EP1491347B1 - Method and system to reduce printing errors in a mail processing device - Google Patents

Abstract

Method for reducing printing errors during printing of mail items has the following steps: generation of encoder impulses corresponding to the relative movement between print and print medium; up and down counting to determine a state with reduced time interval between encoder impulses; execution of direct memory access (DMA) for accessing binary pixel data and; implementation of a print cycle for the data string with subsequent DMA access to the data string during subsequent print cycles. An independent claim is made for an arrangement for reducing printing errors during printing of moving mail items with a moving print head.

Description

The invention relates to a method for reducing printer errors during printing in a mailing machine, according to the preamble of claim 1 and an arrangement according to the preamble of claim 4. The invention is used in franking machines, addressing machines and other printing mail processing equipment.
The invention is related to unpublished German patent applications 10230678.8 entitled "Arrangement for Controlling Printing in a Mailing Machine" and 10230679.6 entitled "Method for Controlling Printing in a Mailing Machine".
In the field of digital printing, the variable print image share is becoming more extensive and the print resolution is getting higher and higher. For example, the variable printed image portion should be flexible for different postal requirements and is changed from impression to impression. At the same time, the controller is burdened with other tasks. To relieve the control by a microprocessor, it has proven to be advantageous to arrange the pixel data belonging to a print image column in the pixel memory so that variable picture elements can be changed by the microprocessor in the available time. Furthermore, it has proven to be advantageous if the responsible for the control of the printing mail processing device or system microprocessor is relieved by a pressure data control in the control of printing. The printing requires a relative movement between a print head and a print carrier, for example a sheet-like object, letter, postcard, packet, franking strip, address label or label.

In US 6,457,901 (DE 10032855 A1) is an apparatus for printing a printing medium proposed in a mailing machine, as preferably in the franking machine ultimail® the patentee Francotyp Postalia AG & Co.KG is used. The print carriers, for example, envelopes, franking strips or comparable franking goods, be post downstream by means of a driven transport drum transported in the transport direction, with a non-driven Counterpressure the print carrier against the transport drum in Pressing direction orthogonal to the transport direction presses. The two Ink cartridges partially protrude into the transport drum and carry two inkjet printheads which are printing on the moving Carry out the print carrier without contact. Both printheads are orthogonal to the transport direction and orthogonal to the pressure direction and with their Nozzles arranged near the edge of the transport drum, the nozzles during a pass of the print carrier through the machine a complete Can produce impression. On the front side of the transport drum are applied close to the perimeter markings that over the Scope of the transport drum are distributed. The markers are for example Reflective lines that come from a reflex light bank or Transmitted light sensor of an encoder detected by the microprocessor and a Control are converted into pressure pulses in the ratio 1: 1. Gradual slow fluctuations in transport speed thus have no influence on a generated print image when the Pressure cycle is changed accordingly in its duration.

According to the printhead manufacturer, a print cycle should be 90% of the time take between two positive encoder edges. The time between the Encoder edges can be measured and the printing cycle time accordingly be adjusted. The set value for the print cycle duration, however, will be the first after its detection with a delay of at least two Encoder clocks effective. By changing the letter thickness and mechanical Vibrations of the encoder sampling system can be erratic Change in the transport speed can be caused. The However, sudden fluctuations in the transport speed have a negative impact on a generated print image, i. despite one subsequent adaptation of the pressure cycle duration to the distances between the encoder edges must be with printer errors due to low Distances between the encoder pulses can be calculated.

The object of the invention is a method and an arrangement to reduce printer errors while printing on to provide moving mail in a mailing machine, the Printer error due to encoder pulses too short a time interval be caused. By a cost-effective solution is also under unfavorable conditions ensure high print quality.

The object is achieved with the features of the method according to the claim 1 or solved with the features of the arrangement according to claim 4.

Assuming that encoder pulses corresponding to the relative movement generated between the printer and a print carrier and one Print data control are supplied with correct encoder pulses at each encoder pulse edge, a period with a sequence of DMA cycles and one print cycle started without printer errors. The print data controller generates print data from the pixel data for parts of each pressure column and triggers a pressure cycle for the latter out. This will cause the microprocessor to control the printing relieved. During each DMA period, the pixel memory becomes binary Transmitted pixel data of a data string for print data control, which the binary pixel data depends on the connected printhead type for a print cycle. The invention works Therefore, from the consideration that it is printed for the evaluation security-relevant postal data would be unfavorable if complete Printing cycles, i. entire printing gaps are missing or those to a pressure column associated pixels are printed offset over several printing columns because the encoder pulses precede the print cycles. That can can be reduced by following complete print cycles immediately. It was found that the failure has only a few pixels of individual pressure gaps are no decisive obstacle to the evaluation of security imprints with a print image resolution of> 300 dpi represents when the falling pixels are distributed across the column, i. Not occur immediately after each other. Assuming that a reduction the encoder period only occurs intermittently, is at several consecutive Encoder pulses with too short a time interval Print cycle aborted if the encoder period is greater than a period of time with DMA cycles. By subsequent adjustment of the printing cycle time the distances of the encoder edges are also ensured that this case remains a temporary exception.

a) Erzeugen von Encoderimpulsen entsprechend der Relativbewegung zwischen Drucker und Druckträger,

b) Vorwärts- und Rückwärtszählen zur Auswertung eines Zustandes mit verringerten zeitlichen Abstandes benachbarter Encoderimpulse,

c) Durchführung von direkten Speicherzugriffen (DMA) für einen Datenstring von binären Pixeldaten und

d) Durchführung eines Druckzyklusses für den vorgenannten Datenstring, wobei während des Druckzyklusses weitere direkte Speicherzugriffe (DMA) für einen nächsten Datenstring erfolgen und wobei nach der Durchführung der direkten Speicherzugriffe (DMA) für den nächsten Datenstring und in Abhängigkeit vom zeitlichen Abstand der Encoderimpulse einer Anzahl von aufeinanderfolgenden Encoderimpulsen der Druckzyklus vollständig ausgeführt wird, solange der Mittelwert der Encoderperiode die eingestellte Dauer eines Druckzyklusses nicht unterschreitet oder wobei bei verringertem zeitlichen Abstand der Encoderimpulse der Anzahl von Encoderimpulsen die Durchführung des Druckzyklusses zum Ausdrucken von binären Pixeldaten eines vorherigen Datenstrings vorzeitig abgebrochen wird.

The procedure for reducing printer errors during printing in a mailing machine includes the following steps:

a) generating encoder pulses according to the relative movement between the printer and the print carrier,

b) forward and backward counting for evaluation of a state with reduced time interval of adjacent encoder pulses,

c) Performing direct memory accesses (DMA) for a data string of binary pixel data and

d) carrying out a printing cycle for the aforementioned data string, wherein during the printing cycle further direct memory accesses (DMA) are carried out for a next data string and wherein after carrying out the direct memory accesses (DMA) for the next data string and depending on the time interval of the encoder pulses of a number of successive encoder pulses, the printing cycle is carried out completely, as long as the average value of the encoder period does not fall below the set duration of a printing cycle, or with a reduced time interval of the encoder pulses of the number of encoder pulses, the execution of the printing cycle for printing binary pixel data of a previous data string is prematurely terminated.

The print data control is connected on the one hand via a bus with the microprocessor and with a pixel memory for the transmission of the print data and on the other hand with an encoder which detects the movement of the print carrier. At least one print head is connected to the print data controller in a manner known per se via a driver circuit. The print data controller, in a printhead-type dependent sequence, provides the required data to the respective drive circuitry of an associated printhead. It is envisaged that the print data controller has an evaluation unit and a logic for error reduction of printer errors, which encoder pulses are supplied and which contains a resettable encoder clock counter whose count is incremented within a count range smaller than or equal to an upper limit with each leading edge of an encoder pulse whose count is decremented with each print cycle start and the output side outputs a digital count, the latter being evaluated by the evaluation unit with respect to exceeding a set within the count range target value, when not exceeding the aforementioned target value, each print cycle completely executed and a running print cycle when exceeding the aforementioned target value the condition is aborted that all direct memory accesses (DMA cycles) to the pixel memory are prepared, which prepare the next print cycle ,
The print data control consists, for example, of at least one pixel data processing unit, a DMA controller, an address generator and a printer controller. The pixel data for each current print cycle is alternately stored in one of two buffers of the pixel data preparation unit. The pixel data for the next print cycle are loaded via DMA into a second buffer of the pixel data processing unit, while the pixel data for the current print cycle are transmitted to the respective driver circuit of the associated print head. The address generator for this purpose provides read addresses for pixel data of the pixel data processing unit and the printer controller organizes the serial output of the pixel data to the respective driver circuit in a print cycle in response to the encoder pulses. The printer controller according to the invention has a logic for error reduction of printer errors, which are caused by encoder pulses too short a time interval. The aforesaid logic includes a resettable encoder clock counter whose count is incremented within a count range less than the upper limit of each leading edge of an encoder pulse, the count of which is decremented at each print cycle start, and whose outputs are coupled to first and second comparators, the first comparator being reached of the zero count prevents decrement and the second comparator checks to reach the upper limit. The encoder clock counter provides the output side of the aforementioned count and is operatively connected to the address generator, the latter aborts the current print cycle when all pixel data for the next print cycle via DMA have already been loaded into the buffer of the pixel data processing unit and the counter value in the count range of the encoder clock counter a predetermined Setpoint has exceeded.

Figur 1a,

Impuls/Zeit-Diagramm für korrekte Encoderimpulse ohne eine Fehlerreduzierung,

Figur 1b,

Impuls/Zeit-Diagramm für Encoderimpulse, wobei die Encoderimpulse in einem zu geringen zeitlichen Abstand auftreten, und ohne Fehlerreduzierung,

Figur 1c,

Impuls/Zeit-Diagramm für Encoderimpulse ohne Fehlerreduzierung, wobei die Encoderimpulse im zu geringen zeitlichen Abstand über einen längeren Zeitabschnitt auftreten,

Figur 1d,

Impuls/Zeit-Diagramm für korrekte Encoderimpulse mit Fehlerreduzierung,

Figur 1e,

Impuls/Zeit-Diagramm für Encoderimpulse mit einem zu geringen zeitlichen Abstand und mit Fehlerreduzierung,

Figur 1f,

Impuls/Zeit-Diagramm für Encoderimpulse mit Fehlerreduzierung, wobei die Encoderimpulse einen zu geringen zeitlichen Abstand über einen längeren Zeitabschnitt aufweisen,

Figur 2,

Blockschaltbild für die Pixeldatenaufbereitung durch eine Druckdatensteuerung,

Figur 3,

Ausschnitt aus der Schaltungsanordnung nach Fig.2 mit einer Pixeldatenaufbereitungseinheit für den zweiten Druckkopf,

Figur 4,

Druckersteuerung,

Figur 5,

Flussplan zur Ablaufsteuerung der Druckersteuerung,

Figur 6,

Logik zur Druckerfehlerreduzierung,

Figur 7a,

Flussplan zur Druckerfehlerreduzierung,

Figur 7b,

Impuls/Zeit-Diagramm für Druckerfehlerreduzierung,

Figur 8,

Blockschaltbild eines-Adressengenerators,

Figur 9,

Blockschaltbild einer DMA-Steuerung,

Figur 10,

Flussplan zur DMA-Steuerung,

Figur 11,

Flussplan zur Adressengenerierung,

Figur 12,

Tabelle zur Adressengenerierung,

Figur 13,

Flussplan der Ausgaberoutine.

Advantageous developments of the invention are characterized in the subclaims or are presented in more detail below together with the description of the preferred embodiment of the invention with reference to FIGS. Show it:

FIG. 1a,

Pulse / time diagram for correct encoder pulses without error reduction,

FIG. 1b,

Pulse / time diagram for encoder pulses, where the encoder pulses occur too short a time interval, and without error reduction,

FIG. 1c,

Pulse / time diagram for encoder pulses without error reduction, wherein the encoder pulses occur in too short a time interval over a longer period of time,

FIG. 1d,

Pulse / time diagram for correct encoder pulses with error reduction,

FIG. 1e,

Pulse / time diagram for encoder pulses with too short a time interval and with error reduction,

FIG. 1f

Pulse / time diagram for encoder pulses with error reduction, wherein the encoder pulses have too short a time interval over a longer period of time,

FIG. 2,

Block diagram for the pixel data processing by a pressure data control,

FIG. 3,

2 shows a section of the circuit arrangement according to FIG. 2 with a pixel data processing unit for the second printhead, FIG.

FIG. 4,

Printer control,

FIG. 5,

Flowchart for the flow control of the printer control,

FIG. 6,

Logic for printer error reduction,

FIG. 7a,

Flowchart for printer error reduction,

FIG. 7b,

Pulse / time diagram for printer error reduction,

FIG. 8,

Block diagram of an address generator,

FIG. 9,

Block diagram of a DMA controller,

FIG. 10,

Flow plan for DMA control,

FIG. 11,

Flowchart for address generation,

FIG. 12,

Table for address generation,

FIG. 13,

Flowchart of the output routine.

FIG. 1a shows a pulse / time diagram for correct encoder pulses without error reduction. Upon the occurrence of each encoder pulse edge, a print cycle and a DMA period are started with DMA cycles when previously a previous print cycle has been completed. The latter is the case with correct encoder pulses.
When printing in a mailing machine, for example, a postage meter, a mail piece is moved in the transport direction under or on at least one print head along. In this case, the print image is generated from columns in a vertical arrangement to the transport direction, each print head with a time-offset printing an associated part of the same column. The time offset in the control of both printheads in order to print one and the same column results, on the one hand, from the distance between the two printheads in the transport direction and, on the other hand, from the transport speed of the print carrier in the transport direction. Since the two print heads are arranged offset in the transport direction to each other and each printhead has at least 300 nozzles, which are also arranged in two spaced in the transport direction nozzle rows, pixels are printed in four spaced pressure columns at the same time driving both printheads at any time. The control of all 300 nozzles of each printhead takes place within a printing cycle in which 22 groups of 14 pixel data are successively transferred to the printhead and printed.
The letter movement is detected by an encoder. After each positive edge of the encoder signal, a print cycle is started, provided that before the occurrence of this encoder edge of the triggered by the previous edge print cycle is completed (Figure 1a). The latter is not the case, however, if a sequence of encoder pulses occurs at too short a time interval because of too short a pulse width or pause (FIG. 1b).

FIG. 1b shows a pulse / time diagram for encoder pulses, in which the third to sixth encoder pulses in too short a time Distance occur. So not every encoder pulse will trigger a print cycle out. This causes some encoder pulses, especially the third one and fifth encoder pulse, now no print cycle and no DMA period trigger because the time between the second to sixth encoder pulses too short. The causes are different. The to Insertion of an encoder disk can result in an error of up to 10% exhibit. Also needs additional errors such as one sudden change in transport speed can be assumed, due to changes in letter thickness, bumps and mechanical Vibrations between encoder disk and encoder scanning caused become. Despite an adjustment of the pressure cycle time to the Distances between the encoder edges must therefore with printing errors calculated by too small distances between the encoder pulses if no measures are taken to reduce the error.

FIG. 1c shows a pulse / time diagram for encoder pulses without error reduction, wherein due to too short a pulse width and / or pause, the second to seventh encoder pulses have too short a time interval, so that over a longer period of time, only each one second print cycle and time period is triggered with DMA cycles. For printing, in particular a 2D barcode, printed image errors can be very annoying if the authenticity of a franking imprint is to be detected therewith.
In the case of the occurrence of encoder pulses in too short a time interval only the next following encoder pulse triggers a print cycle, ie in Figure 1b, the fourth and sixth encoder pulse and in Figure 1c, the fourth, sixth and eighth encoder pulse. Thus, the distance between the printed dots in the transport direction is about twice as large as intended. Also, the pixels belonging to a print column are offset over several print columns offset as dots. The print quality deteriorates to the extent that encoder pulses occur in too short a time interval.

FIG. 1d shows a pulse / time diagram for correct encoder pulses with error reduction. The time between the encoder edges is from the Microprocessor then determines the print cycle duration the information the printhead manufacturer adjusts accordingly with a delay of at least two encoder clocks becomes effective. According to printhead manufacturer should a print cycle as possible 90% of the time between two take positive encoder edges. Together with the aforementioned Control of the pressure cycle duration, which for encoder pulses with large Time interval is effective, the invention is a counter for Encoder pulses used with too short a time interval. Of the Encoder clock counter is used for error reduction. He will be at the beginning of each Pressure cycle reset to the value 'zero' and at each Encoder pulse increments, i. set here to the value 'one'. To the start of the DMA controller and the address generator, wherein the latter triggers the print cycle for the pixel data of a data string, the encoder clock counter is decremented, i. here to the value 'zero' set.

FIG. 1e shows a pulse / time diagram for encoder pulses too short a time interval and with error reduction. Of the time interval between the first two encoder pulses is greater as a print cycle. From the second to the sixth encoder pulse is the time interval between adjacent encoder pulses, however significantly smaller than a print cycle. From the sixth encoder pulse is the time interval between adjacent encoder pulses again larger and larger than a print cycle. According to the invention, it is provided that depending on the reduced time interval a number of encoder pulses the distance of the printing cycles from each other is reduced until the subsequent printing cycle for printing of binary pixel data of a subsequent data string immediately a complete print cycle to print binary Pixel data from a previous data string, with the decrease to the same extent as the encoder pulses the pressure cycles run ahead.

FIG. 1f shows a pulse / time diagram for encoder pulses Error reduction, whereby the encoder pulses are too short a pulse width and / or pause over a longer period of time. In the Figure 1f performs a count value 2, which exceeds a target value Z = 1, at seventh encoder pulse to cause the current print cycle to go around the hatched area is aborted prematurely and a new one Printing cycle begins.

FIG. 2 shows the block diagram of the preferred circuit arrangement for pixel data processing by a print data controller. One first and a second print head 1 and 2 are each via a driver unit (Pen Driver Board) 11 and 12 with a print data control 4 connected, which in the case of a direct memory access on the input side 16 bit assumes parallel binary print image data from a BUS 5 and on the output side serially binary print image data to the driver units 11 and 12 delivers. Via the BUS 5 are at least one microprocessor 6, a pixel memory 7, a nonvolatile memory 8 and a read-only memory 9 connected in terms of address, data and control. An encoder 3 is connected to the print data controller 4 for caching binary pixel data and printing the image columns trigger each printhead with a clock frequency of max. 6.5 KHz is operated. One for a franking imprint postal secured and controllable via an associated power unit (pen driver board) Inkjet printhead mounted in an HP 51645A type ink cartridge The company Hewlett Packard is located in the European Patent Application EP 1 176 016 A2, which is entitled: Arrangement and method for data tracking for warm-up cycles of Inkjet printheads. The print data controller 4 has a first and second pixel data processing unit 41 and 42 and the associated Controls such as DMA controller 43, address generator 44 and Printer control 45. The controls are included for error reduction of printer errors, an evaluation unit 453 and a logic 452, which the encoder pulses e are supplied. It is intended that one Printer controller 45 at least with the DMA controller 43 and with the Address generator 44 and that the address generator 44 with the Pixel data processing unit 41, 42 is connected in terms of control.

The printer controller 45-is connected directly to the microprocessor 6 via the BUS 5 and via an interrupt signal control line I. The DMA controller 43 is connected to the microprocessor 6 via a control line for DMA control signals DMA _ACK , DMA _REQ .

FIG. 3 shows a section of the circuit arrangement according to FIG. 2 with the pixel data processing unit 42 for the second print head, with the DMA controller 43 for a direct memory access (DMA) as well as with the address generator 44 and the printer controller 45. The encoder 3 supplies a signal e and is connected to the inventive logic 452 of the printer controller 45. The latter is connected to the DMA controller 43 directly via control lines 46 for first DMA control signals (DMA start and DMA busy), the DMA controller 43 being supplied with the DMA start signal by the printer controller 45, and with the DMA Control 43 outputs the DMA busy signal with the value 'zero' to the printer controller 45 to signal that the DMA cycles have ended. The DMA controller 43 is also connected to the address generator 44 via a DMA busy signal control line 50.
The printer controller 45 is connected to the microprocessor 6 via the bus 5 and via an interrupt signal control line 47, to the address generator 44 via a control line for supplying an address generator start signal AG-start and to the DMA controller 43 via a switching line 49 for a switching signal SO connected. At least the microprocessor 6, the pixel memory 7, the nonvolatile memory 8 and the read-only memory 9 are connected in terms of address, data and control via the bus 5. It is provided that the printer controller 45 has means for generating and outputting a switching signal SO in order to drive the pixel data processing unit 42 thereby selecting the pixel data from the respectively first or the respective second of the two latches 412 and 422 for transmission to the driver unit 12 become. The driver unit 12 can thereby be supplied in groups to the binary pixel data of a further data string. The DMA controller 43 is supplied from the printer controller 45, the switching signal SO. The DMA controller 43 has means for generating and outputting selection signals Sel_2.1, Sel_2.2 as a function of the switching state of the switching signal SO in order to buffer the binary pixel data into the respectively first or the respective second of the two latches 421 or 422, wherein in a transfer of pixel data from the respective one of the two latches to the driver unit 12, the respective other latches for buffering a data string are successively selected by the selection signals. The binary pixel data is provided to the pixel data conditioning unit 42 in the DMA periods in a data string manner, respectively.
The DMA controller 43 outputs the DMA busy signal with the value 'zero' to the printer controller 45 to signal that all direct memory accesses have occurred, ie, the period is completed with DMA cycles. The printer controller 45 is connected to the address generator 44 via at least one control line for supplying a start signal (AG start). The address generator 44 has - in a manner not shown here - a unit for generating read addresses and means for forming an address read signal AR. If a number of address read signals AR has been generated for an address group A, the address generator 44 outputs a print start signal PS via the control line 48 to the printer controller 45.
After transmission of a 22nd data set all of the 300 binary pixel data are transmitted, which requires a ½ inch inkjet print head for printing per print cycle. The other half of the print image is printed by the first printhead. The pixel data processing unit for the first printhead is constructed in a similar manner.
The printer controller 45 is connected to the microprocessor 6 via a control line 47 for an interrupt signal I and via the bus 5. At least the microprocessor 6, the pixel memory 7, the non-volatile memory 8 and read-only memory 9 are connected in terms of address, data and control via the BUS 5. The printer controller 45 has means for generating and outputting a switching signal SO and is connected via a control line to the DMA controller 43 and to the pixel data processing unit 42. The latter is controlled in order to select one of the latches 412, 422 for the transmission of pixel data to the driver unit 12 by means of a switching signal SO. The latter can thereby be supplied in groups to the binary pixel data of an already stored data string. The switching signal SO is supplied to the DMA controller 43 to select the other one of the latches 412 and 422 for pixel data loading. The DMA controller 43 has means for generating and outputting selection signals Sel_2.1, Sel_2.2 as a function of the switching state of the switching signal SO in order to buffer the binary pixel data into the respectively first or the respective second of the two latches 421 or 422, wherein in a transfer of pixel data from the respective one of the two latches to the driver unit 12, the respective other latches for buffering a data string are successively selected by the selection signals.
The two pixel data processing units are each connected to the input side of the bus 5 but there only to the low-order 16 bits of the data bus. In the following exemplary embodiments, the terms "data word" and "wordwise" should always be understood to mean a 16-bit wide data word if the data word width is not expressly specified additionally. The pixel data for a ½ inch printhead requires only half the space (at most 320 bits from each data string) in the pixel memory 7, from which this pixel data is provided to the pixel data conditioning unit 42. A data string for both printheads therefore requires that buffering of 20 * 16 bit data words each time be performed twice, for example in the respective first buffer. For successive data strings, the first and second latches 421 and 422 are alternately selected by the selection signals Sel-2.1 and Sel-2.2.
It is envisaged that the DMA controller 43 is control connected to the microprocessor 6 and to the latches 421 and 422, that the DMA controller 43 comprises means for generating and outputting address write signals AW, which in a DMA access to the in Pixel memory 7 stored binary pixel data allow their writing to the latches 421, 422 of the pixel data processing unit 42. The DMA controller 43 supplies a 5-bit address write signal AW for wordwise addressing. The latter is in each case at a separate address input of the first and second buffer memories 421 and 422 for pixel data for the second print head. The DMA controller 43 provides a first select signal Sel_2.1 for pixel data for the second printhead and asserts a separate control input of the first pixel data for the second printhead 421. The DMA controller 43 provides a second select signal Sel_2.2 for pixel data for the second printhead and asserts a separate control input of the second pixel data latch 422 for the second printhead.
It is provided that the address generator 44 has at least means for generating and outputting address signals AR, AP and control signals WR, LD, PS, and wherein the address signals AR, AP and control signals WR, LD of the pixel data processing unit 41, 42 for selecting the cached binary pixel data and their grouping in a predetermined order.
Each pixel data processing unit has two latches, a selector for selecting the binary pixel data, and a shift register for parallel / serial conversion of the binary pixel data provided in a new order. The address read signals AR generated by the address generator 44 are supplied to the latches and the selector of the pixel data processing unit. The primitive address signals AP and the write control signal WR generated by the address generator 44 are supplied to the selector, and a load signal LD is supplied to the shift register. From the printer controller 45, a start signal AG-start is supplied to the address generator 44. The address generator 44 now provides an address read signal AR for selecting the data word with the pixel data destined for the second printhead. For word-by-word addressing, the higher-order bits of the address read signal AR are applied to a separate address input of the first and second buffer memories 421 and 422. The four least significant bits of the address read signal AR are applied to an address input of a second selector 423 and allow addressing within the 16-bit data word. The parallel data outputs of the first and second pixel data for the second printhead 421 and 422 are applied to a first and second input of the selector 423, controlled by the address generator 44 at its output a 14 bit parallel data signal to the parallel data input of a shift register 424 for Pixel data for the second printhead. The shift register 424 is controlled by a shift clock signal SCL of the printer controller 45, and outputs a serial data output signal SERIAL DATA OUT 2. The address generator 44 also generates a primitive address AP for controlling the selector 423 and a write signal WR. The address generator 44 outputs a load signal LD to the shift register 424 and a print start signal PS to the printer controller 45. The latter outputs the Latch and Print2 signals for the control of the Pen Driver Board 12. The printer controller 45 is connected via a control line for the output of the switching signal SO to a corresponding control input of the DMA controller 43 and to the selector 423 of the pixel data processing unit 42. The printer controller 45 has evaluation means for evaluating the address and control signals transmitted via BUS 5, which are evaluated with regard to the occurrence of a print command. The printer controller 45 generates at least the DMA start, AG start and SO signals, stores the latter in registers, and communicates with the DMA controller 43 via DMA start, DMA busy, and SO signal control lines. The SO signal is generated only upon receipt of a print command and triggered by the print command is output from the printer controller 45, a first control signal DMA start to the DMA controller 43, the latter then generates a request signal DMA _REQ and sends to the microprocessor 6. The microprocessor has an internal DMA controller (not shown) which, in the case of a direct memory access, applies a specific address to the pixel memory (RAM) 7, thereby enabling word-wise transmission of binary pixel data via BUS 5 to the latches. An address write signal AW is supplied to the latches by the DMA controller 43 for this purpose. The microprocessor 6 can read out from the pixel memory 7, for example, a 16-bit wide data word with pixel data via DMA and transmit it to the print data control unit. The microprocessor 6 sends an acknowledgment signal DMA _ACK to the DMA controller 43 to synchronize the generation of the address write signal AW in the DMA controller 43 with the DMA cycle of the microprocessor 6. For each DMA cycle, a 16-bit wide data word with binary pixel data enters a buffer. Each of the four latches can provide a total of 320 bits for further data conditioning after each 20 DMA cycles. To achieve a print resolution of 600 dpi, two of the four buffers each are used for storage during the DMA cycles. When wordwise storing and reading pixel data for the second printhead, the two latches 421 and 422 alternate. Therefore, during the DMA cycles, DMA controller 43 supplies first and second select signals Sel_2.1 or Sel_2.2 alternately for word-wise storing pixel data for the second printhead. By way of example, the DMA controller 43 supplies a first selection signal Sel_2.1 and an address write signal AW for the alternate and wordwise storage of pixel data for the second print head. The number of pixels desired for each print image column requires a maximum of 40 data words to buffer 16 bits in two out of four latches. In the DMA controller 43, circuit means are provided for outputting the second control signal DMA-busy and for realizing at least one cycle counter for a predetermined number of 16-bit data words.

In the same manner-but not shown in detail-the binary pixel data for the first printhead is word-delivered via BUS 5 and applied to a corresponding data input of the first and second latches 411 and 412 for pixel data for the first printhead. The first pixel data processing unit 41 for the first printhead-not shown in detail-also includes first and second latches 411 and 412, which are each connected on the input side to the low-order 16 bits of the data bus of the bus 5. The address write signal AW supplied by the DMA controller 43 is also applied to each of a separate address input of the first and second buffer memories 411 and 412 for pixel data for the first printhead. The DMA controller 43 supplies a first select signal Sel_1.1 for pixel data for the first printhead and asserts a separate control input of the first pixel data stager 411 for the first printhead. The DMA controller 43 provides a second select signal Sel_1.2 for pixel data for the first printhead and asserts a separate control input of the second pixel data for the first printhead.
The address read signal AR provided by the address generator 44 is also applied again to a separate address input of the first and second buffer memories 411 and 412 for pixel data for the first print head and to a first selector 413. The parallel data outputs of the first and second pixel pixel data latches 411 and 412 are applied to first and second inputs of the selector 413, controlled by the address generator 44 at its output, a 14 bit parallel data signal to the parallel data input of a shift register 414 for Provides pixel data for the first printhead. The shift register 414 is controlled by the shift clock signal SCL of the printer controller 45 and outputs a serial data output signal Serial data out 1.
The printer controller 45 outputs a shift clock SCL to the pixel data for the first printhead shift register 414 and signals Latch and Print1 for the control of the pen driver board 11. The printer controller 45 is connected to a corresponding control input of the DMA controller 43 and to the pixel data processing unit 41 via a control line for outputting the signal SO.
The cycle counter of the DMA controller 43 is a word counter for a predetermined number of 16-bit data words, which is started by a DMA start signal. The DMA control is, for example, part of an application-specific circuit (ASIC), the cycle counter is connected on the one hand with the aforementioned means for generating and output of address write signals AW and the other with means for generating and output of selection signals, the latter - in a manner not shown - Have at least one output means and comparison means. For example, a first comparison means controls the output means as a function of the SO signal in order to reach a first predetermined number of 16-bit data words for the first pixel data processing unit 41 specific selection signal Sel_1.1 or Sel_1.2 and after reaching the first predetermined number of 16-bit data words to output a for the second pixel data processing unit 42 specific selection signal Sel_2.1 or Sel_2.2. Upon reaching a second predetermined number of 40 * 16-bit data words, the first or second comparing means receives a signal which is applied to the cycle counter to terminate the counting of DMA cycles. From the signal, a DMA-busy signal with the value 'zero' is generated and output via a register.
While the pixel data for a data string is loaded by direct memory access (DMA) into the respective first latches 411 and 421 and stored therebetween, the respective second latches 412 and 422 can be read out. By means of the special address generator 44 and the selectors 413, 423, the binary pixel data are read from these latches in the order required by the printheads, collected in groups and then transferred by means of shift registers 414, 424 serially to the two printheads. At least one half of the print image column is printed by the first print head and at least one other half print image column is printed by the second print head.
By this solution, binary pixel data can be stored in the pixel memory in an optimal order, which relieves the microprocessor in the printed image change. Data transfer via DMA also relieves the load on the microprocessor.
The error-reducing encoder pulses logic 542 is a component of the printer controller 45 and the associated evaluation unit (not shown) forms part of the address generator 44. The logic 542 then transmits a more than two-bit wide digital count signal ENC to the address generator 44 on the output side Comparator contains.

In the printer controller 45, a data string counter is further realized (not shown in detail), wherein each data string has a maximum of the above-mentioned number of 40 * 16-bit data words. After the binary pixel data taken from a data string has been printed, the data string counter is incremented when the LH edge of the encoder clock occurs. When a predetermined setpoint U on data strings is reached, printing of the print image is ended.
The entire print data control can preferably be realized with an application specific circuit (ASIC) or programmable logic, such as Spartan II 2.5V FPGA from XILINX ( www.xilinx.com ).

4 shows a block diagram of the printer controller 45, which has a logic 542 for error reduction, which will be explained below with reference to FIG 6 in more detail. The aforementioned logic 542 is operatively connected to a sequencer and processing unit 451 to which a printer control logic 450 and an input / output unit 454 are connected. The printer control logic 450 comprises a number of blocks - not shown in detail - but at least the following blocks, such as a data string count V 4503 for the data string number V reached on printing, a register 4504 for the data string setpoint U, a comparator 4506 for U = V, a memory 4505 for Y and other parameters, a shift pulse generator 4507, a latch pulse generator 4508, a first print pulse generator 4509, and a second print pulse generator 4510.
In the first block 451 with the sequence control and processing unit are - not shown - an encoder filter and an encoder controller included, which provides a start signal for printing a column, an interrupt request with each rising edge of the encoder triggers and the correct time Transmission of print data within a print column supported. The encoder filter suppresses spikes on the encoder signals. In the first block 451 - in a manner not shown - yet another counter and further evaluation circuits included, the counter is a system clock for determining the period of encoder pulses is supplied. The e-input BUS-I / O and other registers of the I / O unit 454 communicate with the microprocessor. A microprocessor-controlled control of the pressure cycle duration causes - in a manner not shown - an adjustment of the printing cycle to a predetermined period of time, preferably about 90% of the time between two positive encoder edges.
The input / output unit 454 also has a number of blocks - not shown in detail - but at least the following blocks, a BUS input / output unit 4541, an input signal 4542 for the encoder signal e, an input 4543 for the DMA busy signal, a DMA start signal register 4544, an AG busy signal input 4545, an AG start signal register 4546, a switching signal SO register 4547, a PS signal input 4550; Output 4551 for the I signal, an output 4553 for the shift-clock signal, an output 4554 for the latch pulse signal, a Print1 pulse output 4555 and a Print2 pulse output 455x. To output the digital count ENC, the input / output unit 454 has an ENC output 4560.
If, for example, the evaluation unit 453 shown in dashed lines is likewise a component of the printer controller 45, then the output signal of the evaluation unit 453 is likewise transmitted to the address generator 44, but only as a bit-wide binary abort signal that is present at the output 4560 is provided.
The printer controller 45 can also be realized in the already explained or in an alternative embodiment, wherein each printer controller 45 - independent of the embodiment - has a data string counter 4503 and is connected to the encoder 3. After each printed data string, the value V of the data string counter is incremented when the encoder clock occurs, whereby the printing of the print image is terminated when a predetermined setpoint U of the data string counter is reached.

FIG. 5 shows a flowchart for the flow control of the printer controller. After switching on in step 101, a step 102 is reached and in routine 100 of the sequence control, the signals reduction signal DEC_ENC, reset signal Encoderzähler_Reset and selection signals Sel_1.1, Sel_1.2, Sel_2.1, Sel_2.2 are set to the value 'zero' , In a first query step 103, a data word transmitted via the bus is evaluated with regard to the occurrence of a command for printing start. If the latter has not yet been issued, then it branches into a waiting loop. On the other hand, after the start of printing in a step 104, the column count value V is set to the value 'zero'. The switching signal SO is set to the value 'one' and output. In a second interrogation step 105, the encoder signal e is now evaluated with regard to the occurrence of an LH edge. If the latter has not yet occurred, a branch is made to a waiting loop. On the other hand, in a step 106, a signal DMA start is output and a subroutine 300 is started which sets certain selection signals Sel_1.1, Sel_1.2, Sel_2.1 or Sel_2.2 to the value 'one' in order to convert the binary pixel data into the Latch the pixel data processing units 41 and 42 to take over what will be explained in more detail later with reference to FIG.
In a third interrogation step 107, the DMA busy signal is now evaluated as to whether it has been set to the value 'zero'. If the latter is not yet the case, then it branches into a waiting loop. However, if the DMA-busy signal has been set to the value 'zero', then in step 108 a signal Encoderzähler_Reset: = 1 is generated. Subsequently, a fourth interrogation step 109 is reached, in which the digital output signal ENC of the encoder clock counter is evaluated with respect to a value deviating from the value 'zero'. For the unambiguous time interval of the encoder pulses, the printing cycle is set to a predetermined period of time between two positive encoder edges. This results in a distance between the individual pressure cycles due to the fourth interrogation step 109 (FIG. 1d). If the digital output signal ENC is not equal to the value 'zero', ie equal to the value 'zero', then a branch is made to a waiting loop.
When an LH edge of an encoder signal occurs, the encoder clock counter is incremented and the digital output signal ENC now becomes unequal to the value 'zero'. When an LH edge occurs, a branch is thus made from the fourth interrogation step 109 in a step 110, in which the switching signal SO is logically negated and then output.
Subsequently, in step 111, the address generator is activated and a subroutine 400 is started, which generates for the pixel data processing units 41 and 42 specific read addresses AR and control signals such as the switching signal SO, the primitive address AP, the write signal WR and a load signal LD. In step 112, a DMA start signal is output and DMA control is enabled to restart the aforementioned subroutine 300. Both subroutines 300 and 400 are in parallel with each other. Then, in a step 113, the print cycle start is signalized by generating a short pulse or by setting the signal DEC_ENC to the value 'one' in a first substep 113a and the signal DEC_ENC to the value 'zero' in a second substep 113b. Subsequently, in a fifth interrogation step 114, it is evaluated whether the address generator has finished its subroutine 400 and whether the DMA busy signal has been set to the value 'zero'. If the former or the latter is still not the case, then a branch is made in a waiting loop. However, if the address generator has completed its subroutine 400 and the DMA busy signal has been set to the value 'zero', then a step 115 is reached. In step 115, the data string count is incremented by a value 'one' to V: = V + 1. In a sixth interrogation step 116, it is evaluated whether the column count value V has reached a limit value U. If this is not the case, then the fourth query step 109 is branched back. Otherwise, the system branches back to the first interrogation step 103 via a step 117, and the routine restarts when a print start command is detected in the first interrogation step 103. In the aforementioned step 117, a reset signal Encoderzähler_Reset: = 0 is generated and reset the encoder counter to the value 'zero'.

In the exemplary embodiment shown in FIG. 4, there is a separate second block 452 which is operatively connected to the first block 451 of the printer controller 45 containing the sequence control and processing unit. The aforementioned block 452 contains the logic for printer error reduction and is part of the printer controller 45. The connection comprises a plurality of - not shown - lines for analog and / or digital electrical signals. The second block 452, which generates a parallel n-bit wide digital output signal ENC representing a count value, serially feeds an encoder signal e filtered by the first block 451. The connection of the second block 452 to the ENC output of the I / O unit 454 is also via the first block 451 of the printer controller 45. The block 451 is connected to the I / O unit 454 of the printer controller 45. After powering up, the processing unit of the first block 451 sets the decrease signal DEC_ENC and the reset signal Encoder_counter_Reset to the value 'zero'. The printer controller 45 includes a third block 450 with the printer control logic. The latter has parameter Y in parameter memory 4505 as an upper limit for the printer error reduction logic.
Alternatively, a separate second block may be omitted (not shown) if the printer error reduction logic is part of the first block 451 of the printer controller 45.

FIG. 6 is a block diagram of the logic for printer error reduction shown. First and second AND gates 4521, 4522 are output with the inputs of an encoder clock counter 4523 connected, having a first input CLK_down, a second input CLK_up and a third input ENC_RESET for the reset signal Encoderzähler_Reset has. The decrease signal DEC-ENC becomes supplied via a first input of the first AND gate 4521. With its second input is the output K of a first digital comparator 4524 connected. In a simple embodiment (not shown) is used as the first digital comparator 4524 an OR gate, the the digital output signal ENC of the encoder clock counter 4523 linked and at ENC ≠ 0 output a logic signal 'one' on the output side.

The filtered encoder signal e is supplied in series via a first input of the second AND gate 4522, to whose second input the output L of a second digital comparator 4525 is connected.
At the respective first inputs of the first and second digital comparators 4524, 4525, the count value is in digital form in that a binary value is output at the, for example, n = 4 outputs Q1, Q2, Q3, Q4 of the encoder clock counter 4523. The encoder clock counter 4523 is shown in the embodiment as a 4-bit counter, but should not be limited to this alone. Rather, embodiments other than n-bit counters are possible. In the preferred embodiment, both digital comparators are the same. For example, n Exclusive OR gates are linked together on the output side via an OR gate, wherein each Exclusive OR gate compares one digit of the n-bit number with a corresponding position of the n-bit comparison number.
At the respective second inputs of the first digital comparator 4524 is an n-bit wide digital data signal with the binary value 'zero'. At the respective second inputs of the second digital comparator 4525 is an n-bit wide digital data signal with the corresponding binary values for an upper limit 'Y', which is supplied via the first block 451 of the printer controller 45. The first block 451 of the printer controller 45 also provides the encoder count_reset signal and the DEC ENC signal. The first digital comparator 4524 operates in a manner known per se and outputs a logic signal, for example TTL. For example, a binary value 'one' at the output K indicates that the condition ENC ≠ 0 is fulfilled. The DEC-ENC signal supplied via a first input of the first AND gate 4521 is consequently switched through to the first input CLK_down of the encoder clock counter 4523 in order to decrement the count value. The second digital comparator 4525 outputs a logic signal having the binary value 'one' at the output L when the condition ENC ≠ Y is satisfied. The filtered encoder signal e supplied via a first input of the second AND gate 4522 is then turned on to the second input CLK_up of the encoder clock counter 4523 to increment the count value. The first and second digital comparators 4524 and 4525 respectively output a TTL signal with the value 'zero' if the aforementioned respective condition is not fulfilled. In this case, the AND gates 4521 and 4522 are disabled.

FIG. 7 a shows a flowchart for reducing printer errors. The routine 700 is started in step 701. After the start, first in step 702 the count is reset to the value 'zero'. In the following interrogation step 703, in the expectation of an activation by means of a reset signal encoder counter reset with the value 'one', the process is continuously branched back to step 702.
The reset signal encoder counter reset is set to the value 'one' after a first DMA cycle by the processing unit of the block 451 shown in FIG. 4, as can be seen from step 108 of FIG. The aforesaid reset signal is applied to the third input ENC_RESET of the encoder clock count 4523 shown in FIG. If an encoder count reset signal of value 'one' is detected in query step 703, then a second interrogation step 704 is reached in which it is determined whether an encoder LH edge is applied to the second AND gate 4522 via a first input while the second comparator 4525 outputs on the output side a value 'one', with which the latter signals that the condition ENC ≠ Y is fulfilled.
The second AND gate 4522 outputs a value 'one' on the output side to the second input CLK_up of the encoder clock counter, which results in the step 705 being incremented by the value 'one' so that the count value ENC: = ENC +1 is achieved. Subsequently, a branch is made to a third interrogation step 706.
If, however, no encoder LH edge is supplied via a first input to the second AND gate 4522 and the condition ENC ≠ Y is not fulfilled, then the second query step 704 branches to the third query step 706.
In the third interrogation step 706, the first comparator 4524 determines that the count ENC is not equal to the count value 'zero'. In this case, a branch is made to the fourth interrogation step 707. Otherwise, the first query step 703 is branched back. In the fourth interrogation step 707, it is checked whether the DEC ENC signal supplied via the first input of the first AND gate 4521 has the value 'one'. In this case, the first AND gate 4521 outputs a value 'one' on the output side to the first input CLK_down of the encoder clock counter, which means that in the following step 708 the count value is decremented by the value 'one', so that the count value ENC : = ENC - 1 is reached. Otherwise, the fourth interrogation step 707 branches back to the first interrogation step 703. Subsequently, branching back from step 708 to the first query step 703.

FIG. 7b shows a pulse / time diagram for the printer error reduction. After the start, in the routine 100 of the printer controller, the reset signal Encoder_zähler_Reset and the decrease signal DEC_ENC are set to the value 'zero'. Several encoder pulses occur in too short a time interval over a longer period of time. In the pulse / time diagram of Figure 7b, however, the aforementioned period is longer than in the pulse / time diagram of Figure 1f.
After a first encoder pulse occurs, a first period begins at time t ₀ to load the pixel data for the first print cycle into the latches via DMA cycles. At the time t ₁ , the aforementioned encoder clock counter 4523 is then enabled by the reset signal Encoderzähler_Reset: = 1 set to the value 'one' (FIG. 5, step 108). With the count value 'zero', the condition ENC ≠ Y is fulfilled, so that the aforementioned encoder clock counter is incremented by the value 'one' with each positive encoder edge of the encoder signal e. At time t ₂ , a second encoder pulse occurs and the aforementioned encoder clock counter is incremented and now has the count value 'one'. The output K of the first comparator 4524 (FIG. 6) then outputs the binary value 'one' which is applied to the input of the first AND gate 4521. The output Q1 or K remain set to the binary value 'one' until the time t ₃ . Depending on the time interval between adjacent encoder pulses, either the incrementing is continued as long as a predetermined upper limit Y is not reached, or the count value can be decremented upon the occurrence of a print cycle. At time t ₃ , a second period is started with DMA cycles and a first print cycle for the pixel data of a data string loaded in the preceding first DMA period with DMA cycles. A pulse-shaped decrease signal DEC_ENC generated by the printer controller 45 is output to the logic 452, resulting in decrementing and the result that the count value again has the value 'zero'.
From the time t ₃ , the first-mentioned buffer of the print data controller 4 is read out for printing a data string. At the same time, the pixel data for a second print cycle are loaded by DMA into the respective other buffer of the print data controller 4. At time t ₄ , a third encoder pulse already occurs before a third print cycle can be started. The outputs Q1 and K remain set to the binary value 'one' until time t ₅ . At time t ₅ , a second print cycle and a third DMA period are started. The time interval of the subsequent times t ₆ and t ₇ , t ₈ and t ₉ and t ₁₀ and t _{11 increases to the} extent that at time t _{11 of} the seventh encoder pulse before the sixth DMA period is effective. In such a case, the encoder clock counter is no longer decremented to the same extent as it is incremented. As a result, the count increases. Despite a current print cycle, the first comparator 4524 checks the aforementioned logic 452 as to whether the counter contents are nonzero. If this is the case, it is waited for a reduction signal DEC_ENC. When a print cycle is started, the printer controller outputs a decrease signal DEC_ENC. At time t ₁₁ , the encoder clock counter is still incremented before decrementing. The outputs Q1 and K remain set to the binary value 'one' from time t ₁₀ to time t ₁₁ and then change to the value 'zero'. At the same time, the output Q2 changes to the binary value 'one', ie the counter content has risen to the count value 'two' at time t ₁₁ . This counter state stops only until the time t ₁₂ , since at the time t ₁₁ also a reduction signal DEC_ENC with the value 'one' has been delivered to the encoder clock counter. Output Q2 returns to binary 'zero' and output Q1 returns to binary 'one' according to the decremented 'one' count. At time t ₁₃ , the eighth encoder pulse is output and there are comparable conditions as at time t ₂ with the exception that the encoder clock count before incrementing is now at the aforementioned value 'one'. Also, the eighth encoder pulse at time t ₁₃ is effective before the seventh DMA period with DMA cycles. Output Q2 returns to binary 'one' and output Q1 returns to binary 'zero' according to the incremented count 'two'. This counter state stops only until the time t ₁₄ , since at the time t ₁₃ also a decrementing signal DEC_ENC with the value 'one' was delivered to the encoder clock counter.
After printing each pixel group, the address generator 44 in the subroutine 400 checks to see if the count has reached a set point Z. If this is the case and the data for the next print cycle is loaded into the buffer by DMA, the current print cycle is aborted. In FIGS. 1f and 7b, a count value-which exceeds the setpoint value Z = 1-causes the current print cycle at the seventh or seventh and eighth encoder impulses to be prematurely terminated by the region shown hatched and a new print cycle to begin.
At time t ₁₅ , the sixth print cycle is ended and a seventh print cycle begins. Therefore, a decrease signal DEC_ENC of value 'one' is output to the encoder clock counter, and the output Q1 returns to the binary value 'zero' corresponding to the decremented count value 'zero'. This counter state will only last until time t ₁₆ , since the ninth encoder pulse has been delivered. Thus, it is again incremented and the output Q1 changes back to the binary value 'one' corresponding to the incremented count 'one'. This continues until a predetermined number of printing cycles has been completed. The count of the encoder clock counter is decremented with each print cycle start (Figures 1d, 1e, 1f, and 7b). This process provides a number of advantages. All short-interval encoder pulses start printing cycles if the encoder pulse period is greater than the DMA cycle with DMA cycles. Only with a large number of consecutive encoder pulses of very short time interval is a slight impairment of the print quality to be expected. A resulting offset of the pixel data belonging to a print column is at most equal to the spacing of a print column in the illustrated exemplary embodiment. This ensures that the pixels belonging to each print column are not printed offset across multiple print columns. Since only one running print cycle is aborted prematurely, which is hatched in FIGS. 1f or 7b, fewer pixels are correspondingly missing in the print image than in the extreme case when a complete print cycle is missing.

FIG. 8 shows a block diagram of an embodiment of the address generator. It is provided that the address generator 44 has an input / output logic 444, an evaluation unit 442 and a read address generation unit 441, the latter having a first counter 4410 for the primitive address and an associated first comparator 4411 for comparing a count value P of the primitive address with a first setpoint supplied by a first setpoint register 4412. The unit 441 comprises a second counter 4413 for an address group and an associated second comparator 4414 for comparing a count A of the address group with a second set value supplied by a second setpoint register 4415 and a scheduler 4401. The latter works in conjunction with a calculation unit 4402 for the parameter C, a WR signal generator 4403, an LD signal generator 4404, a PS signal generator 4405, with the aforementioned counters 4410 and 4413, with the comparators 4411, 4414, 4418, with the registers 4412, 4415 4417 and with a AG-busy buzzer 4416.
The evaluation unit 442 comprises an inverter 4420 for the DMA busy signal, an AND gate 4423, a register 4422 for at least a third setpoint value Z and a third comparator 4421 to which the count value of the encoder clock counter is supplied for comparison with at least one third setpoint Z. becomes. The inverter 4420 and the comparator 4421 are each connected to an input of the AND gate 4423 on the output side. Normally, when only ENC values from 'zero' to 'two' occur, the third setpoint Z = 1. The third comparator 4421 may be simply constructed by checking that neither ENC values are 'zero' nor 'one'. occur. For example, on the one hand, the logic signals output at the Q outputs of the encoder clock counter are linked via an OR gate and on the other hand n Exclusive OR gates on the output side via a second OR gate, whereby each Exclusive OR gate in each case has one position of the n-gate. Bit number with a corresponding position of the n-bit comparison number of the setpoint Z = 1 compares and the OR gate outputs are connected via an AND gate, which outputs the output signal of the third comparator 4421. The AND gate 4423 outputs an abort signal BO having the value 'one' on the output side when the negated DMA busy signal and the output signal of the third comparator 4421 have the value 'one'.
It is further contemplated that the input / output logic 444 of the address generator 44 has an input 4450 for the DMA busy signal, an input 4451 for the ENC signal, an input 4444 for receiving the address generator start signal and a register 4445 for the signal to be sent Addressgeneratorbusysignal comprises.
Alternatively, the printer controller may already include the above-mentioned comparison of the ENC signal with the setpoint value Z, in order to generate an abort signal BO as a result of the comparison. The I / O unit 454 of the printer controller 45 then includes a BO output instead of the ENC output 4560. The input / output logic 444 of the address generator 44 need only contain a BO input instead of the ENC input 4451 which has the abort signal BO instead of the count ENC is supplied. This eliminates the control line 50, the DMA busy input 4450 and the evaluation 442 in the address generator 44. A corresponding evaluation unit 453 is instead realized in the printer controller 45 to perform the above comparison of the ENC signal with the setpoint Z and the result the comparison to generate an abort signal for the printing cycle that is supplied to the address generator 44.

The operation of the address generator 44 will be described in more detail below. After a formation of the address read signal AR and after an incrementation of a count value P for the primitive address by the value 'one', the comparison is made in the first comparator 4411, wherein after successively generates a number of read addresses, exceeding the first setpoint or an overflow of the Counter 4410 for the primitive address is triggered, a load signal LD is output and a subroutine for output is started, the counter 4413 for an address group is incremented by the value 'one', wherein falling below the second setpoint in the comparison in the second comparator 4414, a reset of the count value P of the primitive address to the value 'one' and generating a subsequent read address associated with a further address group and wherein upon the occurrence of a predetermined condition, a running print cycle is aborted prematurely. The latter is the case when, on the one hand, the third setpoint value is exceeded during comparison in the third comparator 4421 and, on the other hand, a DMA busy signal with the value 'zero' is output by the DMA controller, which indicates that all DMA cycles to access the pixel memory 7 are prepared, which prepare the next printing cycle.
The process controller 4401 is provided with a parameter C calculating unit 4402 having a signal generator 4403 for generating a write signal WR, a signal generator 4404 for generating a load signal LD, and another signal generator 4405 for generating a print start signal PS to cause the pixel data to be printed out a data string and connected to the busy signaling device 4416. The input / output logic 444 also has a register 4446 for the delivery of the primitive address AP, a register 4447 for the write signal WR, a register 4448 for the load signal LD, a register 4449 for the output of the address read signal AR and a register 4440 for the output of the Pressure start signal PS on.

FIG. 9 shows a block diagram of a DMA controller. The DMA controller 43 has at least one scheduler 4301, a word counter 4302, a setpoint register 4303, an input / output logic 4304, a memory 4305, a comparator 4306, and a shift register 4307, which are interconnected to perform DMA cycles. Integrated into the flow control 4301 is a further processing unit to which the aforementioned blocks 4302 to 4307 are connected in circuit and which has further comparators (not shown). In order to ensure the signal flow within the print data controller 4 and between the DMA controller 43 and the microprocessor 6, it is further provided that - not shown - the input / output logic 4304 at least one input 43042 for the received DMA start signal and Registers 43043 to 43046 for the select signals to be transmitted, a DMA busy signal register 43047, a request signal DMA _REQ register 43048, an input acknowledgment signal DMA _ACK input 43049, an input 43050 for the switching signal (SO) and register 43051 for the address write signal AW.

FIG. 10 shows a flowchart for DMA control. Such a subroutine 300 is called when a DMA start signal is output from the printer controller 45 to the DMA controller 43 (step 301). In a step 302 of the subroutine 300, a word count W is set to the value 'zero'. A DMA-busy signal is set to the value 'one' and transmitted to the printer controller 45. In a further step 303 of the subroutine 300, a DMA request _signal DMA _REQ having a value 'zero' is transmitted to the microprocessor 6. The latter transmits an acknowledgment signal DMA _ACK to the DMA controller 43. In a first interrogation step 304 of the subroutine 300, a non-receipt of the acknowledgment signal DMA _ACK branches to a waiting loop with a value 'zero'. From the first interrogation step 304 of the subroutine 300, on receipt of the acknowledgment signal DMA _ACK , a value 'zero' is jumped to a second interrogation step 305, the state of the switching signal SO being determined. If the switching signal SO has the state equal to one, then a branch is made to a third interrogation step 306. Otherwise, the switching signal SO has the state equal to 'zero' and a branch is made to a fourth interrogation step 309. In the third interrogation step 306, it is checked whether the word counter has a value W smaller than twenty. For this case (W <20), a branch is made to a step 307. In step 307, the first selection signal for the first print head Sel_1.1. switched to the value 'one' and the address write signal AW receives the current value W of the word counter. In the subsequent step 312, the pixel data are transferred to the latches of the pixel data modification units 41, 42. Subsequently, in step 313, all selection signals are switched to the value 'zero' and a DMA request _signal DMA _REQ with a value 'one' is transmitted to the microprocessor 6.
Then, in step 314, the word count W is incremented with the value 'one'. In a subsequent query step 315, it is checked whether the word counter has a value W less than forty. For this case, in which the word counter has such a value W <40, a branch back to a step 303. Otherwise, a branch is made to a step 316 to output a signal DMA busy having the value 'zero' before the end (step 317) of the subroutine 300 is reached. Otherwise, if it is determined in the third query step 306 that the word count W is not less than twenty, then a branch is made to a step 308 in which the first selection signal for the second print head Sel_2.1. is switched to the value 'one' and the address write signal AW is replaced by the value 'twenty' reduced current value W of the word counter. In the subsequent step 312, the pixel data are taken back into the buffer.
In the aforementioned fourth interrogation step 309, it is also checked whether the word counter has the predetermined value W <20, it being previously determined in the interrogation step 305 that the binary switching signal SO does not have the value equal to one. If the word counter has the predetermined value W <20, then in step 310, the second selection signal for the first printhead Sel_1.2. switched to the value 'one' and the address write signal AW receives the current value W of the word counter. In the subsequent step 312, the pixel data are taken back into the buffer.
Otherwise, if the word counter does not have the predetermined value W <20, the fourth interrogation step 309 branches to a step 311 in which the second selection signal for the second printhead Sel_2.2 is switched to the value 'one' and the address write signal AW the by the value 'Twenty' reduced actual value W of the word counter receives. In the following step 312, the pixel data are taken back into the buffer.

FIG. 11 shows a flowchart for address generation. The addresses of stored binary pixel data begin at both printheads with the start address zero, which is generated in the following manner for the address read signal AR. After the start in step 401, the initial values are called in step 402, A: = 1 for a counter of the address group, P: = 1 for a counter of the primitive address AP and C: = 255 for a counter of the address read signal AR. In the first query step 403, it is asked if the numerical value P of the counter of the primitive address is equal to the value one. If so, the second query step 404 is reached. Here it is determined whether the counter A has reached the value 8 or 9 or 15 or 16. If this is the case, then the step 406 is executed and the numerical value C is subtracted from the numerical value C of the counter of the address read signal AR. In the subsequent third interrogation step 418, it is determined that the numerical value C of the counter of the address read signal AR is greater than or equal to the value zero and then branches to step 419 for the output of the address read signal AR. Otherwise, a branch is made to step 420 in order to add a numerical value 512 to the negative numerical value. After steps 419 and 420, steps 425, 426 and 427 are traversed.
In step 425, the numerical value for the counter of the primitive address AP is output. Then, in step 426, a write signal WR for the binary pixel data entry is output to a collection register. In step 427, the numerical value for the numerator of the primitive address AP is incremented by the value one. Then, a fourth interrogation step 428 is reached and it is determined that the numerical value P of the counter of the primitive address AP has not yet reached the limit value 15. Subsequently, branching back to the first query step 403.
In the first query step 403, it is now determined that the numerical value P of the counter of the primitive address is not equal to the value one and branches to the fifth interrogation step 407. If the numerical value P is odd, then the program branches to the sixth query step 408, in which it is checked whether the counter of the address group has the value 8 or 15. If this is the case, then a branch is made to a step 409 and the numerical value 3 is added to the numerical value C of the counter of the address read signal AR. Otherwise, branching is made from the sixth interrogation step 408 to a step 410, and the numerical value 47 is added to the numerical value C of the counter of the address read signal AR.
If, however, the numerical value P is even, then the fifth query step 407 branches to the seventh query step 415, in which it is checked whether the counter of the address group has the value 8 or 15. If this is the case, then a branch is made to a step 416 and the numerical value 41 is added to the numerical value C of the counter of the address read signal AR. Otherwise, a branch is made from the seventh interrogation step 415 to a step 417 and the numerical value C is subtracted from the numerical value C of the counter of the address read signal AR.
From the aforementioned steps 405, 406, 409, 410, 416 and 417, the third interrogation step 418 is reached again and it is determined whether the numerical value C of the counter of the address read signal AR is greater than or equal to the value zero. After steps 419 and 420, steps 425, 426 and 427 are again run until the fourth interrogation step 428 is reached, in which it is determined whether the numerical value P of the counter of the primitive address AP has already reached the limit value 15. If this is the case, then a branch is made to a step 429 and a charging signal for loading the shift register is output. In order to print out the pixel data, a subroutine 500 is started in step 430 in which, inter alia, a shift clock signal SCL is applied to the shift register in order to serially output the pixel data from the latter. Subsequently, in step 431, the value of the counter of the address group is incremented by the value one. Then, an eighth query step 432 is reached in which it is determined whether the numerical value A of the counter of the address group has already reached the limit value 23. If this is not the case, then a ninth query step 433 is reached, in which it is determined whether the DMA busy signal with the value 'zero' already exists and whether the numerical value ENC of the counter of the address group exceeds the third setpoint Z. If this is not the case, then the first query step 403 branches back. If, however, the third setpoint value Z is exceeded, then a signal AG-busy 0 is output in step 434 and the subroutine 400 is stopped in step 435.
In an alternative variant, with a corresponding evaluation circuit in the printer controller 45, need only be determined in the ninth query step 433, whether an abort signal already exists.

FIG. 12 shows a table for generating addresses, whose address read signals AR for 22 address groups are generated by the aforementioned routine 400. The address generator 44 preferably generates the address values as a binary number and applies them to the pixel data processing units 41, 42. For example, a binary number may be represented as a hexadecimal number or a decimal number, which requires less space. Only for this reason and for better understanding are decimal numbers entered in the table. The routine 400 first generates a primitive address P: = 1 and a binary number zero as an address read signal AR for a first address group A 1. Then, one after the other until the primitive address P: = 14, a corresponding binary number as address read signal AR for the first address group A: = 1 generated. The address read signal AR (address read) is thus generated for 14 binary numbers per address group. One after the other corresponding binary numbers are generated as address read signal AR for 22 address groups. Each address read signal AR is used to access a binary pixel data in the buffer.
The driver units 11 and 12 ignore the binary pixel data obtained at address values A = 1 with P = 2, A = 7 with P = 13, A = 8 with P = 1 and with P = 14, A = 15 with P = 1 and with P = 14, A = 16 with P = 2 as well as those at A = 22 with P = 14 are read. The address values greater than 500 must therefore not be able to be generated completely as a binary number. To provide binary pixel data, all address values greater than 299 are generated but also not needed for printing.
Routine 400 is executed until all print image columns have been printed or aborted. It has already been explained that the rows of nozzles of a printhead become active alternately for printing image columns. While one of the latches is being loaded with binary pixel data via direct memory access, the other latches are read out to transfer rendered groups of binary pixel data to the driver units. The mutual repetition of the routine 400 and further subsequent steps are caused by the printer controller 45, which under the control of a signal e of the encoder 3 also generates the print signals Print 1 or Print 2.

FIG. 13 shows a flowchart of the output routine 500. At the start thereof, the address generator 44 outputs a print start signal PS to the printer controller 45. For a full print cycle, the issue routine is called twenty-two times as a subroutine in the course of subroutine 400 to drive the shift registers in the print data controller 41, 42 and around the driver units 11, 12. After the start in step 501, a step 502 is reached and a shift clock SCL is generated in order to shift the pixel data loaded in the shift register via the serial data output to the respective drive unit 11, 12. Subsequently, in step 503, a latch signal is generated and output to the driver units 11, 12. Then, in step 504, the print signals Print1, Print2 are generated and output to the driver units 11, 12, and in step 505, the subroutine 500 is stopped.
In each data string exists a first and second number of data words, respectively, containing binary pixel data for a first and a second inkjet printhead 1, 2, respectively. Each inkjet printhead 1, 2 prints one-half of each print image column, with odd-numbered pixels on at least one half of a first print image column and even-numbered pixels on at least one half through the first and second nozzle rows of each inkjet printhead a second printed image column to be printed. The first or second number of data words in the data string respectively contain the binary pixel data for both nozzle rows of the first and second inkjet printhead, wherein in each data word of each data string only the first and second pixel data are contained for printing a first or further print image column, so that one of the print image columns is completely printed only after printing out the pixel data, for example, three data strings or at least one further data string. From each data string, the row of nozzles respectively in the transport direction in the first direction is supplied with the binary pixel data for the odd-numbered pixels of the first print image column, while the nozzle row in the transport direction in the second direction is already supplied with the binary pixel data for the even-numbered pixels the subsequent further print image column is supplied. Each print image column half is printed by the first and second nozzle rows of each ink jet print head, each print image column half being completed in time after printing with the second nozzle row by printing with the respective first nozzle row.
The first print image column is thus printed at a distance from the second print image column in the transport direction, with both print image columns being further in some print head types and very close to each other in other types. To increase the horizontal print image resolution, in particular to 600 dpi, it is provided that within the distance further print image columns are in the transport direction. This increases accordingly the number U of data strings which are stored in the pixel memory for a print image. If the parameters Z and Y stored in the printer controller 45 and in the address generator 44 are greater than those mentioned in the aforementioned embodiment, discontinuation of current printing cycles will only take place after an even greater number of consecutive encoder pulses have occurred with a reduced time interval , It must then be a larger offset - as shown - accepted. Pixels associated with a print column would then also be printed across a correspondingly larger number of print columns.

The blocks shown in the block diagrams of Figures 2, 3, 4, 8 and 9 are in the specific embodiment in a positive digital logic executed. Since the type of logic is basically arbitrary, there is one Variety of suitable design variants. The hardware realization is in a ansich known manner, for example by an ASIC or advantageous by a FPGA (Field Programmable Gate Array) possible. The Invention is not limited to the present embodiment. So Obviously, other other embodiments of the invention may be developed or are used, the same basic idea of the Starting from the invention, which comprises the appended claims become.

Claims (10)

1. A method for reducing printer errors during the printing on moving postal items in a postal-item processing apparatus, characterized by the following steps:

   a) Generation of encoder pulses according to the relative movement between printer and printing matter;

   b) Counting-up and -down for analyzing a condition with a reduced time interval adjacent encoder pulses;

   c) Performance of direct memory accesses (DMA) for a data string of binary pixel data; and

   d) Performance of a printing cycle for the above-mentioned data string, wherein, during the printing cycle, further direct memory accesses (DMA) are made for a next data string and wherein, after completing the direct memory accesses (DMA) for the next data string and depending on the time interval of the encoder pulses of a number of successive encoder pulses, the printing cycle is completely performed as long as the mean value of the encoder period does not fall below the preset duration of a printing cycle or wherein, in case of a reduced time interval the encoder pulses of the number of encoder pulses, the performance of the printing cycle for the printing of binary pixel data of a preceding data string is aborted early.
2. A method according to claim 1, characterized in that, depending on a reduced time interval of the encoder pulses of a number of encoder pulses, the interval of the printing cycles is reduced until the subsequent printing cycle for the printing of binary pixel data of a subsequent data string follows immediately after a completed printing cycle for the printing of binary pixel data of a preceding data string, wherein the reduction is made to such an extent as the encoder pulses are ahead of the printing cycles.
3. A method according to claim 1, characterized in that, for the non-reduced time interval of the encoder pulses, the printing cycle is set to a predefined period between two positive encoder edges in such a manner that the individual printing cycles are distanced from one another and that, for detecting a reduced time interval of the encoder pulses, a count value is incremented upon every encoder pulse and the count value is decremented upon every start of a printing cycle and that, on the one hand subject to a condition and on the other hand upon a count value that exceeds a predefined set-point (Z), the current printing cycle is aborted early and a new printing cycle is started, said condition being that all direct memory accesses (DMA) to the pixel memory (7) that prepare the next printing cycle have been completed.
4. An arrangement for reducing printing errors during the printing in a postal-item processing apparatus comprising a print-data control unit (4) for processing pixel data during the printing on a moving postal item with at least one print head, said print-data control unit being connected via a bus (5) with at least one microprocessor (6), with a pixel memory (7) and with an encoder (3), characterized in that the print-data control unit (4) has an evaluation unit (442 or 453) and a logic (452) for the reduction of printer errors that encoder pulses (e) are fed to and that includes a resettable encoder clock pulse counter (4523) the count value of which is incremented within a counting range of less than to equal to an upper limit value (Y) with every leading edge of an encoder pulse, the count value of which is decremented with every printing cycle start and that supplies, at the output side, a digital count value (ENC), the latter being analyzed by the evaluation unit (442 or 453) as to an exceeding of a set-point (Z), wherein, when the set-point (Z) is not exceeded, every printing cycle is completely performed and, when the set-point (Z) is exceeded, the printing cycle is aborted subject to the condition that all direct memory accesses (DMA) to the pixel memory (7) preparing the next printing cycle have been completed.
5. An arrangement according to claim 4, characterized in that the print-data control unit (4) comprises a DMA control (43), an address generator (44), a printer control (45) and the logic (452) for the error reduction of printer errors that encoder pulses (e) are fed to and that, at the output side, supplies a digital count value (ENC) to an address generator (44) and that said address generator (44) has an evaluation unit (442) for aborting a running printing cycle, said evaluation unit (442) comprising a comparator (4421) and a register (4422) for storing the set-point (Z).
6. An arrangement according to claims 4 to 5, characterized in that the logic (452) for the error reduction of printer errors includes a resettable encoder clock pulse counter (4523) the count value of which is incremented within a counting range of less than to equal to an upper limit value (Y) with every leading edge of an encoder pulse, the count value of which is decremented with every printing cycle start and the outputs of which are connected to a first comparator (4524) that prevents a decrementing upon reaching the zero count-value and to a second comparator (4525) that checks for reaching the upper limit value (Y) and, at its output side, supplies the respective count-value and that is in an operating connection with the address generator (44), wherein the latter aborts the running printing cycle when all pixel data for the next printing cycle have already been loaded via direct memory accesses (DMA) into an intermediate memory of a pixel-data processing unit (41, 42) and when the count-value of the encoder clock pulse counter that is within the counting range exceeds a preset set-point (Z).
7. An arrangement according to claims 4 to 6, characterized in that the logic (452) for the error reduction of printer errors includes a first and a second AND gate (4521, 4522) the outputs of which are connected to the inputs of a clock pulse counter (4523); that the clock pulse counter (4523) has a first input (CLK_down), a second input (CLK_up) and a third input (ENC_RESET) for a reset signal (encoder counter_reset); that a reduction signal (DEC-ENC) is fed through a first input of the AND gate (4521) to the second input of which the output (K) of the first digital comparator (4524) is connected; that an encoder signal (e) is serially fed through a first input of the second AND gate (4522) to the second input of which the output (L) of the second digital comparator (4525) is connected; that at the respective first inputs of the first and second digital comparator (4524, 4525), there is applied the count-value that is issued at the outputs (Q1, Q2, Q3, Q4) of the clock pulse counter (4523) as an n-bit broad binary value; and that, at the second inputs of the first digital comparator (4524), there is applied a comparative value "zero" and, at the second inputs of the second digital comparator (4525), there is applied an upper set-point (Y) as an n-bit broad data signal.
8. An arrangement according to claims 4 to 7, characterized in that the clock pulse counter (4523) is designed as an n-bit counter.
9. An arrangement according to claims 4 to 7, characterized in that, through a first block (451) of the printer control (45) of the print-data control unit (4), there are supplied a filtered encoder signal (e), the reset signal (encoder counter_reset) and the reduction signal (DEC-ENC) as well as the upper set-point (Y).
10. An arrangement according to claims 4 to 9, characterized in that the print-data control unit (4) is designed as a specific application-related circuit or as a programmable logic.

Images