~ectiYe
element spacing is ten mils ~ith the vertical sinusoidal
scanning being plus or minus 5 mils with a carriage
moeiQn of ten mils and the printing elements producing a
,
.
~::
~ ~7~
dot of sixteen mils diameter ~the "dot" could, of
course, also be square or other convenient shapes even
though a round dot is sho~n ;n the f;gures3. The
intra--dat scanning could be triangularv square wave or
any shape in-between, but s;nusoidal or triangular is
nearly op-t;mum and~ therefore, preferred. Assuming the
refire speed ~as six milliseconds, the carriage speed
would be 13 ;nch~second~ The printing speed for high
quality letters (ten pitch) would be 130 cps, which
would be the same as a thirty-si~ pin printhead. The
savings would be in using a twelve pin pr;nthead with
the resulting elimination of the assoc;ated element
drivers and electronics. In the case of a thermal
printhead, however, the entire printhead would need to
be rotated or moved vertically to obtain the desired
scann;ng motion.
L;ne Thermal Printer or Electrostat;c Printer
Using Synchronized Intra~dot Scanning
. . . _ _ _ . . . ~ _
A line thermaL printer utilizes synchronized
intra-dot scanning by impart;ng a sinusoidal horizontal
motion to the pr;nthead as the paper is fed vertically,
thus giving the desired path to the printing eLementsO
The printing elements would be two rows of printing
elements. An example of such a line thermal printer
that could be modified in accordance ~ith the present
invention is shown on pages 4 and 5 of the
above-mentioned Kyocera brochure. The printing elements
would produce a dot sixteen m;ls in diameter tor square~
etc. as ment;oned above). The effective spacing of the
3Q printing elements is ten mils and the paper advances ten
mils ~hrough one sinusoidal cycle. The effective matrix
of such a configuration ;s 300/inch horizontaLly and
10001inch vertically
Assuming the refire speed is eight milliseconds, the
paper speed ~ould be 1"/second and the print speed would
,~ .
.. ' . ' , .
. ~ ~
''-" ' : '
, . - :
- l3 -
be 360 l;nes/minute. The savin~s would be the reduction
in printing elements from 2000-4000 elem~nts to 1000
elements for a 10" printlineO
Electrostatic Printer
An electrostati ~ cording to the present
invention operates in much the same way as the
previously described devices; that is, the printhead
oscillating horizontally as the paper moves vertically~
An example Df such a printer that could be modified ;n
accordance with the present invention is shown ;n
Versatek 8ulletin No. 525-2, April 1984, describlng its
V-80 pr;n~ers. The printhead would have a sinuso;dai
motion as the paper is fed vertically~ The printing
eLements wo~ld produce dots of sixteen mils in
diameter~ The effective spacing of the pr;nting
elements is ten mils and the paper advances ten mils
through one sinusoidal cycle. The e~fective matr1x is
300/inch hori~ontally and 1000~inch verticalLy. The --
intra-dot scanning could be triangular, square wave or
almost any shape in-between, although s;nusoidal or
triangular is, aga;n, nearly optimumy and therefore,
preferred. The major saving would be the reduction in
elements and the;r associated drivers from 2Dû0-4000
elements and drivers to 1000 elementsO
~
.
Intra-dot scann;ng can improve Let~er quaLity
print;ng by a factor of two over conventional methods in
addition to improving the vertical matrix size. Using
the technique~ draft printing is nearly doubled. Also,
the intra-dot scanning allows the printer to compensate
for continuous paper motion el;minat;ng the rap;d paper
motion at the end of eaGh shuttle motion. This is
accomplished by mod;fying the matrix patterns to take t
into account continuous paper mot;on. Different
'
.
'
~'
5~
character PROMS are used for each character pos~tion and
directions (onLy four combinations in a s;xty-eight pin
printer). When th;s feature is used, throughput is
;ncreased as much as 50%~ For example~ a sixty-eight
p;n ~13.6" line length) pr;nter sould print 2300
l;neslm;nute wh;le a t~elve pin (8" l;ne length~ printer
could pr;nt 730 l;nes/m;nute. It also greatly reduces
the cost of the paper handling mechanism in the printer.
Laser_Printers Using
~ Scannin~
In the case o~ a laser printer employing the present
invention a technique sim;lar to that employed ~;th CRT
display ~to be described shortly) ;s used; that ;s, the
beam is deflected vertically as the beam is swept
horizontally so the beam makes a sinusoidal track across
the drum. An example of such a printer that can be
modified to utilize the present invention is shown in
"Printout", Vol. IX, No. 6, June 1985, and "Printout"~
Vol. IX, No. 5, May 1985. If the dot size were ten ,
mils, the modulaticn would be swch that the beam
deflec~ed plus or minus 3~5 mils vert;cally while the
beam was moving horizontally 1/144~o The beam could be
deflected by a vibrating m;rror in the light path~ In
such case, the effective matrix would be 1500/inch
horizon~ally by 480/inch Yertically.
The advantage of this method is that a larger size
laser beam can be used, which could result in faster
printing by a factor of t~o or three times. The
intra-dot ssanning could be triangular~ square ~ave or
almost any shape in-between although sinusoidal is
nearly optimum, and therefore, preferred.
~h Emitting _ de (LED) or
Liqu;d Cr~t-L D;~d- ~L~ Pri~er
In LED tor LCD) printers tsuch as produced by Richo
or IBM), synchronized intra-dot scanning would again
`', : . ' ' ~', : '
,5 ~ Jr~~3
~ l~ ~
prov;de an advantage~ An example of such a printer that
could be mod;fied ;n accordance w;th the present
;nvention is shown in "Printout", Vol. IX, No~ 10,
October, 1985. Two rows of LED's (or lCD's) are
empLoyed. Each LED or LCD makes a dot on the drun o~
sixteen mils diameter and the effective spacing is
1/96". The path the LED or LCD elements make on the
drum is due to a s;nusoidal motion ;mparted to the LD
or LCD elements that move pLus or minus 50D m;ls
s;nuso;dally as the drum moves 1/144". The e~fective
matr;x is 480/inch horizontally and 1500/inch vert;cally.
The advantage of this over conventional LED or LCD
pr;nters is that it uses one-third or one-fourth the LED
or LCD eLements that the convent;onal LED or LCD
pr;nters use for the same matrix definition~ The
print;ng speed would be increased by a factor of two or
three. The intra-dot scann;ng couLd be tr;angular~
square wave or almost any shape ;n-between although
s;nusoidal is nearly optimum~ and once~ again therefore, ~.0 preferred.
Application to Imag;n~ S~stems
While the present invent;on has been described
initially in the context of ;ts use in dot matrix
pr;nt;ng, the basic pr;ncipals and points of novelty
involved can be utilized equally in image scanning as
well as image print;ng, as ~ill now be discussed.
Scanner for Graphic_ Input to Synchronized
~ Intra dot Scannin~ _r _ters
ln order to input graphic data into synchronized
scanning printers, it is necessry to scan the graphic
data in exactly the same way as the dots are printed.
Scanner for Serial Dot Matrix Printer
or Serial Thermal Printer Us;ng
ning
In th;s case, -he scanning head consists of tweLve
.
'
- l6 -
elements arranged ;n the same manner as the pins 12 in
the pr;nthead 10 of Figure 1~ In the scanning caseO the
eLements would be photo diodes~ for example~ ra~her than
print pins. The elements are made to move vertically
sinusoidally ~s the he3d moves hori~ontally so that the
head scans plus or minus f;ve mils 35 the head moves
1/~6" mils horizontally. The scanning could be
accomplished mechanically by moving or rotating the
scann;ng head~ or~ it could be done optically.
During the scanning, the effect is to p;ck up
horizontal lines as the scanner is moving vertically and
vertical lines as the scanner moves horizontally. This
type of scanner thus has a matrix def;nition of 960/inch
horizontally by 300/inch verticaLly~
The intra-dot~(or inter-dot) scann;ng ~ust be
identical to that used in print;ng~ Commerçial optical
scanning systems wh;ch could be modified to utilize the
present ;nvention are described in Richo ~ulletin, IS30
No. 8506-TA-8506, 8401, April, 1984.
Scanner S stem for L~ne Thermal or CCD Pr;nter
Using Synchronized Intra-dot Scanning
.~
In this case, the scanning head should match the
printhead both in geometry and mot;on. As a result, the
sranning elements should be as previously shown and ~he
motion should be as previously described for the
applicable associated pr;nthead.
Converting Standard Raster Graphics to
Synchronized Intra-dot Scanni_~ Format
- Assuming that standard raster graphics are in
30 sufficient detail to warrant it, the conversion to
synchronized intra-dot scanning format can be made as
part of a sof~ware program. ~asically, the method i5 as
follows. The s~andard raster graph;cs (with a dot size
to match the matrix size) ;s put ;nto a bit map windo~
35 ~hich progresses as the processin~ is completed. The
.. "
~ ~ .
.
s~
l7 -
sof~ware moves a Larger doe (as used in synchronized
scanner format) through the bit map on a scanning path
that matches the synchronized scanning format until a
match is obtained. So, basicallyf the software does
wha~ ~as proposed in the physicaL scanner in the two
previous paragraphs~ In the same way, it is poss;ble to
convert from intra-dot scanning to inter-dot scanning.
Facs;mile System_Using ~y~
Intra-Dot Scann;ng
Th;s system uses a synchroni~ed scanner for
transm;tting a synchronized scanner printer for
reception. A good combination would be the synchronized
scanner described above with the thermaL line printer~
The major advantage is an increase in matrix definition
(resuLt;ng in Letter quaLity output) with increased
printer speed and no increase in transmission time using
: 3 modified group 3 or 4 compression system. (For
descript;on of such compression system, reference shouLd
l be had to EIA standards).
n~er- and Intra-dot Scanning
As ment;oned earl;er herein, the present invention
~ is as applicable to CRT type d;spLays as it ;s to
; print;ng on paper. In fact, it is in this area that the
present ;nvention has, perhaps, its greatest potential.
Such use will now be described~
The electron beam is verticalLy deflected as it
I scans horizontally as shown ;n F;gure ~. Assuming a
spot size of fi~teen mils with a spacing between scans
of sixty m;Ls, in moving along its path 41, the electron
beam would be deflected sinusoidaLly plus or minus
~hirty mils while ~he beam was moving horizontally
fifteen mils. The vertical beam deflect;on sho~n in
Figure 6 is a sawtooth. This could very well be
accomplished by a sinusoidaL deflection that was tilted
.
,
: ', .
^~L~
~ l8 -
to give a sawtooth-like pa~hO A sinuso;dal vertical
deflection will give very sim;lar results. An alternate
preferred method is sho~n in F;gure 6a us;ng stepped
square wave modulat;on ~hich be~ter lum;nesce
uniFormityO The hor;zontal scan;frequency ;s 15.6 KHz
with about a ten ~;lliseco~d retrace. The vert;cal
scar1n;ng frequency ;s s;xteen MHz. The beam deflect;on
of plus or m;nus th;rty mils g;ves a non-interlaced
display~ This requ;res a sixty-four MHz video. The
result is a 1024 x 863 display~
In using an interlac~d display, the vert;cal
deflection is reduced to plus or m;nus fifteen mils.
This reduces the video to thirty-two MH2 and the resul~
is still a 1024 x 860 display.
~y increas;ng the dot access frequency (consequently
decreas;ng memory access time) the matr;x definition in
both directions is ;ncreased. Thus, with 2 sixty-four -~
MHz dot access frequency, the result ;s a 2048 x 1720
display using interlaced scanning, although mini~um line
width remains fifteen mils; however, the increased
matrix s;ze w;ll perm;t letter quality fonts of the same
quaLity as shown in Fi~ure 4 and eliminate the "jaggies"
in d;agonaLs and cur~ed lines in the graph;cs mode. j~
Inter- and Intra-dot Scanning
Inter- and ;ntra-dot scanning in a CRT can be
acco~pl;shed e;ther magnetically or electrostatic3lly,
d~pend;ng on the type of CRT employed. If magnetically~
a separate small deflection coil 42 is mounted
immediately after the electronic lens 44. It is
preferable to resona~e ~he coil 42 with a capacitor 46
a~ the vertical scanning rate as illus~rated in F;gures
7 and 8~ I~ is not efficient to use the vertical
deflection coils (not sho~n~ of the CRT itself. The
- coil ~2 is ~oun~ed horizontally if sine waves are
desired and off horizontal if pseudo-sawtooth scannirg
'- '
. ~ ' .
as sho~n in Figure 9 is desired. The angle from
horizontal 5~) is picked to match the horizontal
scanning rate.
A preferable way to obtain inter and intra-dot
scannlng in an electrostatic CRT is to use small
electrostatic deflection plates 4g ;mmediately after the
electronic focusirlg lens 44 as shown in Figures 10 and
11~ These plates 48 are placed vert;cally if sinuso;dal
scanning is ~anted, or off vertical if pseudo-sa~tooth
scann;ng is the desired result. Figure 12 illustrates
pseudo-square Aave scanning if the sine wave driving the
electrostatic plates is limited. This scanning is
useful in some applications.
High Definition Pictures ~TV)
If the p;cture ;s scanned in the same way as the
monitor, definition of the resulting images can be
improved in the same fashion~ This could be very useful
in all the cases where current TV standards are not
involved~ For example: photographic ima~es of aLl
types, closed circuit TV, map projections, etc.
Vertical Def~nition Enhancement
~Apparent) of Televi_i_n Picture t
A television picture is produced every one-thirtieth
of a second by scanning each alternate line on a
25 television tube (CRT~ during the first one-sixtieth of a li
second (i.el~ one-half of the picture); and, during the
subsequent one-sixtieth of a second, scanning the lines
intermedia~e ~hose scanned in the previous one-sixt1eth,
with ~he second half of the pic~ureO This is a
well-known and much-used technique known as
'interLacing"~
To improve the apparent vertical definition, it is
desirable to scan all lines everyone-sixtieth of a
s~cond~ The bLank intermediate scan lines could then be
filled ;n by memoriz;ng a preceding line, comparin~ the
.
.. . .
: -
.
~.~75;~
~ o
data ;n that memory with the succeeding llne~ and then
fill;ng the intermed;ate l;ne ~;th averaged,
;nterpolated, or otherw;se der;ved ;nformation. A
simple, less effective method normally employed simply
repeats the scan l;ne twice.
There is much interest in improving the quality of
ord;nary TV as transmitted by today's standards~ The
major opportunity for qual;ty improvement results from
elim;nating the above~descr;bed ;nterlace, which causes
problems when the human eye moves vertiGally ~saccades)
over the picture. For example, if you store an entire
frame including the interlace~ combine the two, and then
put the comb;ned picture on the tube at approx;mately 60
Hz~ you obta;n a much better qual;ty p;ctureu The major
disadvantage of th;s is the expense~ Storing an ent;re
frame and comb;n;ng ;t is still expensive to do
electron;cally. Doubl;ng the hori7intal scan frequ~ncy
is also expens;ve.
Another method that has been tried ;s to store one
scan l;ne and then repeat the stored scan line ln tha
;nterlaced l;ne below. Th;s gives improvement, but not
as much as the case where you combine the entire frame7
Further, it is st;ll relatively expens;ve because you
must still double the horizontal scan frequency~
Using inter-dot scanning in various forms, one can
dupl;cate or come close to these above-descr;bed
;mprovements; but, in a much less expensive way. The
present invention achieves this not by doubling the
hori~ontal scan to produce a complete picture every
one-s;xtieth of a second; but, by a technique der;ved
from ;nter-dot scanning in which the scanning electrode
beam is caused to oscillate vertically with an amplitude
equal to the space between adjacent scan lines and equal
to the elec~ron beam diameter produc;ng the trace. By
this means, the electron beam can produce picture
- ~ '
75~
information for one scan trace directly from the
broadcast signal while on the peaks of its vertical
oscillation and from deriYed informat;on from the
preceding ~nd succeeding scan l;nes wh;le in the troughs
of the oscillat;on. In th;s way, a single scan can
produce ad~acent scan traces simultaneously, ~;th the
electron beam being modulated to vary the intens;t;es as
described~
The derived~ intermediate signal may be generated by
storing, on the fly, one Line of scan information and
comparing the s;gnal from the output of the store with
the s;gnal on the input o~ the store ~the store being
dynamic ;n that the stored ;nformation proceeds serially
from the input to the output, i~e. a FIF0 store) and
deriving ;ntermediate scan information from th;s.
Hence, the compared information will always be of
sisnals synchronized vertically from preced;ng and
succeeding scan lines.
The vertical oscillation of the electron beam may be
of a sawtooth form whereby the scan line information of
the preceding and intermediate scan lines are
synchronized vertically one above the other essentially
at the top and bottom of the vertical of the sawtooth
form. The oscillation may be produced by the techniques
of Figures 7 and 8 or Figures 10 and 11, as prev10usly
described in detailO More detailed explanations ~ill
now be provided~
Vertically scan the electron beam 40 sinuso;dalLy as
in Figure 13 plus or minus one-half dot diameter at
t~ice the maximum video frequency tsay seven to ten
MHz)~ This produces a dot that is the same spot size
horizontally, but tu;ce as high vertically. O~her than
t~is, leave th~ monitor the same as before. This does
the same th;ng as storing one scan line and repeating
.
~ 22 ~
the scan line in the interlaced line belo~. ~owever, lt
does not need to store a line nor does it double the
horizontal frequency. The disadvantage of this scheme
is that the top and bottom horizontal lines ~ill act as
follows. Assume top of horizontal contrast shift starts
on an interlaced line. Extending donn the
non-interlaced line above gives no intensity where there
should be such~ The interlaced frame line gives the
proper ;ntensity. So, the net result is that the top of
hori20ntal edges are at one-half of proper intensity and
blinking at the thirty H2 interlace rate. The same
effect takes place at the bottom of the horizontal edge,
The diagonal edges develop a sort of f~z7y/jagyy effect
for the same reason. This technique ~;ll, ho~ever, fill
;n uertical l;nes. The apparent improvement takes place
because much less of the picture is blinking at the
thirty Hz ;nterlace rate~
Method 2 - Increased Picture $morovement
Vertically scan the beam s;nuso;dally as in Figure
13 plus or minus one-half diameter at the maximum video
frequency in a synchronized manner during the hor;zontal
scan. Store one scan line on a CCD ch;p (such as the
Fairchild CCD321A). Derive the interlace line by
averaging the first and second ;nterlaced lines;combine
the ori~inal in the derived inter`laced line so that a
video synchronized signal that records the first line
and the f;rst interlaced line on the face of the tube ;n
one hori~ontal scan is obtained. This process is ~,
repeated for the entire frame for one sixtieth of a
second. When the interlaced frame comes up, ths
; forego;ng process is repeaeed. The superimposed p;cture
resulting therefrom should be much improved.
Just how much better the picture actually is depends
on the quality of the interpola~ion between the t~o
scans. The simplest scheme is to average the intensi~y
'
. .
.
:
,
~ 5~
23 -
oF adjacent vertical pix~ls. The d;sadvantage of th;s
scheme ;s similar to Method 1, described above. The
effect on horizontal and diagonal l;nes is the same as
Method 1~ except the edge ;ntensity ;s three-fourths of
the proper value tas opposed to one-half); and, it still
blinks at the ;nterlace rate of th;rty H~
Method 3 - Maximum P;cture Improvement
The best technique is to compare three adjacent
horizontal pixels on one scan ~;th three adjacent
hor;zontal pixels immediately belo~ on the second scan
as sho~n in Figure 14. I~ p;xel H1 ;s the same
;ntensity as pixel B2~ we make B12 the same intens1ty.
If B1 and ~2 are not the same ;ntensity, we make B12 the
average of ~he two intens;ties unless A1 intensity is
the same as C2, ;n ~hich case we make ~12 the same
intensity as A1 or C2. In a similar manner, ;f A2 is
the same intensity as C1~ we make 812 the same
intens;ty. The advantage of th;s method over Method 1
;s that most diagonal edges are continued almost
perfectly. Horizontal fuzziness should be about the
same as Method 2.
In Methods 2 and 3, special circuitry is needed to
interpolate between horizontal scans. Figures 15 and 16
show in block d;agram form two circuits tha~ could b~
employed to perform this interpolation. figure 15 shows
a CCD32A chip 50 connec~ecl wi~h a multiplexer 52 to
provide th~ video signal synchronized with the vertical
pseudo-square wave scanning for Method 2c The CCD321A
chip S0 ;s connected in series mode providing 910 bits
of analog shift register~ After one horizontal sweep is
stored in the CCD321A, the timing signals are such that
the correspond;ng b;t of the subsequent horizontaL scan
signal is going into the CCD321A as the corresponding
bit of the previous horizontaL scan signal is going out
.
.
:,,, ' :, . ,
.~ -, .. . .
' '. :
'
7~5~
-- 24 --
of it~ By ~verag;ng the correspond;ng b;ts, one can
obta;n the ;nterpolated ;nterlaced s;gnal that ;s
desired~ The video signal to the CRT at 54 is the
result;ng of the multiplexer 52 multiplexing between the
prev;ous horizontal scan signal and the ;nterpolated
interlaced s;gnal at the ;nter-dot scanning frequency.
It should be noted that when us;ng this techn;que~ the
s;gnals appear on line 54 to the CRT one hor;zontal scan
line late.
To accompl;sh Method 3, the more complicated c1rcu1t
shown in F;gure 16 must be used. In th;s case, a
spec;al modified CCD321A ch;p 50 wh;ch makes available
the f;rst three bits of the incoming s;gnal and the last
three bits in memory ;s used. Here, again, the timing
is such that there are three corresponding bits of two
adjacent horizontal scan signals. These are fed to a
special logic chip 56 which includes three comparators
and which includes logic to perform ~he steps of Method
3~ The output is aga;n multiplexed by the multiplexer --
52 ~t the scanning freq~ency and fed to the CRT on
output line 54. HereO aga;n, the picture displayed is
one scan line late.
Method ~
Still another embodiment of the invention is shown
in Fig. 17 ~hich is a schematic circuit diagram showing
a modif;ed digital TV system. With the advantage of 1/2
frame and full frame digital storage, Fig~ 17 shows ho~
to eLiminate any interpolation and elim;nate the
;nterlaced scan us;ng a standard TV transmission.
The video s;gnal goes into a 1/2 frame store and ~e
multiplex the incoming signal ~;th the signal delayed
1/2 frame as shown by the diagram at ~he bottom of
Figure 17. This gives a non-interlaced picture with no
interpolation. If we add the interpolation used by
Method 2 or 3, we can double the number of displayed
.' ' ~ ' .:
..
- 25
l;nes ~ithout interlace by us;ng the scan pattern shown
on F;gure 6A still only using standard TV transm;ssions
and only 15.6 KHz horizontal scan.
Pro~ection_TV
Projection TV us;ng three tubes of different colors
will operate as ~escribed above as there are no
complicati~ns oF color shadow masks or trinitron
aperature ~rilles.
~d~
In order to make sure that we don't have to match
the vertical scan ~ith the shadow mask holes, ue must
use a vertical scanning frequency twice the frequency
which the horizontal scan intercepts the shadow mask~
Th;s insures that the beams hit the mask as well as they
do w;thout any vertical scanning. Its disadvantage is
that the final video amplifier must be quadruple the
band ~idth and the beam current must be larger to
compensate for the shorter time the beam is in the
shadow mask. If we use the pseudo-square wa~e scanning
trequ;rin~ eLectrostat;c deflection as in Figure 10 and
11) as iLlustrated in F;gure 1Z, ~e minimize the
;ncrease injbeam current required. With double the
shado~ mask frequency, ~e get half the beam current on
~he phosphor in one horizontal scan as compared to ~he
current if we were not inter-dot scanning; however,
since we pain~ each line twice as much, the beam current
required remains about the same.
Tr ini ~ , ~ [~
When using a trin;tron type color tube, ~he
inter-dot scanning ~requency should be equal to tw;ce
the frequency at w~ich the hori20ntal scan lntercepts
the aperature grill. Pseudo-square wave modulation as
;n f;gure 12 should also be used again. Once again, we
~et half the beam current pn the phosphor in each
~, . .
.~ . , ~ - ~ . . .
~' , '. ,.
5;~4
- 26 --
horizontal scan; but, because we have tw;ce the scan,
there is no appreciabLe increase ln beam current
required.
Wherefore, having thus described my invent~on~ I
5 cla;m: ;
10
'
.
: 30
.
.
