Characteristic table generation system, characteristic table generation method and characteristic table generation program

Abstract

The present invention provides a characteristic table generation system and the like capable of improving user convenience. The characteristic table generation system according to one embodiment of the present disclosure is a system for generating a predicted voltage characteristic table that specifies a predicted characteristic curve, and includes: a data acquisition unit that acquires, as input data, predetermined parameters and a measured viscosity characteristic table that specifies a measured characteristic curve between viscosity and temperature of a liquid; a conversion coefficient generation unit that uses a first analysis method, which is a predetermined analysis method, to generate conversion coefficients based on the predetermined parameters, wherein the first analysis method uses the predetermined parameters as explanatory variables and uses, as target variables, conversion coefficients when performing conversion processing from the measured characteristic curve to the predicted characteristic curve; and a table generation unit that performs the conversion processing using the measured viscosity characteristic table and the conversion coefficients generated by the conversion coefficient generation unit to generate the predicted voltage characteristic table.

Description

Technical Field

[0001] The present disclosure relates to a characteristic table generating system, a characteristic table generating method, and a characteristic table generating program. Background Art

[0002] Liquid jet recording apparatuses including a liquid jet head are used in various fields, and various types of liquid jet heads have been developed (for example, see Patent Document 1).

[0003] Prior art literature

[0004] Patent Literature

[0005] Patent Document 1: Japanese Patent Application Laid-Open No. 2012-187850. Summary of the Invention

[0006] Problems to be solved by the invention

[0007] Such a liquid ejecting head is required to improve user convenience, and it is desirable to provide a characteristic table creation system, a characteristic table creation method, and a characteristic table creation program that can improve user convenience.

[0008] Solutions to Problems

[0009] A characteristic table generation system involved in one embodiment of the present disclosure is a system for generating a predicted voltage characteristic table, wherein the predicted voltage characteristic table specifies a predicted characteristic curve between temperature and a voltage value applied to an ejection portion that ejects liquid, wherein the voltage value represents a peak value of a pulse of a driving signal having one or more pulses and is based on a predetermined reference value. The characteristic table generation system comprises: a data acquisition unit that acquires, as input data, predetermined parameters and a measured viscosity characteristic table that specifies a measured characteristic curve between viscosity and temperature of a liquid; a conversion coefficient generation unit that uses a first analysis method that is a predetermined analysis method to generate a conversion coefficient based on the predetermined parameters, the first analysis method using the predetermined parameters as explanatory variables and using a conversion coefficient when performing a conversion process from the measured characteristic curve to the predicted characteristic curve as a target variable; and a table generation unit that performs the conversion process using the measured viscosity characteristic table and the conversion coefficient generated by the conversion coefficient generation unit to generate the predicted voltage characteristic table.

[0010] A characteristic table generation method involved in one embodiment of the present disclosure is a method for generating a predicted voltage characteristic table, wherein the predicted voltage characteristic table specifies a predicted characteristic curve between temperature and a voltage value applied to an ejection portion for ejecting liquid, wherein the voltage value represents a peak value of a driving signal having one or more pulses, based on a predetermined reference value. The characteristic table generation method includes the following steps: obtaining predetermined parameters and a measured viscosity characteristic table specifying a measured characteristic curve between viscosity and temperature of a liquid as input data; generating a conversion coefficient based on the predetermined parameters using a first analysis method as a predetermined analysis method, wherein the first analysis method uses the predetermined parameters as explanatory variables and uses the conversion coefficient when performing the conversion process from the measured characteristic curve to the predicted characteristic curve as a target variable; and performing the conversion process using the measured viscosity characteristic table and the generated conversion coefficient to generate the predicted voltage characteristic table.

[0011] A characteristic table generation program involved in one embodiment of the present disclosure is a program for generating a predicted voltage characteristic table, wherein the predicted voltage characteristic table specifies a predicted characteristic curve between temperature and a voltage value applied to an ejection portion for ejecting liquid, wherein the voltage value represents a peak value of a driving signal having one or more pulses, based on a predetermined reference value. The characteristic table generation program causes a computer to execute the following steps: obtaining predetermined parameters and a measured viscosity characteristic table specifying a measured characteristic curve between viscosity and temperature of a liquid as input data; generating a conversion coefficient based on the predetermined parameters using a first analysis method as a predetermined analysis method, wherein the first analysis method uses the predetermined parameters as explanatory variables and uses the conversion coefficient when performing the conversion process from the measured characteristic curve to the predicted characteristic curve as a target variable; and performing the conversion process using the measured viscosity characteristic table and the generated conversion coefficient to generate the predicted voltage characteristic table.

[0012] Effects of the Invention

[0013] According to the characteristic table generation system, characteristic table generation method, and characteristic table generation program according to one embodiment of the present disclosure, user convenience can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 It is a schematic perspective view showing an example of the schematic configuration of a liquid jet recording apparatus according to one embodiment of the present disclosure.

[0015] Figure 2 Yes Figure 1 The diagram is a schematic diagram of an example of the schematic configuration of a liquid ejecting head shown.

[0016] Figure 3 This is a functional block diagram showing a configuration example of a characteristic table generating system according to an embodiment.

[0017] Figure 4 Yes Figure 3 A physical block diagram of an example configuration of an information processing device is shown.

[0018] Figure 5 Yes Figure 3 、 Figure 4 A block diagram showing a detailed configuration example of a machine learning model.

[0019] Figure 6 2 is a timing chart schematically showing a configuration example of a drive signal.

[0020] Figure 7 It is a block diagram showing a schematic configuration example of a liquid jet recording apparatus according to a comparative example.

[0021] Figure 8 This is a diagram showing an example of viscosity information related to a comparative example.

[0022] Figure 9 It is a diagram showing an example of various characteristic curves according to a comparative example.

[0023] Figure 10 This is a flowchart showing an example of conversion processing according to the embodiment.

[0024] Figure 11 It is a diagram showing an example of various characteristic curves according to the embodiment.

[0025] Figure 12 This is a diagram showing an example of predetermined parameters according to the embodiment.

[0026] Figure 13 Yes Figure 12 An example of the importance analysis results of each parameter is shown in the figure.

[0027] Figure 14 This is a flowchart showing a process of generating a characteristic table according to the embodiment.

[0028] Figure 15 This is a diagram showing an example of predicted values ​​and measured values ​​using machine learning according to an embodiment.

[0029] Figure 16A Yes Figure 15 The graph shows an example of the correspondence between SVM predicted values ​​and measured values.

[0030] Figure 16B Yes Figure 15 The diagram shows an example of the correspondence between RF predicted values ​​and measured values.

[0031] Figure 17A It means that only Figure 13 A diagram showing an example of the correspondence between RF predicted values ​​and measured values ​​for some of the parameters shown.

[0032] Figure 17B It means that only Figure 13 A diagram showing an example of the correspondence between RF predicted values ​​and measured values ​​for some of the parameters shown.

[0033] Figure 17C It means that only Figure 13 A diagram showing an example of the correspondence between RF predicted values ​​and measured values ​​for some of the parameters shown.

[0034] Figure 18 This is a block diagram showing an example of the configuration of a machine learning model according to variant example 1.

[0035] Figure 19 This is a diagram showing an example of predetermined parameters according to Modification 1.

[0036] Figure 20 Yes Figure 19 An example of the importance analysis results of each parameter is shown in the figure.

[0037] Figure 21 This is a diagram showing an example of predicted values ​​and measured values ​​using machine learning according to Modification Example 1.

[0038] Figure 22A Yes Figure 21 The graph shows an example of the correspondence between SVM predicted values ​​and measured values.

[0039] Figure 22B Yes Figure 21 The diagram shows an example of the correspondence between RF predicted values ​​and measured values.

[0040] Figure 23 It means that only Figure 20 A diagram showing an example of the correspondence between RF predicted values ​​and measured values ​​for some of the parameters shown.

[0041] Figure 24 This is a block diagram showing an example of the configuration of a machine learning model according to variant example 2.

[0042] Figure 25 This is a diagram showing an example of predetermined parameters according to Modification 2.

[0043] Figure 26 Yes Figure 25 An example of the importance analysis results of each parameter is shown in the figure.

[0044] Figure 27 This is a diagram showing an example of predicted values ​​and measured values ​​using machine learning according to Modification Example 2.

[0045] Figure 28A Yes Figure 27 The graph shows an example of the correspondence between SVM predicted values ​​and measured values.

[0046] Figure 28B Yes Figure 27 The diagram shows an example of the correspondence between RF predicted values ​​and measured values.

[0047] Figure 29A It means that only Figure 26 A diagram showing an example of the correspondence between RF predicted values ​​and measured values ​​for some of the parameters shown.

[0048] Figure 29B It means that only Figure 26 A diagram showing an example of the correspondence between RF predicted values ​​and measured values ​​for some of the parameters shown.

[0049] Figure 29C It means that only Figure 26 A diagram showing an example of the correspondence between RF predicted values ​​and measured values ​​for some of the parameters shown.

[0050] Figure 30 This is a block diagram showing a configuration example of a characteristic table creation system according to Modification 3.

[0051] Figure 31 This is a block diagram showing a configuration example of a characteristic table creation system according to Modification 4.

[0052] Figure 32 This is a block diagram showing a configuration example of a characteristic table creation system according to Modification 5.

[0053] Figure 33 This is a block diagram showing a configuration example of an information processing unit according to Modification Example 6. DETAILED DESCRIPTION

[0054] Hereinafter, the embodiments of the present disclosure will be described in detail with reference to the accompanying drawings.

[0055] 1. Embodiment (Example in which an information processing unit is provided in an information processing device outside a liquid jet recording device)

[0056] 2. Modification

[0057] Modification 1 (Example of Further Setting a Machine Learning Model with Voltage Sensitivity as the Target Variable)

[0058] Modification 2 (Example of Further Setting a Machine Learning Model with Voltage Offset as the Target Variable)

[0059] Modification 3 (Example in which the Information Processing Unit is Provided in a Server External to the Liquid Jet Recording Apparatus)

[0060] Modification 4 (Example in which the Information Processing Unit is Provided Inside the Liquid Jet Head in the Liquid Jet Recording Device)

[0061] Modification 5 (Example in which the Information Processing Unit is Provided Outside the Liquid Jet Head in the Liquid Jet Recording Device)

[0062] Modification 6 (Example in which a signal generating unit is further provided in the information processing unit)

[0063] 3. Other variations

[0064] 1. Implementation Method

[0065] [A. Overall Structure of Printer 1]

[0066] Figure 1 The printer 1 is an inkjet printer that records (prints) images or characters on recording paper P, a recording medium, using ink 9 described later.

[0067] like Figure 1 As shown, the printer 1 includes a pair of transport mechanisms 2a and 2b; an ink tank 3; an ink supply tube 30; an inkjet head 4; and a scanning mechanism 6. These components are housed within a housing 10 having a predetermined shape. In the drawings used in this specification, the proportions of the components are appropriately altered to make them recognizable.

[0068] Here, the printer 1 corresponds to a specific example of a "liquid jet recording device" in the present disclosure, and the inkjet head 4 (the inkjet heads 4Y, 4M, 4C, and 4K described later) corresponds to a specific example of a "liquid jet head" in the present disclosure. Furthermore, the ink 9 corresponds to a specific example of a "liquid" in the present disclosure.

[0069] like Figure 1 As shown, the conveying mechanisms 2a and 2b are each a mechanism for conveying recording paper P along a conveying direction d (X-axis direction). Each of these conveying mechanisms 2a and 2b includes a grid roller 21, a pressure roller 22, and a drive mechanism (not shown). The drive mechanism rotates the grid roller 21 about its axis (in the Z-X plane) and is comprised of, for example, a motor.

[0070] (Ink Pot 3)

[0071] The ink tank 3 is a tank that contains ink 9. In this example, as the ink tank 3, Figure 1 As shown, four types of tanks are provided, each containing ink 9 of four colors: yellow (Y), magenta (M), cyan (C), and black (K). Specifically, an ink tank 3Y is provided for containing yellow ink 9, an ink tank 3M is provided for containing magenta ink 9, an ink tank 3C is provided for containing cyan ink 9, and an ink tank 3K is provided for containing black ink 9. Within the housing 10, these ink tanks 3Y, 3M, 3C, and 3K are arranged side by side along the X-axis direction.

[0072] Note that the ink tanks 3Y, 3M, 3C, and 3K have the same configuration except for the color of the ink 9 they contain, and therefore are collectively referred to as ink tanks 3 in the following description.

[0073] (Inkjet head 4)

[0074] The inkjet head 4 is a head that ejects (discharges) droplet-shaped ink 9 from a plurality of nozzles (nozzle holes Hn) described later to the recording paper P to record (print) images or characters. Figure 1 As shown, four types of heads are also provided, each of which ejects the four colors of ink 9 contained in the ink tanks 3Y, 3M, 3C, and 3K, respectively. Specifically, an inkjet head 4Y ejects yellow ink 9, an inkjet head 4M ejects magenta ink 9, an inkjet head 4C ejects cyan ink 9, and an inkjet head 4K ejects black ink 9. Within the housing 10, these inkjet heads 4Y, 4M, 4C, and 4K are arranged side by side along the Y-axis direction.

[0075] In addition, each inkjet head 4Y, 4M, 4C, 4K has the same structure except the color of the ink 9 used, so it is collectively referred to as the inkjet head 4 in the following description. In addition, the detailed structure example of the inkjet head 4 will be described later ( Figure 2 ).

[0076] The ink supply tube 30 is a tube for supplying ink 9 from the ink tank 3 to the inkjet head 4. The ink supply tube 30 is formed of, for example, a flexible hose having a degree of flexibility that can follow the operation of the scanning mechanism 6 described below.

[0077] (Scanning mechanism 6)

[0078] The scanning mechanism 6 is a mechanism for causing the inkjet head 4 to scan along the width direction (Y-axis direction) of the recording paper P. Figure 1 As shown, the scanning mechanism 6 includes: a pair of guide rails 61a and 61b extending in the Y-axis direction; a carriage 62 movably supported on these guide rails 61a and 61b; and a driving mechanism 63 that moves the carriage 62 in the Y-axis direction.

[0079] The drive mechanism 63 includes a pair of pulleys 631a and 631b disposed between the guide rails 61a and 61b; an endless belt 632 wound between the pulleys 631a and 631b; and a drive motor 633 for rotating the pulley 631a. Furthermore, the four inkjet heads 4Y, 4M, 4C, and 4K described above are arranged side by side along the Y-axis on the carriage 62.

[0080] Furthermore, the scanning mechanism 6 and the aforementioned transport mechanisms 2 a and 2 b constitute a moving mechanism for relatively moving the inkjet head 4 and the recording paper P.

[0081] [B. Detailed Structure of Inkjet Head 4]

[0082] Next, refer to Figure 2 , a detailed structural example of the inkjet head 4 is described.

[0083] Figure 2 An example of the general configuration of the inkjet head 4 is schematically shown.

[0084] like Figure 2 As shown, the inkjet head 4 includes a nozzle plate 41 , an actuator plate 42 , and a driving unit 49 .

[0085] The nozzle plate 41 and the actuator plate 42 correspond to a specific example of the “injection portion” in the present disclosure.

[0086] (Nozzle Plate 41)

[0087] The nozzle plate 41 is a plate made of a film material such as polyimide or a metal material, such as Figure 2 As shown, there are a plurality of nozzle holes Hn (refer to Figure 2 These nozzle holes Hn are formed side by side at predetermined intervals on a straight line (along the X-axis direction in this example).

[0088] (Actuator Plate 42)

[0089] The actuator plate 42 is made of a piezoelectric material, such as PZT (lead zirconate titanate). Multiple channels (not shown) are provided on the actuator plate 42. These channels function as pressure chambers for applying pressure to the ink 9 and are arranged parallel to each other at predetermined intervals. Each channel is defined by a driving wall (not shown) formed of a piezoelectric body, forming a concave groove in cross-section.

[0090] These channels include ejection channels for ejecting ink 9 and dummy channels (non-ejection channels) that do not eject ink 9. In other words, the ejection channels are filled with ink 9, while the dummy channels are not. Furthermore, each ejection channel communicates with the nozzle holes Hn in the nozzle plate 41, while each dummy channel does not communicate with the nozzle holes Hn. These ejection channels and dummy channels are arranged alternately side by side along a predetermined direction.

[0091] The inner side surfaces opposing each other in the above-mentioned drive wall are respectively provided with drive electrodes (not shown). Among the drive electrodes, there is a common electrode (common electrode) provided on the inner side surface facing the discharge channel and an active electrode (individual electrode) provided on the inner side surface facing the pseudo channel. These drive electrodes are electrically connected to the drive circuit in the drive substrate (not shown) via a plurality of lead-out electrodes formed on a flexible substrate (not shown). Thus, a drive voltage Vd (drive signal Sd) is applied to each drive electrode from the drive circuit including the drive unit 49 via the flexible substrate.

[0092] (Driver 49)

[0093] The driving unit 49 applies the driving voltage Vd (driving signal Sd) to the actuator plate 42 to expand or contract the ejection channel, thereby ejecting the ink 9 from each nozzle hole Hn (performing an ejection operation) (see FIG. Figure 2 Specifically, the driving unit 49 performs such an ejection operation using a driving signal Sd generated by the signal generating unit 48 described later.

[0094] [C. Overall Structure of the Characteristics Table Generating System 5]

[0095] Next, refer to Figures 3 to 6 Next, an overall configuration example of a characteristic table generating system 5 including a printer 1 having the above-described inkjet head 4 will be described.

[0096] Figure 3 A block diagram (functional block diagram) shows an example of the configuration of the characteristic table generating system 5 according to this embodiment. Figure 4 The block diagram (physical block diagram) shows Figure 3 The configuration example of the information processing device 7 (described later) shown in FIG. Figure 5 The block diagram shows Figure 3 、 Figure 4 A detailed configuration example of the machine learning model 74 is shown.

[0097] In addition, since the characteristic table generation method according to this embodiment is embodied in the characteristic table generation system 5 according to this embodiment, it will be described together below. This also applies to the modified examples (modifications 1 to 6) described later.

[0098] The characteristic table generating system 5 is a system for generating a predicted voltage characteristic table TPvp (a characteristic table that defines a predicted characteristic curve CPvp between a voltage value Vp representing a peak value of a pulse of a drive signal Sd based on a predetermined reference value and an ambient temperature Ta) described later. Figure 3 As shown, the characteristic table generating system 5 includes a printer 1 having the inkjet head 4 and an information processing device 7. The printer 1 and the information processing device 7 are connected to each other via a network 50.

[0099] Furthermore, such a network 50 is a network that uses, for example, a standard communication protocol (TCP / IP) used on the Internet for communication. Alternatively, such a network 50 may be a secure network that uses a communication protocol unique to that network for communication. Furthermore, such a network 50 is, for example, the Internet, an intranet, or a local area network. The connection between such a network 50 and the printer 1 and information processing device 7 may be a wired LAN (Local Area Network) such as Ethernet (registered trademark), a wireless LAN such as Wi-Fi (registered trademark), or a mobile phone line.

[0100] (Information Processing Device 7)

[0101] The information processing device 7 is a device located outside the printer 1 and is composed of, for example, a PC (Personal Computer). Figure 3 As shown in the (Functional Block Diagram), the information processing device 7 includes an input unit 71, a display unit 72, an information processing unit 73 and a machine learning model 74.

[0102] In addition, such information processing device 7 corresponds to a specific example of the “external device” in the present disclosure.

[0103] The input unit 71 receives instructions from the outside (e.g., a user) and outputs the received instructions to the information processing unit 73. Such an input unit 71 is composed of, for example, a keyboard or a mouse. Alternatively, the input unit 71 may be composed of, for example, a touch panel provided on the display unit 72 (display surface) of the information processing device 7.

[0104] The display unit 72 displays an image based on the video signal output from the information processing unit 73. The display unit 72 is configured using various display types (e.g., a liquid crystal display, a CRT (cathode ray tube) display, an organic EL (electroluminescence) display, etc.).

[0105] The information processing unit 73 is a part that performs various information processing, such as Figure 3As shown in FIG. 7 , the data acquisition unit 731, the conversion coefficient generation unit 732 and the table generation unit 733 are provided. Figure 4 As shown in the (physical block diagram), such an information processing unit 73 is composed of a control unit 75, a storage unit 76 and a network IF (Interface) 77. Figure 4 In the example of FIG, the input unit 71 , the display unit 72 , the control unit 75 , the storage unit 76 , and the network IF 77 are connected to each other via a bus 70 .

[0106] like Figure 3 As shown, the data acquisition unit 731 acquires the following data (input data) via the input unit 71 or the network 50. Specifically, the data acquisition unit 731 acquires the measured viscosity characteristic table TMvi (a characteristic table defining a measured characteristic curve CMvi between the viscosity Vi of the ink 9 and the ambient temperature Ta) and the predetermined parameter Pr as input data.

[0107] like Figure 3 As shown, the conversion coefficient generating unit 732 uses a predetermined analysis method (first analysis method) based on the parameter Pr acquired by the data acquiring unit 731 to generate a predetermined conversion process described later (refer to the conversion process from the above-mentioned measured characteristic curve CMvi to the above-mentioned predicted characteristic curve CPvp: Figure 10 、 Figure 11 ) when the conversion coefficient Kc. This established analysis method refers to an analysis method that uses the above-mentioned parameter Pr as an explanatory variable and the above-mentioned conversion coefficient Kc as a target variable. In addition, in the example of this embodiment, Figure 3 、 Figure 4 As shown, the conversion coefficient generation unit 732 generates the conversion coefficient Kc based on the parameter Pr using an analysis method using the machine learning model 74 described below.

[0108] As described above, such a machine learning model 74 is a prediction model obtained by performing machine learning with the parameter Pr as the explanatory variable and the conversion coefficient Kc as the target variable. Figure 5 As shown, when the parameter Pr (explanatory variable) is input, the machine learning model 74 generates (predicts) the conversion coefficient Kc (target variable) based on the learning result and outputs the generated conversion coefficient Kc.

[0109] Examples of analysis methods (prediction methods) using the machine learning model 74 include support vector machines (SVM), random forests (RF), and multiple regression analysis.

[0110] like Figure 3As shown, the table generation unit 733 performs the above-described predetermined conversion process using the measured viscosity characteristic table TMvi acquired by the data acquisition unit 731 and the conversion coefficient Kc generated by the conversion coefficient generation unit 732, thereby generating a predicted voltage characteristic table TPvp. The predicted voltage characteristic table TPvp generated in this manner is supplied via the network 50 to the signal generation unit 48, described later, within the inkjet head 4 of the printer 1.

[0111] In addition, the details of each process in such an information processing section 73 (the data acquisition section 731 , the conversion coefficient generation section 732 , and the table generation section 733 ) will be described later.

[0112] Figure 4 The control unit 75 shown in FIG. 1 includes a CPU (Central Processing Unit) and a GPU (Graphics Processing Unit), and is configured to execute various programs stored in the storage unit 76. Specifically, for example, Figure 4 As shown, the control unit 75 executes a program 730 stored in the storage unit 76. This program 730 is a program for executing the various processes in the information processing unit 73 (data acquisition unit 731, conversion coefficient generation unit 732, and table generation unit 733). Specifically, this program 730 is a program for causing the computer (control unit 75) to execute the various functions in the information processing unit 73 (data acquisition unit 731, conversion coefficient generation unit 732, and table generation unit 733).

[0113] The storage unit 76 is a part that stores various programs executed by the control unit 75 or various data. Figure 4 As shown, the storage unit 76 stores the aforementioned program 730 as an example of such various programs, and stores the aforementioned machine learning model 74 as an example of such various data. Such a storage unit 76 is configured using, for example, a RAM (Random Access Memory), a ROM (Read Only Memory), an auxiliary storage device (such as a hard disk), and the like.

[0114] like Figure 4 As shown, the network IF 77 is a communication interface for communicating with the printer 1 via the network 50 .

[0115] (Signal generation unit 48)

[0116] Here, in Figure 3 In the illustrated example, the inkjet head 4 includes a signal generating section 48 in addition to the aforementioned nozzle plate 41 , actuator plate 42 , and driving section 49 .

[0117] As described above, the signal generating unit 48 generates a drive signal Sd having one or more pulses (pulse width Wp, voltage value Vp indicating a peak value) using the predicted voltage characteristic table TPvp generated by the table generating unit 733 in the information processing device 7 .

[0118] Here, Figure 6 (A)~ Figure 6 (C) schematically shows the configuration example of such a drive signal Sd in a timing diagram. Figure 6 (A)~ Figure 6 In (C), the horizontal axis represents time t, and the vertical axis represents the drive voltage Vd (positive voltage in this example) in the drive signal Sd.

[0119] first, Figure 6 The drive signal Sd shown in (A) has one pulse (pulse Pa), which is an example of a so-called "one drop." This pulse Pa is provided during the on-time between the rising and falling timings, and has a pulse width Wpa1 and a voltage value Vp1 as an example of the aforementioned pulse width Wp and voltage value Vp.

[0120] On the other hand, as a pulse applying the so-called "multi-pulse method", Figure 6 The drive signal Sd shown in (B) has the following two pulses (pulses Pa and Pb) (an example of a so-called "two-fall" pattern). Specifically, two pulses Pa and Pb are provided as such pulses (ON periods). Furthermore, an OFF period ("OFF1") is provided between these two pulses Pa and Pb. Furthermore, as an example of the aforementioned pulse width Wp and voltage value Vp, pulse Pa has a pulse width Wpa2 and a voltage value Vp2, and pulse Pb has a pulse width Wpb2 and a voltage value Vp2.

[0121] Similarly, as a pulse to which the above-mentioned "multi-pulse method" is applied, Figure 6 The drive signal Sd shown in (C) has the following three pulses (pulses Pa, Pb, and Pc) (an example of a so-called "three-fall" pattern). Specifically, three pulses Pa, Pb, and Pc are provided as such pulses (ON periods). Furthermore, an OFF period ("OFF1") is provided between pulses Pa and Pb, and an OFF period ("OFF2") is provided between pulses Pb and Pc. Furthermore, as an example of the aforementioned pulse width Wp and voltage value Vp, pulse Pa has a pulse width Wpa3 and a voltage value Vp3, pulse Pb has a pulse width Wpb3 and a voltage value Vp3, and pulse Pc has a pulse width Wpc3 and a voltage value Vp3.

[0122] Furthermore, each of the pulses Pa, Pb, and Pc in the drive signal Sd is a positive pulse that expands the aforementioned discharge channel during a high-level (High) state and contracts the discharge channel during a low-level (Low) state.

[0123] Here, the signal generating unit 48 (described in detail later) sets the pulse width Wp and voltage value Vp of each of these pulses (pulses Pa, Pb, and Pc), and generates the drive signal Sd using the pulses having the set pulse width Wp and voltage value Vp. Specifically, the signal generating unit 48 uses the predicted voltage characteristic table TPvp described above to determine the voltage value Vp of the pulses, and generates the drive signal Sd using the pulses having the determined voltage value Vp.

[0124] Here, the voltage value Vp mentioned above corresponds to a specific example of the "peak value" in the present disclosure. In addition, the "pulse" mentioned above does not only refer to, for example, Figure 6 The rectangular wave shown also refers to a concept including waveforms such as trapezoidal waves, triangular waves, and step waves, and the same applies to the following.

[0125] [Actions and effects]

[0126] (A. Basic Operation of Printer 1)

[0127] In the printer 1, the recording operation (printing operation) of an image or a character on the recording paper P is performed as follows. Figure 1 The four types of ink tanks 3 (3Y, 3M, 3C, 3K) shown are each sufficiently filled with ink 9 of the corresponding color (four colors). The ink 9 in the ink tanks 3 is then filled into the inkjet head 4 via the ink supply tubes 30 .

[0128] In this initial state, when the printer 1 is operated, the grid rollers 21 in the conveying mechanisms 2a and 2b rotate, respectively, and the recording paper P is conveyed between the grid rollers 21 and the pinch roller 22 along the conveying direction d (X-axis direction). Simultaneously with this conveying operation, the drive motor 633 in the drive mechanism 63 rotates the pulleys 631a and 631b, respectively, thereby operating the endless belt 632. As a result, the carriage 62 reciprocates along the width direction (Y-axis direction) of the recording paper P, guided by the guide rails 61a and 61b. At this time, the inkjet heads 4 (4Y, 4M, 4C, 4K) appropriately eject four colors of ink 9 onto the recording paper P, thereby recording images, characters, etc., on the recording paper P.

[0129] (B. Detailed Operation of the Inkjet Head 4)

[0130] Next, the detailed operation (ejection operation of the ink 9) in the inkjet head 4 will be described. Specifically, in the inkjet head 4, the ejection operation of the ink 9 using the shear (sharing) mode is performed as follows.

[0131] First, the driving unit 49 applies a driving voltage Vd (driving signal Sd) to the driving electrodes (common electrode and active electrode) in the actuator plate 42 (see Figure 2 、 Figure 3 Specifically, the driving unit 49 applies a driving voltage Vd to each driving electrode disposed on a pair of driving walls defining the ejection channel. As a result, the pair of driving walls are deformed so as to protrude toward the dummy channel adjacent to the ejection channel.

[0132] At this point, the drive wall bends and deforms in a V-shape, centered at the middle of the depth direction of the drive wall. This bending deformation of the drive wall causes the discharge channel to deform as if expanding. This bending deformation, caused by the piezoelectric thickness slip effect of the pair of drive walls, increases the volume of the discharge channel. This increased volume of the discharge channel allows ink 9 to be directed into the discharge channel.

[0133] The ink 9 thus guided into the ejection channel transforms into a pressure wave and propagates within the ejection channel. Furthermore, at the time (or near the time) when this pressure wave reaches the nozzle hole Hn of the nozzle plate 41, the drive voltage Vd applied to the drive electrode becomes 0 (zero) V. This causes the drive wall to recover from the aforementioned curved and deformed state, and the ejection channel volume, which had temporarily increased, returns to its original state.

[0134] In this way, while the volume of the ejection channel returns to its original state, the pressure inside the ejection channel increases, and the ink 9 in the ejection channel is pressurized. As a result, the ink 9 in droplet form is ejected to the outside through the nozzle hole Hn (toward the recording paper P) (see Figure 2 、 Figure 3 In this manner, the ink 9 in the inkjet head 4 is ejected (discharged), and as a result, an image or text is recorded (printed) on the recording paper P.

[0135] (C. Generating the property table, etc.)

[0136] Next, in addition to the reference Figures 1 to 6 In addition, refer to Figures 7 to 17C , and the comparison example ( Figures 7 to 9 ) for comparison, while explaining in detail the generation action (generation processing) of the characteristic table (the aforementioned predicted voltage characteristic table TPvp) in the characteristic table generation system 5.

[0137] (C-1. Comparative Example)

[0138] Figure 7A block diagram shows a schematic configuration example of a printer 101 as a liquid jet recording apparatus according to a comparative example. The printer 101 of the comparative example includes the aforementioned nozzle plate 41, actuator plate 42, signal generator 48, and driver 49 in an inkjet head (not shown).

[0139] However, in the printer 101 of this comparative example, unlike the printer 1 of the embodiment, the signal generating unit 48 sets the voltage value Vp using the viscosity information Iv described below instead of the predicted voltage characteristic table TPvp described above.

[0140] Figure 8 An example of viscosity information Iv related to such a comparative example is shown. Figure 8 In FIG. 1 , an example of the correspondence between the ambient temperature Ta and the viscosity Vi (actual value) of the ink 9, the voltage value Vp (actual value) in the pulse of the drive signal Sd, and the difference ΔV (=Vi-Vp) between the viscosity Vi and the voltage value Vp (including information on viscosity information Iv) is shown. In other words, in this Figure 8 In the example, the characteristic curve between the viscosity Vi (measured value) and the ambient temperature Ta (measured characteristic curve CMvi), the characteristic curve between the voltage value Vp (measured value) and the ambient temperature (measured characteristic curve CMvp), and the characteristic curve between the above-mentioned differential value ΔV and the ambient temperature Ta are shown respectively.

[0141] In addition, the above-mentioned ambient temperature Ta corresponds to a specific example of “temperature” in the present disclosure.

[0142] Incidentally, the "viscosity Vi of the ink 9" referred to herein refers to the static viscosity, and the same applies hereinafter. The viscosity Vi of the ink 9 can be measured using, for example, a rotational viscometer, a vibration viscometer, or a viscometer employing another measurement method, such as a capillary or falling ball viscometer (a viscometer capable of measuring static viscosity).

[0143] In this comparative example, first, for example, the viscosity Vi of the ink 9 is detected relative to the change in the ambient temperature Ta (for example, multi-point measurement at 5 or more points) to obtain the following information: Figure 8 Viscosity information Iv shown. In addition, for example, Figure 8 As shown in FIG. 1 , it is known that the change in viscosity Vi of the ink 9 relative to the ambient temperature Ta and the change in voltage value Vp (voltage value Vp for obtaining the ejection speed as a reference) relative to the ambient temperature Ta show similar change characteristics. Thus, for example, Figure 8 As shown in FIG. 1 , the difference value ΔV between the viscosity Vi and the voltage value Vp exhibits a substantially constant value and does not depend on the ambient temperature Ta.

[0144] Moreover, if Figure 7As shown, the signal generating unit 48 in this comparative example utilizes the similarity of the temperature change characteristics and subtracts the difference value ΔV (negative value) calculated in advance from the value of the viscosity Vi at a certain ambient temperature Ta (refer to the viscosity information Iv), thereby obtaining the voltage value Vp that can obtain the reference discharge speed. That is, the signal generating unit 48 in this comparative example uses the relationship of Vp = (Vi - ΔV) (refer to Figure 8 )Calculate the voltage value Vp at a certain ambient temperature Ta.

[0145] However, the characteristic curve between the voltage value Vp and the ambient temperature Ta (the aforementioned measured characteristic curve CMvp) generally exhibits different slopes depending on the number of pulses included in the drive signal Sd, the type and function of each pulse (e.g., the type and function of each pulse when including auxiliary pulses). Therefore, in this comparative example, manual measurement is essentially required to obtain the measured characteristic curve CMvp. Furthermore, under limited conditions (e.g., the aforementioned "one-step drop" based on the ejection speed), derivation of the measured characteristic curve CMvp can be performed without actually measuring it.

[0146] Thus, for example, it is necessary to actually measure each type of pulse number included in the drive signal Sd to obtain the measured characteristic curve CMvp. Therefore, the user of the printer 101 of this comparative example needs a lot of time and effort, which increases the workload and cost.

[0147] Here, Figure 9 An example of various characteristic curves (measured characteristic curve CMvp and measured characteristic curve CMvi) involved in the comparative example is shown. Figure 9 The measured characteristic curve CMvp shown in FIG. 1 shows the cases where the number of pulses (the aforementioned number of drops) is 1 drop (denoted as "1d"), 3 drops (denoted as "3d"), 7 drops (denoted as "7d"), and 9 drops (denoted as "9d"). Figure 9 Each measured characteristic curve CMvp shown in FIG. 1 shows a voltage value Vp based on a predetermined reference value. Figure 9 The measured characteristic curves CMvp shown in the figure respectively show the voltage value Vp (referred to as "Vj reference") which can be used to obtain the ejection speed as a reference when the ink 9 is ejected, and the voltage value Vp (referred to as "DV reference") which can be used to obtain the drop volume (DV) of the ink 9 as a reference. Figure 9 The driving waveforms for the various characteristic curves shown in FIG. 1 are for all conditions (number of drops) and include the case of “common driving” described later.

[0148] In this Figure 9In the example shown, as described above, the slope of the measured characteristic curve CMvp differs depending on the type of pulse number (falling number) or the type of the predetermined reference value (Vj reference or DV reference). Figure 8 If a single measured characteristic curve CMvp is repeatedly used when generating the drive signal Sd, as in the viscosity information Iv of the comparative example shown in FIG, the accuracy of setting the voltage value Vp will be reduced due to differences in the slope corresponding to the type of pulse number, the type of predetermined reference value, the type / function of each pulse, etc. In other words, it will be difficult to accurately set the voltage value Vp (peak value) of the pulses in the drive signal Sd.

[0149] Specifically, in this comparative example, for example, only a single voltage characteristic table can be created based on the measured characteristic curve CMvi (for example, the "single drop" case based on the ejection speed, as described above). Furthermore, as described above, obtaining the measured characteristic curve CMvp for each condition (for each type of pulse number, etc.) requires substantial effort in actual measurement. Due to these circumstances, the method of this comparative example may compromise user convenience due to factors such as reduced accuracy in setting the voltage value Vp and increased workload on the user.

[0150] (C-2. This embodiment)

[0151] Therefore, in this embodiment, the aforementioned information processing unit 73 (program 730) uses the aforementioned established characterization method to generate the conversion coefficient Kc for the conversion processing described below, and uses this conversion coefficient Kc to generate (automatically generate) the aforementioned characteristic table (predicted voltage characteristic table TPvp that specifies the predicted characteristic curve CPvp) at any time.

[0152] Here, Figure 10 An example of the conversion process described later in this embodiment is shown in a flowchart (which is similar to the conversion process described later). Figure 14 Specific example of the processing in step S13 in FIG. 1 ). In addition, Figure 11 The various characteristic curves involved in this embodiment (executing Figure 10 Specifically, in the Figure 11 In FIG. 1 , an example of various characteristic curves (such as the aforementioned measured characteristic curve CMvi and the preliminary characteristic curve CPvp0 of the aforementioned predicted characteristic curve CPvp) showing the corresponding relationship between the viscosity Vi [mPa] or the voltage value Vp of the ink 9 and the ambient temperature Ta [°C] is shown. In addition, for convenience, Figure 11The preliminary characteristic curve CMvp0 shown in FIG is a characteristic curve obtained by performing a predetermined process (processing to set the voltage value Vp=0 at a predetermined reference temperature Tr described later) on the measured characteristic curve CMvp described above, so as to facilitate comparison (comparison of slopes) with the preliminary characteristic curve CPvp0 described above. Figure 12 An example of the predetermined parameter Pr involved in this embodiment is shown. Figure 13 Expressed Figure 12 An example of the importance analysis results of each parameter Pr shown in FIG. Figure 12 In FIG, the values ​​of each parameter Pr are shown for 6 samples ("Sample 1" to "Sample 6"). Figure 13 The importance in the importance analysis results shown is an indicator (contribution rate) that measures how much the division of the feature value contributes to the classification of the target, and is calculated using a predetermined calculation formula based on the so-called Gini impurity.

[0153] (About conversion processing)

[0154] First, for example Figure 10 、 Figure 11 As shown, in this embodiment, the conversion process using the conversion coefficient Kc is a process of converting the measured characteristic curve CMvi (the characteristic curve between the measured value of the viscosity Vi and the ambient temperature Ta) into the predicted characteristic curve CPvp (the characteristic curve between the predicted value of the voltage value Vp and the ambient temperature Ta). Figure 11 As shown in the example, it can be seen that the preliminary characteristic curve CPvp0 obtained during such conversion processing is highly consistent (approximately consistent) with the preliminary characteristic curve CMvp0 for the aforementioned measured characteristic curve CMvp (the characteristic curve between the measured value of the voltage value Vp and the ambient temperature Ta).

[0155] Here, refer to Figure 10 、 Figure 11 A specific example of such conversion processing (conversion processing from the measured characteristic curve CMvi to the predicted characteristic curve CPvp) will be described.

[0156] In this conversion process, first, a multiplication process (CMvi×Kc) is performed to multiply the measured characteristic curve CMvi by the conversion coefficient Kc ( Figure 10 Then, at a predetermined reference temperature Tr (at Figure 11In the example of Tr = 40 ° C), the multiplication result in step S131 is subtracted to generate the above-mentioned preliminary characteristic curve CPvp0 (preliminary characteristic curve between the predicted value of voltage value Vp and ambient temperature Ta) (step S132). That is, through such preliminary processing (processing of steps S131 and S132), the conversion coefficient Kc is used to generate, for example, the measured characteristic curve CMvi. Figure 11 In addition, the execution order of each process of step S131 and S132 during such a preliminary process may be, for example, the same as Figure 10 The example shown is executed in the reverse order (step S132 is executed first, and then step S131 is executed).

[0157] Then, to become Figure 9 In the manner of the voltage value Vp under the aforementioned (Vj reference or DV reference), a predetermined voltage offset ΔVp is added to the voltage value Vp in the preliminary characteristic curve CPvp0 (CPvp0+ΔVp), and a final predicted characteristic curve CPvp is generated (step S133). That is, the voltage value Vp after adding the voltage offset ΔVp (the voltage value Vp in the predicted characteristic curve CPvp) corresponds to the voltage value Vp that can obtain the reference ejection speed when ejecting ink or the voltage value Vp that can obtain the reference droplet amount of ink 9. In this way, the final predicted characteristic curve CPvp is generated, and the process ends. Figure 10 A series of conversion processes are shown.

[0158] Incidentally, a specific conversion formula when performing such conversion processing is expressed by the following formula (1) using the above-mentioned conversion coefficient Kc.

[0159] H=(H0×e(E / kT)) / Kc……(1)

[0160] H: The viscosity value of ink 9 after conversion

[0161] H0: constant

[0162] T: Absolute temperature (ambient temperature Ta)

[0163] E: activation energy

[0164] k: Boltzmann constant

[0165] In addition, the formula (1) without the conversion coefficient Kc is called the Arrhenius formula (law), which is generally known to people. In addition, in the formula (1), the Arrhenius formula is divided by the conversion coefficient Kc. This is to use (the viscosity value of ink 9 / the measured value of the voltage value Vp) for calculation when the analysis method using the machine learning model 74 is used. Therefore, for example, when the analysis method using the machine learning model 74 is performed, when (the measured value of the voltage value Vp / the viscosity value of ink 9) is used for calculation, the formula obtained by multiplying the Arrhenius formula by the conversion coefficient Kc becomes the conversion formula when performing the above-mentioned conversion process. That is, it can be said that any of these can be used as the conversion formula when performing the conversion process.

[0166] (Regarding the given parameter Pr)

[0167] Here, as the above-mentioned predetermined parameter Pr, Figure 12 As shown, the following contents (a) to (l) are given as an example.

[0168] (a) Number of drops (pulses): Figure 6 The number of pulses contained in a unit period of the driving signal Sd is equivalent to;

[0169] (b) Presence of common drive ("0": not present; "1": present; "2": special value): so-called common drive (a driving method in which the pulses of the drive signal Sd are set so that the volume of the discharge channel during discharge of the ink 9 includes a change that is more contracted than a reference value);

[0170] (c) Head type: a symbol indicating the type of inkjet head 4;

[0171] (d) Ink type: Type of ink 9 classified according to the main solvent of ink 9 ("Oil": ink 9 with oil-based solvent; "Sol": ink 9 with organic solvent; "UV": UV (ultraviolet) curing ink; "WB": water-based ink 9);

[0172] (e) (Vj reference or DV reference): indicates Figure 9 Which of the following parameters is used?

[0173] (f) Head level value: This corresponds to the voltage value Vp at a predetermined ejection speed when a predetermined test liquid is ejected from the inkjet head 4. This is a value specific to the inkjet head 4 (unit: [V]).

[0174] (g) Viscosity value at reference temperature Tr: viscosity value of the ink 9 at the reference temperature Tr when the ink 9 is heated and used (unit: [mPa]);

[0175] (h) Voltage sensitivity Vr during ejection (Vj reference or DV reference): a value corresponding to the change in voltage per unit of ejection speed or droplet volume of the ink 9 when the ink 9 is ejected at the reference temperature Tr (unit: [m / s / V] or [pl / V]);

[0176] (i) Surface tension of ink 9 (unit: [mN / m]);

[0177] (j) specific gravity value of the ink 9 (or a physical property value obtained using the specific gravity value of the ink 9 (for example, the density of the ink 9 or the speed of sound in the ink 9 ));

[0178] (k) voltage offset ΔVp;

[0179] (1) The target value of the ejection speed of the ink 9 or DV (droplet volume).

[0180] In addition, according to Figure 13 As shown in FIG. 1 , an example of the importance analysis result of each parameter Pr as an explanatory variable is shown. When generating the conversion coefficient Kc using the machine learning model 74, the following parameters have a relatively high importance (contribution rate) among the parameters Pr shown in (a) to (l). That is, in Figure 13 In the example shown, the importance is higher in the order of (j) specific gravity value of ink 9 (or the above-mentioned physical property value), (a) drop number, (g) viscosity value at the reference temperature Tr, (k) voltage offset ΔVp, (h) voltage sensitivity Vr during ejection, (l) ejection speed of ink 9 or target value of DV, (d) ink type, (b) presence or absence of common drive, (i) surface tension value of ink 9, (f) head grade value, (e) (Vj reference or DV reference), and (c) head type.

[0181] Therefore, in this embodiment, as the predetermined parameter Pr, it is preferable to include at least the specific gravity value (or the above-mentioned physical property value) of the ink 9, which is the most important among them. In addition, it can be said that as the predetermined parameter Pr, it is preferable to further include Figure 13 Among the remaining parameters shown in the figure, the parameters with relatively high importance (as an example, the 2nd to 5th most important parameters) are at least one of (a) the number of drops, (g) the viscosity value at the reference temperature Tr, (k) the voltage offset ΔVp, and (h) the voltage sensitivity Vr during discharge.

[0182] (Details on the generation process of the characteristic table, etc.)

[0183] Here, Figure 14 The generation process of the characteristic table (predicted voltage characteristic table TPvp) according to this embodiment is shown in a flowchart. Figure 14Among the series of processes (steps S10 to S16 described later), steps S11 to S13 described later correspond to the generation of the predicted voltage characteristic table TPvp, and steps S14 and S15 described later correspond to the generation of the drive signal Sd.

[0184] In this Figure 14 In the series of processes shown, the information processing unit 73 (program 730) first determines, as a starting stage, whether it is necessary to generate (update) a predicted voltage characteristic table TPvp defining the predicted characteristic curve CPvp (step S10). If it is determined that generation of the predicted voltage characteristic table TPvp is necessary (step S10: Yes (Y)), the process proceeds to the generation process of the predicted voltage characteristic table TPvp (steps S11 to S13) described below. On the other hand, if it is determined that generation of the predicted voltage characteristic table TPvp is not necessary (step S10: No (N)), the process proceeds to step S15, described below, to generate the drive signal Sd using the pulse having the voltage value Vp (peak value) at the current stage.

[0185] In addition, as cases where the predicted voltage characteristic table TPvp needs to be generated, for example, the following cases can be cited. That is, for example, a case where a predetermined time has passed, or an ink cartridge of the ink tank 3 is installed, or a predetermined operation signal from the user is input to the printer 1, or a case where the non-dispensing period (idle period) of the ink 9 is longer than a predetermined time, etc. In addition, for example, a case where the color or type of the ink 9 in the ink tank 3 is changed, or a case where a different type of inkjet head 4 is installed in the printer 1 can be cited. Furthermore, as Figure 12 As shown, there is also the case where, for example, at least one of the parameters Pr is changed.

[0186] (Steps S11 to S13: Processing of Generating Predicted Voltage Characteristics Table TPvp)

[0187] Next, in the process of generating the predicted voltage characteristic table TPvp (steps S11 to S13), the data acquisition unit 731 first acquires the following data (input data). Specifically, the data acquisition unit 731 acquires the predetermined parameter Pr and the measured viscosity characteristic table TMvi, which specifies the measured characteristic curve CMvi between the viscosity Vi of the ink 9 and the ambient temperature Ta, as input data using the aforementioned method (step S11).

[0188] Next, the conversion coefficient generator 732 generates a conversion coefficient Kc based on the parameter Pr using a predetermined analysis method (first analysis method) that uses the parameter Pr obtained in step S11 as an explanatory variable and the aforementioned conversion coefficient Kc as a target variable (step S12). Specifically, in this embodiment, the conversion coefficient generator 732 generates the conversion coefficient Kc based on the parameter Pr using an analysis method that utilizes the aforementioned machine learning model 74.

[0189] Then, the table generating unit 733 performs the aforementioned predetermined conversion processing (see Figure 10 、 Figure 11 ), thereby generating a predicted voltage characteristic table TPvp (step S13). In this way, as described above, a predicted voltage characteristic table TPvp is generated that specifies the predicted characteristic curve CPvp between the voltage value Vp (peak value) of the pulse of the drive signal Sd and the ambient temperature Ta.

[0190] (Steps S14 and S15: Generation of Drive Signal Sd)

[0191] Next, in the process of generating the drive signal Sd (steps S14 and S15), the signal generating unit 48 first uses the predicted voltage characteristic table TPvp generated in step S13 and generates the predicted voltage characteristic table TPvp by the aforementioned method (see Figure 6 ) The voltage value Vp (peak value) in the pulse of the drive signal Sd is obtained (step S14). Specifically, the voltage value Vp of the pulse is obtained by applying the current ambient temperature Ta to the predicted voltage characteristic table TPvp.

[0192] Then, the signal generating unit 48 generates the aforementioned pulse using the voltage value Vp obtained in step S14 and a pulse width Wp set in advance, for example. Figure 6 (A) to Figure 6 (C) shows a driving signal Sd (step S15).

[0193] Incidentally, the pulse width Wp is calculated based on, for example, the on-pulse peak value (AP) of the pulse. This AP corresponds to half the natural vibration period of the ink 9 within the ejection channel (1AP = (natural vibration period of the ink 9) / 2). Furthermore, when the pulse width Wp is set to AP, the ejection velocity (ejection efficiency) of the ink 9 is typically maximized when ejecting a single drop of ink 9 (1 drop ejection). Furthermore, this AP is determined, for example, by the shape of the ejection channel or the physical properties of the ink 9 (such as specific gravity).

[0194] In addition, based on such AP, the pulse width Wp is set, for example, by the following process. Figure 6(A) to Figure 6 For the example of the driving signal Sd shown in (C) (the examples of so-called "1 fall", "2 falls", and "3 falls"), the signal generating unit 48 sets the pulse width Wp as follows. Figure 6 (A) to Figure 6 In the example (C), the signal generating unit 48 sets the pulse width Wp so that each pulse width Wp satisfies the relationship expressed by the following formulas (2) and (3) with AP. However, the present invention is not limited to the examples expressed by these formulas (2) and (3), and each pulse width Wp can be set as appropriate.

[0195] (1.25×AP)≤(Wpa1, Wpa2, Wpa3, Wpb2, Wpb3, Wpc3)≤(1.75×AP)……(2)

[0196] (Wpa1)≥(Wpa2, Wpb2)≥(Wpa3, Wpb3, Wpc3)……(3)

[0197] (Step S16: Ink 9 Ejection Operation)

[0198] Next, the driving unit 49 applies the driving signal Sd generated in step S15 to the actuator plate 42 in the inkjet head 4 to eject the ink 9 from the nozzle holes Hn (step S16).

[0199] Above, end Figure 14 A series of processing shown.

[0200] (C-3. ​​Example of prediction results using machine learning)

[0201] Here, Figure 15 An example of the predicted value and the measured value using machine learning according to this embodiment is shown. Figure 16A express Figure 15 An example of the correspondence between the SVM predicted value (mentioned above) and the measured value is shown. Figure 16B express Figure 15 An example of the correspondence between the RF predicted value (mentioned above) and the measured value is shown. 17A to 17C Respectively indicate that only Figure 13 An example of the correspondence between RF predicted values ​​and measured values ​​for some of the parameters shown. Figure 17A Shown only use Figure 13 An example of such a correspondence relationship is shown in the case of the parameter with the highest importance (specific gravity value of ink 9) among the parameters shown. Figure 17B Shown only use Figure 13An example of such a correspondence relationship is shown in the case of the two parameters with the highest and second highest importance (specific gravity value and drop number of ink 9). Figure 17C Shown only use Figure 13 This is an example of such a correspondence relationship when the two parameters with the highest importance (specific gravity value of ink 9 and voltage sensitivity Vr) are the first and fifth most important parameters among the parameters shown. Figure 16A 、 Figure 16B The use of Figure 12 An example of the correspondence between predicted values ​​and measured values ​​for all parameters shown.

[0202] In addition, Figure 15 In the example shown, for the aforementioned 6 samples ("Sample 1" to "Sample 6"), the measured values ​​of the aforementioned conversion coefficient Kc and the predicted values ​​of the conversion coefficient Kc using machine learning (SVM predicted values ​​and RF predicted values) are respectively shown. Figure 16A 、 Figure 16B 、 17A to 17C In the example shown, the (x, y) coordinates of many (562) samples are plotted, with the predicted value of the conversion coefficient Kc (SVM predicted value or RF predicted value) as the variable x and the measured value of the conversion coefficient Kc as the variable y. Figure 16A 、 Figure 16B 、 17A to 17C , an example of a formula (for example, a linear function formula determined using the least squares method) that expresses the trend of the correlation between these variables x and y is also shown.

[0203] First, in Figure 15 In the example shown, it can be seen that both the SVM prediction value and the RF prediction value can predict the measured value with good accuracy. Figure 16A 、 Figure 16B 、 17A to 17C In each example shown, since the slope of the linear function formula is approximately "1" and the intercept of the linear function formula is approximately "0", the predicted value (SVM predicted value and RF predicted value) and the measured value have the following relationship. That is, it can be seen that the predicted value and the measured value have a sufficient correlation to the extent that it is practical when printing using the predicted value. Figure 16B 、 17A to 17C From the examples shown (examples of RF predicted values), it can be seen that the predicted values ​​(RF predicted values) have a good agreement accuracy (prediction accuracy) with the measured values. Figure 16B The example shown (the example using all parameters above) is relatively higher than 17A to 17CThe examples shown above (the case where only some parameters are used). Furthermore, as mentioned above, Figure 17A Only the most important parameter (the specific gravity of ink 9) is used in Figure 17B Use the first and second most important parameters in Figure 17C The two parameters with the highest importance are used. However, for example, Figure 12 、 Figure 13 When the combination of other parameters shown in and the specific gravity value of ink 9 is used as an explanatory variable, Figure 17B 、 Figure 17C The same is true for Figure 17A The case (the case of using only the specific gravity value of ink 9) is high.

[0204] (C-4. Action / Effect)

[0205] As described above, the characteristic table generation system 5 of this embodiment uses the aforementioned predetermined analysis method (first analysis method) to generate the conversion coefficient Kc based on the predetermined parameter Pr. Furthermore, the predicted voltage characteristic table TPvp is generated by performing the aforementioned conversion process using the previously described measured viscosity characteristic table TMvi and the conversion coefficient Kc. Thus, in this embodiment, the following is achieved.

[0206] That is, a predicted voltage characteristic table TPvp is automatically generated each time, indicating a predicted characteristic curve CPvp between a predetermined voltage value Vp (peak value) and an ambient temperature Ta. This reduces workload and costs compared to, for example, obtaining characteristic curves between the voltage value Vp and the ambient temperature Ta by actually measuring these characteristic curves (the aforementioned measured characteristic curve CMvp) (e.g., obtaining characteristic curves by actually measuring each type of pulse number included in the drive signal Sd), as in the aforementioned comparative example. Furthermore, as previously mentioned, the characteristic curves between the voltage value Vp and the ambient temperature Ta (the measured characteristic curve CMvp) generally have different slopes and other characteristics depending on the type of pulse number included in the drive signal Sd or the type / function of each pulse. Therefore, automatically generating the predicted voltage characteristic table TPvp each time results in the following. That is, compared to, for example, repeatedly using a single characteristic curve, the voltage value Vp (peak value) of the pulses in the drive signal Sd can be set with greater accuracy.

[0207] For these reasons, in this embodiment, the efficiency of obtaining the characteristic curve (voltage characteristic table) between the voltage value Vp and the ambient temperature Ta can be improved, and the accuracy of setting the voltage value Vp (peak value) of the pulses in the drive signal Sd can be easily improved. As a result, in this embodiment, user convenience can be improved.

[0208] In addition, in this embodiment, for example, the following effects can be obtained.

[0209] Since the characteristic curve between the voltage value Vp and the ambient temperature Ta can be easily obtained, it is easy to perform voltage control to keep the ejection speed of the ink 9 or the amount of droplets approximately constant even when the number or type of pulses or the type / function of each pulse are different.

[0210] As in the aforementioned comparative example, since an expensive evaluation device (temperature controller, etc.) used to obtain the measured characteristic curve CMvp is not required, the cost can be reduced.

[0211] Furthermore, in this embodiment, the predetermined parameter Pr includes at least the specific gravity of the ink 9 or a physical property value obtained using the specific gravity of the ink 9 (e.g., the density of the ink 9 or the speed of sound in the ink 9). Therefore, when generating the conversion coefficient Kc, which serves as the target variable, using the predetermined analysis method using the parameter Pr as an explanatory variable, the conversion coefficient Kc is generated using the parameter with the highest importance (contribution) (the specific gravity of the ink 9 or the physical property value). This improves the accuracy of generating the conversion coefficient Kc (the prediction accuracy of the predicted characteristic curve CPvp). As a result, the predicted voltage characteristic table TPvp can be generated with high accuracy.

[0212] Furthermore, in this embodiment, when the parameter Pr further includes at least one of the aforementioned parameters: the number of dips (pulses), the viscosity of the ink 9 at the reference temperature Tr, the voltage offset ΔVp, and the voltage sensitivity Vr of the ink 9, the following is achieved. Specifically, the conversion coefficient Kc is generated using the aforementioned parameter (the aforementioned at least one parameter) of relatively high importance (contribution) in addition to the specific gravity of the ink 9 or the aforementioned physical property values. This further improves the accuracy of generating the conversion coefficient Kc (the prediction accuracy of the predicted characteristic curve CPvp). As a result, the predicted voltage characteristic table TPvp can be generated with even greater accuracy.

[0213] Furthermore, in the present embodiment, a method using the machine learning model 74 is used as the predetermined analysis method (first analysis method), and therefore the predicted voltage characteristic table TPvp can be generated easily and accurately.

[0214] Furthermore, in this embodiment, the conversion process includes the aforementioned preliminary process and addition process, and as described above, the voltage value Vp (the voltage value Vp in the predicted characteristic curve CPvp) after adding the voltage offset ΔVp corresponds to a voltage value that can be used to obtain the reference discharge speed or droplet volume. Therefore, the following is achieved. That is, the voltage offset ΔVp can be used to easily set the reference discharge speed or droplet volume, further improving user convenience.

[0215] Furthermore, in this embodiment, the characteristic table generation system 5 further includes a signal generation unit 48. Therefore, the predicted voltage characteristic table TPvp generated by the table generation unit 733 is used to determine the voltage value Vp (peak value) of the pulses in the drive signal Sd. Pulses having this voltage value Vp are then used to generate the drive signal Sd. Therefore, the ink 9 ejection operation is performed using the thus generated drive signal Sd, making it possible to easily improve the ejection characteristics of the ink 9. Consequently, user convenience can be further enhanced.

[0216] Furthermore, in this embodiment, the aforementioned data acquisition unit 731, conversion coefficient generation unit 732, and table generation unit 733 are each located outside the printer 1 (inside the information processing device 7). This allows the predicted voltage characteristic table TPvp to be automatically generated within the information processing device 7 while maintaining the existing configuration of the inkjet head 4 and printer 1. This further improves user convenience.

[0217] <2. Modifications>

[0218] Next, modifications (modifications 1 to 6) of the above embodiment will be described. Components identical to those in the above embodiment are denoted by the same reference numerals, and description thereof will be omitted as appropriate.

[0219] [Variation 1]

[0220] In the above embodiment, the case of setting a machine learning model 74 (first analysis method) with the conversion coefficient Kc as the target variable is described, but in the following variant example 1, an example of further setting a machine learning model (second analysis method) with the aforementioned voltage sensitivity Vr as the target variable is described.

[0221] (constitute)

[0222] Figure 18 A block diagram shows an example of the configuration of a machine learning model (machine learning model 74A) according to Modification 1. Figure 19 The predetermined parameter Pr according to the first modification is shown (see Figure 18 ). Figure 20 Expressed Figure 19 An example of the importance analysis results of each parameter Pr shown in FIG. Figure 19 , the values ​​of the parameters Pr are shown for 6 samples ("Sample 1" to "Sample 6").

[0223] in addition, Figure 21 An example of predicted values ​​and measured values ​​using machine learning according to Modification 1 is shown. Figure 22A Expressed Figure 21 An example of the correspondence between the SVM predicted value (mentioned above) and the measured value is shown. Figure 22B Expressed Figure 21 An example of the correspondence between the RF predicted value (described above) and the measured value is shown. Figure 23 Indicates that only Figure 20 An example of the correspondence between RF predicted values ​​and measured values ​​for some of the parameters shown. Figure 23 In the figure, only the use of Figure 19 The most important parameter among the parameters shown is ( Figure 20 On the other hand, in the case of the target value of the discharge speed or DV, an example of such a correspondence relationship is given. Figure 22A 、 Figure 22B The use of Figure 19 An example of the correspondence between predicted values ​​and measured values ​​for all parameters shown.

[0224] In addition, regarding these Figure 21 、 Figure 22A 、 Figure 22B 、 Figure 23 Details of the implementation Figure 15 、 Figure 16A 、 Figure 16B 、 17A to 17C The same is true for . That is, Figure 21 In the example shown, for Figure 19 The six samples ("Sample 1" to "Sample 6") shown in FIG. 1 respectively correspond to the measured values ​​of the voltage sensitivity Vr and the predicted values ​​of the voltage sensitivity Vr using machine learning (SVM predicted values ​​and RF predicted values). Figure 22A 、 Figure 22B 、 Figure 23 In the example shown, the (x, y) coordinates of many (562) samples are plotted, with the predicted value of the voltage sensitivity Vr (SVM predicted value or RF predicted value) as the variable x and the measured value of the voltage sensitivity Vr as the variable y. Figure 22A 、 Figure 22B 、 Figure 23 , an example of a formula (for example, a linear function formula determined using the least squares method) that expresses the trend of the correlation between these variables x and y is also shown.

[0225] First, in this first modification, the voltage sensitivity Vr is generated using a predetermined analysis method (second analysis method) based on the predetermined parameter Pr acquired by the data acquisition unit 731. This predetermined analysis method is an analysis method that uses the parameter Pr as an explanatory variable and the voltage sensitivity Vr as a target variable.

[0226] Specifically, for example Figure 18 As shown, in this modification 1, the voltage sensitivity Vr is generated based on the parameter Pr using an analysis method using a machine learning model 74A. As described above, the machine learning model 74A is a prediction model obtained by performing machine learning using the parameter Pr as an explanatory variable and the voltage sensitivity Vr as a target variable. Figure 18 As shown, when the machine learning model 74A receives the parameter Pr (explanatory variable) as input, it generates (predicts) the voltage sensitivity Vr (target variable) based on the learning results and outputs the generated voltage sensitivity Vr. A specific example of the analysis method (prediction method) using the machine learning model 74A is the same as that described in the embodiment.

[0227] Here, as a specific example of the predetermined parameter Pr in the modification 1, Figure 19 As shown, the parameters described in the embodiment and shown in the following (a) to (g) and (i) to (l) can be cited.

[0228] (a) Number of drops (pulses)

[0229] (b) Presence of common drive

[0230] (c) Header type

[0231] (d) Type of ink

[0232] (e) (Vj standard or DV standard)

[0233] (f) Top grade value

[0234] (g) Viscosity value at reference temperature Tr

[0235] (i) Surface tension value of ink 9

[0236] (j) Specific gravity of ink 9

[0237] (l) Target value of discharge speed or DV

[0238] (k) Voltage offset ΔVp

[0239] In addition, if based on Figure 20As an example of the importance analysis results of the parameters Pr as explanatory variables shown in FIG, when the voltage sensitivity Vr is generated using the machine learning model 74A, the following parameters have a relatively high importance (contribution rate) among the parameters Pr shown in (a) to (g) and (i) to (l). That is, in Figure 20 In the example shown, the target value of the discharge speed or DV (1) is a relatively important (highest) parameter. Therefore, in this first modification, it is preferable to include the target value of the discharge speed or DV (1) as the predetermined parameter Pr.

[0240] In addition, Figure 21 In the example shown, it can be seen that both the SVM prediction value and the RF prediction value can predict the actual value with good accuracy. Figure 22A 、 Figure 22B 、 Figure 23 In each of the examples shown, the slope of the aforementioned linear function formula (the formula that expresses the trend of the correlation between the variables x and y) becomes approximately "1", and the intercept of the linear function formula becomes approximately "0". Based on this, in this modification 1, regarding the voltage sensitivity Vr as the target variable, the predicted value (SVM predicted value and RF predicted value) and the measured value become the following relationship. That is, as in the case of the aforementioned embodiment, it can be seen that the predicted value and the measured value have a sufficient correlation to the extent that it is practical when printing using the predicted value. Furthermore, if we compare Figure 22B 、 Figure 23 The examples shown (examples of RF predicted values) show that the predicted values ​​(RF predicted values) have a good agreement accuracy with the measured values. Figure 22B The example shown (the example using all parameters above) is relatively higher than Figure 23 The example shown (the case where only some of the parameters mentioned above are used).

[0241] (Action / Effect)

[0242] In the first modification having such a configuration, basically the same operations and effects as those of the embodiment can be achieved.

[0243] Furthermore, in this first modification, the voltage sensitivity Vr (one of the predetermined parameters Pr in the embodiment) described in the embodiment is also generated using a predetermined analysis method (a second analysis method), resulting in the following. This eliminates the need to obtain a measured value for this voltage sensitivity Vr, further reducing workload and costs. Furthermore, compared to, for example, repeatedly using a single value, the value of this voltage sensitivity Vr can be set with greater precision. This further improves user convenience.

[0244] Furthermore, in this first modification, similarly to the embodiment, the predetermined analysis method (second analysis method) is a method using the machine learning model 74A, and thus the voltage sensitivity Vr as the target variable can be easily and accurately predicted.

[0245] [Variation 2]

[0246] In the above embodiment, a case where a machine learning model 74 is provided with the conversion coefficient Kc as the target variable (a first analysis method) is described. However, in the following Modification 2, an example is described where a machine learning model is further provided with the voltage offset ΔVp as the target variable (a third analysis method). Furthermore, for example, this Modification 2 and the above Modification 1 may be used in combination.

[0247] (constitute)

[0248] Figure 24 A block diagram shows an example of the configuration of a machine learning model (machine learning model 74B) according to Modification 2. Figure 25 The predetermined parameter Pr according to the second modification is shown (see Figure 24 ). Figure 26 Expressed Figure 25 An example of the importance analysis results of each parameter Pr shown in FIG. Figure 25 , the values ​​of the parameters Pr are shown for 6 samples ("Sample 1" to "Sample 6").

[0249] in addition, Figure 27 An example of predicted values ​​and measured values ​​using machine learning according to Modification Example 2 is shown. Figure 28A Expressed Figure 27 An example of the correspondence between the SVM predicted value (mentioned above) and the measured value is shown. Figure 28B Expressed Figure 27 An example of the correspondence between the RF predicted value (mentioned above) and the measured value is shown. Figures 29A to 29C Respectively, only use Figure 26 An example of the correspondence between RF predicted values ​​and measured values ​​for some of the parameters shown. Figure 29A Shown in the Figure 26 An example of such a correspondence relationship is shown in the case of the parameter with the highest importance (viscosity value at reference temperature Tr) among the parameters shown. Figure 29B Shown in the Figure 26 This is an example of such a correspondence relationship when the two parameters with the highest and second highest importance (viscosity value at reference temperature Tr and presence or absence of common drive) are used among the parameters shown. Figure 29C Shown in the Figure 26 This is an example of such a correspondence relationship when the seventh most important parameter (surface tension value of ink 9) is used among the parameters shown. Figure 28A 、 Figure 28B The use of Figure 25 An example of the correspondence between predicted values ​​and measured values ​​for all parameters shown.

[0250] In addition, regarding these Figure 27 、 Figure 28A 、 Figure 28B 、 Figures 29A to 29C Details of the implementation Figure 15 、 Figure 16A 、 Figure 16B 、 17A to 17C The same is true for . That is, Figure 27 In the example shown, for Figure 25 The six samples ("Sample 1" to "Sample 6") shown in FIG. 1 respectively correspond to the measured values ​​of the voltage offset ΔVp and the predicted values ​​of the voltage offset ΔVp using machine learning (SVM predicted values ​​and RF predicted values). Figure 28A 、 Figure 28B 、 Figures 29A to 29C In the example shown, the (x, y) coordinates of many (562) samples are plotted, with the predicted value of the voltage offset ΔVp (SVM predicted value or RF predicted value) as the variable x and the measured value of the voltage offset ΔVp as the variable y. Figure 28A 、 Figure 28B 、 Figures 29A to 29C , an example of a formula (for example, a linear function formula determined by the least squares method) that expresses the trend of the correlation between these variables x and y is also shown.

[0251] First, in this second modification, the voltage offset ΔVp is generated using a predetermined analysis method (third analysis method) based on the predetermined parameter Pr acquired by the data acquisition unit 731. This predetermined analysis method uses the parameter Pr as an explanatory variable and the voltage offset ΔVp as a target variable.

[0252] Specifically, for example Figure 24 As shown, in this modification 2, the voltage offset ΔVp is generated based on the parameter Pr using an analysis method using a machine learning model 74B. As described above, the machine learning model 74B is a prediction model obtained by performing machine learning using the parameter Pr as an explanatory variable and the voltage offset ΔVp as a target variable. Figure 24As shown, when the machine learning model 74B receives the parameter Pr (explanatory variable) as input, it generates (predicts) the voltage offset ΔVp (target variable) based on the learning results and outputs the generated voltage offset ΔVp. A specific example of the analysis method (prediction method) using the machine learning model 74B is the same as that described in the embodiment.

[0253] Here, as a specific example of the above-mentioned predetermined parameter Pr in the second modification, Figure 25 As shown, the parameters described in the embodiment and modification example 1, which are shown in the following (a) to (j) and (l), can be cited.

[0254] (a) Number of drops (pulses)

[0255] (b) Presence of common drive

[0256] (c) Header type

[0257] (d) Type of ink

[0258] (e) (Vj standard or DV standard)

[0259] (f) Top grade value

[0260] (g) Viscosity value at reference temperature Tr

[0261] (h) Voltage sensitivity Vr during discharge (Vj reference or DV reference)

[0262] (i) Surface tension value of ink 9

[0263] (j) Specific gravity of ink 9

[0264] (l) Target value of discharge speed or DV

[0265] In addition, if based on Figure 26 As an example of the importance analysis results of the parameters Pr as explanatory variables shown in FIG, when the voltage offset ΔVp is generated using the machine learning model 74B, the following parameters have a relatively high importance (contribution rate) among the parameters Pr shown in (a) to (j) and (l). That is, in Figure 26 In the example shown, the importance is higher in the order of (g) viscosity value at the reference temperature Tr, (b) presence or absence of common drive, (f) head grade value, (c) head type, (j) specific gravity value of ink 9, (h) voltage sensitivity Vr during ejection, (i) surface tension value of ink 9, (l) target value of ejection speed or DV, (a) drop number, (d) ink type, and (e) (Vj reference or DV reference).

[0266] Therefore, in this second modification, the predetermined parameter Pr preferably includes at least one of the values ​​of (g) the viscosity value at the reference temperature Tr and (b) the presence or absence of the common drive, which are relatively important (for example, the first and second most important values). In addition, the predetermined parameter Pr preferably further includes Figure 26 Among the remaining parameters shown in , the parameters with relatively high importance (as an example, the 3rd to 7th most important) are parameters, namely, at least one of (f) head grade value, (c) head type, (j) specific gravity value of ink 9, (h) voltage sensitivity Vr during ejection, and (i) surface tension value of ink 9.

[0267] In addition, Figure 27 In the example shown, it can be seen that both the SVM prediction value and the RF prediction value can predict the actual value with good accuracy. Figure 28A 、 Figure 28B 、 Figures 29A to 29C In each of the examples shown, the slope of the aforementioned linear function formula (the formula that expresses the trend of the correlation between the variables x and y) is also approximately "1", and the intercept of the linear function formula is also approximately "0". Based on this, in this modification example 2, regarding the voltage offset ΔVp as the target variable, the predicted value (SVM predicted value and RF predicted value) and the measured value have the following relationship. That is, as in the case of the aforementioned embodiment or modification example 1, it can be seen that the predicted value and the measured value have a sufficient correlation to the extent that it is practical when printing using the predicted value. Furthermore, if we compare Figure 28B 、 Figures 29A to 29C The examples shown (examples of RF predicted values) show that the predicted values ​​(RF predicted values) have a good agreement accuracy with the measured values. Figure 28B The example shown (the example using all parameters above) is relatively higher than Figures 29A to 29C The examples shown (the case where only some of the parameters mentioned above are used).

[0268] (Action / Effect)

[0269] In the second modification of this structure, basically the same effects as those of the embodiment can be obtained.

[0270] Furthermore, in this second variation, the voltage offset ΔVp (a parameter used in the addition process described in the embodiment) described in the embodiment is also generated using a predetermined analysis method (the third analysis method), resulting in the following. This eliminates the need to obtain a measured value for this voltage offset ΔVp in advance, further reducing workload and costs. Furthermore, compared to, for example, repeatedly using a single value, the voltage offset ΔVp value can be set with greater precision. This further improves user convenience.

[0271] Furthermore, in this second modification, as in the embodiment or the first modification, the predetermined analysis method (third analysis method) is a method using the machine learning model 74B, so that the voltage offset ΔVp as the target variable can be predicted easily and accurately.

[0272] [Variation 3]

[0273] (constitute)

[0274] Figure 30 A block diagram illustrates an example configuration of a characteristic table generation system 5A according to Modification 3. This characteristic table generation system 5A includes a printer 1 having an inkjet head 4, an information processing device 7A, and a server 8 located external to the printer 1. The printer 1, information processing device 7A, and server 8 are connected to each other via a network 50. Specifically, this characteristic table generation system 5A corresponds to a system in which the characteristic table generation system 5 of the embodiment is modified by including the information processing device 7A in place of the information processing device 7 and further including the server 8.

[0275] In addition, in this modification 3, the server 8 corresponds to a specific example of the “external device” in the present disclosure.

[0276] like Figure 30 As shown in FIG. 1 , the information processing device 7A comprises a bus 70, an input unit 71, a display unit 72, a control unit 75, a storage unit 76A, and a network IF 77 as a physical block. Figure 4 The information processing device 7 of the illustrated embodiment corresponds to a device having a storage unit 76A in place of the storage unit 76. Unlike the storage unit 76, the storage unit 76A does not store the program 730 and the machine learning model 74 described in the embodiment. Therefore, the information processing device 7A corresponds to, for example, a general-purpose PC.

[0277] like Figure 30As shown in FIG. 1 , the server 8 includes a bus 80, a control unit 85, a storage unit 86, and a network IF 87 as a physical block. The control unit 85, the storage unit 86, and the network IF 87 are connected to each other via the bus 80. The control unit 85 and the network IF 87 have the same physical blocks as those in the embodiment ( Figure 4 ) has the same structure as the control unit 75 and network IF77 in the embodiment. In addition, the storage unit 86 also has the same structure as the embodiment ( Figure 4 ) has the same structure as the storage unit 76. Figure 30 As shown in FIG. 8 , the storage unit 86 stores the program 730 and the machine learning model 74 described in the embodiment. Figure 30 As indicated in brackets, in addition to such machine learning model 74, the machine learning models 74A and 74B described in Modifications 1 and 2 may be provided, and the same applies to Modifications 4 to 6 described later.

[0278] Thus, in the characteristic table generating system 5A of the third modification, unlike the characteristic table generating system 5 of the embodiment, the aforementioned predicted voltage characteristic table TPvp is generated in the server 8 instead of the information processing device 7A. Moreover, the predicted voltage characteristic table TPvp generated in this way is as follows: Figure 30 As shown, the signal is supplied from the server 8 to the signal generating unit 48 in the inkjet head 4 in the printer 1 via the network 50 .

[0279] (Action / Effect)

[0280] In the third modification having such a configuration, the characteristic table creation system 5A as a whole can also basically operate and achieve the same effects as the characteristic table creation system 5 of the embodiment.

[0281] Furthermore, in this third modification, the aforementioned data acquisition unit 731, conversion coefficient generation unit 732, and table generation unit 733 (the aforementioned program 730) are each located outside the printer 1 (within the server 8), resulting in the following configuration. Specifically, as in the aforementioned embodiment, the predicted voltage characteristic table TPvp can be automatically generated within the server 8 while maintaining the existing configuration for the inkjet head 4 and printer 1. Furthermore, in this third modification, as described above, the information processing device 7A can also utilize an existing (general-purpose) configuration, for example, utilizing the server 8 functioning as a cloud server, thereby achieving the same effects as in the embodiment. Consequently, in this third modification, user convenience can be further improved.

[0282] [Variation 4]

[0283] (constitute)

[0284] Figure 31A block diagram illustrates an example configuration of a characteristic table generation system 5B according to Modification 4. This characteristic table generation system 5B includes a printer 1B having an inkjet head 4B and the aforementioned information processing device 7A. The printer 1B and information processing device 7A are connected to each other via a network 50. Specifically, this characteristic table generation system 5B corresponds to a system in which the characteristic table generation system 5 of the embodiment is modified by replacing the information processing device 7 with the aforementioned information processing device 7A, and replacing the printer 1 and inkjet head 4 with a printer 1B and inkjet head 4B.

[0285] The printer 1B corresponds to a specific example of a “liquid jet recording device” in the present disclosure. The inkjet head 4B corresponds to a specific example of a “liquid jet head” in the present disclosure.

[0286] like Figure 31 As shown, in this fourth variation, the aforementioned information processing unit 73 (data acquisition unit 731, conversion coefficient generation unit 732, and table generation unit 733), in other words, the aforementioned program 730, is located within the inkjet head 4B. Furthermore, the aforementioned machine learning model 74 is also located within the inkjet head 4B. That is, in this fourth variation, unlike the embodiment and third variation, the information processing unit 73 (program 730) and the machine learning model 74 are each located within the inkjet head 4B built into the printer 1B.

[0287] (Action / Effect)

[0288] In the fourth modification configured in this manner, the characteristic table creation system 5B as a whole can also basically operate and achieve the same effects as those of the characteristic table creation system 5 of the embodiment.

[0289] Furthermore, in this fourth variation, the data acquisition unit 731, conversion coefficient generation unit 732, and table generation unit 733 are each provided within the printer 1B. This results in the following: Unlike the embodiment and third variation, there is no need to separately prepare these data acquisition unit 731, conversion coefficient generation unit 732, and table generation unit 733 within an external device (information processing device 7 or server 8). This allows the printer 1B to automatically generate the predicted voltage characteristic table TPvp, further improving user convenience.

[0290] Furthermore, in this fourth variation, the data acquisition unit 731, conversion coefficient generation unit 732, and table generation unit 733 are each provided within the inkjet head 4B built into the printer 1B. This allows the printer 1B itself, while maintaining its existing configuration, to automatically generate the predicted voltage characteristic table TPvp within the inkjet head 4B itself. This further improves user convenience.

[0291] [Variation 5]

[0292] (constitute)

[0293] Figure 32 A block diagram illustrates an example configuration of a characteristic table generation system 5C according to Modification 5. This characteristic table generation system 5C includes a printer 1C having the aforementioned inkjet head 4 and the aforementioned information processing device 7A. The printer 1C and information processing device 7A are connected to each other via a network 50. Specifically, this characteristic table generation system 5C corresponds to a system in which the characteristic table generation system 5 of the embodiment is modified by including the aforementioned information processing device 7A in place of the information processing device 7 and a printer 1C in place of the printer 1.

[0294] Note that the above-described printer 1C corresponds to a specific example of the “liquid jet recording apparatus” in the present disclosure.

[0295] like Figure 32 As shown, in this modification example 5, the aforementioned information processing unit 73 (data acquisition unit 731, conversion coefficient generation unit 732 and table generation unit 733), in other words, the aforementioned program 730 and the modification example 4 ( Figure 31 ) is also provided in the printer 1C. In addition, the aforementioned machine learning model 74 is also provided in the printer 1C in the same manner as in the modification 4. However, as Figure 32 As shown, in this modification example 5, unlike the modification example 4, the information processing unit 73 (program 730) and the machine learning model 74 are both arranged outside the inkjet head 4 in the printer 1C.

[0296] (Action / Effect)

[0297] In the modified example 5 having such a configuration, the characteristic table creation system 5C as a whole can also basically operate and obtain the same effects as those of the characteristic table creation system 5 of the embodiment.

[0298] Furthermore, in this fifth modification, similar to the fourth modification, the data acquisition unit 731, the conversion coefficient generation unit 732, and the table generation unit 733 are each provided within the printer 1C. This allows the following configuration: Similarly to the fourth modification, the printer 1C itself can automatically generate the predicted voltage characteristic table TPvp, thereby further improving user convenience.

[0299] [Variation 6]

[0300] (constitute)

[0301] Figure 33A block diagram shows an example configuration of an information processing unit 73D (program 730D) according to Modification 6. Information processing unit 73D according to Modification 6 corresponds to the configuration of information processing unit 73 (comprising a data acquisition unit 731, a conversion coefficient generation unit 732, and a table generation unit 733) described in the embodiments, etc., with the aforementioned signal generation unit 48 further provided. In other words, program 730D according to Modification 6 corresponds to a program that further incorporates the functions of the various processes performed by signal generation unit 48 in program 730 described in the embodiments, etc.

[0302] The configuration of this information processing unit 73D (program 730D) corresponds to a configuration in which, in addition to providing the information processing unit 73 (program 730) within the external device (information processing device 7 or server 8) of the printer 1 as in the embodiment or modification 3, the configuration or functions of the signal generating unit 48 are further provided. In other words, unlike in these embodiments or modification 3, the configuration or functions corresponding to the signal generating unit 48 are provided within the external device (information processing device 7 or server 8) of the printer 1, rather than within the printer 1.

[0303] (Action / Effect)

[0304] In the modified example 6 having such a configuration, basically the same effects as those of the embodiment can be obtained.

[0305] Furthermore, in this sixth modification, the configuration and functions of the signal generator 48 are further incorporated into the information processing unit 73D (program 730D). Therefore, the operation of the signal generator 48 (the operation of generating the drive signal Sd) can also be centrally executed within the information processing unit 73D (program 730D). As a result, user convenience can be further improved.

[0306] <3. Other Modifications>

[0307] As mentioned above, although embodiment and modification were given and this disclosure was demonstrated, this disclosure is not limited to these embodiment etc., Various modifications are possible.

[0308] For example, in the above-mentioned embodiments, the configuration examples (shape, configuration, number, etc.) of the components in the printer and the inkjet head are specifically cited for description, but the configuration is not limited to the configuration described in the above-mentioned embodiments, and other shapes, configurations, numbers, etc. may also be used. Specifically, for example, in the above-mentioned embodiments, a shuttle printer in which the inkjet head moves is described as an example, but the invention is not limited to this example, and for example, a single-pass printer in which the inkjet head is fixed may also be used. In addition, in the above-mentioned embodiments, the case where the ink tank is housed in a predetermined housing is described as an example, but the invention is not limited to this example, and the ink tank may also be configured outside the housing. Furthermore, in the above-mentioned embodiments, the case where the signal generating unit is mainly provided in the inkjet head is described as an example, but the invention is not limited to this example, and the signal generating unit may also be provided outside the inkjet head in the printer.

[0309] In addition, various types of heads can be used as the structure of the inkjet head. That is, it can be a so-called side-jet type inkjet head that ejects ink 9 from the center of the extension direction of each ejection channel in the actuator plate. Alternatively, it can be a so-called edge-jet type inkjet head that ejects ink 9 along the extension direction of each ejection channel. Furthermore, the printer method is not limited to the methods described in the above embodiment, and various methods such as thermal (thermal on-demand) or MEMS (Micro Electro Mechanical Systems) methods can be used.

[0310] Furthermore, while the above embodiments and the like illustrate a non-circulating inkjet head in which the ink 9 is not circulated between the ink tank and the inkjet head, the present invention is not limited to this example. Specifically, the present disclosure can also be applied to a circulating inkjet head in which the ink 9 is circulated between the ink tank and the inkjet head.

[0311] Furthermore, while the above-described embodiments and other embodiments specifically illustrate the generation process of a characteristic table (predicted voltage characteristic table TPvp) or a drive signal Sd, the present invention is not limited to the examples presented in the above-described embodiments and other embodiments, and other methods may be used to generate the characteristic table or drive signal Sd. Specifically, while the above-described embodiments and other embodiments illustrate the use of a machine learning model as an example of the predetermined analysis method (the first to third analysis methods), the present invention is not limited to this method, and other analysis methods may be used. Furthermore, the predetermined parameter Pr is not limited to the various parameters presented in the above-described embodiments and other embodiments, and other parameters may be added (or substituted) for the analysis method. Furthermore, while the above-described embodiments and other embodiments illustrate the generation of the drive signal Sd by setting (automatically adjusting) both the pulse width Wp and the voltage value (peak value) Vp of the pulse, the present invention is not limited to this example. That is, the drive signal Sd may be generated by setting only the pulse width Wp or the voltage value Vp of the pulse, for example, or by adjusting only the pulse width Wp. Furthermore, in the above embodiments, the case where all the voltage values ​​Vp in a plurality of pulses are the same is described as an example. However, for example, the voltage values ​​Vp in the plurality of pulses may be different (at least some of the voltage values ​​Vp may be different). In this case, the generation process of the predicted voltage characteristic table TPvp described in the above embodiments may be performed by using each of the plurality of voltage values ​​Vp as an explanatory variable.

[0312] Furthermore, in the above-described embodiments, the pulses (pulses Pa, Pb, and Pc) that expand the volume within each discharge channel are described as pulses that expand during a high-level (High) state (positive pulses), but this is not limited to this case. Specifically, pulses that expand during a high-level state and contract during a low-level (Low) state are not limited to this case. Conversely, pulses that expand during a low-level state and contract during a high-level state (negative pulses) are also possible. Furthermore, even with such negative pulses, such "common drive" can be applied as long as the method achieves the same function as the previously described "common drive."

[0313] Furthermore, for example, a pulse to assist droplet ejection may be additionally applied during the OFF period immediately following the ON period. Examples of such pulses include pulses for shrinking the volume within each ejection channel and pulses (auxiliary pulses) for retracting a portion of ejected droplets. Furthermore, the pulse (main pulse) applied immediately before the latter auxiliary pulse may have a pulse width equal to or less than the peak value (AP) of the ON pulse. The addition of such pulses to assist droplet ejection does not affect the present disclosure described thus far.

[0314] In addition, the series of processing described in the above embodiments and the like can be performed by hardware (circuit) or by software (program). In the case of being performed by software, the software is composed of a group of programs for causing a computer to execute various functions. Each program can be pre-loaded onto the above-mentioned computer and used, or it can be installed onto the above-mentioned computer from a network or a recording medium and used. In addition, as a recording medium (non-transient computer-readable recording medium) for recording such programs, various media such as FLOPPY (registered trademark) disks, CD (Compact Disk)-ROMs, DVD (Digital Versatile Disc: Digital Versatile Disc)-ROMs, and hard disks (Hard Disks: Hard Disks) can be cited.

[0315] Furthermore, in the above-described embodiments, a printer 1 (inkjet printer) is used as a specific example of the "liquid jet recording device" in this disclosure. However, this disclosure is not limited to this example, and the disclosure can also be applied to devices other than inkjet printers. In other words, the "liquid jet head" (inkjet head) of this disclosure can also be applied to devices other than inkjet printers. Specifically, for example, the "liquid jet head" of this disclosure can also be applied to devices such as facsimile machines and on-demand printers.

[0316] Furthermore, the various examples described so far can be applied in any combination.

[0317] Furthermore, the effects described in this specification are merely examples and not limitations, and other effects may also be achieved.

[0318] In addition, the present disclosure may also have the following configurations.

[0319] (1) A characteristic table generating system is a system for generating a predicted voltage characteristic table, the predicted voltage characteristic table defining a predicted characteristic curve between temperature and a voltage value applied to an ejection portion for ejecting liquid, the voltage value representing a peak value of a drive signal having one or more pulses, with respect to a predetermined reference value;

[0320] The characteristic table generating system comprises:

[0321] a data acquisition unit that acquires, as input data, predetermined parameters and a measured viscosity characteristic table that defines a measured characteristic curve between viscosity and temperature of the liquid;

[0322] a conversion coefficient generating unit that generates the conversion coefficient based on the predetermined parameter using a first analysis method that is a predetermined analysis method, the first analysis method using the predetermined parameter as an explanatory variable and using a conversion coefficient when performing a conversion process from the measured characteristic curve to the predicted characteristic curve as a target variable; and

[0323] The table generating unit generates the predicted voltage characteristic table by performing the conversion process using the actually measured viscosity characteristic table and the conversion coefficient generated by the conversion coefficient generating unit.

[0324] (2) A characteristic table generating system according to (1) above, wherein:

[0325] The predetermined parameters include at least the specific gravity of the liquid or a physical property value obtained using the specific gravity of the liquid.

[0326] (3) A characteristic table generating system according to (2) above, wherein:

[0327] The conversion process includes:

[0328] a preliminary process of generating a preliminary characteristic curve representing a relationship between a voltage value and a temperature from the measured characteristic curve using the conversion coefficient; and

[0329] adding a voltage offset to the voltage value in the preliminary characteristic curve to generate the predicted characteristic curve;

[0330] The predetermined parameters further include at least one of the following parameters:

[0331] a falling number corresponding to the number of the pulses included in a unit period of the drive signal;

[0332] The viscosity value of the liquid at the reference temperature;

[0333] the voltage offset; and

[0334] The voltage sensitivity of the liquid corresponds to the amount of change per unit voltage in the discharge speed of the liquid or the amount of droplets of the liquid when the liquid is ejected at a reference temperature.

[0335] (4) A characteristic table generating system according to (3) above, wherein:

[0336] A second analysis method is used as a predetermined analysis method, which uses a target value of the discharge speed of the liquid or the amount of liquid droplets as an explanatory variable and a voltage sensitivity of the liquid as a target variable.

[0337] The voltage sensitivity of the liquid is generated based on a target value of a discharge speed of the liquid or a droplet amount of the liquid.

[0338] (5) The characteristic table generating system according to any one of (1) to (4) above, wherein the conversion process includes:

[0339] a preliminary process of generating a preliminary characteristic curve representing a relationship between a voltage value and a temperature from the measured characteristic curve using the conversion coefficient; and

[0340] adding a voltage offset to the voltage value in the preliminary characteristic curve to generate the predicted characteristic curve;

[0341] The voltage value after adding the voltage offset corresponds to the following voltage value:

[0342] When the liquid is ejected from the ejecting portion, a voltage value serving as the predetermined reference value can be obtained for a discharge speed of the liquid serving as a reference, or a voltage value serving as the predetermined reference value can be obtained for a droplet amount of the liquid serving as a reference.

[0343] (6) A characteristic table generating system according to (5) above, wherein:

[0344] As a description variable, include at least one of the following parameters:

[0345] the viscosity value of the liquid at a reference temperature; and

[0346] Indicates whether there is a common drive parameter in the drive signal,

[0347] Then, the voltage offset is generated based on the explanatory variables including the at least one parameter using a third analysis method that is a predetermined analysis method and uses the voltage offset as a target variable.

[0348] (7) A characteristic table generating system according to (6) above, wherein:

[0349] As the explanatory variables in the third analysis method, at least one of the following parameters is further included:

[0350] a head level value corresponding to the voltage value at a predetermined ejection speed when a predetermined test liquid is ejected from the ejection portion, and the head level value being a value unique to the liquid ejecting head having the ejection portion;

[0351] a parameter indicating the type of the liquid ejecting head;

[0352] the specific gravity of the liquid;

[0353] voltage sensitivity of the liquid, which corresponds to a change in voltage per unit area in a discharge speed of the liquid or a droplet volume of the liquid when the liquid is ejected at a reference temperature; and

[0354] The surface tension value of the liquid.

[0355] (8) A characteristic table generation system according to any one of (1) to (7) above, wherein the predetermined analysis method is a method using a machine learning model that receives the predetermined parameters as input and outputs the conversion coefficient.

[0356] (9) The characteristic table generating system according to any one of (1) to (8) above, further comprising:

[0357] The signal generating unit obtains a peak value of the pulse using the predicted voltage characteristic table generated by the table generating unit, and generates the driving signal using the pulse having the obtained peak value.

[0358] (10) A characteristic table generating system according to any one of (1) to (9) above, wherein the data acquisition unit, the conversion coefficient generating unit, and the table generating unit are respectively arranged in an external device, and the external device is located outside a liquid ejection recording device having a liquid ejection head having the ejection unit.

[0359] (11) A characteristic table generating system according to any one of (1) to (9) above, wherein the data acquisition unit, the conversion coefficient generating unit, and the table generating unit are respectively provided in a liquid ejection recording device having a liquid ejection head having the ejection unit.

[0360] (12) A characteristic table generating system according to (11) above, wherein:

[0361] The data acquisition section, the conversion coefficient generation section, and the table generation section are respectively provided in the liquid ejecting head.

[0362] (13) A method for generating a characteristic table, wherein the method generates a predicted voltage characteristic table, the predicted voltage characteristic table defining a predicted characteristic curve between temperature and a voltage value applied to an ejection portion for ejecting liquid, the voltage value representing a peak value of a drive signal having one or more pulses, with respect to a predetermined reference value.

[0363] The method for generating a characteristic table comprises the following steps:

[0364] As input data, predetermined parameters and a measured viscosity characteristic table defining a measured characteristic curve between viscosity and temperature of the liquid are respectively acquired;

[0365] generating the conversion coefficient based on the predetermined parameter using a first analysis method as a predetermined analysis method, the first analysis method using the predetermined parameter as an explanatory variable and using the conversion coefficient when performing a conversion process from the measured characteristic curve to the predicted characteristic curve as a target variable; and

[0366] The conversion process is performed using the actually measured viscosity characteristic table and the generated conversion coefficient, thereby generating the predicted voltage characteristic table.

[0367] (14) A characteristic table generating program is a program for generating a predicted voltage characteristic table, the predicted voltage characteristic table defining a predicted characteristic curve between temperature and a voltage value applied to an ejection portion for ejecting liquid, the voltage value representing a peak value of a drive signal having one or more pulses, with respect to a predetermined reference value, the characteristic table generating program causing a computer to execute the following steps:

[0368] As input data, predetermined parameters and a measured viscosity characteristic table defining a measured characteristic curve between viscosity and temperature of the liquid are respectively acquired;

[0369] generating the conversion coefficient based on the predetermined parameter using a first analysis method as a predetermined analysis method, the first analysis method using the predetermined parameter as an explanatory variable and using the conversion coefficient when performing a conversion process from the measured characteristic curve to the predicted characteristic curve as a target variable; and

[0370] The conversion process is performed using the actually measured viscosity characteristic table and the generated conversion coefficient, thereby generating the predicted voltage characteristic table.

[0371] (15) A recording medium, a non-transitory computer-readable recording medium having recorded thereon a program for generating a predicted voltage characteristic table, wherein the predicted voltage characteristic table defines a predicted characteristic curve between temperature and a voltage value applied to an ejection portion for ejecting liquid, wherein the voltage value represents a peak value of one or more pulses in a drive signal with respect to a predetermined reference value, wherein the recording medium has recorded thereon a program for generating the characteristic table that causes a computer to execute the following steps:

[0372] As input data, predetermined parameters and a measured viscosity characteristic table defining a measured characteristic curve between viscosity and temperature of the liquid are respectively acquired;

[0373] generating the conversion coefficient based on the predetermined parameter using a first analysis method as a predetermined analysis method, the first analysis method using the predetermined parameter as an explanatory variable and using the conversion coefficient when performing a conversion process from the measured characteristic curve to the predicted characteristic curve as a target variable; and

[0374] The conversion process is performed using the actually measured viscosity characteristic table and the generated conversion coefficient, thereby generating the predicted voltage characteristic table.

[0375] (Description of labels)

[0376] 1, 1B, 1C printer; 10 housing; 2a, 2b conveying mechanism; 21 grid roller; 22 pinch roller; 3 (3Y, 3M, 3C, 3K) ink tanks; 30 ink supply tubes; 4 (4Y, 4M, 4C, 4K), 4B inkjet heads; 41 nozzle plate; 42 actuator plate; 48 signal generation unit; 49 drive unit; 5, 5A, 5B, 5C characteristic table generation system; 50 network; 6 scanning mechanism; 61a, 61b guide rail; 62 carriage; 63 drive mechanism; 631a, 631b pulley; 632 endless belt; 633 drive motor; 7, 7A information processing device; 70 bus; 71 input unit; 72 display unit; 73, 73D information processing unit; 730, 730D program; 731 data acquisition unit; 732 conversion coefficient generation unit; 733 table generation unit; 74, 74A, 74B machine learning Model; 75 control unit; 76 storage unit; 77 network IF; 8 server; 80 bus; 85 control unit; 86 storage unit; 87 network IF; 9 ink; P recording paper; d conveying direction; Hn nozzle hole; Sd drive signal; Vd drive voltage; Vr voltage sensitivity; Ta ambient temperature; Tr reference temperature; Iv viscosity information; Vi viscosity; ΔV differential value; ΔVp voltage offset; Wp, Wpa1, Wpa2, Wpa3, Wpb2, Wpb3, Wpc3 pulse width; Vp, Vp1, Vp2, Vp3 voltage value (peak value); Pa, Pb, Pc pulse; Pr parameter; Kc conversion coefficient; TMvi measured viscosity characteristic table; TPvp predicted voltage characteristic table; CMvi, CMvp measured characteristic curve; CPvp0 preliminary characteristic curve; CPvp predicted characteristic curve; t time.

Claims

1. A system for generating a characteristic table, the system generating a predicted voltage characteristic table, the predicted voltage characteristic table defining a predicted characteristic curve between temperature and a voltage value applied to an ejection portion ejecting a liquid, the voltage value representing a peak value of a drive signal having one or more pulses, relative to a predetermined reference value; The characteristic table generating system comprises: a data acquisition unit that acquires, as input data, predetermined parameters and a measured viscosity characteristic table that defines a measured characteristic curve between viscosity and temperature of the liquid; a conversion coefficient generating unit that generates a conversion coefficient based on the predetermined parameter using a first analysis method as a predetermined analysis method, the first analysis method using the predetermined parameter as an explanatory variable and using the conversion coefficient when performing a conversion process from the measured characteristic curve to the predicted characteristic curve as a target variable; and The table generating unit generates the predicted voltage characteristic table by performing the conversion process using the actually measured viscosity characteristic table and the conversion coefficient generated by the conversion coefficient generating unit.

2. The characteristic table generating system according to claim 1, wherein: The predetermined parameters include at least the specific gravity of the liquid or a physical property value obtained using the specific gravity of the liquid.

3. The characteristic table generating system according to claim 2, wherein: The conversion process includes: a preliminary process of generating a preliminary characteristic curve representing a relationship between a voltage value and a temperature from the measured characteristic curve using the conversion coefficient; and adding a voltage offset to the voltage value in the preliminary characteristic curve to generate the predicted characteristic curve; The predetermined parameters further include at least one of the following parameters: A falling number corresponding to the number of the pulses included in a unit period of the driving signal; The viscosity of the liquid at the reference temperature; the voltage offset; as well as The voltage sensitivity of the liquid corresponds to the amount of change per unit voltage in the discharge speed of the liquid or the amount of droplets of the liquid when the liquid is ejected at the reference temperature.

4. The characteristic table generating system according to claim 3, wherein: A second analysis method is used as a predetermined analysis method, which uses a target value of the discharge speed of the liquid or the amount of liquid droplets as an explanatory variable and a voltage sensitivity of the liquid as a target variable. The voltage sensitivity of the liquid is generated based on a target value of a discharge speed of the liquid or a droplet amount of the liquid.

5. The characteristic table generating system according to any one of claims 1 to 4, wherein: The conversion process includes: a preliminary process of generating a preliminary characteristic curve representing a relationship between a voltage value and a temperature from the measured characteristic curve using the conversion coefficient; and adding a voltage offset to the voltage value in the preliminary characteristic curve to generate the predicted characteristic curve; The voltage value after adding the voltage offset corresponds to the following voltage value: When the liquid is ejected from the ejection portion, The predetermined reference value may be a voltage value that can obtain a reference discharge speed of the liquid, or a voltage value that can obtain a reference droplet amount of the liquid.

6. The characteristic table generating system according to claim 5, wherein: As a description variable, include at least one of the following parameters: the viscosity value of the liquid at a reference temperature; and Indicates whether there is a common drive parameter in the drive signal, Then, a third analysis method is used as a predetermined analysis method using the voltage offset as a target variable. The voltage offset is generated based on the explanatory variables including the at least one parameter.

7. The characteristic table generating system according to claim 6, wherein: As the explanatory variables in the third analysis method, at least one of the following parameters is further included: a head level value corresponding to the voltage value at a predetermined ejection speed when a predetermined test liquid is ejected from the ejection portion, and the head level value being a value unique to the liquid ejecting head having the ejection portion; a parameter indicating the type of the liquid ejecting head; the specific gravity of the liquid; voltage sensitivity of the liquid, which corresponds to a change in voltage per unit area in a discharge speed of the liquid or a droplet volume of the liquid when the liquid is ejected at a reference temperature; and The surface tension value of the liquid.

8. The characteristic table generating system according to any one of claims 1 to 4, wherein: The predetermined analysis method is a method using a machine learning model that receives the predetermined parameters as input and outputs the conversion coefficient.

9. The characteristic table generating system according to any one of claims 1 to 4, further comprising: The signal generating unit obtains a peak value of the pulse using the predicted voltage characteristic table generated by the table generating unit, and generates the driving signal using the pulse having the obtained peak value.

10. The characteristic table generating system according to any one of claims 1 to 4, wherein: The data acquisition unit, the conversion coefficient generation unit, and the table generation unit are respectively provided in an external device. The external device is located outside the liquid jet recording apparatus that incorporates the liquid jet head having the ejection portion.

11. The characteristic table generating system according to any one of claims 1 to 4, wherein: The data acquisition unit, the conversion coefficient generation unit, and the table generation unit are each provided in a liquid jet recording device having a built-in liquid jet head having the ejection unit.

12. The characteristic table generating system according to claim 11, wherein: The data acquisition section, the conversion coefficient generation section, and the table generation section are respectively provided in the liquid ejecting head.

13. A method for generating a characteristic table, the method comprising generating a predicted voltage characteristic table, the predicted voltage characteristic table defining a predicted characteristic curve between temperature and a voltage value applied to an ejection portion ejecting liquid, the voltage value representing a peak value of a drive signal having one or more pulses, with respect to a predetermined reference value. The method for generating a characteristic table comprises the following steps: As input data, predetermined parameters and a measured viscosity characteristic table defining a measured characteristic curve between viscosity and temperature of the liquid are respectively acquired; generating a conversion coefficient based on the predetermined parameter using a first analysis method as a predetermined analysis method, the first analysis method using the predetermined parameter as an explanatory variable and the conversion coefficient when performing a conversion process from the measured characteristic curve to the predicted characteristic curve as a target variable; as well as The conversion process is performed using the actually measured viscosity characteristic table and the generated conversion coefficient, thereby generating the predicted voltage characteristic table.

14. A program for generating a characteristic table, the program being for generating a predicted voltage characteristic table, the predicted voltage characteristic table defining a predicted characteristic curve between temperature and a voltage value applied to an ejection portion ejecting a liquid, the voltage value representing a peak value of a drive signal having one or more pulses, relative to a predetermined reference value, the program causing a computer to execute the following steps: As input data, predetermined parameters and a measured viscosity characteristic table defining a measured characteristic curve between viscosity and temperature of the liquid are respectively acquired; generating a conversion coefficient based on the predetermined parameter using a first analysis method as a predetermined analysis method, the first analysis method using the predetermined parameter as an explanatory variable and the conversion coefficient when performing a conversion process from the measured characteristic curve to the predicted characteristic curve as a target variable; and The conversion process is performed using the actually measured viscosity characteristic table and the generated conversion coefficient, thereby generating the predicted voltage characteristic table.

Images