The invention relates to a method of adjusting alignment
positions of ink dots printed with at least one printhead that is mounted on a moving
carriage of an ink jet printer, comprising the steps of printing ink dots on a testchart
while the carriage moves over the testchart with a predetermined speed, measuring
a relative dislocation of the ink dots, and correcting the alignment of the ink
dots by adjusting the timing of activation and/or the position of the printhead
in accordance with the measured result.
An ink jet printer typically has one or more printheads
mounted on a carriage that is moved over a recording medium in a main scanning direction
Y. Thus an image swath consisting of a certain number of pixel lines, corresponding
to the number of nozzles of the printhead, is printed during each pass of the carriage,
and adjoining swaths of the image are printed in subsequent passes of the carriage,
while the recording medium is intermittently advanced in a sub-scanning direction
X normal to the main scanning direction Y. In order to obtain a good image quality
at the transition from one swath to the other, the ink dots that are printed in
different passes have to be aligned correctly in the sub-scanning direction.
However, when an ink droplet is expelled from a nozzle
of a printhead, it has to travel a certain distance until it impinges on the recording
medium. Since the printhead is moving in the main scanning direction, the ink droplet
undergoes a certain speed-dependent aberration in that direction. This may lead
to an alignment error between two ink dots that are printed in different passes.
For example, when the printer is to be operated in a bi-directional print mode,
i.e., a mode in which ink dots are printed not only in a forward pass but also a
return pass of the carriage, the aberration depends on the direction of travel of
the carriage. The activation timings of the printhead, and hence the positions at
which the pertinent nozzles are fired, must therefore be adjusted carefully, so
that the different aberrations in the forward pass and the return pass are compensated
If the image is printed with a plurality of printheads
mounted on the same carriage, not only the timings but also the positions of the
printheads on the carriage must be adjusted in order to make sure that the ink dots
printed with different printheads have the correct positions relative to one another.
A high quality multi-color printer is preferably equipped
with at least two printheads per color, and the printheads for the different colors
are arranged mirror-symmetrically. Then, one set of color printheads is used only
during the forward pass, and the other one only during the return pass. This has
the advantage that the ink dots of different colors will always be superposed in
the same sequence, irrespective of the direction of travel of the carriage, so that
the color composition will always be the same. However, if the printheads for the
same color are not aligned correctly, the ink dots printed with these printheads
in the forward and return passes of the carriage will be dislocated relative to
one another, so that a thin line extending in the sub-scanning direction X will
In a conventional method for checking and adjusting the
alignment of the ink dots, the printer is used for printing a test pattern onto
a testchart. In that print process, the operating conditions and parameters of the
printer are the same as in a normal print operation. In particular, in view of the
aberration effect, it is important that the test pattern is printed with a well
defined speed of the carriage. Then, the test pattern on the testchart may be inspected
visually, e.g. with a microscope, or the positions of the ink dots on the testchart
may be measured with an electro-optical sensor, in order to provide the data that
are needed for correcting the activation timings and/or the printhead positions,
A difficulty encountered in detecting the alignment of
the ink dots is caused by the fact that, when a nozzle of an ink jet printhead is
fired, it normally does not just expel a single droplet, but it first expels a relatively
large droplet which is followed by one or more smaller droplets, the so-called satellites.
Since the aberration of the satellites is different from that of the main droplet,
the corresponding dots formed on the recording medium or the testchart are shifted
relative to one another in the main scanning direction, which makes it difficult
to detect the exact position of the dot.
It is an object of the invention to provide a method adjusting
alignment positions of ink dots, which can be performed with a simple measuring
equipment and reduces errors that may be caused by satellites.
In order to achieve this object, according to the invention,
said predetermined speed, with which the carriage is moved when the test chart is
printed, is smaller than a nominal speed with which the carriage is moved over a
recording medium in a print process, and a misalignment of the ink dots that will
be printed when the printhead is moved with the nominal speed is calculated from
said measured dislocation, said predetermined speed and said nominal speed.
The invention takes advantage of the effect that the satellites
tend to be absorbed in the main dots when the speed of the carriage is reduced.
Thus, by printing the testchart with a reduced carriage speed, errors resulting
from the satellites can largely be eliminated. However, due to the reduced speed
of the carriage, the aberration of the ink droplets is different from the aberration
occurring in a normal print process. According to the invention, this problem is
solved by calculating back from the measured aberration of the ink dots to the true
aberration that will occur in the normal print process. As a result, the alignment
of the ink dots can be detected with improved accuracy. When a sensor, e.g. an opto-electronic
sensor is used for measuring the positions of the ink dots, it is not necessary
to employ a complicated and expensive high-resolution sensor that would be capable
of resolving the satellites and/or a satellite-induced distortion of the shape of
the ink dots on the testchart.
Optional features and further developments of the invention
are indicated in the dependent claims.
An apparatus suitable for carrying out the method according
to the invention is defined in claim 4, and claim 5 defines a printer in which the
ink dots printed in the normal print mode are aligned in a specific way.
Preferably, the speed of the carriage used for printing
on the testchart is reduced to such an extent that the satellites are almost completely
absorbed in the main dots, so that the measured position of the ink dot corresponds
to the position of the center of the main dot. Then, it is particularly easy to
adjust the alignment of the printheads in such a way that the main dots printed
with different printheads or in different passes are exactly aligned in the sub-scanning
direction X. It has been found that, in terms of image quality, this type of alignment
is superior to an alignment configuration in which the "centers of mass" of the
ink dots, including the satellites, would be aligned. More particularly, a thin,
one pixel-wide line extending in sub-scanning direction X appears sharper to the
human eye when only the main dots are aligned, regardless of the satellites.
When the method according to the invention is applied to
a printer that shall be used (at least among others) for bi-directional printing,
the test pattern on the testchart is printed while the carriage moves reciprocatingly
in the main scanning direction Y, so that the effects of aberrations in opposite
directions can be detected on the testchart. The alignment of the ink dots may be
corrected either by mechanically adjusting the positions of the printheads on the
carriage or by electronically adjusting the timings with which the nozzles of the
printheads are fired.
Preferred embodiments of the invention will now be described
in conjunction with the drawings, wherein:
- Figs. 1 and 2
- are diagrams illustrating a multi-pass print mode of an ink jet printer;
- Figs. 3 to 5
- are enlarged views of test patterns printed with different alignment conditions
of the printheads;
- Figs. 6 and 7
- are diagrams explaining the effect of a carriage speed on the positions of printed
ink dots; and
- Fig. 8
- is a block diagram of an apparatus suitable for carrying out the method according
to the invention.
Fig. 1 schematically shows a carriage 10 of an ink jet
printer. A number of printheads 12, 14 are mounted on the carriage 10. Although
only two printheads 12, 14 have been shown in the drawing, it shall be assumed here
that the printer is a full color printer having additional printheads intervening
between the two shown printheads 12, 14 and being used for printing the colors cyan,
magenta and yellow, whereas the printheads 12, 14 are used for printing with black
Each printhead 12, 14 has a row of nozzles 16 arranged
in a sub-scanning direction X in which a sheet of a recording medium 18 is advanced
step-wise. The carriage 10 is moved across the recording medium 18 in a main scanning
direction Y normal to the sub-scanning direction X.
In Fig. 1, the carriage 10 moves from left to right, and
the printhead 12 is active, so that some of its nozzles 16 print pixels or ink dots
20 onto the recording medium 18. It is observed that the ink dots 20 form pixel
lines which are separated by gaps 22 having a width of one pixel.
As has only been shown symbolically in Fig. 1, the carriage
10 has a position detector 24 which cooperates with a ruler 26 for detecting the
position of the carriage in the main scanning direction Y. Thus, nozzles of the
printheads can be fired at appropriate timings for printing the ink dots 20 at the
correct positions, in accordance with the image information to be printed. The ruler
26 defines a pixel raster which is symbolized here by raster marks 28 arranged with
a pitch corresponding to exactly the width of one pixel, e.g. 42,33 µm for
an image resolution of 600 dpi.
In Fig. 2, the recording medium 18 has been shifted one
step in X-direction, and the carriage 10 performs a return pass from right to left
in the drawing. During this pass, the printhead 12 is inactive, while all the nozzles
of the printhead 14 are active to print ink dots 30. Some of the dots 30 fill the
gaps between the pixel lines that have been printed in the previous pass. In the
lower part of the printed image, the dots 30 form pixel lines with gaps that will
be filled in during the next pass of the carriage from left to right.
The two printheads 12, 14 must be aligned relative to one
another with high precision. Ideally, the positions of the printheads 12, 14 on
the carriage 10 and/or the timings at which the nozzles of these printheads are
fired should be so adjusted that the (circular) ink dots 20 and 30 are exactly aligned
with one another in the sub-scanning direction X. In practice, however, the ink
dots 20 and 30 do not have an exact circular shape, but are accompanied by satellites
20a and 30a, as has been shown in Fig. 3. These satellites are due to the fact that,
each time an ink droplet has been expelled from a nozzle, at least one smaller ink
droplet is formed and will reach the surface of the print substrate a short time
later. Since the carriage 10 is moving, the satellites are shifted from the main
dots to opposite sides, depending on the direction of movement of the carriage.
When the main dot 20 and its satellite 20a are inspected
visually, without using a microscope, or when the dot position is measured with
a sensor that does not have an extremely high resolution, the main dot and the satellite
appear as a single dot, and the location thereof will be given by the "center of
mass" 32 of the main dot and the satellite.
Thus, when the measured dot positions are used for alignment
of the printheads, the result will be that the centers of mass 32 are aligned, as
is shown in Fig. 3.
However, experience has shown that a single-pixel line
gives a sharper impression if the ink dots are not aligned with their centers of
mass 32, as in Fig. 3, but instead are aligned with the centers 34 of their main
dots, as shown in Fig. 4. A misalignment &Dgr;Y of an individual ink dot 30' has
also been shown (exaggeratedly) in Fig. 4.
The invention provides a method of achieving the alignment
pattern of Fig. 4 without having to measure the dot positions with high resolution.
To this end, a test pattern of ink dots 20, 30, as shown in Fig. 5, is printed on
a testchart 36, with a reduced carriage speed. That is, the speed of the carriage
is reduced in both, the forward pass and the return pass. As a result, the aberration
of the satellites 20a, 30a becomes smaller, and the satellites are completely or
almost completely absorbed in their main dots. Thus, the apparent center of mass
will coincide with the geometric center 34 of the main dot, so that the desired
alignment may be achieved on the basis of the apparent centers of mass. However,
the reduced carriage speed has also an effect on the aberration of the ink dots
20, 30, so that the measured dislocation &Dgr;Y' of the ink dot 30' will be different
from the true misalignment &Dgr;Y in a print process under normal conditions.
It is possible, however, to calculate the true misalignment
&Dgr;Y from the measured dislocation &Dgr;Y', as will be explained in conjunction
with Figs. 6 and 7.
As can be seen in Fig. 1, the raster marks 28 are offset
from the actual positions of the printed ink dots 20 by a half pitch, i.e. a half
pixel width. The distance between the nozzles 16 of the printheads 12 and 14 is
an integral multiple of the pixel width. When the carriage 10 moves to the right,
as in Fig. 1, the nozzles of the printhead 12 are fired each time the position detector
24 passes a raster mark 28. The shift of the ink dots 20 by a half pixel width is
due to an aberration of the ink droplets on their way from the nozzle to the recording
medium 18. When the carriage 10 is moved with the same speed in reverse direction,
as in Fig. 2, the nozzles of the printhead 14 are also fired when the position detector
24 passes a raster mark 28, so that the ink dots 30 are also shifted by a half pixel
width and will thus be aligned with the ink dots 20.
In Fig. 6, the distance between two adjacent raster marks
28(i) and 28(i+1) has been indicated as d. In the forward pass of the carriage,
a signal to fire the nozzles is output when the position detector passes the raster
mark 28(i), while the carriage 10 travels to the right with a speed Vc.
Due to an inevitable time delay t in the electronics for energizing the nozzles
of the printhead 12, the nozzles will have traveled a distance t*Vc until
an ink droplet is actually expelled from the nozzle. A droplet (and its satellite)
moves towards the surface of the recording medium 18 with a speed Vd
and thus travels along a path P20. Thus, the position where the ink dot 20 is formed
on the recording medium 18 is dependent on the speeds Vc the Vd
and on the height h of the nozzle relative to the recording medium.
In the return pass (Fig. 2), the same holds true for the
ink dots 30 that are expelled from the nozzles of the printhead 14, and these ink
dots travel along a path P30. If the printheads 12, 14 are not aligned correctly,
the ink dots 20 and 30 will show the misalignment &Dgr;Y.
Fig. 7 is a corresponding diagram for the test print process,
wherein the speed of the carriage 10 is reduced to V'c and the ink dots
are printed on the testchart 36. When the alignment of the printheads 12, 14 is
the same as in Fig. 6, the resulting dislocation of the ink dots 20 and 30 will
A simple calculation shows that the actual misalignment
&Dgr;Y of the ink dots is related to the measured dislocation &Dgr;Y' by the
Thus, when the carriage speeds Vc and Vc'
are known and the dislocation &Dgr;Y' (as in Fig. 5) is measured, the misalignment
&Dgr;Y can be calculated, and the printheads 12, 14 can be adjusted in order to
correct this misalignment. It is observed that the time delay t, the droplet speed
Vd and the height h do not appear in the above equation, which means
that these quantities need not be known for carrying out the calculation. It should
also be observed that the quantities &Dgr;Y' and &Dgr;Y should be considered
as vectors, i.e. they may also assume negative values.
In this specific embodiment, the alignment pattern of Fig.
4 can be obtained by appropriately adjusting the distance between the printheads
12 and 14 and by adapting the timing control for the printheads such that the nozzles
are fired right at the moment when the position detector 24 passes a raster mark
28. In a more general case, an alignment correction will involve a change in the
timing control for the printheads.
Of course, the adjustment of the printheads achieved in
the way described above will also be beneficial in a single-pass print mode, wherein
the printheads 12 and 14 are used for bi-directional printing of subsequent stripes
of an image, or in a case where the printhead 14 is used as a spare printhead for
compensating nozzle failures in the other printhead 12 or vice versa.
In case of a printer having only a single printhead (per
color) and adapted for bi-directional printing, the dislocation &Dgr;Y' can be
detected, and the misalignment &Dgr;Y can be calculated in an analogous way, and
the alignment correction will the be achieved by delaying or advancing the timings
at which the nozzles are fired in the forward and return passes of the printhead.
Fig. 8 is a block diagram of an apparatus 38 that can be
connected to a printer 40 for carrying out the alignment procedure described above.
A control unit 42 of the apparatus 38 is connected to the printer 40 and measures
or reads the nominal carriage speed Vc that has been programmed in the
printer 40. Then, the control unit 42 controls the printer 40 to reduce the carriage
speed to V'c. Using this reduced carriage speed V'c, the printer
40 prints the test pattern onto the testchart 36.
The apparatus 38 further comprises a (low resolution) opto-electrical
sensor 44 for measuring the dislocation &Dgr;Y' of the ink dots on the testchart
36, a processor 46 for calculating the misalignment &Dgr;Y, and an output unit
48 for outputting the misalignment &Dgr;Y.
Optionally, the output unit 48 may be configured to control
the printer 40, so that the calculated misalignment is printed-out by the printer
40, e.g., directly on the testchart 36. As an alternative, the output unit 48 may
be configured to re-program a timing control unit 50 of the printer 40 in such a
way that the timings, at which the nozzles of the printheads 12, 14 are fired, are
appropriately advanced or delayed relative to the timings when the position sensor
24 passes the raster marks 28, so that the misalignment is corrected electronically.