JP2007152626A - Recording apparatus and recording control method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a recording apparatus and a recording control method capable of accurately landing ink droplets from a recording head on a recording medium and performing high quality recording.
In a recording apparatus that records an image by ejecting ink droplets from a recording head onto a recording medium while reciprocating a carriage on which the recording head is mounted, the following control is performed. That is, in the reciprocal movement of the carriage, the position of the carriage is detected, and the landing position of the ink droplets ejected from the recording head on the recording medium is shifted in the carriage moving direction. At this time, the shift amount of the landing position is determined with reference to the correction table in which the correction value of the landing position is set according to the detected carriage position.
[Selection] Figure 3

Description

  The present invention relates to a recording apparatus and a recording control method, and more particularly to a recording apparatus and a recording control method for performing recording by mounting a recording head according to an ink jet system, for example.

  A serial type ink jet recording apparatus has a recording head mounted on a carriage, and ejects ink droplets onto a recording medium such as paper while reciprocating the carriage to form a desired image. Such a recording apparatus needs to eject ink droplets at an optimal timing from a recording head mounted on a carriage. If there is an error in the timing of ejecting ink droplets, it becomes an error in the landing position, and the image to be printed cannot be faithfully formed on the paper surface.

  Therefore, techniques for correcting such landing errors are conventionally known.

According to this technique, a correction value for accelerating or delaying the ejection timing is input in each of the carriage forward scanning and the backward scanning, and the ejection timing is adjusted based on the correction value in each scanning. Fine-tune the position. Further, the optimum value of the correction value is found based on the landing error value of the actually formed image (adjustment pattern). The recognition error value of the adjustment pattern may be recognized by a person or automatically by the machine. The landing position is affected by parameters such as the distance between the recording head and the paper surface, the carriage speed, the nozzle ejection angle, and the encoder accuracy for detecting the carriage position. Therefore, the position varies in the moving direction of the carriage. This is because each parameter has a property of changing according to the carriage movement direction (hereinafter referred to as Y direction). Therefore, it is necessary to change the parameter of the landing position correction value according to the coordinate value in the Y direction.
Japanese Patent Laid-Open No. 10-329381 JP 2002-120360 A

  However, in the above conventional example, the optimum value of the above parameter cannot be changed according to the Y coordinate, and therefore, when the landing error changes greatly according to the Y coordinate, the image quality deterioration is conspicuous. There is a point.

  For example, the optimal value of the parameter may be “1” at the first coordinate, and the optimal value of the parameter may be “10” at the second coordinate. In this case, according to the conventional technique, it is possible to input the correction parameter “1” and set the landing error at the first coordinate to 0, but the landing error at the second coordinate is “ It can only be 9 ", and optimal landing error correction cannot be realized. For this reason, the image quality near the second coordinate is deteriorated, and ruled line deviation at the time of reciprocal recording is a serious problem.

  The ruled line deviation is, for example, in the reciprocal recording, when the upper half 50% of the ruled line is recorded in the recording in the carriage forward direction, and the lower half 50% of the ruled line is recorded in the recording in the carriage backward direction. This occurs when the landing positions of the lower half 50% are slightly shifted. When this occurs, the recorded ruled lines are not visually attractive, and the recording quality deteriorates.

  Further, for example, in Patent Document 1, a recording apparatus that easily performs recording position alignment between reciprocating scans of a recording head and recording position alignment between a plurality of recording heads without bothering a user or the like, and A recording alignment method has been proposed. However, this technique simply performs recording position alignment (landing error correction) automatically without bothering the user, and an optimum correction value cannot be input at each of the above-described coordinates.

  Further, for example, Japanese Patent Laid-Open No. 2004-228620 proposes to obtain a high-quality image by correcting the amount of deviation of the ink landing position caused by the deviation of the position of the recording head using an electrical signal. However, this technique is also merely related to an optimum parameter selection method, and the landing error correction parameter cannot be changed according to the Y coordinate.

  The present invention has been made in view of the above-described conventional example, and an object thereof is to provide a recording apparatus and a recording control method capable of accurately landing ink droplets from a recording head on a recording medium and performing high-quality recording. And

  In order to achieve the above object, the recording apparatus of the present invention has the following configuration.

  That is, a recording apparatus that records an image by ejecting ink droplets from the recording head onto a recording medium while reciprocating a carriage on which the recording head is mounted, and detects the position of the carriage in the reciprocating movement of the carriage. Detection means; shift means for shifting the landing position of the ink droplets ejected from the recording head on the recording medium in the moving direction of the carriage; and storage means for storing a correction table in which correction values for the landing positions are set. And shift control means for instructing the shift means to shift the landing position while referring to the correction table according to the position of the carriage detected by the detection means.

  Here, the detection means is based on a linear encoder provided along the moving direction of the carriage, a linear encoder sensor provided on the carriage that reads the linear encoder as the carriage moves, and an output signal from the sensor. And a position counter that counts the carriage movement amount.

  Further, the detecting means includes a period counter that counts the period of the output signal from the linear encoder sensor, a period predicting unit that predicts the counted period or the period of the future output signal from the linear encoder, and the predicted value. It is preferable that a trigger generation unit that generates a trigger signal based on a cycle calculated by dividing the cycle is used as a timing trigger that is a unit of an ink discharge cycle from the recording head.

  Furthermore, it is desirable to have an adjustment unit that finely adjusts the ink discharge timing from the recording head by adjusting the ink discharge period from the recording head based on the shift amount of the landing position.

  On the other hand, it is desirable to further include means for generating a signal indicating timing for referring to the correction table corresponding to the value counted by the position counter.

  The recording head has a plurality of ink discharge nozzles, and the plurality of ink discharge nozzles are preferably divided into a plurality of groups and driven in a time division manner in units of the groups.

  According to another invention, there is provided a recording control method for a recording apparatus for recording an image by ejecting ink droplets from the recording head onto a recording medium while reciprocating a carriage on which the recording head is mounted. In the movement, the detection is performed in the detection step for detecting the position of the carriage, the shift step for shifting the landing position of the ink droplet ejected from the recording head on the recording medium in the movement direction of the carriage, and the detection step. And a shift control step of instructing the shift step of a shift amount of the landing position while referring to a correction table in which a correction value of the landing position is set according to the position of the carriage.

  Therefore, according to the present invention, the landing position of the ink droplet that changes in accordance with the position in the moving direction of the carriage is shifted with reference to the correction table, so that printing with high print quality can be performed with less ink droplet landing error. There is an effect that can be.

  Hereinafter, preferred embodiments of the present invention will be described more specifically and in detail with reference to the accompanying drawings.

  FIG. 1 is a partially broken perspective view showing a schematic configuration of a printer equipped with an ink jet recording head (hereinafter referred to as a recording head) which is a typical embodiment of the present invention.

  FIG. 2 is a perspective view showing a detailed configuration of the carriage on which the recording head is mounted and its periphery.

  FIG. 3 is a block diagram showing the control configuration of the printer.

  As shown in FIG. 1, the recording head 1 is mounted on a carriage 200, and the carriage 200 is driven by a carriage motor 204.

  As shown in FIG. 2, the position of the carriage 200 is detected by an optical linear encoder sensor (ENCODER) 206 and an optical linear encoder 207.

  As shown in FIG. 3, each process of the printer is electrically processed by a system controller 214. The system controller 214 includes an SOC (system on chip) 215 that is a custom IC that executes image processing. The SOC 215 includes a CPU 216 for controlling the entire system and an external interface (EXT I / F) 217 for inputting desired image data to be printed from a host PC (HOST PC) 221. A program for controlling the CPU 217 is stored in a flash ROM (FLASH ROM) 218 that is externally attached to the SOC 215. The program is read by a ROM interface (ROM I / F) 219 which is an interface between the flash ROM 218 and the SOC 215. Further, an SDRAM 220 that stores image data and position correction values, a recording head 1 that ejects ink droplets from nozzles based on image data transferred from the system controller 214, and an ENCODER 206 are externally attached to the SOC 215.

  The SOC 215 further includes a memory controller (MEMC) 222 that is an interface unit that performs control to directly access the SDRAM 220. Further, the SOC 215 includes a timing signal generation unit (CRENC) 223 that generates a print reference timing signal serving as a reference timing of the print drive pulse based on an input signal from the ENCODER 206. Further, a print buffer (PBUF) 224 that reads image data from the SDRAM 220 in synchronization with the print reference timing signal from the CRENC 223 is included. Furthermore, an SRAM 225 for buffering image data read from the PBUF 224 is included. Furthermore, a transfer control unit (HSRAMC) 226 that transfers image data stored in the SRAM 225 to the recording head 1 in synchronization with a print reference timing signal generated from the CRENC 223 is included. Furthermore, a drive signal generation unit (HPULSEC) 227 that outputs a head drive pulse to the recording head 1 in synchronization with the print reference timing signal from CRENC 223 is included. The function of each part will be described later.

  As shown in FIG. 2, the printer reciprocates the carriage 200 in the direction of arrow A, and in synchronism with this, the recording head such as recording paper is conveyed in the B direction (conveying direction) shown in FIG. Ink is ejected from 1 to form a desired image on a recording medium.

  FIG. 4 is a diagram showing an output signal waveform from the optical encoder sensor.

  The movement amount of the carriage, that is, the movement amount of the recording head is obtained from the output value of the optical encoder sensor 206. The output signal waveform from the optical encoder sensor 206 is as shown in FIG. 4A when the carriage 200 moves in the forward direction, and shown in FIG. 4B when the carriage 200 moves in the backward direction. It becomes like this. According to the shape of this phase, the position counter built in CRENC 223 counts the count value and determines the position of the carriage 200.

  FIG. 5 is a diagram illustrating the correspondence between encoder signal phases and counter values.

  The CRENC 223 generates a print reference timing signal when the internal position counter value reaches the print area.

  FIG. 6 is a block diagram showing each part of the system controller 214 that is a generation source or a reception destination of the print reference timing signal. As can be seen from FIG. 6, the print reference timing signals are a data set window, a block trigger, a scale trigger, a heat trigger, a heat window, and a data transfer window.

  FIG. 7 is a diagram showing the nozzle arrangement when the recording head 1 is viewed from the ink ejection surface.

  As shown in FIG. 7, nozzle rows corresponding to six ink colors are arranged in the recording head 1, and each color is divided into an A row and a B row. The A row nozzle and the B row nozzle are further divided into an odd (odd) row (O) and an even (even) row (E). That is, in the recording head 1, nozzle rows of 6 colors × 2 × 2 = 24 rows are arranged. Further, as shown in FIG. 7, the carriage 200 is driven in a direction perpendicular to each nozzle row. Although details will be described later, in this printer, the head drive timing of each nozzle row can be driven independently.

<Block division drive>
In an inkjet recording head that discharges ink by converting thermal energy into kinetic energy, all the nozzles in each nozzle row shown in FIG. 7 are not driven simultaneously. This is because, if driven simultaneously, the heater chips of the recording head are driven simultaneously, causing a voltage drop and affecting the print quality. Further, such a driving method is not desirable because the life of the recording head is shortened and the failure rate is increased.

  Therefore, time-division driving of the recording head is conventionally employed. In this method, printing of one nozzle row that should be ejected at the same time is recorded by dividing the nozzles of the nozzle row into a plurality of blocks and driving each block with a short time interval.

  For example, if the number of nozzles in one nozzle row is 640 nozzles, this is divided into 40 blocks. In this case, the number of nozzles in one block is 16 nozzles. Each block is numbered sequentially as block 1, block 2, block 3... Block 40, and is driven in order from block 1. In this case, the maximum number of nozzles that can be driven simultaneously is 16 nozzles, and the voltage drop of the recording head can be minimized to improve the image quality. Further, it is a time-division driving condition that the driving time from block 1 to block 40 is within one column. One column can also be said to be between one dot. In addition, the drive interval of each block must be equal to or greater than the interval in time for data transfer to the recording head.

  For example, in order to drive the 16 nozzles of block 1, image data corresponding to the 16 nozzles must be transferred to the print head, and then the heaters of the print head must be driven simultaneously to drive the 16 nozzles simultaneously. However, if the heater is driven before the transfer of the block 2 is started after the block 1 is driven and the transfer is completed, the print data cannot be recorded faithfully. Therefore, the cycle of the timing signal for driving each block is calculated by CRENC 223 so that it falls within one column, and is generated as a block trigger signal. The block trigger signal is a signal common to each nozzle row and is a serial signal. As the name suggests, this signal indicates the timing and is a trigger signal synchronized with the system CLK. All SOCs are synchronous circuits.

<Quarter split drive, half split drive, full drive>
Further, in the block division driving, how many nozzles of one row are divided is determined by the design of the recording head, and is not determined by the system controller 214. In this embodiment, the recording head 1 is 40-division drive and corresponds to 10-division drive, 20-division drive, and 40-division drive. These are called quarter division driving, half division driving, full driving, and the like.

  In the case of quarter division driving, the period obtained by dividing the correction carriage period by 11 is set as a block trigger period, and in the case of half division driving, the period obtained by dividing the correction carriage period by 22 is set as a block trigger period. Further, in the case of full driving, a period obtained by dividing the correction carriage period into 44 equal parts is set as a block trigger period. In either case, one row of 640 nozzles is divided into 40 blocks and driven in sequence. In the full drive, all 40 blocks are driven during one correction carriage cycle (described later). In the half drive, 20 blocks are driven during one correction carriage cycle, and the remaining 20 blocks are driven in the next one correction carriage cycle. Is done. In the case of quarter division driving, 10 blocks are driven during one correction carriage cycle, 10 blocks are driven during the next one correction carriage cycle, and all 40 blocks are driven during four correction carriage cycles.

<Encoder cycle prediction control>
The recording head 1 may be driven while the carriage 200 is accelerated.

  In this case, since the linear encoder signal does not become a pulse with a constant period, the block trigger signal calculated from the previous period is not necessarily within one column as described above. Therefore, the period fluctuation is predicted to some extent, the next encoder pulse period is predicted from the previous encoder pulse period, and the block trigger pulse period is calculated based on the predicted encoder pulse period.

  FIG. 8 is a diagram showing encoder signals during carriage acceleration and deceleration that perform encoder cycle prediction control. 8A shows an encoder signal when the carriage is accelerated, and FIG. 8B shows an encoder signal when the carriage is decelerated.

  As shown in FIG. 8, it is necessary to count the difference (Tover / Tshort) between the predicted period and the actual period, and to predict the next encoder pulse period from the difference. The predicted encoder pulse period is called a correction carriage period, and the block trigger period is calculated by dividing the correction carriage period by an integer.

  In this embodiment, a 150 dpi linear encoder is used. However, when printing at 600 dpi, an interval obtained by dividing each correction carriage cycle interval into four equals one column interval. Furthermore, the period obtained by dividing the one-column interval into 40 equals becomes the maximum period as the block trigger period. As will be described later, in this embodiment, when 40-division driving is performed, the correction carriage cycle is divided into 44 to form the block trigger cycle.

  Also, during carriage acceleration, as shown in FIG. 8A, when correcting the speed fluctuation of the pulse period, the overlap time is subtracted when the next periodic pulse comes before the pulse interval time calculated from the previous period. The next cycle is calculated. Carriage acceleration pulse control is a function for correcting this because the previous cycle is longer than the next pulse cycle during carriage acceleration recording. Based on the correction carriage cycle, the scaletrg cycle and the blktrg cycle are calculated.

  On the other hand, when the carriage decelerates, as shown in FIG. 8B, when the speed fluctuation of the pulse period is corrected, the overlap time is added when the next periodic pulse comes after the lapse of the pulse interval time calculated from the previous period. To calculate the next cycle. Carriage deceleration pulse control is a function for correcting this because the previous cycle is longer than the next pulse cycle during carriage acceleration recording. Based on the correction carriage cycle, the scaletrg cycle and the blktrg cycle are calculated. Note that scaletrg is a reference scale pulse, which is a pulse that serves as a reference scale for the recording position. Blktrg is a block trigger.

<Scale trigger>
Most commercially available linear encoders are 150 dpi or 300 dpi, and the minimum resolution obtained from the encoder sensor is 600 dpi and 1200 dpi, respectively. Therefore, in order to perform printing at a higher resolution, it is necessary to divide the above-described correction carriage cycle equally, but in the embodiment, a scale trigger that is driven at a cycle obtained by dividing the correction carriage cycle by n is provided. Generate.

  FIG. 9 is a diagram showing how the scale trigger is generated.

  In particular, FIG. 9 illustrates that the correction carriage cycle is calculated by dividing it into n. The value of n is specifically shown. Note that the minimum division driving number of the recording head is 10 division driving.

  The scale trigger (scaletrg) is a signal serving as a reference for printing timing. The period of the scale trigger driven in such a divided period is exactly one column section with 11 pulses in the dynamic registration mode. In the static registration mode, 10 pulses are exactly one column section.

  FIG. 10 is a diagram showing the relationship between the block trigger and the scale trigger.

  The scale trigger cycle is equal to the block trigger cycle during 10-division driving in the static registration mode. In the dynamic registration mode, it is equal to the block trigger period in 11-division driving. The difference from the block trigger is that the block trigger continues to be output in the calculated cycle, whereas the scale trigger outputs only a predetermined number of times in one correction carriage cycle period. The predetermined number of times is equal to the value of n shown in FIG.

  FIG. 10 shows a waveform when the carriage is decelerated. There is a period during which the scale trigger is not output during carriage deceleration. This section is a non-printing area, but since it is actually a very small section, the printing quality is not greatly affected. FIG. 10 is for the dynamic registration mode, and the block trigger resolution is a period obtained by dividing one column by a multiple of 11. In the static registration mode, the block trigger resolution is a period obtained by dividing one column by a multiple of 10.

<Driving before block trigger>
As described above, the block trigger period is basically 11 divisions, 22 divisions, and 44 divisions, but this embodiment can generate block triggers with a finer period. This is a function called front-end driving, and by making the block trigger cycle finer, printing for one column can be executed more faithfully. For example, this function is useful when outputting a print image such as CAD. It is.

  If it halves from the block trigger period of 11 divisions, the block trigger period of 22 divisions, the block trigger period of 44 divisions, the block trigger period of 88 divisions, the block trigger period of 176 divisions, the block trigger period of 352 divisions Is easy to calculate. Although it has been described above that the recording head is driven at the timing of the block trigger, the recording head is actually driven when a block trigger is output and a certain signal is active. One signal is a heat window. The heat window becomes active at the printing timing for each column and is negated when the block trigger is output a predetermined number of times. The predetermined number of times is 40 at full drive, 20 at half drive, and 10 at quarter drive.

<Data transfer to the recording head (data transfer window, data transfer window generation trigger, head drive pulse>
The data transfer window is a signal corresponding to each nozzle row. In this embodiment, there are 24 nozzle rows, so there is a data transfer window corresponding to each. That is, the data transfer window is a 24-bit signal. Transfer of print data to the head is started at the timing when the data transfer window is active and the block trigger is output. The HSRAMC 226 receives a data transfer window and a block trigger from the CRENC 223, and starts transferring image data stored in the SRAM 225 at the reference timing of the block trigger. This transfer is serial transfer. In this case, 16 bits of image data for one block is one unit. The block select 6-bit information is added to this, and the dummy information 2 bits is further added, so that a total of 24 bits of data is serially transferred to the recording head 1. As with the heat window, the data transfer window is active for 40 block triggers during full drive, 20 block triggers during half drive, and 10 block triggers during quarter drive.

  The output timing is the timing of the data transfer window generation trigger shown in FIG. The data transfer window generation trigger starts output after Δ (described later) +2 columns from the start position of the data set window (described later), and the output cycle is usually 11 scale trigger cycles. As will be described later, the output cycle may be a 10 scale trigger cycle or a 12 scale trigger cycle at the timing of executing the landing position correction. As the name suggests, the data transfer window generation trigger is a signal that is the output timing of the data transfer window. When a data transfer window generation trigger is output, the data transfer window becomes active and becomes active for the predetermined period. The heat window is a signal obtained by delaying the data transfer window by two block triggers.

  FIG. 11 shows the relationship between the data set window and the encoder address.

  According to FIG. 11, the data transfer window becomes active at the timing corresponding to the encoder address C, and data transfer to the recording head 1 is started. Further, the heat window becomes active at the timing corresponding to the encoder address D, and the drive of the recording head 1 is started. That is, the print data is transferred from the system controller 214 to the recording head 1 at a timing corresponding to the encoder addresses C to D.

  FIG. 17 is a time chart showing waveforms of signals at the time of data transfer to the recording head.

  The transfer data 24 bits are transferred in synchronization with the data transfer clock (clk). In the transfer, data is latched in a register in the recording head at both edges of the clock signal (clk), and the latched data is input to the heater drive circuit of the head at the timing of the head latch signal. Then, the heater driving circuit corresponding to the active input data is turned on by the input of the head driving pulse, and ink is ejected. The generation of the head drive pulse is processed by the HPULSEC 227 inside the SOC.

  As shown in FIG. 17, the timing of the head drive pulse is determined based on the block trigger signal. Each parameter of PT01, PT02, and PT03 is held in a register inside HPULSEC 227, and a head drive pulse is generated based on the timing of the block trigger signal based on the register value. Of course, these registers can be read and written from the CPU 216.

<Data set window>
The timing at which the encoder address shown in FIG. 11 corresponds to A is the timing at which the data set window opens. The timing of opening the data set window is compared with the value of the readable / writable data set window start position register in CRENC 223, and the data set window is opened when the register value = the position counter value. Thereby, the print start position can be freely set by the CPU 216. A data set window start position register is provided for each nozzle row. When the value of the position counter is equal to each register value, each data set window is opened. If an encoder with a resolution of 150 dpi is used, the position counter has a resolution of 600 dpi. Therefore, the start position of the data set window can be set with a resolution of 600 dpi.

  FIG. 12 is a time chart of each signal related to the data set window (DSWIN).

  It is CRENC 223 that compares the start position of the data set window with the value of the internal position counter and processes opening and closing of the data set window. The start position register is 16 bits of 15: 0, and the start position can be set up to about 109 inches (inch). A value set in the start position register is also called a start address or a print start address.

<Print mode setting register>
Data transfer to the recording head 1 must be changed according to the output resolution and the number of block divisions. The system controller 214 includes a print mode setting register in CRENC 223 of the SOC 215. The print mode setting register is readable / writable from the CPU 216, and can set an arbitrary value.

  FIG. 13 is a diagram showing the configuration of the print mode setting register.

  Data transfer as set in this register is executed for the recording head 1. D10 to D8 are output resolutions. The output resolution is a factor that determines the scale trigger cycle. That is, as described regarding the scale trigger, the value of n shown in FIG. 9 is determined according to the output resolution. There is also an exception for prohibited writing. For example, when D18-D16 = “000”, that is, when the block trigger resolution is 1/11 of the DTWIN generation trigger period, 20-division drive or 10-division drive cannot be performed. "Only can be written. For the same reason, when D18-D16 = "001", only quarter drive and half drive are available, and only D1-D0 = "10" or D1-D0 = "01" can be written.

<Column counter>
FIG. 12 shows a data transfer window generation trigger (DTWIN generation trigger) when printing is performed at a resolution of 4800 dpi. Since the size of an image to be printed can be measured by counting the number of DTWIN generation triggers, CRENC 223 has a counter that counts the number of DTWIN generation triggers. Like the start address, it has a data set window end position register. The value of the column number counter is compared with the value of the data set window end position register. When the column number counter value = the data set window end position register value, the data set window is closed. The end position register is 19 bits of 18: 0, and the end position can be set up to about 109 inches (inch) at the time of 4800 dpi printing. The data set window end position register value is also referred to as an end address or a print end address.

<Static cash register adjustment mode>
The printer of this embodiment has a dynamic registration mode and a static registration mode. The CRENC 223 has a 1-bit register for switching between the static registration mode and the dynamic registration mode. When the value of this register is 0, the dynamic registration mode is set and when it is 1, the static registration mode is set.

  In the dynamic registration mode, the block trigger resolution value determined according to the values of D18 to D16 in the print mode setting register shown in FIG. 13 is a multiple of 11 divisions. On the other hand, in the static registration adjustment mode, the value of the block trigger resolution determined according to the value of D18-D16 of the print mode setting register is a multiple of 10 divisions. The block trigger resolution is the period of the block trigger (blktrg), and FIG. 10 shows the block trigger waveform in the dynamic registration mode.

  The block trigger period in the static registration mode is longer by 11/10 than the block trigger period in the dynamic registration mode, and there is a margin in data transfer to the recording head 1 accordingly. As described above, from one block trigger to the next block trigger, one block of image data and block selection information total of 24 bits must be transferred serially. Therefore, when the carriage 200 is driven at high speed to perform high speed printing, the static registration mode is more advantageous.

  The dynamic registration mode is a mode in which registration adjustment is performed in correspondence with the encoder address, and is a main feature of the present invention. Only in the dynamic registration mode, correction parameters for correcting ink droplet landing errors in the Y coordinate direction that change according to the coordinate (Y coordinate) value in the carriage movement direction are input corresponding to each coordinate. Based on this, the discharge timing is finely adjusted to correspond to each coordinate. The static registration mode is a function for high-speed printing.

<Landing position correction value (Δ1, δdyn)>
Δ1 shown in FIG. 11 indicates a static component of the landing position correction. If the coordinates in the carriage movement direction of the ink ejection position when the landing position correction is not performed are y, the actual ejection position coordinates are expressed by Expression (1).

Y = y + Δ1 + δdyn (y) (1)
Δ1 is a value determined by writing from the CPU 216 to the landing position correction value register of CRENC 223.

  Further, δdyn (y) is always “0” in the static registration mode, and can take other values only in the dynamic registration mode. Further, δdyn (y) has a property as a function of y, and is a variable that changes according to the encoder address.

  FIG. 14 is a diagram showing the relationship between the Δ value and the encoder address. FIG. 14 shows an example of the behavior of Δ expressed as Δ = Δ1 + δdyn (y). The meaning of Δ is the distance that the carriage 200 moves from the opening of the data set window to the start of data transfer to the recording head 1, and one scale trigger is a unit.

  Δ is a register value, which is the value of the landing position correction value register held in CRENC 223. The landing position correction value register is a 16: 0 register of 15: 0 assigned to each nozzle row, and is set in units of scale trigger intervals. That is, the setting unit of Δ is equal to the resolution of the scale trigger. For example, when the printing mode is 600 dpi and the dynamic registration mode, the resolution of the scale trigger is 11 times the resolution of 600 dpi and 6600 dpi, so the setting unit of Δ is 6600 dpi. In the 1200 dpi printing mode and the dynamic registration mode, the resolution of the scale trigger is 11 times the resolution of 1200 dpi and 13200 dpi, so the setting unit of Δ is 13200 dpi. In the 2400 dpi print mode and the static registration mode, the resolution of the scale trigger is 10 times the resolution of 2400 dpi and 24000 dpi, so the setting unit of Δ is 24000 dpi.

  Note that Δ can take a value of −32767 to +32767, but it is basically necessary to consider that the value is always 0 or more. If the print window is opened at a position where Δ is negative, printing must be started with a delay of Δ from the start position of the print window, but this is impossible. Actually, since the landing position correction value is a value indicating how much the actual printing is delayed with respect to the printing start address, if this value is negative, it is regarded as 0 and 0 delay from the start address (ie, Printing without delay). In this way, printing is executed assuming that all of the minus (−) landing position correction values are zero.

  For this reason, it is necessary to set the initial value of the landing position correction value, that is, the initial value (Δ1) of the landing position correction value register to a sufficiently large value so that the minimum value of the correction value is 0 or more.

  Since the actual print position = print start address + Δ (during forward scanning), the print start address needs to be changed according to the value of Δ1. In the first place, Δ1 is a conventional static registration component, and Δ is a value obtained by adding a dynamic component to Δ1. As described above, Δ = Δ1 + y can be expressed, and y is a variable value that is updated by adding a correction value update value acquired from the SDRAM 220 for each ADJTRG (described later) input. The component of Δ1 includes an odd-even registration adjustment value, and the component of Δ includes Δ1, a bidirectional registration adjustment value, and a color registration adjustment value. The nozzles in the odd and even rows cannot refer to different correction value tables, but different values can be set as the initial value Δ1.

  Further, δdyn (y) is a cumulative value obtained when the SOC 215 accesses the correction value update value stored in the SDRAM 220 and adds the value suggested by the correction value update value to the register value of the landing position correction value register.

<ADJWIN, ADJTRG>
FIG. 15 is a time chart showing the timing of updating the correction value (Δ).

  In FIG. 15, the landing position correction value is equal to the value of the ENCADJ unit position correction register and can be monitored from the CPU 216.

  Δ is the value of the landing position correction value register as described above, and the value is determined by writing by the CPU 216. That is, the above Δ1 is determined by the writing of the CPU 216.

  δdyn (y) is a variable that changes when the correction value update value of the SDRAM 220 is read and accumulated at a timing suggested by ADJWIN and ADJTRG. ADJWIN indicating the timing for updating δdyn (y) is a timing signal for starting correction value update. As with the data set window, ADJWIN sets the start position in a register, compares the value of the ADJWIN start position register with the value of the position counter, and ADJWIN becomes active when the two match.

  Further, unlike the data set window, ADJWIN determines the end position by comparison with a position counter. The value of the ADJWIN end position register is compared with the value of the position counter, and ADJWIN is closed when they match. Further, although the data set window is provided for each nozzle row, ADJWIN is one signal. Similarly, ADJTRG is only one signal.

  The CRENC 223 includes a register for setting the ADJTRG signal that is output in a cycle in which the ADJTRG signal is set when the ADJWIN is active, and can be set in units of 300 dpi.

  FIG. 15 shows an ADJTRG waveform when the ADJTRG cycle has a resolution of 300 dpi. ADJTRG is output in synchronization with the edge of the A phase signal. When output at a resolution of 300 dpi, ADJTRG is output in synchronization with both edges of the A-phase signal. As shown in FIG. 15, the landing position correction value register value is updated in synchronization with the output of ADJTRG. However, the value is updated with reference to the SDRAM correction value update table as shown in FIG.

  FIG. 16 is a diagram showing a correction value update table and how correction values are updated.

  The correction value update table shown in FIG. 16 stores either “01” to +1 the landing position correction value register value, “10” to −1, or “11” or “00” that is not changed. ing. When ADJTRG is output, the landing position correction value register value is updated by referring to the correction value update table two bits at a time. Updates are performed one by one.

  As shown in FIG. 14, the value of the correction value update register value from the ADJWIN start position is continuously updated according to the correction value update table of the SDRAM 220, and the value is maintained after the end of ADJWIN. Normally, the ADJWIN start position and end position are set so that the data set window is included from the ADJWIN start position to the ADJWIN end position.

  In the example shown in FIG. 16, the landing position correction value register refers to the value of the correction value update value register 2, and each time ADJTRG is input, the reference destination bit (the value of the correction value update value buffer (d127: 126)) The correction value update value buffer adds the add value of the correction value update table according to the value of the correction value update value buffer, and each time ADJTRG is input, the lower bits are shifted to the upper bits by two bits. Is updated every time ADJTRG is input, and refers to all 128 bits when ADJTRG is input 64 times.The value of the correction value update register is the value of the correction value update table of SDRAM 220 every time ADJTRG is input 64 times. Is read and updated.

  Also, as shown in FIG. 16, the dynamic registration correction value is read from the correction value table of the SDRAM 220 in synchronization with the ADJTRG timing. On the other hand, the buffer referred to by the landing position correction value register is not the correction value update registers 1 to 12 that can be read from the CPU 216, but the value of the correction value update value buffers 1 to 12 that latches the values of the correction value update registers 1 to 12. is there. Since the correction values stored in the correction value update value buffers 1 to 12 become empty when ADJTRG is input 64 times, the correction value at the next address is read from the correction value table of the SDRAM 220 while ADJTRG is input 64 times. It is necessary to complete the lead.

<Correction of landing position>
The position where the head drive pulse shown in FIG. Since the period (PT02) from the input of the block trigger to the output of the head drive pulse is a very short interval, it should be approximated that the output position of the AND (AND) pulse of the block trigger and the heat window is Y. Can do.

  In FIG. 17, if PT00 = PT01 <PT02, a single pulse is generated because no prepulse occurs. The pulse width of the data latch of the recording head is programmable from 1 to 16 clocks. In order to avoid the latch signal, the pulse inversion timing needs to be specified by adding a margin of at least 25 clocks from the falling edge of the block trigger. In this example, 25 clocks × 1 ÷ 166 = 0.15 μs.

  That is, the position where the pulse of the logical product of the block trigger and the heat window is output is shifted by Δ1 + δdyn (y), and ink is ejected at the corrected ejection position Y. Since the block trigger is a trigger that is continuously output, shifting the output position of the logical product pulse of the block trigger and the heat window is equivalent to shifting the position of the heat window. Furthermore, shifting the heat window is equivalent to shifting the data transfer window. The unit for shifting is the scale trigger unit, which is also an integer multiple of the block trigger period. The output position of the data transfer window generation trigger is shifted in units of scale triggers, ink droplets are ejected at the corrected ejection position Y, and landing position correction is executed.

  FIG. 18 is a timing chart of a data transfer window, a heat window, a scale trigger, and a block trigger at the start of printing.

  The data transfer window is generated using the transfer window generation trigger as a reference timing. However, the data transfer window is not opened during the first two pulses from B to C in FIG. 18, that is, from the data set window start position. During that time, the image data to be transferred from the SDRAM 220 to the recording head 1 is read at the timing of the first heat trigger. Δ corresponding to between A and B in FIG. 18 is as described above.

  Since Δ is a scale trigger unit, the discharge position can be shifted in scale trigger units according to the value of Δ. In addition, as described above, Δ must be a positive value. The interval C to D in FIG. 18 is a 2-block trigger interval, and the image data transfer to the recording head 1 is performed serially during this interval, as described above.

  Further, Δ is a variable that changes from the ADJWIN start position to the end position, but FIG. 18 shows that the output start position of the data transfer window generation trigger is delayed by Δ by the value of Δ at the position A. Further, after B in FIG. 18, in order to correct the ejection position, the period of the data transfer window generation trigger is changed and executed. Specifically, at the timing when ADJTRG is output, the correction value update value in the landing position correction value update table is referred to, and the cycle is accelerated or delayed based on that value.

  FIG. 19 is a time chart showing how the data transfer window (DTWIN) is generated.

  As shown in FIG. 19, the MSB2 bit of the correction value update register 2 shown in FIG. 16 is referenced at the output timing of ADJTRG, and the period of the DTWIN generation trigger changes according to the reference value. When the reference value is “00” or “11”, the cycle of the DTWIN generation trigger does not change, and the scale trigger 11 pulse is the cycle. When the reference value is “10”, the DTWIN cycle is set to a 10-scale trigger interval for only one cycle. When the reference value is “01”, the DTWIN cycle is set to 12 scale trigger intervals for only one cycle.

  In this way, the ejection position can be shifted by one pulse of the scale trigger each time ADJTRG is output. As described above, the DTWIN generation trigger continues to be output while the data set window is active. When the print column number counter matches the end address +2, the data set window is closed and the output of the DTWIN generation trigger is stopped. The print column number counter counts the number of DTWIN generation triggers as described above. Since the first two DTWIN generation triggers are not ejected as shown in FIG. 18, the print column count value-2 is the number of columns actually ejected. That is, if the size in the Y direction of the image to be printed is S column, the end address is set so that S = end address. When the relationship of end address = print column count value−2 is established, it is determined that printing is finished, and the data set window is closed.

  In the dynamic registration adjustment mode, the data transfer window (DTWIN) becomes DTWIN ≦ 1 when the DTWIN generation trigger which is an internal signal of ENCADJC becomes “1”. Also, it is held for n periods (n is 10 for quarter drive, 20 for half drive, 40 for full drive) in the corresponding recording mode, and DTWIN ≦ 0 unless the next DTWIN generation trigger is input. Become.

  The DTWIN generation trigger is driven in synchronization with the scale trigger of the corresponding recording mode. On the other hand, the heat trigger is output after delaying the DTWIN generation trigger by one clock. The DTWIN generation trigger is driven when the internal print window signal is “1”.

There are three types of DTWIN generation trigger cycles: “L (long)”, “N (neutral)”, and “S (short)”. Normally, the cycle is driven with a cycle “N (neutral) = 1 column cycle. If there is an ADJTRG input, one cycle is driven in the cycle “L”, cycle “N”, or cycle “S” according to the reference correction value as follows:
Period “L” = corresponding recording mode scale trigger 12 pulses> 1 column Period “N” = corresponding recording mode scale trigger 11 pulses = 1 column Period “S” = corresponding recording mode scale trigger 10 pulses Minute <1 column.

  FIG. 20 is a diagram showing a timing chart at the end of printing.

  Conventionally, the address corresponding to the end address has been set by the value of the position counter. However, when the landing position is corrected, the Y coordinate value for closing the data set window changes according to the value taken by Δ. The method of specifying with a counter is not very desirable. On the other hand, if the end address is not set by the position counter but is set by the number of columns, the end address is uniquely determined by the size of the image data regardless of the value of Δ.

  In the dynamic registration adjustment mode, the data transfer window (DTWIN) is generated from the data set window at the end of printing. The bit display is “1” = ENABLE and “0” = DISABLE. In the dynamic registration adjustment mode, the data set window is closed when the value of the column number counter becomes equal to the value of the print end address + 2. The column number counter counts the number of DTWIN generation triggers.

<Image data read trigger>
As shown in FIGS. 18 to 20, the image data read trigger is output in synchronization with the DTWIN generation trigger.

  As shown in FIG. 6, the image data read trigger is input to the PBUF 224, and the PBUF 224 reads the image data from the SDRAM 220 when the image data read trigger is input. One column of data is read per image data read trigger. Image data is read from the SDRAM 220 using the image data read trigger as a timing trigger. The image data read out to the recording head 1 is transferred in synchronization with the timing of the block trigger when DTWIN is active, which is output in synchronization with the DTWIN generation trigger corresponding to the image data readout trigger. The Therefore, the PBUF 224 may have a memory for buffering image data for three columns. This buffer is constituted by the SRAM 225 inside the SOC 215.

<Method of determining Δ>
FIG. 21 is a diagram for explaining the cause of the landing error.

  FIG. 21A is a diagram showing a bi-directional landing error. There is a difference between the position where the ink droplet ejected when the carriage 200 moves in the forward direction and the position where the ink droplet ejected when the carriage 200 moves in the backward direction. Say error.

  The bidirectional landing error is a value that depends on the carriage moving speed, the distance between the paper surface of the recording medium and the nozzles of the recording head, and the ink ejection speed.

  FIG. 21B is a diagram showing an inter-color landing error. This is a landing error caused by a difference in landing points between different ink colors, that is, between nozzle rows. The intercolor landing error is a value that depends on the ejection speed of the ink droplets and the speed of the carriage.

  There are also landing errors that occur due to factors other than the above two types, but any landing error that occurs due to any factor is the landing error corresponding to the Y coordinate. By changing so as to absorb the error, it is possible to make the landing error as close to zero as possible.

  FIG. 22 is a diagram illustrating an example of landing position correction.

  The landing error is affected by parameters between the carriage moving speed and the recording head and the paper surface of the recording medium, and exhibits a complicated behavior. In order to reduce this landing error to zero, as shown in FIG. 22, it is necessary to perform landing correction whose sign is opposite to that of the landing error.

  However, in reality, it is very difficult to obtain an ideal landing correction value that makes the landing error zero at all coordinates of printing. Therefore, in this embodiment, as shown in FIG. 22, the landing correction value is obtained at a plurality of representative points (A, B, C,... In the figure), and an approximate value is predicted at an intermediate coordinate point. A realistic method is adopted.

  In addition, it is desirable for the plurality of representative points to be equally spaced as indicated by A, B, C... In FIG. 22, but there is no particular limitation thereto.

  However, if the distance between the representative points is too large, the ideal landing correction value and the predicted value are greatly deviated, so that the landing error cannot be brought close to zero. On the other hand, the waveform of the landing error has a frequency component equal to the pitch between the recording head and the paper surface of the recording medium, so it seems that the interval between the representative points is about 1/2 to 1/4 of the pitch. It is. The representative point is preferably near the maximum value of the landing error. From such a viewpoint, it is desirable that the coordinates of the representative point be matched with the peak and valley of the pitch between the recording head and the recording medium. .

  According to the embodiment described above, correction parameters for correcting ink droplet landing errors that change according to the position of the carriage in the movement direction are input corresponding to each coordinate, and the correction value including the input correction value Create a table in SDRAM. Then, by reading the correction value table and finely adjusting the ejection timing corresponding to each coordinate based on the value, the ink droplet landing error can be made as close to zero as possible at all coordinates. Thereby, it is possible to perform printing with high print quality by landing ink droplets at ideal landing positions.

  As a result, the landing error is small at one coordinate, but the landing error is large at another coordinate and the image quality is not beautiful.

  For example, when printing is performed in a state where the recording paper is cockled (a state where the paper surface is wavy), the landing error is also a parameter having a periodicity with respect to the Y coordinate according to the pitch of the cockling. . In the present invention, the correction parameter for absorbing the landing error is input in correspondence with the Y coordinate, and the discharge timing caused by cockling is finely adjusted based on the input parameter in correspondence with each coordinate. It is also possible to eliminate the problem.

  Also, in a printer equipped with a recording head having a large number of nozzles, recording is often performed by ejecting ink droplets even while the carriage is accelerating or decelerating. During such acceleration / deceleration of the carriage, acceleration is applied to the recording head, and the direction of the nozzle changes slightly, so that a landing error that changes according to the Y coordinate value occurs. In the present invention, such an error can be eliminated as well. In addition, since the landing error correction parameter can be input for each color, landing deviation between colors can be similarly corrected in all coordinates.

1 is a partially broken perspective view showing a schematic configuration of a printer equipped with an ink jet recording head which is a typical embodiment of the present invention. FIG. 3 is a perspective view illustrating a detailed configuration of a carriage on which a recording head is mounted and its periphery. FIG. 2 is a block diagram illustrating a control configuration of a printer. It is a figure which shows the output signal waveform from an optical encoder sensor. It is a figure which shows a response | compatibility with the phase of an encoder signal, and a counter value. FIG. 3 is a block diagram illustrating each part of a system controller that is a generation source or a reception destination of a print reference timing signal. FIG. 3 is a diagram illustrating a nozzle arrangement when the recording head is viewed from the ink ejection surface. It is a figure which shows the encoder signal at the time of the carriage acceleration and deceleration which perform encoder period prediction control. It is a figure which shows the mode of a scale trigger production | generation. It is a figure which shows the relationship between a block trigger and a scale trigger. It is a figure which shows the relationship between a data set window and an encoder address. It is a time chart of each signal regarding a data set window (DSWIN). 3 is a diagram illustrating a configuration of a print mode setting register. It is a figure which shows the relationship between (DELTA) value and an encoder address. It is a time chart shown about the timing of correction value (delta) update. It is a figure which shows the mode of a correction value update table and correction value update. 6 is a time chart showing waveforms of signals at the time of data transfer to the recording head. 6 is a timing chart of a data transfer window, a heat window, a scale trigger, and a block trigger at the start of printing. It is a time chart which shows the mode of a data transfer window (DTWIN) production | generation. It is a figure which shows the timing chart at the time of completion | finish of printing. It is a figure explaining the cause of a landing error. It is a figure which shows an example of landing position correction | amendment.

Explanation of symbols

1 Recording head 200 Carriage 200
204 Carriage motor 206 Optical linear encoder sensor (ENCODER)
207 Optical linear encoder 214 System controller 215 SOC
216 CPU
217 External interface (EXT I / F)
218 Flash ROM
219 ROM interface (ROM I / F)
220 SDRAM
221 Host PC
222 Memory Controller (MEMC)
223 Timing signal generator (CRENC)
224 Print buffer (PBUF)
225 SRAM
226 Transfer control unit (HSRAMC)
227 Drive signal generator (HPULSEC)

Claims (7)

A recording apparatus for recording an image by ejecting ink droplets from the recording head onto a recording medium while reciprocating a carriage mounted with the recording head,
Detecting means for detecting the position of the carriage in the reciprocating movement of the carriage;
Shift means for shifting the landing position of the ink droplets ejected from the recording head on the recording medium in the moving direction of the carriage;
Storage means for storing a correction table in which correction values for the landing positions are set;
A recording apparatus comprising: a shift control unit that instructs the shift unit to shift the landing position while referring to the correction table according to the position of the carriage detected by the detection unit. The detection means includes
A linear encoder provided along the moving direction of the carriage;
A linear encoder sensor provided on the carriage for reading the linear encoder as the carriage moves;
The recording apparatus according to claim 1, further comprising: a position counter that counts a movement amount of the carriage based on an output signal from the linear encoder sensor. The detection means further includes
A period counter that counts the period of the output signal from the linear encoder sensor;
A period predicting unit that predicts a period of a future output signal from the linear encoder from the period counted by the period counter;
A trigger generation unit that generates a trigger signal based on a cycle calculated by dividing the cycle predicted by the cycle prediction unit;
The recording apparatus according to claim 2, wherein the trigger signal is a timing trigger that is a unit of an ink discharge period from the recording head.   4. The apparatus according to claim 3, further comprising an adjusting unit that finely adjusts an ink discharge timing from the recording head by adjusting an ink discharge period from the recording head based on a shift amount of the landing position. Recording device.   The recording apparatus according to claim 2, further comprising a reference timing generation unit configured to generate a signal indicating a timing for referring to the correction table corresponding to a value counted by the position counter. The recording head has a plurality of ink discharge nozzles,
The recording apparatus according to claim 1, wherein the plurality of ink discharge nozzles are divided into a plurality of groups and are time-division driven in units of the groups. A recording control method for a recording apparatus for recording an image by ejecting ink droplets from the recording head onto a recording medium while reciprocating a carriage on which the recording head is mounted,
A detection step of detecting the position of the carriage in the reciprocating movement of the carriage;
A shift step of shifting the landing position of the ink droplets ejected from the recording head on the recording medium in the movement direction of the carriage;
A shift control step of instructing the shift amount of the landing position to the shift step with reference to a correction table in which a correction value of the landing position is set according to the position of the carriage detected in the detection step. Recording control method.

Images