Sheet conveying apparatus, printing apparatus, correction information acquiring apparatus, printing system, method of conveying sheets and method of acquiring correction information

[0001] 1. Field of the Invention

[0002] The present invention relates to a printing apparatus and a conveying control method. Specifically, the invention relates to a technique to apply a correction value to correct an error in conveying a printing medium used in an inkjet printing apparatus.

[0003] 2. Description of the Related Art

[0004] An inkjet printing apparatus has a print head that has a fine-nozzle array, and ink is ejected from each nozzle in accordance with printing data. The ejected ink forms dots on the printing medium to form an image. Accordingly, to form a high-quality image, it is important that the dots should be formed on the printing medium at intended positions. The displacement of the dot-formation position has to be avoided as much as possible. Some of the various causes of such displacement deviation are: difference in shape amongst the nozzles of the print head; noise factors, such as the vibrations of the apparatus that occur while the printing is being carried out; and the distance between the printing medium and the print head. The inventors of the present invention have discovered that one of the significant causes for such displacement deviation of the dot-formation position is the lack of accuracy in conveying the printing medium. One of the commonly used conveying unit for the printing medium is a roller (a conveying roller). Conveying the printing medium by a desired distance can be achieved by rotation of the conveying roller by a designated angle with the conveying roller being pressed onto the printing medium. Here, the accuracy in the conveying of the printing medium depends, to a significant extent, on the eccentricity of the conveying roller.

[0005] FIGS. 47, 48A and 48B, and 49 illustrate cross sectional shapes of various conveying rollers. The conveying roller of FIG. 47 has a perfectly-circular cross-sectional shape, and has its central axis aligned exactly with its rotational axis. The conveying roller of FIGS. 48A and 48B has a cross-sectional shape that is not a perfect circle. The conveying roller of FIG. 49 has its rotational axis offset from its central axis.

[0006] Assume such a case as shown in FIG. 47, or, to be more specific, a case where the cross-sectional shape of the conveying roller is a perfect circle and where the central axis of the conveying roller is aligned exactly with its rotational axis. In addition, further assume that the rotational angle to convey the printing medium is uniform. Then, every rotation of the conveying roller by an angle R constantly gives a particular length (L0) in the circumferential directions (length of arc). Accordingly, every position within the conveying roller always gives a uniform amount of conveying the printing medium that is conveyed while being in contact with the conveying roller.

[0007] Contrasting outcomes are obtained by a conveying roller with an ellipsoidal cross-sectional shape as ones shown in FIGS. 48A and 48B. Such a conveying roller gives different amount of conveying even when the conveying roller rotates by the same angle R. This difference in the amount of conveying depends on the rotational position of the conveying roller. To be more specific, for the rotational position shown in FIG. 48A, the printing medium is conveyed by an amount L1 while for another rotational position shown in FIG. 48B, the printing medium is conveyed by an amount L2. Here, the lengths L0, L1, and L2 has such a relationship as L1>L0>L2. That is to say, a periodical variation in amount of conveying the printing medium occurs, and the variation depends on the period of the conveying roller.

[0008] Alternatively, as in the case of FIG. 49, the offsetting of the rotational axis of the conveying roller from the central axis O that is intended to be the rotational axis may sometimes cause the amount of conveying the printing medium to vary periodically in response to the period of the conveying roller. To be more specific, assume cases where the rotational axis is offset from the central axis O and is positioned at either the point A or the point B shown in FIG. 49. In these cases, the same rotational angle α produces different amounts of conveying. Such difference in conveying amount results in a periodical variation in the conveying of the printing medium. Here, the variation depends on the period of the conveying roller.

[0009] The eccentricity of the roller, which has been mentioned above, includes these above-described states. Specifically, included are a state where the roller has a cross-sectional shape that is not a perfect circle, and a state where the conveying roller has its rotational axis offset from its central axis. In the case of an ideal accuracy being achieved in conveying, the image should be printed in such a way as shown in the schematic diagram of FIG. 50A. With the above-mentioned eccentricity, however, the printed image will be an uneven image with stripes that appear periodically in the conveying direction as shown in FIG. 50B while the period is the same as the amount of conveying corresponding to a full rotation of the conveying roller.

[0010] The amount of eccentricity for the conveying roller is usually controlled so as to stay within a certain range. The stricter the standard for the amount of eccentricity is, the lower the yielding of the conveying roller becomes. Accordingly, the printing apparatus thus produced becomes more expensive. For this reason, an excessively strict standard for the amount of eccentricity is not preferable.

[0011] To address the above-mentioned problem, various measures have been proposed. Different correction values for the conveying errors are set for different phases of the conveying roller so that even an eccentric conveying roller can achieve a steady amount of conveying as similar to the case of a conveying roller with a perfectly-circular cross-sectional shape and with its rotation axis being aligned exactly with its central axis (Japanese Patent Laid-Open No. 2006-240055 and Japanese Patent Laid-Open No. 2006-272957). To be more specific, correction to reduce the amplitude of the fluctuation in amount of conveying with a period equivalent to the circumferential length of the conveying roller can be done by applying a periodic function with the same period and reversed polarity.

[0012] Besides the eccentricity that is mentioned above, variations in outer circumference, or outer diameter, of the roller is another important cause for lowering the accuracy in conveying. With such variations in outer diameter of the roller, rotation of a roller by a rotation angle determined for a roller with a reference outer diameter will not produce a predetermined amount of conveying. Specifically, use of a roller with an outer diameter that is larger than the standard outer diameter produces a larger amount of conveying while use of a roller with an outer diameter that is smaller than the standard outer diameter produces a smaller amount of conveying. Accordingly, even when the amplitude of the variation is reduced by the above-described correction, the range of variations which is maximum and which exceeds a certain amount of conveying error causes unevenness that appears in the image. This means that to achieve the printing of a high-quality image without unevenness requires not only the lowering of the influence of the eccentricity but also the lowering of the influence of the variations in the outer diameter of the conveying roller.

[0013] An example of the techniques to achieve the printing of a high-quality image reducing unevenness is disclosed in Japanese Patent Laid-Open No. 2002-273956. In the disclosed technique, the correction value to correct the conveying error caused by the variations in the outer diameter of the conveying roller (correction value for outer-diameter) is acquired. Also acquired is the correction value to correct the conveying error caused by the eccentricity (correction value for eccentricity).

[0014] In some cases, however, in the course of printing images on a single printing medium, different kinds of rollers, and/or different numbers of rollers are used for conveying the printing medium in order to print images on different areas of the printing medium. For example, in some cases, during the printing of the front-end portion of the printing medium, only the conveying roller that supports the front end of the printing medium is used for the conveying of the printing medium without using the roller that supports the opposite end of the printing medium. The conveying error in this case may have a different value from the conveying error in the case of printing the central area of the printing medium. Actually, each roller may have a unique eccentricity. Nevertheless, if correction is done uniformly for the entire areas of the printing medium by using a single correction value for eccentricity and a single correction value for outer-diameter, some of the correction attempts done in such a way may result in inappropriate correction for some areas of the printing medium. As a consequence, the evenness across the areas may possibly be lowered.

[0015] In summary, according to the technique disclosed in the above-mentioned document, the correction value for eccentricity and the correction value for outer-diameter can be acquired by setting the correction values for each phase of the conveying roller. However, in a case where the kinds of rollers used for the conveying of the printing medium are switched from one to another during the printing, unevenness may sometimes appear on the image by continuous and uniform use of the same correction values across printing areas including ones corresponding to the conveying by a different roller used for conveying from the one for which the correction values have been acquired. In this case, the unevenness may appear in the areas corresponding to the conveying by the different roller used for that purpose.

[0016] It is desirable to address one or more of the above-described problems. It is also desirable to carry out correction that reflects the kinds and the number of rollers used for conveying actually involved in the printing of each of the printing areas. It is also desirable to contribute to achievement of the printing of a high-quality image.

[0017] To achieve these objects, an aspect of the present invention provides sheet conveying apparatus comprising: a plurality of rotary conveying elements operable individually or in combination to convey sheets along a conveying path in at least first and second conveying phases; and conveying correction means operable during said first conveying phase to employ first correction information to correct for conveying errors in that phase, and operable during said second conveying phase to employ second correction information, different from said first correction information, to correct for conveying errors in that phase.

[0018] According to the above-described configuration, the conveying control is executed by applying appropriate correction values for correcting conveying errors in accordance with the roller actually involved in the conveying. Consequently, the configuration can contribute to achievement of the printing of a high-quality image.

[0019] Further features of the present invention will become apparent from the following description of exemplary embodiments (with reference to the attached drawings).

[0071] Hereafter, the present invention will be described in detail with reference to accompanying drawings.

(1) Configuration of Apparatus

[0072] FIG. 1 is a schematic perspective view illustrating the entire configuration of an inkjet printing apparatus according to an embodiment of the present invention. When the printing is carried out, a printing medium P is held by and between a conveying roller 1—one of the plural rollers provided in the conveying path—and pinch rollers 2 that follow and are driven by the conveying roller 1. The printing medium P is guided onto a platen 3 by rotations of the conveying roller 1. The printing medium P is conveyed in a direction indicated by the arrow A in FIG. 1 while being supported on the platen 3. Though not illustrated in FIG. 1, a pressing member, such as a spring, is provided to elastically bias the pinch rollers 2 against the conveying roller 1. The conveying roller 1 and the pinch rollers 2 are components of a conveying unit on the upstream side.

[0073] The platen 3 is disposed at the printing position opposite to the face on which ejection openings are formed in a print head 4 provided in the form of an inkjet print head (hereafter the face is referred to as “ejection face”). The platen 3 thus disposed supports the back side of the printing medium P to keep a constant, or a predetermined, distance between the top surface of the printing medium P and the ejection face.

[0074] Once the printing is carried out on the printing medium P that has been conveyed onto the platen 3, the printing medium P is conveyed in the direction A, being held by and between a discharging roller 12 that rotates and spurring rollers 13 that follow and are driven by the discharging roller 12. The printing medium P is thus discharged out onto an output tray 15. The discharging roller 12 and the spurring rollers 13 are components of conveying unit on the downstream side. It should be noted that only a single pair of the discharging roller 12 and the line of spurring rollers 13 is shown in FIG. 1, but that two pairs of them may be provided as will be described later.

[0075] A member 14 is disposed by one of the side ends of the printing medium P, and is used to set the reference line when the printing medium P is conveyed (the member will, therefore, be referred to as “conveying reference member 14”). Any printing medium P, irrespective of the width thereof, is conveyed with the above-mentioned side of the printing medium along the reference line set by the conveying reference member 14. Besides the role of setting the reference line, the conveying reference member 14 may also serve the purpose of restricting the rising-up of the printing medium P towards the ejection face of the print head 4.

[0076] The print head 4 is detachably mounted on a carriage 7 with its ejection face opposing to the platen 3, or the printing medium P. The carriage 7 is driven by a driving source—a motor—to reciprocate along two guide rails 5 and 6. The print head 4 may perform ink-ejection action during the reciprocating movement. The direction in which the carriage 7 moves is orthogonal to the direction in which the printing medium P is conveyed (in the direction indicated by the arrow A). Such a direction is usually referred to as “main-scanning direction” while the direction in which the printing medium P is conveyed is usually referred to as “sub-scanning direction.” The printing of images on the printing medium is carried out by repeating the alternation of main scan (printing scan) of the carriage 7, or the print head 4, and the conveying of the printing medium P (sub scan).

[0077] As the print head 4, for example, a print head that includes an element for generating thermal energy to be used for ejecting ink (an example of such element is a heat-generating resistor element) may be employed. The thermal energy causes a change in the state of the ink (that is, film boiling of the ink occurs). As another example, a print head that includes, as an element for generating energy, an element to generate mechanical energy may be employed. An example of such an element is a piezo element. The mechanical energy thus generated is used for the ejection of the ink.

[0078] The printing apparatus of this embodiment forms an image with pigment inks of ten colors. The ten colors are: cyan (C), light cyan (Lc), magenta (M), light magenta (Lm), yellow (Y), first black (K1), second black (K2), red (R), green (G), and gray (Gray). When a term “K-ink” is used, either the first black (K1) ink or the second black (K2) ink is mentioned. Here, the first and the second black inks (K1 and K2) may, respectively, be a photo black ink that is used to print a glossy image on glossy paper and a matt black ink suitable for matt coated paper without gloss.

[0079] FIG. 2 schematically illustrates the print head 4 used in this embodiment, and the print head 4 is viewed from the side of the nozzle-formed face. The print head 4 of this embodiment has two printing-element substrates H3700 and H3701, in each of which nozzle array for five colors of the above-mentioned ten colors formed. Each of the nozzle arrays H2700 to H3600 corresponds to each one of the ten different colors.

[0080] Nozzle arrays H3200, H3300, H3400, H3500, and H3600 are formed in one of the two substrates—specifically in the printing-element substrate H3700—to perform ink ejection with respective inks of gray, light cyan, the first black, the second black and light magenta being supplied to. Meanwhile, nozzle arrays H2700, H2800, H2900, H3000 and H3100 are formed in the other one of the two substrates—specifically, in the printing-element substrate H3701—to perform ink ejection with respective inks of cyan, red, green, magenta and yellow being supplied to. Each of the nozzle arrays is formed by 768 nozzles arranged in the direction of conveying the printing medium P at intervals of 1200 dpi (dot/inch) and ejects ink droplets each of which is approximately 3 picoliters. Each nozzle has an ejection opening with an opening area of approximately 100 μm².

[0081] The above-described head configuration enables what is termed as “one-pass printing” to be carried out. In this way of printing, the printing on a single area of the printing medium P is completed in a single main scanning. However, what is termed as “multi-pass printing” is also possible for the purpose of improving the printing quality by reducing the negative influence of the nozzles that are formed with lack of uniformity. In this mode of printing, the printing on a single scanning area of the printing medium P is completed by carrying out main scanning plural times. When the multi-pass printing is selected, the number of passes is determined appropriately by taking account of conditions, such as the mode of printing.

[0082] Plural ink tanks corresponding to colors of inks to be used are detachably installed in the print head 4, independently. Alternatively, the inks may be supplied to the print head 4 via respective liquid-supply tubes from the corresponding ink tanks fixed somewhere in the apparatus.

[0083] A recovery unit 11 is disposed so as to be able to face the ejection face of the print head 4. The recovery unit 11 is disposed at a position within the area that the print head 4 can reach when the print head 4 moves in the main scanning direction. The position is located outside of side-edge portion of the printing medium P, or of the platen 3. That is, the position is in an area where no image is to be printed. The recovery unit 11 has a known configuration. Specifically, the recovery unit 11 includes a cap portion for capping the ejection face of the print head 4, a suction mechanism for sucking the inks with the ejection face being capped to force the inks out of the print head 4. A cleaning blade to wipe off the tainted ink-ejection face, among other members, is also included in the recovery unit 11.

[0084] FIG. 3 illustrates an example of the configuration for the principal portion of the control system for the inkjet printing apparatus according to this embodiment. A controller 100 controls each portions of the inkjet printing apparatus according to this embodiment. The controller 100 includes a CPU 101, a ROM 102, an EEPROM 103, and a RAM 104. The CPU 101 performs various arithmetic processing and determination for processing related to the printing action and the like including processing procedures that are to be described later. In addition, the CPU 101 performs the processing related to the print data and the like. The ROM 102 stores the programs corresponding to the processing procedures that are executed by the CPU 101, and also stores other fixed data. The EEPROM 103 is a non-volatile memory and is used to keep predetermined data even when the printing apparatus is switched off. The RAM 104 temporarily stores the print data supplied from the outside, and the print data developed in conformity with the configuration of the apparatus. The RAM 104 functions as a work area for the arithmetic processing performed by the CPU 101.

[0085] An interface (I/F) 105 is provided to connect the printing apparatus to an outside host apparatus 1000. Communications in both directions based on a predetermined protocol is carried out between the interface 105 and the host apparatus 1000. It should be noted that the host apparatus 1000 is provided by a known form, such as a computer. The host apparatus 1000 serves as a supply source of the print data on which the printing action of the printing apparatus of this embodiment is based. In addition, a printer driver—the program to cause the printing apparatus to execute the printing action—is installed in the host apparatus 1000. To be more specific, from the printer driver, the print data and the print set-up information, such as the information on the kind of printing medium P on which the print based on the print data is performed are sent. Also sent therefrom is the control command that causes the printing apparatus to control its action.

[0086] A linear encoder 106 is provided to detect the position of the print head 4 in the main-scanning direction. A sheet sensor 107 is provided in an appropriate position in the path of conveying the printing medium P. By detecting the front end and the rear end of the printing medium P with this sheet sensor 107, the conveying position (sub-scanning position) of the printing medium P can be determined. Motor drivers 108 and 112 and a head-driving circuit 109 are connected to the controller 100. The motor driver 108, under the control of the controller 100, drives a conveying motor 110, which serves as the driving source for conveying the printing medium P. The drive power is transmitted from the conveying motor 110 via a transmission mechanism, such as gears, to the conveying roller 1 and the discharge roller 12. The motor driver 112 drives a carriage motor 114, which serves as the driving source for the movement of the carriage 7. The drive power is transmitted from the carriage motor 114 via a transmission mechanism, such as a timing belt, to the carriage 7. The head-driving circuit 109, under the control of the controller 100, drives the print head 4 to execute the ink-ejection.

[0087] A rotary encoder 116 is mounted on each of the shafts of the conveying roller 1 and the discharge roller 12. Each of the rotary encoders 116 detects the rotational position and the speed of the corresponding roller so as to control the conveying motor 110.

[0088] A reading sensor 120 is provided to serve as detector for detecting the density of the images printed on the printing medium P. The reading sensor 120 may be provided in the form of a reading head mounted on the carriage 7 either along with or in place of the print head 4. Alternatively, the reading sensor 120 may be provided as an image-reading apparatus constructed as a body that is independent of the printing apparatus shown in FIG. 1.

(2) Outline of the Processing

[0089] In the printing apparatus with the above-described configuration, one of the biggest causes for the lowering of the accuracy in conveying is the eccentricity of a roller. The eccentricity of a roller is defined as a state where the rotational axis of a roller is offset from the central axis of the roller, that is, a state in which the axis of the rotational center of a roller deviates from the geometrical central axis of the roller. In addition, the eccentricity is defined as a state where the roller has a cross-sectional shape that is not a perfect circle. The eccentricity of a roller causes a periodical conveying error, and the period depends on the rotational angle from the reference position of the roller. Assume that such eccentricity exists. In this case, even when the roller is rotated by an equal angle, the length in the circumferential direction (lengths of arc) corresponding to the equal-angle rotation varies from one time to another. As a result, an error occurs in the amount of conveying the printing medium P. An error that occurs in this way prevents the formation, in the direction of conveying the printing medium P, of the dots in positions in which the dots are originally supposed to be formed. Dots are formed densely in some areas, and sparsely in others, in the direction of conveying the printing medium P. In summary, unevenness of printing occurs with a period equivalent to the amount of conveying corresponding to a full rotation of the roller.

[0090] Another example of the big causes for the lowering of the accuracy in conveying is a cause that derives from the error in the outer diameter of a roller. Assume that such an error in the outer diameter of a roller exists. In this case, even when the roller is rotated by a rotational angle that has been determined for a roller with a certain reference outer diameter, a predetermined amount of conveying which is supposed to be obtained cannot always be obtained. To be more specific, when a roller with an outer diameter that is larger than the reference outer diameter is used, the amount of conveying becomes larger than what is supposed to be. In this case, white stripes are likely to occur in the printed image. In contrast, when a roller with an outer diameter that is smaller than the reference outer diameter is used, the amount of conveying becomes smaller than what is supposed to be. In this case, black stripes are likely to occur in the printed image.

[0091] In view of what has just been described above, this embodiment of the present invention aims to provide a configuration that is capable of reducing variations in positions of dot formation, which derives from the lack of accuracy in conveying due to such causes as the eccentricity of the conveying roller 1 and of the discharge roller 12 as well as the errors in outer diameter of these rollers. For this purpose, in this embodiment, a first correction value is acquired to reduce the negative influence of the eccentricity of the rollers (hereafter, the first correction value is referred to as “correction value for eccentricity”). In addition, a second correction value is acquired to reduce the negative influence of the outer-diameter error (hereafter, the second correction value is referred to as “correction value for outer-diameter”). Then, these correction values are used to control the rotation of the rollers, or to be more precise, to control the driving of the conveying motor 110 when the printing is actually carried out.

[0092] FIG. 4 is a flowchart illustrating the outline of processing procedures to acquire the correction value for eccentricity and the correction value for outer-diameter. In this procedure, firstly, preparation for the start of printing action including the setting and the feed of the printing medium P is done (step S9). When the printing medium P is conveyed to a predetermined position for the printing, test patterns are printed (step S11). With these test pattern, simultaneous detection of the errors in the amount of conveying caused by both the eccentricity and the outer-diameter error (hereafter, also referred to as “conveying error”) is possible, and detail descriptions of the test patterns will be given later.

[0093] Subsequently, the test pattern is read using the reading sensor 120, and the information on the density of the test pattern is acquired (step S13). Then, on the basis of this density information, the acquiring of the correction value for eccentricity (step S15) and the acquiring of the correction value for outer-diameter (step S17) are carried out in this order.

(3) Test Pattern

[0094] FIG. 5 illustrates an example of the test patterns used in this embodiment. In this embodiment, test patterns used to detect the conveying error caused by the conveying roller 1 and test patterns used to detect the conveying error caused by the discharge roller 12 are formed side by side with each other in a direction, which is corresponding to the direction of conveying the printing medium P, that is, in the sub-scanning direction. Two test patterns are formed side by side with each other in a direction corresponding to the direction of the rotational axis of each roller, that is, in the main-scanning direction. One of the two test patterns is formed in a position near the conveying reference member 14, and the other is formed in a position far from the conveying reference member 14, so as to detect the conveying errors of the corresponding roller in the respective positions. To be more specific, in FIG. 5, a test pattern FR1 is provided to detect the conveying error of the conveying roller 1 in a position near the conveying reference member 14, and a test pattern ER1 is provided to detect the conveying error of the discharge roller 12 in a position near the conveying reference member 14. In addition, a test pattern FR2 is provided to detect the conveying error of the conveying roller 1 in a position far from the conveying reference member 14, and a test pattern ER2 is provided to detect the conveying error of the discharge roller 12 in a position far from the conveying reference member 14.

[0095] Now, some of the reasons why the test patterns for both the conveying roller 1 and the discharge roller 12 are printed will be given in the paragraphs that follow.

[0096] In the printing apparatus according to this embodiment, conveying units are respectively provided at the upstream and the downstream sides, in the direction of conveying the printing medium P, of the position where the printing is executed by the print head 4 (printing position). Accordingly, the printing medium P can be in any one of the following three states: first, the printing medium P is supported and conveyed by the upstream-side conveying unit alone: second, the printing medium P is supported and conveyed by the conveying units on both sides (FIG. 6A); and third, the printing medium P is supported and conveyed by the downstream-side conveying unit alone (FIG. 6B).

[0097] The conveying roller 1 and the discharge roller 12 have their respective main functions that are different from each other. So, the conveying accuracy of the conveying roller 1 frequently differs from that of the discharge roller 12. The main function of the conveying roller 1 is to set the printing medium P, for each stage of the printing scan action, in an appropriate position for the print head 4. Accordingly, the conveying roller 1 is formed with a roller diameter that is large enough to carry out the conveying action with relatively high accuracy. In contrast, the main function of the discharge roller 12 is to discharge the printing medium P with certainty when the printing on the printing medium P is finished. So, most frequently, the discharge roller 12 cannot rival the conveying roller 1 in the accuracy of conveying the printing medium P.

[0098] As evident from what has been described above, when the conveying roller 1 is actually involved in the action of conveying the printing medium P, the conveying accuracy for the conveying roller 1 affects the error of conveying the printing medium P. When, in contrast, only the discharge roller 12 is involved in the action of conveying the printing medium P, the conveying accuracy for the discharge roller 12 affects the error of conveying the printing medium P.

[0099] That is why, in this embodiment, the printing medium P is divided into two areas—an area I and an area II—as shown in FIG. 7. For the printing on the area I, the conveying roller 1 is involved in the conveying action. Meanwhile, the printing medium P is conveyed by the discharge roller 12 alone when the printing is done on the area II. The test patterns are printed while the printing medium P is conveyed by the rollers that are mainly involved in the conveying action for the printing on the respective areas I and II. From each of the test patterns, information on the density is acquired, and thus the correction values that are used in the actual printing of the respective areas are acquired. Incidentally, the printing apparatus according to this embodiment is designed to be capable of printing an image with no margins, that is, “marginless printing” in the front-end portion or in the rear-end portion of the printing medium P. The correction value is usable when the marginless printing is performed in the rear-end portion of the printing medium P. For this reason, acquiring the correction value for the occasion where the printing medium P is conveyed by the discharge roller 12 alone is useful.

[0100] FIG. 6B illustrates a state where the printing apparatus performs an actual printing action with the printing medium P being conveyed by the downstream-side conveying unit alone. In this case, the area where the test patterns used for detecting the conveying error of the discharge roller 12—specifically, the test patterns ER1 and ER2—are printed is limited to the area II. So, to secure an enough area to be used for this purpose, a state shown in FIG. 6C—the state where the printing medium P is conveyed by the downstream-side conveying unit alone—can be artificially created by releasing the pinch rollers 2 when the printing of the test patterns FR1 and FR2 is finished. This releasing may be done manually. Alternatively, the releasing action may be automatically executed by the printing apparatus configured as such.

[0101] When the printing medium P is conveyed by both the conveying roller 1 and the discharge roller 12, the conveying accuracy for the conveying roller 1 has a predominant influence on the conveying error. For this reason, the entire printing area is divided into such two areas as described above. However, the conveying error in a case where the conveying roller 1 alone is involved in the conveying of the printing medium P (printing is performed on the front-end portion of the printing medium P) may differ from the conveying error in a case where both the conveying roller 1 and the discharge roller 12 are involved in the conveying. Then, the area corresponding to both of the above-mentioned cases may be divided further into smaller portions to be processed independently.

[0102] To be more specific, as shown in FIG. 8, the area I can be, firstly, divided into two portions—a portion corresponding to the conveying done by the conveying roller 1 alone and another portion corresponding to the conveying done by both the conveying roller 1 and the discharge roller 12. Then, test patterns are printed individually in both portions, and the density information and the correction values are acquired for each of the portions. In this case, to secure enough space to print test patterns corresponding to the state where the printing medium P is conveyed by the conveying roller 1 alone, the spurring rollers 13 may be designed to be released from the discharge roller 12.

[0103] Now, some of the reasons why the test patterns for each of the conveying roller 1 and the discharge roller 12 are formed both in a position near the conveying reference member 14 and in a position far from the conveying reference member 14 will be given in the following paragraph.

[0104] Assume that each roller is manufactured within a predetermined design tolerance. Even in this case, the conveying error that derives from such factors as the amount of eccentricity and the state of eccentricity may sometimes differ between a position on the side of the printing apparatus near the conveying reference member (a position on the conveying-reference side) and a position on the side thereof far from the conveying reference member (a position on the non-conveying-reference side). Rollers, which are used in a large-scale inkjet printing apparatus that can print on an A3-sized (297 mm×420 mm) or larger printing medium P, tend to have such a difference that is more prominent than those used in other types of apparatus. A possible way to minimize the difference in the conveying error between a position on the conveying-reference side and a position on the non-conveying-reference side is that a single test pattern is printed in the central position in the main-scanning direction, that is, in the longitudinal direction of the roller, and then a correction value is acquired from the information on the density of the test pattern. In this embodiment, however, plural test patterns are printed in the main-scanning direction (for example, two test patterns are printed in this embodiment, but three, or more, are also allowable). Then, having compared those printed test patterns, a correction value is selected so as to reduce most the negative influence of the conveying error on the test pattern that is affected most prominently by the corresponding conveying error (this will be described later).

(4) Details of Test Pattern

[0105] Each of the test patterns shown in FIG. 5 is formed in the following way.

[0106] FIG. 9 is an explanatory diagram for describing the way how the nozzles are used when the test patterns are formed. When the test patterns are formed, by using, amongst the 768 nozzles included in the nozzle array H3500 for the second black ink, for example, a nozzle group NU that consists of a part of the 768 nozzles consecutively formed on the upstream side in the conveying direction and another nozzle group ND that consists of a part of the 768 nozzles consecutively formed on the downstream side in the conveying direction. The nozzle groups NU and ND are located with an in-between distance that is equal to each amount of conveying between every two printing scans multiplied by the number of printing scans done until patch elements, which are to be described later, are laid over each other. In this embodiment, the nozzle group located on the downstream side (the nozzle group ND) is made to be the nozzle group for reference, and 128 nozzles located in a range from the 65th to 193rd nozzle counted from the nozzle located in the most downstream position are used, in a fixed manner, to print plural patch elements for reference RPEs (first patch elements). The nozzle group located on the upstream side (the nozzle group NU) is made to be the nozzle group for adjustment. The number of nozzles, amongst the nozzle group NU, to be used is 128, which is the same number of nozzles to be used amongst those in the nozzle group ND. However, the range of nozzles of the nozzle group NU is shifted by one nozzle during the main scan. In this way, plural patch elements for adjustment APEs (second patch elements) are printed.

[0107] FIGS. 10A to 10E are explanatory diagrams for describing the way how the test patterns, or the patches constituting the test patterns, are formed by using the upstream-side nozzle group NU and the downstream-side nozzle group ND. Firstly, patch elements for adjustment is formed in a main scan at a certain conveying position (that is, by the first main scan), then printing medium P is conveyed by an amount corresponding to 128 nozzles, and thereafter patch elements for adjustment are further formed. When the above-described series of actions are repeated, the first ones of the patch elements for adjustment thus formed reach the position where the downstream-side nozzle group ND is located at the time of the fifth main scan. By forming patch elements for reference at this position, patches that are used to acquire the density information (the kind of patches of the first line) are completed.

[0108] Likewise, at the sixth main scan, the patch elements for adjustment formed at the second main scan reach the position where the downstream-side nozzle group ND is located. By forming patch elements for reference at this position, patches of the second line are completed. Patches of the third line onwards are formed in a similar way, and thus plural lines of patches are completed in the sub-scanning direction.

[0109] The above descriptions show that, to complete the patches, four times of conveying the printing medium P are necessary to be carried out between the scan to form the patch elements for adjustment and the scan to form the patch elements for reference. Accordingly, each of the patches reflects the conveying error caused by the sector of the roller used in the four times of conveying the printing medium P, which are carried out between the scan having formed the patch elements for adjustment and the scan having formed the patch elements for reference.

[0110] FIGS. 11A and 11B illustrate, respectively, a group of patch elements for reference printed by a single main scan and a group of patch elements for adjustment printed likewise. As FIG. 11A shows, the patch elements for reference RPEs are printed neatly in a line in the main-scanning direction. In contrast, FIG. 11B shows that when the patch elements for adjustment APEs are printed, each of the patch elements for adjustment APEs is shifted by a pitch corresponding to one nozzle. The group of patch elements for adjustment APEs includes a reference APE APEr that is printed by using 128 nozzles located in a range from the 65th nozzle to the 193rd nozzles that are counted from the nozzle located in the most upstream position.

[0111] The patch elements for adjustment APEs that are located closer to the conveying reference member 14 than the reference APE APEr are shown at the left side of the reference APE APEr in FIG. 11B. Each such patch element for adjustment APE is printed by using the nozzle group for adjustment NU, but the range of nozzles used to print a patch element for adjustment is shifted, by one nozzle towards the downstream side of the conveying, from the range of nozzles used to print the adjacent patch element for adjustment APE that is located at the right side thereof. The patch elements for adjustment APEs that are located farther from the conveying reference member 14 than the reference APE APEr are shown at the right side of the reference APE APEr in FIG. 11B. Each such patch element for adjustment APE is printed by using the nozzle group for adjustment NU, but the range of nozzles used to print a patch element for adjustment is shifted, by one nozzle towards the upstream side of the conveying, from the range of nozzles used to print the adjacent patch element for adjustment APE that is located at the left side thereof. The range of nozzles is shifted by 3 nozzles for the conveying-reference side and by 4 nozzles for the non-conveying-reference side. When the shifting towards the upstream side is denoted as positive, the range of shifting, as a whole, is from −3 to +4.

[0112] Now, assume that the printing medium P is conveyed between two main scans, without any error, by a distance corresponding to a range of 128 nozzles arranged at a pitch of 1200 dpi (128/1200×25.4=2.709 [mm]). Then, the patch elements for reference RPEs that are printed at the fifth main scan is laid exactly over the reference APE APEr (shifting amount=0) printed at a main scan after the printing medium P is conveyed four times. Note that a positive amount of shifting corresponds to a case where the amount of conveying is larger than the above-mentioned distance while a negative amount of shifting corresponds to a case where the amount of conveying is smaller than the above-mentioned distance.

[0113] FIG. 12 illustrates a test pattern including plural patch elements, or including a group of patches each of which is composed of a patch element for reference and a patch element for adjustment. FIG. 12 illustrates, in an enlarged manner, one of the four test patterns shown in FIG. 5.

[0114] With the reference APE APEr, patch elements for adjustment APEs are printed by with the nozzles actually used for printing being shifted, by one nozzle, from the respective adjacent ones within a range from −3 to +4 nozzles. Accordingly, in each test pattern, 8 patches are formed in the main-scanning direction. In addition, the amount of conveying the printing medium P, in this embodiment, between each two main scans is set at 2.709 mm (as an ideal value). Main scans are repeatedly carried out 30 times in total to form 30 patches across the range in the sub-scanning direction (in the direction of conveying the printing medium P). Accordingly, each test pattern has a length in the sub-scanning direction of 2.709×30=81.27 mm (as an ideal amount). When a roller has, nominally, an outer diameter of 37.19 mm, the above-mentioned length of the test pattern corresponds to more than twice the circumference of the roller.

[0115] A patch column A shown in FIG. 12 includes the reference patch elements for adjustment APErs. Each of patch columns marked with A+1 to A+4 includes patch elements for adjustment APEs printed with the used range of the nozzle group for adjustment NU being shifted towards the upstream side in the direction of conveying the printing medium P from the reference patch elements for adjustment APErs by an amount corresponding to 1 nozzle to 4 nozzles. Each of patch columns marked with A−1 to A−3 includes patch elements for adjustment APEs printed with the used range of the nozzle group for adjustment NU being shifted towards the downstream side in the direction of conveying the printing medium P from the reference patch elements for adjustment APErs by an amount corresponding to 1 nozzle to 3 nozzles.

[0116] Patch rows B1 to B30 are formed with different sectors of the roller used to convey the printing medium P between the scan to form each patch element for adjustment APE and the scan to form the corresponding patch element for reference RPE. Assume that the conveying of the printing medium P after the printing of the patch element for adjustment APE of the patch row B1 is carried out from a reference position of the roller. In this case, for the patch row B1, the sector of the roller used between the scan to print the patch element for adjustment (APE) and the scan to print the patch element for reference (RPE) corresponds to a sector of the roller used to convey the printing medium P four times (0 mm to 10.836 mm) starting from the reference position of the roller. For the patch row B2, the sector of the roller used between the scan to print the patch element for adjustment (APE) and the scan to print the patch element for reference (RPE) corresponds to a sector of the roller used to convey the printing medium P four times (2.709 mm to 13.545 mm) starting from a position away from the reference position by 2.709 mm. Likewise, for the patch row B3, a sector of the roller (5.418 mm to 18.963 mm) is used, while for the patch row B4, another sector of roller (8.127 mm to 21.672 mm). In this way, for the different patch rows, different sectors of the roller are used between the scan to print the patch element for adjustment (APE) and the scan to print the patch element for reference (RPE).

[0117] In addition, patch rows that are adjacent to each other share, partially, a sector of the roller to be used between the scan to print the patch element for adjustment (APE) and the scan to print the patch element for reference (RPE). For example, both of the patch rows B1 and B2 use a common sector of the roller (2.709 mm to 10.836 mm).

[0118] The position of conveying after the printing of the patch element for reference (RPE) of the patch row B1 may be aligned with the reference position of the roller. In the formation of the test pattern, however, no such control as to make the above state accomplished is necessary. Alternatively, the conveying position after the printing of the patch element for reference of the patch row B1 may be printed and may be used as the reference to acquire the relations between the patch rows (positions to be used within a roller) and the conveying error, which relations are to be described later.

(5) Details of Patch

[0119] FIG. 13 illustrates the patch element for reference or the patch element for adjustment in enlarged manner. In FIG. 14, the patch element is illustrated in a further enlarged manner. The patch element is formed in a stair-shaped pattern with print blocks, as base units, each of which has dimensions of 2 dots in the sub-scanning direction and 10 dots in the main-scanning direction. In addition, a certain distance in the sub-scanning direction between each two stair-shaped patterns is secured by taking account of the range for shifting the group of nozzles to be used. In the example shown in FIG. 14, the group of nozzles to be used is shifted by 1 to 4 nozzles towards the upstream side of the conveying direction (+1 to +4) and by 1 to 3 nozzles towards the downstream side in the conveying direction. In response to this, a space of 6 nozzles is secured in the sub-scanning direction.

[0120] In this embodiment, such a patch element as shown in this drawing is printed in the upstream-side nozzle group NU and in the downstream-side nozzle group ND as well. Accordingly, the state of overlaying of the patch element for reference (RPE) and the patch element for adjustment (APE) is changed in response to the degree of conveying errors. As a result, in the test pattern, patches of various densities are formed as shown in FIG. 12.

[0121] Specifically, when the patch element for adjustment (APE) printed by the upstream-side nozzle group NU and the patch element for reference (RPE) printed by the downstream-side nozzle group ND are aligned exactly with each other as shown in FIG. 15A, the density (OD value) becomes low. In contrast, when these patches are misaligned as shown in FIG. 15B, the space that is supposed to be blank is filled, so that the density becomes high.

[0122] The reliability of the test pattern has to be enhanced so that the conveying error can be detected from the information on the density of the test pattern. To this end, it is preferable that the state of the nozzles of the print head 4 be less likely to affect the patches. In nozzles that are used continuously or used under certain conditions, such ejection failure as deflection in the ejection direction (dot deflection) and no ejection of ink may sometimes occur. When such ejection failure brings about a change in the information on the density of the patches, the correction value for conveying error can be calculated only incorrectly. It is, therefore, strongly desirable that patches to be formed are capable of reducing the change in information on the density even with the existence of such ejection failure as mentioned above. The patch element employed in this embodiment can respond such a demand. The reason for this will be described in the following paragraphs by using a simple model.

[0123] The patch element is formed in a pattern with spaces in the sub-scanning direction as shown in FIG. 16A so that the amount of offset in positions can be measured as the information on the density. However, when a particular nozzle does not eject any ink at all, all the area that is supposed to be printed with the particular nozzle becomes blank as shown in FIG. 16B.

[0124] To address the problem, the patch element is formed, as shown in FIG. 17A, of plural print blocks also with spaces placed between two adjacent blocks arranged in the main-scanning direction. In addition, the range of used nozzles is dispersed so that the patterns may not be adjacent to each other amongst print blocks. Thus, the negative influence of a particular nozzle on the pattern can be reduced. Specifically, even when there is ejection failure of a particular nozzle, a blank area, the blank area being produced because the patch elements for reference (RPEs) and the patch elements for adjustment (APEs) are not aligned with one another, is reduced (the example in FIG. 17B has half a blank area of that in FIG. 16B). Accordingly, the density of the patch elements, and eventually, that of the patch itself, can be prevented from being lowered. The pattern in FIG. 17B has an area factor (proportion of the area of the patch pattern to the patch area) that is equal to the area factor of the pattern in FIG. 16B. Here, the sum of the density for each unit area within the pattern or the average value thereof is made to be the density value for the entire area of the pattern. Then, the density value becomes the same even when the patterns are different.

[0125] Note that in this embodiment, the more the patch element for reference (RPE) and the patch element for adjustment (APE) are laid over each other, the smaller the area factor becomes and the lower the density of the patch thus formed becomes. In another allowable configuration, however, the more the patch element for reference (RPE) and the patch element for adjustment (APE) are laid over each other, the larger the area factor becomes and the higher the density of the patch thus formed becomes. In essence, any configuration is allowable as long as the information on the density can change sensitively in response to the degree of overlaying of, or the degree of offsetting (that is, the conveying error) of, the patch element for reference (RPE) and the patch element for adjustment (APE).

[0126] In addition, in this embodiment, each patch element is formed with print blocks arranged in a stair shape. Another arrangement, however, is allowable as long as the print blocks are not continuous in the direction of the scan for printing and as long as the arrangement can effectively reduce the negative influence of the ejection failure. For example, the print blocks may be arranged in a mottled fashion, or at random.

[0127] Moreover, in this embodiment, the matt black ink is used to form the test patterns. Any ink of a different color may be used for this purpose as long as the information on density can be acquired with a reading sensor in a favorable manner. In addition, inks of different colors may be used to print the patch elements for reference (RPEs) and to print the patch elements for adjustment (APEs), respectively.

[0128] Furthermore, regarding the numbers of the nozzle groups to be used and the positions of the nozzles to be used, the respective examples given in the above embodiment are not the only ones. Any number of nozzle groups and any positions of the nozzles are allowable as long as the change in information on density in response to the conveying error can be acquired in a favorable manner and as long as little negative influence is exerted by an ejection fault of the nozzle. To enhance the accuracy in detection of the conveying error caused by the eccentricity of the roller and by the outer-diameter error, the distance between the nozzle group used to print the patch elements for reference (RPEs) and the nozzle group used to print the patch elements for adjustment (APEs) is preferably made larger, and the two kinds of patch elements preferably have the same pattern.

(6) Correction Value for Conveying Error

[0129] In this embodiment, the density of each of the patches constituting the test pattern is measured with the reading sensor 120. In the measurement with the reading sensor 120, the test pattern is scanned with an optical sensor that includes a light emitter and a light detector thereon, and thus the density of each of the patches where the pattern for reference and the pattern for adjustment interfere with each other (FIGS. 15A and 15B) is determined. The density of the patch is detected as the amount of light reflected (intensity of reflected light) when light is emitted onto the patch. The detection operation may be executed only once for each area to be detected, or may be executed plural times to reduce the negative influence of the detection error.

[0130] Following the detection of the density of the patches, the densities of the respective plural patches printed in the main-scanning direction are compared with one another. Then, the error in conveying amount is calculated from the positions of, and from the difference in density between, the least dense patch and the second least dense patch. Here, the density values obtained from the least dense patch is denoted with N1, and the density value obtained from the second least dense patch is denoted with N2. Then, the difference in density (N=N2−N1) is compared with three threshold values T1, T2, and T3 (T1<T2<T3). When N<T1, little difference exists between N1 and N2. In this case, the conveying error is determined as the intermediate value of the offset amount for the least dense patch and the offset amount for the second least dense patch (the offset amount for the least dense patch+the length of ½ nozzles). When T1<N<T2, the difference between N1 and N2 is slightly larger than the difference in the previous case. In the case of T1<N<T2, the conveying error is determined as the value that is shifted further from the above-mentioned intermediate value to the side of the least dense patch by an amount of ¼ nozzles (the offset amount for the least dense patch+the length of ¼ nozzles). When T2<N<T3, the difference between N1 and N2 is even larger than the difference in the previous case. In the case of T2<N<T3, the conveying error is determined as the value of the offset amount for the least dense patch+the length of ⅛ nozzles. When T3<N, the difference in density N is significantly large. In this case, the conveying error is defined as the offset amount for the least dense patch.

[0131] As has been described above, three threshold values are set in this embodiment, and thus the detection of the conveying error is made possible with a unit of 2.64 μm, which is equivalent to the one eighth of the nozzle pitch, 9600 dpi (=1200×8). The processing is executed for each of the plural—30, to be more specific—patch rows that are formed in the sub-scanning direction. Thus, the conveying error is detected for each circumferential length (2.709 mm×4=10.836 mm) that is used in the four-time actions of conveying the printing medium P for each patch rows.

[0132] FIG. 19 is a chart illustrating the relationship between the patch rows B_n (n=1 to 30) and the conveying errors X_n detected from the respective patch rows B_n. In the chart, the horizontal axis shows the value of n and the vertical axis shows the value of conveying error Xn. The plotted values of conveying error X_n correspond to the respective values of n, which in turn correspond to the respective 1 to 30 patch rows B_n.

[0133] In FIG. 19, the value of the conveying error X_n fluctuates depending upon the values of n. This is because different amounts of conveying are produced by different rotational angles from the reference position of the roller, and this difference in the conveying amount derives from the eccentricity of the roller. Note that the fluctuation of the values of conveying error X_n derives from the eccentricity of the roller so that the fluctuation is a periodic one with a period corresponding exactly to a full rotation of the roller.

[0134] In addition, the values of the conveying error X_n, as a whole, are shifted either upwards or downwards in response to whether the outer diameter of the roller is larger or smaller than that for reference. When the outer diameter of the roller is larger than that for reference, the printing medium P is conveyed by an amount that is larger than a predetermined amount of conveying. Accordingly, the conveying errors X_n, as a whole, are shifted upwards in the chart. In contrast, when the outer diameter of the roller is smaller than that for reference, the conveying errors X_n, as a whole, are shifted downwards in the chart.

[0135] For the purpose of reducing the values of the conveying error X_n, it is necessary to reduce the amplitude, which is the fluctuation component of the conveying errors X_n, and to approximate the center value of the fluctuation to zero, that is, to the nominal value of the outer diameter of the roller. To this end, in this embodiment, an appropriate first correction value (correction value for eccentricity) to reduce the amplitude of the conveying errors X_n is acquired, and then a second correction value (correction value for outer-diameter) to approximate the central value of the fluctuation to zero is acquired.

[0136] In the following paragraphs, detailed descriptions of the processing to acquire these correction values will be given. The following descriptions will be given by taking the processing for the conveying roller 1 as an example, but similar processing can be carried out for the discharge roller 12. In addition, though the conveying roller 1 conveys the printing medium P in cooperation with the pinch rollers 2 and the conveying error is determined as an outcome of the combination of these rollers, the descriptions that follow are based, for convenience sake, on the assumption that the conveying error is of the conveying roller 1.

(7) Acquiring Correction Value for Eccentricity

[0137] To begin with, descriptions will be given as to the outline of the conveying control carried out in this embodiment by using the correction value for eccentricity and the correction value for outer-diameter that have been acquired previously. Though the details of this conveying control is to be given later, only the outline thereof will be given beforehand to describe the steps of acquiring the correction value for eccentricity and the correction value for outer-diameter.

[0138] In this embodiment, as shown in FIG. 28, the roller is divided into 110 sectors starting from a position for reference (thus formed are blocks BLK1 to BLK110). Then, a table is prepared to associate the blocks to their respective correction values for eccentricity. FIG. 26 shows an example of such a table. Correction values for eccentricity e1 to e110 are respectively assigned to the block BLK1 to BLK110.

[0139] In the conveying control of this embodiment, the base conveying amount is added with a correction value other than the correction value for eccentricity, that is, the correction value for outer-diameter, and then the rotation of the conveying roller 1 is calculated. In other words, from which of the blocks to which of the blocks the conveying roller 1 rotates is calculated. Then, correction value for eccentricity that corresponds to the blocks passing with this rotation is added. The value thus produced is made to be the final conveying amount, and the conveying motor 110 is driven to obtain this conveying amount.

[0140] As has just been described, to carry out the conveying control of this embodiment, correction values for eccentricity have to be acquired for each of the blocks created by dividing the circumferential length of the roller in 110 sectors, or, to put it other way, for the blocks each of which has a 0.338-mm (=37.19 mm/110) circumferential length of roller.

[0141] In this embodiment, however, the conveying error is detected, from the test pattern, for each circumferential length of roller used to convey the printing medium P four times for each of the patch rows (the length is 10.836 mm). In addition, two adjacent patch rows in the test pattern share part of their respective roller sectors used to carry out their respective four-time actions of conveying the printing medium P. So, following the procedures to be described below, correction values for eccentricity are acquired from the test pattern for the respective blocks of the roller, each of which blocks has a circumferential length (0.338 mm) formed by dividing the circumferential length of the roller into 110 sectors.

[0142] Incidentally, the period of the eccentricity appears in the form of a periodic function with period equivalent to the circumferential length of the roller. So, a periodic function having a periodic component that is equivalent to the circumferential length of the roller and having a polarity that is opposite to that of the function of the conveying error is to be obtained firstly in this embodiment (hereafter, such a function will be referred to as “correction function”). Then, the distance from the reference position of the roller is assigned to the correction function. Accordingly, the correction value for eccentricity is acquired for each of the blocks formed by the division into 110 sectors.

[0143] The correction function in this embodiment is obtained by selecting a combination of an amplitude A and an initial phase θ that are capable of reducing most the conveying error caused by the eccentricity of the roller—that is, the amplitude component of the conveying error X_n shown in FIG. 19—for a sine function, y=A sin (2π/L×T+θ). Here, L is the circumferential length of the roller (specifically, 37.19 mm for the conveying roller 1), and T is the distance from the reference position of the roller. Four different values—specifically, 0, 0.0001, 0.0002, and 0.0003—can be set for the amplitude A, while 22 different values—specifically, −5 m×2π/110 (m=0, 1, 2, 3, . . . , 21)—can be set for the initial phase θ. In summary, 66 different combinations of the amplitude and the initial phase without including the case of the amplitude A=0 are selectable in this embodiment, and 67 different combinations are selectable when the case of the amplitude A=0 is included. Amongst these different combinations, an optimum combination of the amplitude A and the initial phase θ for correcting the eccentricity of the roller is selected.

[0144] FIG. 18 illustrates an example of arithmetic processing procedure for finding the correction value for eccentricity.

[0145] Firstly, in step S21, a determination is made to judge whether an arithmetic processing is necessary to acquire the correction value for eccentricity, and this determination has to precede the acquirement of the correction value for eccentricity from the correction function. For example, when the conveying error caused by the eccentricity is smaller than a certain threshold value, such arithmetic processing to acquire the correction value for eccentricity is judged to be unnecessary. If this is the case, the amplitude of the correction function is set at zero, and the procedure is finished. In the embodiment, the procedure for determining the necessity of the arithmetic processing to acquire the correction value for eccentricity will be given in the following paragraphs.

[0146] Firstly, the average value X_n(ave) of the conveying errors X_n (n=1 to 30) shown in FIG. 19 is obtained, and the differences X_n′ between this average value X_n(ave) and the conveying errors X_n are calculated. FIG. 20 is a chart illustrating the relationship between the value of n and the difference X_n′ with the values of n on the horizontal axis and with the differences X_n′ on the vertical axis. Then, the absolute value |X_n′| of each of the differences X_n′ is squared, and the sum of this squared values Σ|X_n′|² is calculated. When the sum Σ|X_n′|² thus calculated is smaller than the certain threshold value mentioned above, a determination that the correction value for eccentricity is unnecessary is made.

[0147] In contrast, when the sum Σ|X_n′|² thus calculated is larger than the certain threshold value mentioned above, the operational flow advances to the processing to acquire the correction function to correct the eccentricity of the roller. In a step S24, a correction function having an amplitude A and an initial phase θ that are optimum to correct the eccentricity of the roller is calculated. An example of the way to calculate this correction value will be given in the following paragraphs.

[0148] Firstly, for each of all the combinations (66 combinations without the case of the amplitude A=0) of the amplitude A and the initial phase θ in the above-described sine function, the values are obtained by assigning, to the variable T of the sine function, the 34 different values starting from 2.709 to 92.117 at the intervals of 2.709.

[0149] For example, values y₁, y₂, and y₃ are obtained respectively by assigning 2.709, 5.418, and 8.128 to the variable T of the above-mentioned sine function with a certain amplitude A and a certain initial phase θ. The calculation continues until a value y₃₄ is obtained by assigning 92.117 to the variable T. The processing has to be done for all the 66 different combinations of the amplitude A and the initial phase θ without the case of the amplitude A=0.

[0150] Then, four successive values of y in a certain combination of the amplitude A and the initial phase θ are added together to produce 30 integrated values Y_n′. For example, y₁′=y₁+Y₂+y₃+y₄, and y₂′=y₂+y₃+y₄+y₅. In this way, values from y₁′ to y₃₀′ are calculated. The processing has to be done for all the 66 different combinations of the amplitude A and the initial phase θ.

[0151] Note that the values y₁, y₂, y₃, and y₄ are obtained by assigning, respectively, 2.709, 5.418, 8.128, and 10.836 to the variable T, where T is the distance from the reference position of the roller. Accordingly, in the sine function having a certain combination of the amplitude A and the initial phase θ, the value y₁′ obtained by adding the values y₁ to y₄ together is a value that corresponds to a sector of the roller starting from the reference position and ending with the 10.836-mm position. Likewise, in the sine function having a certain combination of the amplitude A and the initial phase θ, the value y₂′ obtained by adding the values y₂ to y₅ together is a value that corresponds to a sector starting from the 2.709-mm position and ending with the 13.545-mm position.

[0152] Subsequently, for each of the combinations of the amplitude A and the initial phase θ, the integrated values y_n′ are added to the respective differences X_n′ between the conveying errors X_n and the average value. For example, y₁′ is added to x₁′, and y₂′ is added to X₂′. The following additions are carried out similarly until y₃₀′ is added to X₃₀′. Thus obtained are addition values X_n″. Then, the absolute value of each of the addition values X_n″ is squared, and the sum of this squared values Σ|X_n″|² is calculated. FIG. 21 shows a graph illustrating the relationship between the value of n and the squared absolute value |X_n″|² the addition values while the values of n are on the horizontal axis and the values of |X_n″|² are on the vertical axis. By summing up the squared absolute values |X_n″|² corresponding to the respective values of n in this graph, the sum of the Σ|X_n″|² of the addition values Xn squared can be calculated.

[0153] In accordance with a procedure that is similar to the one described above, the sum Σ|X_n″²| of the squared absolute value of the addition values Xn is obtained for each of the all the 66 different combinations of the amplitude A and the initial phase θ. Then, one of the 66 combinations is selected so as to minimize the value of the square sum Σ|X_n″|². What can be obtained in this way is a correction function that can reduce most the conveying error caused by the eccentricity of the roller, that is, the amplitude component of the conveying error X_n. After that, the correction value for eccentricity for each block formed by dividing the roller into 110 sectors can be acquired by assigning the distance from the reference position for each of the blocks to the variable T of the correction function.

[0154] According to the above-described method of acquiring the correction value for eccentricity, the correction value for eccentricity for an area of the roller that is associated with the distance from the reference position of the roller can be obtained even with a test pattern, such as the one of this embodiment, in which: the conveying error X_n detected from each of the patch rows corresponds to a circumferential length of the roller corresponding to plural times of the conveying action for the printing medium P; and two adjacent patch rows share part of the sectors of the roller that are used to print the respective patch elements for reference and to print the respective patch elements for adjustment.

[0155] Subsequently, in step S25 in FIG. 18, a determination is made to judge whether there are plural test patterns in the main-scanning direction.

[0156] When only a single test pattern is printed in the main-scanning direction, a correction function is determined on the basis of the information on the density obtained from the test pattern so as to have an optimum combination of the amplitude A and the initial phase θ to correct the eccentricity. Then the correction value is arithmetically operated using the correction function (step S27).

[0157] Even for a roller manufactured within a predetermined design tolerance, the conveying error that derives from the amount and the state of eccentricity of the roller may sometimes vary between on the conveying-reference side and on the non-conveying-reference side of the printing apparatus. To address this phenomenon, two test patterns can be printed in the main-scanning direction in this embodiment. Accordingly, for each of the patterns, an optimum combination of the amplitude A and the initial phase θ to correct the eccentricity is obtained. Then, in step S29, the two combinations thus obtained are compared to determine whether the two combinations are the same or different. When the two combinations are the same, the correction value is arithmetically operated on the basis of the correction function with the common amplitude A and the common initial phase θ (step S31).

[0158] In contrast, there may be cases where the combination of the amplitude A and the initial phase θ on the conveying-reference side is different from the combination thereof on the non-conveying-reference side. In this case, selected is the combination of the amplitude A and the initial phase θ that minimizes the larger one of the values of square sum Σ|X_n″|² for the conveying-reference side and the non-conveying-reference side. The reason why such a way of selection is employed is avoiding the following inconvenience. It is possible to select the combination of the amplitude A and the initial phase θ that minimizes the smaller one of the values of square sum Σ|X_n″|² for the conveying-reference side and the non-conveying-reference side. Such selection may cause an unfavorable situation in which the conveying error caused by the eccentricity of the roller cannot be limited within the range of the design tolerance. When the combination of the amplitude A and the initial phase θ on the conveying-reference side is different from the combination thereof on the non-conveying-reference side, the processing described in the following paragraphs is carried out.

[0159] Firstly, for each of the three amplitude conditions (specifically, A=0.0001, A=0.0002, and A=0.0003), the square sum Σ|X_n″|² are plotted while the initial phase θ is changed. The plotting is done both for the conveying-reference side and for the non-conveying-reference side. The two curves thus obtained and representing the respective sides are compared with each other. From the two curves, sections of one of the two curves that have larger values than the values of the corresponding section of the counterpart curve are selected. The operation is schematically illustrated in FIGS. 22A and 22B.

[0160] FIGS. 22A and 22B illustrate the curves each of which obtained by plotting the square sum Σ|X_n″|² with the initial phase θ varying for each of the side near the conveying-reference and the side far from the conveying-reference. FIG. 22A is of a case where the curve for the conveying-reference side crosses the curve for the non-conveying-reference side. In this case, the sections represented by a thick solid line are the sections where the values of the square sum Σ|X_n″|² on the curve are larger than the corresponding values on the counterpart curve. FIG. 22B, on the other hand, illustrates a case where the curve for the conveying-reference side does not cross the curve for the non-conveying-reference side. In this case, the whole sector of one of the two curves constantly has the larger values of the square sum Σ|X_n″|², and is accordingly shown by a thick solid line in FIG. 22B.

[0161] Subsequently, within the selected sector, or sectors, having larger values of the square sum Σ|X_n″|² (shown by the thick solid line in FIGS. 22A and 22B), the value of the initial phase θ that makes the value of the square sum Σ|X_n″|² the lowest is selected as the optimum value under the amplitude condition of the case. When the two curves cross each other as shown in FIG. 22, one of the intersecting points that has the lowest value of the square sum Σ|X_n″|² is selected as the optimum value under the amplitude condition of the case. In the case shown in FIG. 22B, the value of the initial phase θ at the lowest-value point on the thick solid line is selected as the optimum value under the amplitude condition of the case.

[0162] The operation described above is carried out for each of the amplitude conditions. Then, the values of the square sum Σ|X_n″|² corresponding to the respective initial values determined individually for the amplitude conditions are compared with one another. Thereafter, the amplitude A and the initial phase θ of a case where the value of the square sum Σ|X_n″|² is the lowest are selected as the optimum values. After that, the correction value is arithmetically operated on the basis of the correction function having the optimum amplitude A and the optimum initial phase θ (step S33).

[0163] As has been described thus far, in this embodiment, the optimum values of the amplitude A and of the initial phase θ are obtained from a single test pattern or plural test patterns and then a correction function having such optimum values is determined. Then, on the basis of this correction function, the correction value for eccentricity is acquired.

[0164] In the above description, the correction value for eccentricity for each of the sectors formed by dividing the roller into 110 parts (blocks BLK1 to BLK110) is acquired while the correction values for eccentricity are associated with the respective distances from the reference position of the roller to the respective sectors. Note that this is not the only way to acquire the correction values for eccentricity. For example, the correction values for eccentricity may be acquired while the correction values for eccentricity are associated with the respective rotational angles from the reference position of the roller to the respective sectors.

[0165] In this embodiment, the rotary encoder 116 attached to the conveying roller 1 outputs 14080 pulses per rotation, for example. Then, the 14080 pulses are divided into groups each of which has 128 pulses so as to suit for the 110 sectors. Thus, the current position of the roller can be detected in accordance with the pulses outputted from the rotary encoder 116. Then, for each of the 110 sectors (blocks), the correction value for eccentricity is associated with the rotational angle from the reference position of the roller. Subsequently, an eccentricity-correction-value table is formed by setting these correction values for eccentricity (step S35) in the table. Storing these set values in, for example, the EEPROM 103 (see FIG. 3), makes it possible to keep these values even when the apparatus itself is switched off. Updating the set values is also made possible according to this configuration.

(7) Acquiring Correction Value for Outer-diameter

[0166] Besides the reduction of the conveying error caused by the eccentricity of the roller, the reduction of the conveying error caused by the outer-diameter error of the roller is effective for reducing the conveying error in total. The latter processing is the outer-diameter correction. Hereafter, descriptions will be given as to the way of acquiring the correction value for outer-diameter to use that processing and as to the reason why the acquiring of the correction value for eccentricity has to precede the processing for acquiring the correction value for outer-diameter.

[0167] FIG. 23 illustrates an example of arithmetic processing procedure to acquire the correction value for outer-diameter.

[0168] Firstly, contents of the eccentricity-correction-value table are applied to the conveying errors X_n detected from the respective patch rows of the test patterns, and the values thus obtained are denoted as Y_n (step S41). Then, the average value of Y_n are calculated and denoted as Y_n (ave) (step S43). Note that, as has been described above, each of the conveying errors X_n is the conveying error for the circumferential length of the roller corresponding to the four-time conveying of the printing medium P. Accordingly, before being applied to the conveying errors, the correction values for eccentricity in the eccentricity-correction-value table have to be integrated so as to be suitable for the conveying errors X_n thus obtained.

[0169] Subsequently, a determination is made to judge whether there are plural test patterns in the main-scanning direction (step S45). When there is only a single test pattern printed in the main-scanning direction, the difference between a target value (the value of the roller with dimensions that are exactly equal to the nominal ones and, accordingly, without any conveying error) and the average value Y_n(ave) are calculated. Then, on the basis of the calculated differences, the correction value for outer-diameter is determined (step S47).

[0170] Here, when the difference obtained by subtracting the average value Y_n(ave) from the target value is positive, the roller has a circumferential length that is longer than the roller with dimensions equal to exactly nominal ones. To put it other way, even a single conveying action using the roller conveys the printing medium P more than the amount that is supposed to be conveyed. Accordingly, in this case, a correction value (correction values for outer-diameter) is determined in step S47 so as to make the average value Y_n(ave) equal to the target value.

[0171] On the other hand, when plural test patterns (two test patterns in this embodiment) are printed in the main-scanning direction, the average values Y_n(ave) obtained from the respective test patterns are added up to find the average value thereof (step S49). The difference between this average value thus obtained and the target value is used to produce determine the correction values for outer-diameter (step S51). This correction value for outer-diameter can also be stored in the EEPROM 103 (see FIG. 3).

[0172] Now, description will be given in the following paragraphs as to the reason why the acquiring of the correction values for eccentricity has to precede the acquiring of the correction values for outer-diameter.

[0173] In this embodiment, emphasis is put on the achievement of a high-accuracy conveying-error correction without sacrificing the versatility of the test pattern and of the printing method. Assume that a test pattern used here has a length in the sub-scanning direction that is equal to an integral multiplication of the circumferential length of the roller. With such a test pattern, acquiring high-accuracy conveying-error correction values is possible even when the order of the acquiring of the correction values for eccentricity and the acquiring of the correction values for outer-diameter is reversed.

[0174] The test pattern used in this embodiment, however, has an 80-mm length in the sub-scanning direction. When a roller with a nominal outer circumference of 37.19 mm is used, the 80-mm length exceeds an integral multiplication of the roller with the nominal outer circumference (exceeds the amount of two full rotations of the roller). Hence, in this embodiment, the conveying error is detected from the area, within the test pattern, corresponding to the two full rotations of the conveying roller and detected from the excess area corresponding to a small, beginning part of the third rotation.

[0175] Note that it is, in fact, difficult to form a test pattern with its length in the sub-scanning direction that is precisely equal to an integral multiplication of the circumferential length of the roller. In addition, the tolerance of the outer diameter of the conveying roller 1 may sometimes cause fluctuations in the period of the eccentricity of the conveying roller 1. It is, therefore, rather preferable that the test pattern have a larger length in the sub-scanning direction than an integral multiplication of the nominal circumferential length of the conveying roller 1. Nevertheless, when the test pattern has a length in the sub-scanning direction that is not equal to an integral multiplication of the circumferential length of the roller, or to put it other way, when the conveying error is detected from the test pattern including an excess area, such inconveniences as described in the following paragraph may possibly occur.

[0176] In FIG. 24, conveying errors (X_n) acquired from the test pattern in this embodiment are plotted. The area marked by a circle in FIG. 24 corresponds to the excess area. As has been described before, the correction value for outer-diameter is used to correct the amount of the conveying error for each rotation of the conveying roller 1, and is calculated by the average of the values of the conveying error. Acquiring a precise correction value for outer-diameter, however, is problematic when the eccentricity of the roller causes significantly large deviation, from the average value, of the conveying error for the excess area.

[0177] In this embodiment, to reduce the negative influence caused by the part of the excess area, the correction value for eccentricity is acquired. Then, after the correction value for eccentricity is applied, the arithmetic processing of the correction value for outer-diameter is carried out. Accordingly, a variation in conveying error in the excess area is suppressed. As a result, it is possible to reduce a difference between the conveying error and the average of the values of the conveying error, so that the influence of the eccentricity can be reduced.

[0178] FIG. 25 shows examples of correction values acquired through the processing, firstly, of the correction value for eccentricity and then through the processing of the correction value for outer-diameter as well as examples of correction values acquired through the two processing carried out in the reverse order. Here, for the sake of simplicity, outcomes of calculation on the test pattern FR1 on the conveying-reference side are compared.

[0179] Firstly, assume that the correction values are calculated in an order in which the processing for the correction value for outer-diameter precedes the processing for the correction value for eccentricity. In this case, when the average value Yn(ave) is calculated in a state shown in FIG. 24, the value becomes 9.31 μm. After the correction value for outer-diameter acquired on the basis of this value of 9.31 μm is reflected, an operation of eccentricity correction is carried out. In this case, a value of 0.0003 is selected for the amplitude A. Meanwhile, a value of n=13 is selected for the initial phase θ. In contrast, assume that the calculation of the correction value for eccentricity precedes the calculation of the correction values for outer-diameter, as in the case of this embodiment. In this case, a value of 0.0003 is selected for the amplitude A. Meanwhile, a value of n=13 is selected for the initial phase θ. Then, while the correction value for eccentricity is applied, the value of Yn(ave) is calculated. The resultant value becomes 8.74 μm (on the basis of this value Y_n(ave) of 8.74 μm, the correction value for outer-diameter is acquired). The comparison of the procedures in different orders makes it clear that the correction values for eccentricity are the same but that the correction values for outer-diameter are different from each other.

[0180] Note that, here, the theoretical figure of the correction value for outer-diameter is 8.54 μm when the correction value for outer-diameter is calculated by extracting the value of Xn corresponding to two full rotations of the roller from the state in FIG. 24. Accordingly, as in the case of this embodiment, when the acquiring of the correction value for eccentricity precedes the acquiring of the correction values for outer-diameter, the correction value for outer-diameter can be acquired with the deviations from the theoretical figure being made smaller.

(8) Control of Conveying

[0181] As has been described above, in this embodiment, the rotary encoder 116 attached to the conveying roller 1 outputs 14080 pulses for each rotation. Then, in this embodiment, the 14080 pulses are divided into 110 circumferential sectors each of which has 128 pulses starting from the reference position of the rotary encoder 116. Subsequently, a table for storing the correction values for eccentricity acquired through the arithmetic processing for correction values for eccentricity is formed so as to make the correction values for eccentricity correspond to the respective above-mentioned circumferential sectors.

[0182] FIG. 26 shows an example of the table thus formed. Correction values for eccentricity e1 to e110 are allocated so as to correspond to the respective blocks BLK1 to BLK110 each of which has a rotational angle corresponding to 128 pulses of the rotary encoder 116. These correction values for eccentricity are reflected in the control of the conveying in a way described in the following paragraphs.

[0183] FIG. 27 shows an example of the procedure for the control of the conveying. FIG. 28 is an explanatory diagram for describing the operation corresponding to this procedure. Note that the procedure shown in FIG. 27 is executed for the purpose of determining the amount of conveying the printing medium P (sub scan) between every two printing scans, and can, accordingly, be done either during a printing scan or after the completion of a printing scan.

[0184] Firstly, in a step S61, the base amount of conveying is loaded. The base amount of conveying is a theoretical value of the sub-scanning amount between every two consecutive printing scans. Then, in a step S63, the base amount of conveying is added with a correction value other than the correction value for eccentricity, that is, the correction value for outer-diameter. Moreover, in a step S65, a calculation is executed so as to find to what position the conveying roller 1 rotates from the current rotational position in response to the resultant value of the above-mentioned addition. In the example shown in FIG. 28, the conveying roller 1 rotates from a position within the block BLK1 to a position within the block BLK4.

[0185] After that, in a step S67, the correction values for eccentricity corresponding to the blocks that are passed by during the rotation of this time are added. To be more specific, in the example shown in FIG. 28, the blocks BLK2 and BLK3 are passed by during the rotation, so that the correction values for eccentricity e2 and e3 are added. The resultant value from the addition is made to be the final amount of conveying, and then the conveying motor 110 is driven to obtain this amount of conveying (step S69).

[0186] Note that only the correction values for eccentricity for the blocks that are passed by are configured to be added in this embodiment, but another configuration is possible. In accordance with the position within the current block before the rotation (i.e. block BLK1) and the position within the block after the rotation (i.e. block BLK4), the correction values for eccentricity for these blocks are converted appropriately, and the values thus converted can be used for the addition. Nevertheless, the simple use of the correction values of the respective blocks that are passed by can be processed with more ease and in shorter time than such a fine-tune recalculation of the correction value can.

[0187] The correction values thus far described are those for the conveying roller 1, but the correction values for the discharge roller 12 can be obtained in a similar way and can be stored in the EEPROM 103. The stored correction value for the discharge roller 12 can be used when the roller, or rollers, used for the conveying is switched to the discharge roller 12 alone.

(9) Ways of Acquiring Correction Values

[0188] The correction value for eccentricity and the correction value for outer-diameter may be acquired on the basis of the information on density obtained by scanning the test pattern with a reading sensor 120 mounted, along with the print head 4, on the carriage 7. Alternatively, the correction value for eccentricity and the correction value for outer-diameter may be acquired on the basis of the information on density obtained by scanning the test pattern with a reading sensor 120 provided in the form of a reading head and mounted, in place of the print head 4, on the carriage 7.

[0189] FIG. 29 shows an example of the processing procedure corresponding to the configurations described above. At the start of this procedure, the printing medium P is set (step S101), and test patterns such as ones shown in FIG. 5 are printed (step S103). Then, the printing medium P with the test patterns formed thereon is set in the apparatus again, and the operation of reading the test patterns is executed to acquire the information on density (step S105). After that, on the basis of the information on density, the correction value for eccentricity and the correction value for outer-diameter are acquired in this order (steps S107 and S109), and then these correction values are stored (or updated) in the EEPROM 103 (step S111).

[0190] In a case where the printing apparatus has no built-in reading sensor (including a case where the printing apparatus are configured as a multi-function apparatus having a scanner apparatus unit integrated therewith), the printing medium P with the test patterns printed thereon is set in an outside scanner apparatus to carry out the reading.

[0191] FIG. 30 shows another example of the processing procedure corresponding to the configurations described above. The difference that this procedure has from the one described above is the provision of a process (step S125) in which the printing medium P with the test patterns formed thereon is set in an outside scanner apparatus followed by the inputting of the information on density thus read.

[0192] In addition, the arithmetic operation for the correction values may be executed not as a process done on the printing-apparatus side but as a process done by a printer driver operating within the host apparatus 1000 provided in the form of a computer connected to the printing apparatus.

[0193] FIG. 31 shows an example of the processing procedure in this case. In this procedure, the printing medium P with the test patterns formed thereon is read using an outside scanner apparatus, and the information on density thus read is then supplied to the host apparatus 1000 to operate arithmetically the correction values. The printing apparatus awaits the imputing of the correction values (step S135). In a case where such an input is actually done, the correction values are stored (updated) in the EEPROM 103 (step S111).

[0194] The above-described processes may be executed either in response to the instruction given by the user. Alternatively, the user may delegate a serviceman to do the processes on behalf of the user, or the user may carry the apparatus in the service center to do the job. In any case, storing the correction values in the EEPROM 103 enables the correction values to be updated when it is necessary. As a result, the deterioration with age of the roller can be addressed properly.

[0195] However, assume a case where the deterioration with time is not a real problem, and where no update is necessary. In this case, a default value for the correction value may be determined in an inspection process done before the printing apparatus is shipped from the factory. Then, the default value thus determined is stored in the ROM 102, which is installed in the printing apparatus. In this sense, “the method of acquiring the correction value for the conveying-amount error” characterized: by an arithmetic operation for the correction value for eccentricity; and by a determination of the correction value for outer-diameter that follows the above-mentioned arithmetic operation, is not necessarily carried out within the printing apparatus, but can also be carried out using an apparatus, or an inspection system, that is provided independently of the printing apparatus.

(10) Other Modifications

[0196] The above-described embodiment and the modified examples thereof described in various places in the course of the descriptions are not the only ways of carrying out the present invention.

[0197] For example, in the configuration described above, the conveying roller 1 and the discharge roller 12 are respectively provided on the upstream side and on the downstream side in the direction of conveying the printing medium P. The printing medium P is conveyed by various conveying units since the printing medium P is loaded till the printing is finished. Assume that units other than the two rollers mentioned above are involved in the conveying, and that the conveying errors caused by the eccentricity or the variation in the outer diameter of each unit may possibly affect the printing quality. If this is the case, a conveying-error correction value can be acquired for each of the rollers in consideration independently or in combination with others. Also in this case, in a similar way to the one employed in the case described above, test patterns are printed firstly, and then an correction value for eccentricity and an correction value for outer-diameter are acquired on the basis of the information on density of the test patterns. In summary, the printing of the test patterns and the acquiring of the correction values can be carried out in accordance with the number of and the combination of the conveying units involved in the conveying at the time when the printing is actually done. In this way, an even and high-quality printing is possible on all over the printing medium P.

[0198] For example, in a case where only a single roller is used to convey the printing medium P, the conveying is always carried out by the single roller alone. As a result, there are only one kind of the printing of the test patterns and one kind of the conveying-error correction value. When two rollers are used in the conveying, the processes to be done can be divided, as in the above-described case, into a case where the conveying roller 1 is involved in the conveying and a case where the discharge roller 12 alone is involved in the conveying. In addition, the processes to be done can also be carried out by further dividing the former of the two resultant cases above into a case where the conveying roller 1 alone is involved in the conveying and a case where the conveying roller 1 is involved in the conveying in cooperation with the discharge roller 12. In a case of three rollers, the processes to be done can be divided into five, at the maximum, cases (areas) in a similar manner. In general terms, when the conveying is carried out by n rollers (n≧2), the processes to be done can be divided into 3+½[n(n−1)] areas at the maximum.

[0199] In addition, in the example described above, the correction value for eccentricity and the correction value for outer-diameter are acquired for the discharge roller 12 as well. Suppose, however, a case where the discharge roller 12 is made of rubber. Rubber is a material, which is susceptible to the changes in environment and to the deterioration with age, and where reflecting the correction value for eccentricity for the discharge roller 12 may have few, if any, effects. If this is the case, the arithmetic operation for or the application of the correction value for eccentricity for the discharge roller 12 can be omitted.

[0200] Moreover, in the example described above, the patch elements for adjustment (the second patch elements) are printed using a part of the nozzle arrays that is located on the upstream side in the conveying direction. Alternatively, for example, as shown in FIG. 32, a printing medium P with patch elements for adjustment RPEs' printed thereon in advance may be used. Then, patch elements for reference APEs are printed using, fixedly, a particular nozzle group of all the nozzle arrays, and thus the formation of the test patterns is completed. After that, on the basis of the test pattern thus formed, processes to acquire the correction values are carried out. Note that the patch elements printed in advance may be the patch elements for reference RPEs', and that the patch elements for adjustment APEs may be printed in the later process.

[0201] Furthermore, given in the descriptions provided above are only examples of: the number of color-tones (color, density and the like) of the inks; the kind of the inks; the number of nozzles; ways of setting the range of nozzles actually used and ways of setting the amount of conveying the printing medium P. Likewise, various numerical values given in the descriptions above are also just examples of those that can be used.

[0202] The correction values obtained in the way described above are applicable to the control of conveying the printing medium at the time of actual printing.

[0203] In a first embodiment of the control of conveying a printing medium, a range including all of the 768 nozzles installed in each of ejection units of a print head is used to carry out the printing of areas of a printing medium. Note that this way of using nozzles is employed particularly in the printing on a printing medium, such as a plain paper sheet, where the bleeding of the ink occurs much. Accordingly, on such printing medium, none of the stripes formed at the connecting portions of two consecutive printing scans, the difference in the amount of ejected ink between nozzles, and the difference in the accuracy of landing position of ink is noticeable. This way of using nozzles, however, is not the only way that can be employed for the present invention. In addition, most of printing apparatuses have plural, selectable printing modes, such as “fine (high-quality printing)”, “standard” and “fast (high-speed printing).” In accordance with the selected mode, the printing apparatus can perform what is termed as one-pass printing (printing on a single scanning area on the printing medium is completed by a single scan) or what is termed as multi-pass printing (printing on a single scanning area on the printing medium is completed by multiple scans). For example, as shown in FIG. 33A, when the fast mode is selected, the printing apparatus is made to perform the printing for a single scanning area by a single printing scan (one-pass printing). In the standard-quality printing mode, four printing scans are carried out to complete the printing (4-pass printing) as shown in FIG. 33B while in the high-quality printing mode, eight printing scans are carried out to complete the printing (8-pass printing) as shown in FIG. 33C. Whichever mode is selected, this embodiment is applicable.

[0204] Subsequently, the application of the eccentricity correction and of the outer-diameter correction will be described in detail.

[0205] As has been described above, what should be reduced in this embodiment is the displacement of dot-formation positions that derives from the insufficient conveying accuracy caused by the error of eccentricity and the error of outer-diameter of the conveying roller 1 and of the discharge roller 12. In the printing of some areas of the printing medium P, both of the conveying roller 1 and the discharge roller 12 are not involved. Here, amongst the areas on the printing medium P, the area on which the printing is done with the conveying roller 1 alone being used for the conveying of the printing medium P is defined as a first area (FIG. 34A). This corresponds to a first conveying phase of the printing medium. The area on which the printing is done with both of the conveying roller 1 and the discharge roller 12 being used for the conveying is defined as a second area (FIG. 34B). This corresponds to a second conveying phase of the printing medium. The area on which the printing is done with the discharge roller 12 alone being used for the conveying is defined as a third area (FIG. 34C). This corresponds to a third conveying phase of the printing medium. In this embodiment, different correction information (correction values for eccentricity and of the correction values for outer-diameter) is provided for each of the first to third conveying phases. This means that different correction values may sometimes be applied even to the same roller in response to the rollers actually involved in the conveying of the print medium. The reason for this is that the states of conveying the printing medium P differ from each other between the printing on the first area for which the conveying is done by the conveying roller 1 alone and the printing on the second area for which the conveying is done by both of the conveying roller 1 and the discharge roller 12. Hence, different correction values for eccentricity and different correction values for outer-diameter have to be applied to optimize the conveying by the same conveying roller 1. In at least one conveying phase (the second conveying phase) the rotary conveying elements (rollers) undergo more than one full rotation. In all the conveying phases, the correction information covers plural sectors of a roller.

[0206] FIG. 35 illustrates the relationships between the printing area and both of the correction value for eccentricity and the correction value for outer-diameter that are applied to the corresponding one of the areas in this embodiment. A first correction value for eccentricity and a first correction value for outer-diameter are used as the correction value for eccentricity and the correction value for outer-diameter for the first area (first conveying phase). A second correction value for eccentricity and a second correction value for outer-diameter are used as the correction value for eccentricity and the correction value for outer-diameter for the second area (second conveying phase). A third correction value for eccentricity and a third correction value for outer-diameter are used as the correction value for eccentricity and the correction value for outer-diameter for the third area (third conveying phase). In this way, different correction values are applied to the respective printing areas depending on which conveying mechanism (the conveying roller 1 alone is a first conveying mechanism; the combination of the roller 1 and the discharge roller 12 is a second conveying mechanism; and the roller 12 alone is a third conveying mechanism) is actually used in the conveying phase concerned, so that the conveying accuracy is made higher than otherwise and higher-quality printing can be obtained.

[0207] More specific descriptions will follow. As shown in FIG. 36, in the second area, the conveying roller 1 and the discharge roller 12 cooperate to convey the printing medium P. Accordingly, following the procedure described with reference to FIGS. 18 and 23, a correction value for eccentricity (a first type of correction information) and a correction value for outer-diameter (a second type of correction information) are acquired for each of the conveying roller 1 and the discharge roller 12 by using the test patterns formed in the second printing area. The conveying control is then carried out by applying the correction values thus acquired to each of the rollers 1 and 12 in accordance with the procedure described with reference to FIG. 28. In other words, individual correction information for the roller 1 and individual correction information for the roller 12 are employed to correct for conveying errors of those rollers in the second conveying phase. In the first area, the conveying roller 1 alone is involved in the conveying of the printing medium P, and the discharge roller 12 is not involved in the conveying of the printing medium P. Accordingly, following the procedure described with reference to FIGS. 18 and 23, a correction value for eccentricity and a correction value for outer-diameter are acquired for the conveying roller 1 by using the test patterns formed in the first printing area. The conveying control is then carried out in accordance with the procedure described with reference to FIG. 28. In the third area, the discharge roller 12 alone is involved in the conveying of the printing medium P, and the conveying roller 1 is not involved in the conveying of the printing medium P. Accordingly, following the procedure described with reference to FIGS. 18 and 23, a correction value for eccentricity and a correction value for outer-diameter are acquired for the discharge roller 12 by using the test patterns formed in the third printing area. The conveying control is then carried out by applying the correction value for eccentricity and the correction value for outer-diameter acquired for the discharge roller 12 in accordance with the procedure described with reference to FIG. 28. Note that the test patterns formed in the first to the third areas as shown in FIG. 8 can be used as the respective test patterns used to acquire the first to the third correction values for eccentricity and to acquire the first to the third correction values for outer-diameter.

[0208] As has been described thus far, the conveying control carried out by applying different correction values for eccentricity and different correction values for outer-diameter to the respective areas enhances the conveying accuracy for the printing on the front-end portion and on the rear-end portion of the printing medium P. As a result, the printing of high-quality image can be achieved.

[0209] In the above-described example, the correction value for eccentricity and the correction value for outer-diameter are changed in response to which one, or both, of the conveying roller 1 and the discharge roller 12 is/are involved in the conveying of the printing medium P. In this modified example, however, the common correction value for eccentricity or outer-diameter can be acquired in the areas where the different roller is involved.

[0210] FIG. 37 illustrates a modified example of the relationship between each printing area and both of the eccentricity correction and the outer-diameter correction in the first embodiment. In this modified example of the first embodiment, while the printing of the second area is being carried out, the printing medium P is conveyed by both of the conveying roller 1 and the discharge roller 12. In this event, however, the outer-diameter component of the conveying roller 1 has more dominant influence than that of the discharge roller 12. Accordingly, in the correction of outer-diameter for both of the first and the second areas, a single, common correction value for outer-diameter is used to carry out the correction of outer-diameter. To put it other way, in the correction of outer-diameter for the first area, the outer diameter correction value for the second area is acquired to carry out the conveying control.

[0211] In addition, each of the first and the third printing areas has a length, in the sub-scanning direction, that is comparable to a full rotation of the conveying roller 1 (=37.2 mm). Accordingly, the unevenness of the roller-period pitch caused by the eccentricity of the conveying roller 1 is less noticeable. For this reason, any correction value for eccentricity is acquired neither for the first area nor for the third area. Only the correction value for outer-diameter is acquired to carry out the conveying control.

[0212] In a second embodiment of the control of conveying a printing medium, the printing on the front-end portion and on the rear-end portion of the printing medium P is carried out by reducing appropriately the range of nozzles to be used for the purpose.

[0213] In some cases of the printing on the front-end portion or the rear-end portion of the printing medium P, either one of the conveying roller 1 and the discharge roller 12 is not involved in the printing on the printing medium P. FIGS. 34A and 34B illustrate examples of these cases. When the printing medium P is supported and conveyed by only one of the conveying roller 1 and the discharge roller 12 as in the above-mentioned cases, the flatness of the printing medium P is not secured at a sufficient level. As a result, the distance between the print head and the end portion that is not supported (hereafter, also referred to as “head-to-paper distance”) varies by not a small amount, and the degradation in the conveying accuracy is more likely to occur. In short, the head-to-paper distance is in a quite unstable state. The slipping of the printing medium P or the like is more likely to occur in a case where the printing medium P is conveyed using only one roller than in the case of two rollers being involved in the conveying. This is just an example of such degradation in the conveying accuracy. Incidentally, assume that the printing is carried out on the central portion of the printing medium P in the second area, which is shown in FIG. 34B. In this case, while the printing medium P is supported and conveyed by both the conveying roller 1 and the discharge roller 12, the printing scan is executed by ejecting the ink. Here, the ink is ejected at timing corresponding to a predetermined head-to-paper distance that is supposed to be kept when the printing medium P is on the platen 3. The ink ejected at appropriate timing form dots on the printing medium P. When the dots thus formed are arranged with an appropriate pitch, an image is formed properly. However, in the first and the third areas, that is, in the front-end portion and in the rear-end portion, the unstable head-to-paper distance—the head-to-paper distance varies to a greater extent within the printing width—destabilizes the positions of the dots formed on the printing medium P. Consequently, such harmful effects as white or black stripes and granular impression sometimes appear in the image thus formed. To prevent such degradation in image quality, in the printing apparatus of this embodiment, the range of nozzles to be used is reduced and the printing width of the print head is restricted during the printing on the front-end portion and the rear end portion of the printing medium P. To put it other way, the range of nozzles to be used is reduced and, at the same time, the amount of conveying the printing medium P is reduced. Thus, the variation in the head-to-paper distance is reduced so that the harmful effects to the image are lowered down to the minimum. Note that in the printing apparatus of this embodiment, the restriction on the range of nozzles to be used is imposed when the printing is carried out on the most front-end portion of the printing medium P, that is, one of the target portions of the marginless printing function. Such restriction on the range of nozzles to be used may be imposed on the entire front-end portion of the printing medium P.

[0214] FIG. 38 is a schematic top plan view of the platen 3. The printing medium P is conveyed from the lower side to the upper side in the FIG. 38, that is, in the conveying direction indicated by the arrow. Accordingly, the conveying roller 1 and the discharge roller 12 are respectively disposed on the lower side and on the upper side in FIG. 38.

[0215] A nozzle array HN is provided in the print head 4. In FIG. 38, only the nozzle array corresponding to the ink of a particular, single color is illustrated for the sake of simplicity. In the platen 3 that supports the printing medium, there is formed an opening corresponding to the area that is scanned with the nozzle array HN. Inside the opening, plural ribs P001 are formed as being raised so as to support the printing medium P. Ink absorber P002 is provided to receive the ink that goes off the edge, such as front, rear, and side edges, of the printing medium P when the marginless printing is carried out.

[0216] The ribs P001 are formed inside the opening of the platen 3. Specifically, a plural number of the ribs P001 are formed in each of the end portions on both upstream and downstream sides in the conveying direction. The distance between the line of the ribs P001 formed in the end portion on the upstream side and the line of the ribs P001 formed in the end portion on the downstream side is wider than the length corresponding to the maximum number of nozzles (768 nozzles in this embodiment) used for the printing of the central portion of the printing medium P. Accordingly, the ribs P001 are not stained by the ink that runs off from the right and left side edges of the printing medium P.

[0217] Also inside the opening, a plural number of the ribs P001 are disposed in the substantially central portion in the direction of conveying the printing medium P so as to support the printing medium P. These ribs P001 in the central portion are disposed so as not to be stained by the ink that runs off from the front and the rear edges as well as from the right-side and the left-side edges. The disposition of the ribs P001 and the maximum number of the nozzles that can be involved in the printing on the front-end portion and the rear-end portion of the printing medium P are determined appropriately by taking account of the relationship between the disposition of the ribs P001 and the number of nozzles.

[0218] FIGS. 39A to 39D illustrate the printing areas at the time when the printing is carried out by the printing apparatus of this embodiment. In the printing apparatus of this embodiment, printing without any margin (marginless printing) is performed on an A4 printing medium P (294 mm×210 mm).

[0219] FIG. 39A illustrates the area in the front-end portion of the printing medium P. As shown in FIG. 34A, printing is performed on the area shown in FIG. 39A before the front end of the printing medium P starts to be supported by the discharge roller 12. FIG. 39B illustrates the area in the central portion of the printing medium P. As shown in FIG. 34B, printing is performed on the area shown in FIG. 39B with the printing medium P being supported by both the conveying roller 1 and the discharge roller 12. FIG. 39C illustrates the area where the printing is performed around the time when the printing medium P is removed from the conveying roller 1. FIG. 39D illustrates the area in the rear-end portion of the printing medium P. As shown in FIG. 34C, printing is performed on the area shown in FIG. 39D after the rear end of the printing medium P is released from the conveying roller 1.

[0220] The printing on the front-end portion of the printing medium P is performed using 192 nozzles located in a range from the 64th nozzle to the 255th nozzle counted from the nozzle located at the most downstream position of the nozzle array HN shown in FIG. 39A. Such restriction on the range of nozzles to be used when printing is performed on the front-end portion of the printing medium P prevents the ink from being ejected onto the ribs P001.

[0221] FIG. 40 illustrates the relationship between the printing scans and the range of nozzles to be used in the printing on the front-end portion shown in FIG. 39A. At first (corresponding to the left-hand side portion of the drawing), 192 nozzles located in a range from the 64th nozzles to the 255th nozzles counted from the nozzle located at the most downstream position are used for the scan. When the scan is done once, the printing medium P is conveyed by an amount corresponding to 48 nozzles (=192/4), and then another scan is done using 192 nozzles on the downstream side. Every two scans are followed by the conveying of the printing medium P by an amount corresponding to 48 nozzles. Printing is performed by repeating the scan and the conveying.

[0222] Firstly, in the head-end portion of the printing medium P, printing on the front-end portion of a 37.2-mm length is performed using the 192 nozzles positioned on the most downstream side of the nozzle array HN. Then, when the front end of the printing medium P is supported by the discharge roller 12 as shown in FIG. 34B, the range of nozzles to be used is broadened gradually.

[0223] FIG. 41 illustrates the relationship between the printing scans and the range of nozzles used in the printing performed on a shifting area. On the shifting area, printing is performed when the front end portion of the printing medium P starts to be supported by the discharge roller 12. The printing on the shifting area is performed while the range of the nozzles to be used is gradually widened. Specifically, use of the 192 nozzles is use of a part of all the 768 nozzles. This partial use of the nozzles has to be changed into the total use of the 768 nozzles. To this end, the range of nozzles to be used is widened while being shifted. As shown in FIG. 41, the scan at the time when the printing medium P starts to be supported by the discharge roller (corresponding to the left-hand side portion of the drawing) is executed using the 192 nozzles located in a range from the 64th nozzles to the 255th nozzles counted from the nozzle located at the most downstream position. Then, gradually, the range of the nozzles to be used is widened to the upstream side by 32 nozzles. Scans and the conveying by an amount corresponding to 48 nozzles are repeatedly carried out while the range of the nozzles to be used is widened. The scans are performed with the range of the nozzles to be used being widened until the range covers all the 768 nozzles.

[0224] Once the range of nozzles to be used reaches all 768 nozzles, printing in the central portion of the printing medium P is performed using the entire nozzle array HN as shown in FIG. 39B. Note that the ribs P001 are disposed appropriately also for this case. To put it other way, no ribs are disposed corresponding to any standard-sized printing medium, for example. For this reason, no ink is ejected onto any ribs P001.

[0225] Subsequently, when printing is performed on a portion of the printing medium P near the rear end portion thereof, the range of nozzles to be used is gradually reduced. The timing for starting the restriction on the range of the nozzles used for the printing in the rear-end portion of the printing medium P can be determined on the basis of the timing when a PE sensor E0007 detects the rear edge of the printing medium P. To be more specific, firstly, the point of time when the rear edge of the printing medium P is released from the position where the rear edge has been nipped by the conveying roller 1 and the pinch rollers 2 (rear-edge removal time) can be identified on the basis of the above mentioned detected timing. Then, the point of time thus identified can be used to find the timing for starting the above-mentioned restriction.

[0226] FIG. 42 illustrates the relationship between the printing scan and the range of the nozzles to be used in the printing in another shifting area located near the rear-end portion of the printing medium P. The printing in this shifting area is performed while the range of the nozzles to be used is gradually reduced. Specifically, the range of the nozzles to be used is shifted and gradually reduced from the use of all the 768 nozzles to the use of 384 nozzles—just a part of all the 768 nozzles—located in the range from the 320th to the 703rd counted from the nozzle located at the most downstream position. As shown in FIG. 42, when the PE sensor E0007 detects the rear edge of the printing medium P (corresponding to the left-hand side portion of the drawing), all the 768 nozzles are used for the scan. Then, the range of the nozzles to be used is gradually reduced to the downstream side by 32 nozzles. Scans and the conveying by an amount corresponding to 48 nozzles are repeatedly carried out while the range of the nozzles to be used is reduced. The scans are performed with the range of the nozzles to be used being reduced until the range includes only 384 nozzles.

[0227] After the range of the nozzles to be used is reduced down to the 384 nozzles, the rear-end portion of the printing medium P is released from the conveying roller 1.

[0228] FIG. 43 illustrates the relationship between the printing scans and the range of the nozzles to be used in the printing at the time when the rear-end portion of the printing medium P is removed from the conveying roller 1 as shown in FIG. 39C. The relieving of the printing medium P from the binding by the conveying roller 1 and the pinch rollers 2 produces an impact, which may in turn cause unevenness in the product print. The occurrence of such unevenness has to be prevented. When the rear edge of the printing medium P is released from the conveying roller 1, the printing medium P is sometimes biased to the discharging direction, so that the conveying amount becomes larger than a predetermined amount. Accordingly, in the printing on such areas, the number of nozzles to be used is kept at 384 but the positions of the nozzles to be used are shifted. In this embodiment, when the rear edge is released, the amount of the conveying is made to correspond to 160 nozzles, and the printing is carried out using 384 nozzles located in the range from the 192nd nozzle to the 575th nozzle counted from the nozzle that is located at the most downstream position after the positions of the nozzles to be used have been shifted to the upstream side by 144 nozzles.

[0229] Printing of several scans is carried out after the rear edge of the printing medium P is released from the conveying roller 1. Then, as shown in FIG. 44, the range of the nozzles to be used is gradually reduced from the 384 nozzles to 192 nozzles located in the range from the 512th nozzle to the 703rd nozzle counted from the nozzle located at the most downstream position. Scans and the conveying by an amount corresponding to 16 nozzles are repeatedly carried out while the range of the nozzles to be used is reduced. The scans are performed with the range of the nozzles to be used being reduced until the range includes only 192 nozzles.

[0230] Then, once the printing medium P is supported by the discharge roller 12 alone as shown in FIG. 34C, 192 nozzles on the upstream side of the nozzle array HN are used for the printing as shown in FIG. 39D.

[0231] FIG. 45 illustrates the relationship between the printing scans and the range of the nozzles to be used in the printing on the rear-end portion shown in FIG. 39C. At first (corresponding to the left-hand side portion of the drawing), 192 nozzles located on the upstream side are used for the scan. When the scan is done once, the printing medium P is conveyed by an amount corresponding to 48 nozzles (=192/4), and then another scan is done using 192 nozzles on the upstream side. Every two scans are followed by the conveying of the printing medium P by an amount corresponding to 48 nozzles. Printing is performed by repeating the scan and the conveying.

[0232] As has been described above, the printing in this embodiment is carried out by changing the range of the nozzles to be used in response to the printing areas on which the printing is actually carried out.

[0233] As has been described above, in this embodiment, the range of the nozzles to be used in the printing on the first and the third areas is reduced. Accordingly, in the printing on the first and the third areas, only a narrower range of the nozzles is used than otherwise. In addition, each of the printing area has a length, in the sub-scanning direction, that is comparable to a full rotation of the conveying roller 1 (=37.2 mm). Accordingly, the unevenness of the roller-period pitch caused by the eccentricity of the conveying roller 1 is not noticeable. For this reason, any correction value for eccentricity is acquired neither for the first area nor for the third area. Only the correction value for outer-diameter is acquired to carry out the conveying control.

[0234] In addition, as in the case of the above-described modified example of the first embodiment, while the printing in the second area is being carried out, the conveying of the printing medium P is under the more dominant influence of the outer-diameter component of the conveying roller 1 than the influence of the outer-diameter component of the discharge roller 12. Accordingly, in the correction of outer-diameter for both of the first and the second areas, a single, common correction value for outer-diameter may be used to carry out the correction of outer-diameter correction. To put it other way, in the correction of outer-diameter correction for the first area, as in the case of the outer diameter correction for the second area, the correction values for outer-diameter for both of the conveying roller 1 and the discharge roller 12 are acquired to carry out the conveying control.

[0235] The printing medium P is conveyed by various rollers since the printing medium is loaded from the convey tray until the printing is finished. Each of these rollers involved in the conveying has a unique amount of eccentricity, and may sometimes affect the quality of the image printed. Such problem can be addressed in the following way. Firstly, the surface of the printing medium P is divided into areas in accordance with the number of the rollers involved in the conveying and in accordance with combinations of these rollers. Then, patterns from which the conveying errors can be detected are printed in the respective areas. The printing medium P is conveyed while reflecting correction values that are calculated from their respective patterns is reflected.

[0236] For example, in a case where only a single, rotational body for conveying—the conveying roller 1—is used to convey the printing medium P, the conveying is always carried out by the conveying roller 1 alone. As a result, there is only one area on the surface of the printing medium P. Alternatively, as in the cases of the first and of the second embodiments, there are two rotational bodies for conveying—the conveying roller 1 and the discharge roller 12. In this case, the surface of the printing medium P can be divided into three areas: a first area on which the printing is carried out with the printing medium P being conveyed by the conveying roller 1 alone; a second area on which the printing is carried out with the printing medium P being conveyed by both the conveying roller 1 and the discharge roller 12; and a third area on which the printing is carried out with the printing medium P is conveyed by the discharge roller 12 alone. In a case of three rotational bodies for conveying, the surface of the printing medium P can be divided into five, at the maximum, cases (areas) in a similar manner. In general terms, when the conveying is carried out by n rotary bodies for conveying (n≧2), the surface of the printing medium P can be divided into 2+½[n (n−1)] areas at the maximum.

[0237] Accordingly, in an allowable configuration, test patterns from which the amount of eccentricity can be detected are printed in the 2+½[n(n−1)] areas formed by dividing the surface of the printing medium P in accordance with the number n of rotational bodies to be used in the conveying. Correction values of eccentricity obtained by calculations using the patterns are reflected when the printing is carried out on the respective predetermined areas.

[0238] When correction of this kind is executed, 2+½[n(n−1)] test patterns as shown in FIG. 5 are printed. Note that the test patterns may be printed across plural pages. In addition, when the test patterns are printed, the following point should be taken into consideration. When the printing is carried out as in the case of ordinary printing, the structure of the printing apparatus sometimes imposes restriction on the length in the sub-scanning direction of the test pattern corresponding to the conveying carried out by at least one of the relevant rotational bodies for conveying. If this is the case, by releasing the pinch rollers, as described with reference to FIG. 6C, the test pattern corresponding to the target rotational body can be printed.

[0239] Moreover, during the process of switching the kind of and the number of the rotational bodies for conveying to be used for the actual conveying, the areas are switched through a shifting process. Specifically, such shifting process appears, for example, when the front end of the printing medium enters a rotational body for conveying, or when the rear end of the printing medium is released from a rotational body for conveying. In the shifting process, since there are conveying-error factors that depend on the outer-diameter component and the phase of the rotational body for conveying, the quality of image in the area corresponding to the shifting process (shifting portion) may sometimes be unstable. It is therefore very likely to achieve high-quality images by acquiring correction values for eccentricity individually for the shifting portions.

[0240] FIG. 46 illustrates the relationship between the printing areas and the correction values for eccentricity to be applied to the respective printing areas, and between the printing areas and the correction values for outer-diameter applied likewise according to this embodiment. The correction values that are similar to those in the first and the second embodiments are applied to the first to the third areas. The correction value for eccentricity and the correction value for outer-diameter for a fourth area—a transition (shifting) area where the printing medium P enters the discharge roller 12—are a fourth correction value for eccentricity and a fourth correction value for outer-diameter. The correction value for eccentricity and the correction value for outer-diameter for a fifth area—a transition (shifting) area where the printing medium P is removed from the conveying roller 1—are a fifth correction value for eccentricity and a fifth correction value for outer-diameter. To put it other way, the eccentricity correction in the fifth area is preferably executed using correction values based on the correction values for the second and third areas which are the areas before and after the rear-end of the printing medium P is removed from the conveying roller 1. Similarly, the correction value for the fourth area, where the front-end of the printing medium P enters the discharge roller 12, is determined based on the correction value for the second area (in which the printing medium P is conveyed by cooperation with the conveying roller 1 and the discharge roller 12) and on the correction value for the first area (in which the printing medium P is conveyed by only the conveying roller 1). In other words, in the FIG. 46 embodiment, there is a series of three successive conveying phases made up of a phase before a sheet starts or stops being conveyed by a given one of the plurality of rollers and a phase after the sheet starts or stops being so conveyed and a transition phase intermediate between the other two phases of the series. The correction information for the transition phase is different from the correction information for each of the other two phases. For example, the correction information for the transition phase may be derived from the correction information for each of the other two phases in the manner described in the next paragraph. Incidentally, in the FIG. 46 embodiment, there are 5 conveying phases but still only 3 conveying mechanisms as in the preceding embodiments. Thus, there does not have to be one-to-one correspondence between conveying phases and conveying mechanisms. The FIG. 46 embodiment uses 5 different sets of correction information for correcting for conveying errors in 5 different conveying phases including the transition phases (corresponding to the fourth and fifth areas).

[0241] A correction value used as a correction value for eccentricity for each transition phase (transition area such as the fourth or fifth area) may be formed by combining the eccentricity component of each of the rotational bodies for conveying that are used in the conveying corresponding to the printing areas respectively preceding and following the transition area in consideration. For example, a polynominal correction function, such as
y=a sin(2π/L1+θ)+b sin(2π/L2+φ),
is effective. Here, a is the amplitude component of the eccentricity correction function in the second area (central area of the printing medium), b is the amplitude component of the eccentricity correction function in the third area (rear end area of the printing medium), L1 is the circumferential length of the roller that is used in the second area (central area of the printing medium), L2 is the circumferential length of the roller that is used in the third area (rear end area of the printing medium), θ is the initial phase of the eccentricity correction function in the second area (central area of the printing medium) and φ is the initial phase of the eccentricity correction function in the third area (rear end area of the printing medium). In addition, the correction of outer-diameter for the transition area is preferably executed using a correction value formed by combining the correction values for outer-diameter of each of the areas that are adjacent to the transition area.

[0242] One embodiment of the present invention can provide a printing apparatus comprising: a plurality of rollers (1, 12) for conveying a printing medium (P), conveying controller for controlling the conveying of the printing medium (P) on the basis of a correction value to correct a conveying error for each one of the plurality of rollers (1, 12), wherein the conveying controller controls the conveying of the printing medium (P) on the basis of at least one correction value to correct the conveying error for at least one of the plurality of rollers (1, 12) that is actually involved in the conveying.

[0243] Another embodiment of the present invention can provide a printing apparatus comprising: a plurality of rollers (1, 12) for conveying a printing medium (P), and conveying controller for controlling the conveying of the printing medium (P) on the basis of a correction value to correct a conveying error for each one of the plurality of rollers (1, 12), wherein the printing medium (P) is divided into a plurality of areas in the conveying direction, and the conveying controller controls the conveying of the printing medium (P) on the basis of different correction values for the respective areas.

[0244] Another embodiment of the present invention can provide a method of controlling the conveying of a printing medium (P) employed for a printing apparatus that includes a plurality of rollers (1, 12) for conveying the printing medium (P), comprising: a step of controlling the conveying of the printing medium (P) on the basis of a correction value to correct a conveying error for each of the rollers (1, 12), wherein the conveying is controlled on the basis of at least one correction value to correct the conveying error for at least one of the plurality of rollers that is actually involved in the conveying of the printing medium (P).

[0245] This method of controlling the conveying of a printing medium (P) may further comprise the steps of: forming, on the printing medium (P), a test pattern (ER1, 2, FR1, 2) used to detect the conveying error for each of the plurality of rollers (1, 12); and acquiring, using the test pattern, the correction value to correct the conveying error.

[0246] While the present invention has been described with reference to exemplary embodiments, it is to be understood that the invention is not limited to the disclosed exemplary embodiments. The scope of the following claims is to be accorded the broadest interpretation so as to encompass all such modifications and equivalent structures and functions.

[0247] This application claims the benefit of Japanese Patent Application No. 2007-103307, filed Apr. 10, 2007, which is hereby incorporated by reference herein in its entirety.