Control System for Computer Controlled Identification of Bottles

The present invention relates to a system for automatically identifying the mold source of a bottle and providing a running history of the identity and status of an entire procession of bottles.

It is known in the art to automatically identify the contents or source of production of a container by a variety of surface markings placed on the exterior thereof. For example, nonreflective stripes may be placed on the exterior of a metal can to connote the contents of the can, as described in U.S. Pat. No. 2,647,799. A plurality of optical source-detector pairs is used to identify the markings on the can to provide an indication of its contents. Also, it is known to inspect a transparent bottle by forming projections on the bottom of the bottle as described in U.S. Pat. No. 3,301,396. The projections are counted by a plurality of mechanical sensing arms. The total number of projections identifies the mold source of the bottle.

Such identification systems are unduly time-consuming and require a multiplicity of inspection equipment. The disadvantages of such systems are dicussed in greater detail in application Ser. No. 467,702 entitled "Method And Apparatus For Identifying A Bottle" filed May 6, 1974 assigned to the assignee herein and incorporated herein by reference. In application Ser. No. 467,702, there is described a method and apparatus for identifying a transparent bottle provided with one or more sectors spanning a circumferential zone with prism-like timing marks thereon using a laser beam and sensor. The sensor generates a series of digital pulses in response to the passage of the laser beam through the timing marks. The inspection technique is simple, fast and reliable and requires a minimal amount of inspection equipment. Only a single inspection station is required to identify each bottle in an entire procession. The mold source of the bottle is rapidly identified by a hard-wired electronic control circuit described in the application. The circuit automatically computes the mold number of a bottle and compares the computed mold number with predetermined mold numbers corresponding to molds known to be producing defective bottles. The circuit activates an ejector mechanism for ejecting the bottle under inspection when the computed mold number matches any of the stored mold numbers.

A hard-wired control system is relatively expensive and inflexible compared to a programmable computer control system for the same purpose, see Callin and Baskin, Microcompressor System Design, Digital Design (February 1975), Hernandez, Mini/Micro I/O Interfaces, Digital Design (April 1975), and Davidoro, Microcomputers--What They Mean to Your Company, Intel Application Note (1975). Thus, the number of logical functions which can be performed by the hard-wired circuit is inherently limited whereas the programmable computer control system offers increased system capability. For example, in the hard-wired circuit disclosed in application Ser. No. 467,702, the computed mold number for each bottle cannot be stored for later use. As a result, it is not possible to provide a history of the identity and status of more than one inspected bottle at a time. Stated otherwise, the information output and control capacity of the hard-wired circuit is relatively limited. Moreover, any modification to a programmable computer control can be effected simply and quickly by modifying existing programs whereas modification of hard-wired systems is more cumbersome and slow.

Briefly, the present invention is directed to a computer controlled system for identifying a bottle provided with one or more sectors spanning a circumferential zone with prism-like timing marks thereon based on the passage of a laser beam through each of the sectors and the generation of a series of digital pulses by a sensor in response thereto, including generating a series of digital clock pulses; time stamping the leading and trailing edges of the sensor pulses with the clock pulses and storing the time stamp information; digitally computing the pulse widths and inter-pulse spacings of the sensor pulses in terms of the time stamp information; digitally comparing the spans of the sectors in terms of the time stamp information with the inter-pulse spacings of the sensor pulses; and digitally weighting the sensor pulses to identify the mold source of the bottle based on the comparison of the spans of the sectors with the inter-pulse spacings of the sensor pulses.

A primary advantage of the present invention is its versatility in providing a large number of logic functions with reduced hardware requirements by the use of software programs.

Another advantage of the present invention is its compactness and reliability.

A further advantage of the present invention is the provision of a running log of the identity and status of each bottle in an entire procession of bottles.

A still further advantage of the present invention is the collection and storage of data for a plurality of bottle inspections and the analysis and collation thereof for immediate use by the operator.

For the purpose of illustrating the invention, there is shown in the drawings a form which is presently preferred; it being understood, however, that this invention is not limited to the precise arrangements and instrumentalities shown.

FIG. 1 is a bottom view of the bottom of a bottle having one or more sectors spanning a circumferential zone with prism-like timing marks thereon.

FIG. 2 is a block diagram of the present invention.

FIG. 3 is a diagram of certain waveforms generated in the present invention for the distribution of timing marks shown in FIG. 1.

FIG. 4 is an electrical schematic of the signal processing circuit shown in FIG. 2.

FIG. 5 is an electrical schematic of the noise immunity circuit shown in FIG. 4.

FIG. 6 is a flow diagram of the Console program in the present invention.

FIG. 7A is a flow diagram of the initial portion of the Interrupt program in the present invention.

FIG. 7B is a flow diagram of the Time Stamp portion of the program in the present invention.

FIG. 7C-7F is a flow diagram of the Computation portion of the program in the present invention.

FIG. 7G is a flow diagram of the Output portion of the program in the present invention.

FIG. 8 is a schematic diagram of the general architecture of the microcomputer system shown in FIG. 2.

Referring to the drawings, wherein like numerals indicate like elements, there is shown in FIG. 1 the bottom of a bottle provided with sectors spanning a circumferential zone thereof. The bottle is transported through an inspection area by a linear conveyor (not shown) as described in U.S. Pat. No. 3,557,950 for "Photo-Electric Crack Detector For Glass Bottles" assigned to the assignee herein and incorporated herein by reference. There are fourteen sectors of uniform widths positioned about a circumferential zone of the bottom of the bottle, see FIG. 1. The sectors are provided with prism-like timing marks 100 which project outwardly from the bottom of the bottle. The timing marks are prismshaped to bend a laser beam directed at the circumferential zone toward a sensor 10, shown in block form in FIG. 2, as described in detail in application Ser. No. 467,702 entitled "Method And Apparatus For Identifying A Bottle" assigned to the assignee herein. The S and T sectors are provided with two or more closely spaced timing marks whereas each of the sectors located between the S and T sectors, hereinafter the "intermediate sectors," are provided with no timing marks or at most one timing mark. The timing marks 100 are distributed in the intermediate sectors symmetrically with respect to a diameter connecting the S and T sectors. Thus, in traversing the circumferential zone from either the S to the T sector or the T to the S sector, the same sequence of timing marks and blank sectors is observed.

The distribution of the timing marks 100 in the intermediate sectors in the circumferential zone as shown in FIG. 1 represents a digital binary number which is equivalent to the decimal number of the mold which produced the bottle. The absence of the timing mark in an intermediate sector connotes a binary "0" while the presence of a timing mark in a sector connotes a binary "1". By changing the distribution of the timing marks in the intermediate sectors, a multiplicity of binary numbers may be defined as explained in greater detail in application Ser. No. 467,702.

The sensor 10, FIG. 2, is located in proximity to the bottle as it is being conveyed to the inspection area so that a laser beam bent by a timing mark will be incident on the sensor whereas a laser beam passing through a sector without a timing mark will miss the sensor. A particular configuration of the laser, bottle and sensor is discussed in detail in application Ser. No. 467,702.

In response to the passage of the laser beam through a timing mark 100, sensor 10 generates a digital pulse, indicating a binary "1". Sensor 10 generates no signal when the laser beam passes through an intermediate sector having no timing mark therein, indicating a binary "0". Each intermediate sector is weighted according to a binary code so that the sequence in which the laser beam is detected by sensor 10 determines a binary number which may be decoded into its decimal equivalent as will be described in greater detail herein below.

Referring to FIG. 2, the system disclosed herein includes an activate control 12 and a signal processing circuit 18, each of which are hard-wired. The activate control 12 comprises one or more cams and relays for providing an indication that the bottle is being rotated in the inspection area, that is, that an inspection run is beginning or ending. The activate control 12 is described in greater detail in U.S. Pat. No. 3,557,950. The signals generated by activate control 12 and sensor 10 are processed by signal processing circuit 18. Signal processing circuit 18 provides the input signals which are operated on by programmed microcomputer 22 which is the heart of the system. In the preferred embodiment herein microcomputer 22 is an Intel MCS-80 microcomputer system using an Intel 8080 microprocessor or central processing unit for performing the arithmetic and control operations described hereinafter. The micro-processor includes typical computer components such as an instruction register, program counter, flag flip-flops and an arithmetic logic circuit. The microcomputer 22, includes, in addition to the micro-processor, an input/output section such as the Intel 8212 8bit input/output port, a memory such as the Intel 8100 series random access memory, a master clock for synchronizing the computer operations such as the Intel 8224 clock generator and driver, and a PL/M software compiler for programming the microcomputer. The composition of the microcomputer is well-known and is described in the literature. For example, a detailed description may be found in the "Intel 8080 Microcomputer Systems Manual" (January 1975). A preferred architecture is generally shown in FIG. 8. Other arrangements will be obvious to a person of ordinary skill in the microcomputer art, given the disclosure herein.

The program for operating the microcomputer 22 may be inputted to the microcomputer by a standard model ASR 33 teletype (TTY) and punched tape, FIG. 2. Alternatively, the program may be stored in a read only memory portion of the microcomputer such as the Intel 8300 series MOS read only memory. The program includes a main portion, hereinafter the "Console program," and an externally activated portion, hereinafter the "Interrupt program." The Console program controls the operation of microcomputer 22 in conjunction with the teletype (TTY) keyboard and printer 23, FIG. 2. Thus, information can be inputted to the input/output section of the microcomputer by means of the teletype keyboard and information stored in the microcomputer can be printed out by the teletype.

Referring to FIGS. 2 and 3, the clock is free-running and emits a high frequency train of clock pulses which are used to time stamp or locate the leading and trailing edges of the signal processing circuit output 20 for use by the microcomputer as will be described in further detail herein below. The signal processing circuit output 20 consists of three digital signals: Data signal 24, Interrupt signal 26 and Activate signal 28, see FIG. 4. The data signal 24 is generated in response to the output of sensor 10. The clock pulses, hereinafter "time stamps" for the Data signal 24 are shown in FIG. 3 for a sample run using a bottle having the distribution of timing marks 100 shown in FIG. 1. The time stamps correspond to the numbers of the clock pulses generated internally by the clock in microcomputer 22. Based on the time stamps for the leading and trailing edges of the pulses in signal 24, the center lines of the pulses are computed to coincide with the clock pulses numbered 980, 1090, 1485, 1685, 2360 and 2480. The S and T sectors are located by comparing the inter-pulse spacings of the pulses in signal 24. Thus, the timing marks 100 in the S and T sectors, FIG. 1, are closely spaced as compared to the distance between timing marks in consecutive intermediate sectors. For the sample run recorded in FIG. 3, the center lines of the pulses in signal 24 due to the timing marks in the S and T sectors are 110 (1090-980) and 120 (2480-2360) clock pulses apart whereas the center lines of consecutive pulses due to the timing marks in the intermediate sectors are 200 (1685-1485) clock pulses apart. The S and T sector pulses, therefore, can always be located by comparing their inter-pulse spacings, in terms of the time stamps, to the inter-pulse spacing of consecutive pulses from the intermediate sectors. The particular sequence of computations performed in discriminating between S and T sector pulses and the intermediate sector pulses is described more fully herein below with respect to the computer program disclosed in flow diagram form in FIGS. 6 and 7A-7G.

Referring to FIG. 6, the main or Console program is entered through the START terminal when power is initially applied to the system. Control passes immediately to the INITIALIZATION routine wherein the random access memories are confirmed as empty and information located in a paper tape or read only memory, such as the Interrupt program described hereinafter, is transferred to the appropriate locations in the random access memories. The INITIALIZATION routine is not per se the subject of the invention and would be obvious to effect by one of ordinary skill in the microcomputer art given the disclosure herein.

Control is transferred from the INITIALIZATION routine through Junction 1 to the FLAG CONSOLE step in which a Flag Console flip-flop in the central processor unit is inspected to determine whether information is available from keyboard 23 of a standard ASR-33 tletype terminal (TTY) connected to the input/output section of the system, see FIG. 2. The Flag Console flip-flop is set by the keyboard operator by depressing a key on keyboard 23. If the Flag Console flip-flop has not been set, then the program returns to Junction 1 and the flip-flop is again inspected. Once information is available at the keyboard, the Flag Console flip-flop is set and control passes through Junction 12 to the READ CONSOLE step.

In the READ CONSOLE step, the microcomputer input/output ports connected to the keyboard are inspected to determine the nature of the command inputted at the keyboard. Control first passes to the IF CHAR=R branching step. In the preferred embodiment, if the keyboard character R has been depressed, control passes to the INPUT MOLD NO. terminal wherein the keyboard operator manually inputs the numbers of the molds which are known to be producing defective bottles. Control then passes to the STORE MOLD NO. step wherein the collection of mold numbers previously inputted through the keyboard and stored in the microcomputer memory is updated by the current input. The program then enters the PRINT MOLD FILE terminal which causes the teletype printer 23 to print out the updated file of stored mold numbers. Control is then transferred to Junction 1, the Flag Console flip-flop is again inspected, and the above steps are repeated.

If the keyboard character R has not been depressed, control passes to the IF CHAR=P branching step. If the character P has been depressed, control passes to the INPUT MOLD NO. terminal wherein the numbers of molds which in the past produced defective bottles but which are now producing acceptable bottles is inputted through the keyboard to the microcomputer input/output section. The program then enters the CLEAR MOLD NO. step in which the inputted mold numbers are erased from the microcomputer memory. Control then passes to the PRINT MOLD FILE terminal which causes the teletype printer 23 to print the updated file of mold numbers. The program then returns to Junction 1.

If neither keyboard character R nor P has been depressed, the program enters the IF CHAR=S branching terminal. If character S has been depressed, control passes directly to the PRINT MOLD FILE step and the file of mold numbers in the microcomputer memory is printed without any updating. The program then returns to Junction 1.

If keyboard characters R, P and S have not been depressed, control passes to the IF CHAR=D branching step. If character D has been depressed, the program enters the INPUT DATE terminal wherein the current date is inputted by the keyboard operator to the microcomputer input/output section. This date is printed out on all reports printed by teletype printer 23. Control then passes to the STORE DATE step wherein the inputted date is stored in the microcomputer memory. The program then returns to Junction 1.

If keyboard characters R, P, S and D have not been depressed, the program enters branching step IF CHAR=A. If character A has been depressed, the program enters the PRINT REPORT terminal wherein all information stored in the microcomputer memory which relates to the quantity of bottles rejected and the mold numbers of those bottles is printed out by the teletype. In particular, the report will include both the quantities of bottles rejected and accepted, the quantity of bottles rejected as a result of the file of mold numbers in the microcomputer memory, the quantity of bottles rejected as a result of on-line detection as will be explained more fully hereinafter in connection with the Interrupt program, and the number of bottles which were inspected but for which no mold numbers could be identified. The program then returns to Junction 1.

If keyboard characters R, P, S, D and A have not been depressed, control passes to the IF CHAR=T branching step. If character T has been depressed, the program enters the SET TIME STAMP FLAG terminal wherein the Time Stamp Flag flip-flop in the central processor unit is set. The Time Stamp Flag flip-flop is used for diagnostic purposes as will be seen in the description of the Interrupt program following. Control then passes to the PRINT TIME STAMPS terminal which causes the keyboard to print all time stamps at which signals 20, FIG. 2, have been received by the central processor unit. The function of the time stamps will be described more fully hereinafter in connection with the Interrupt program. The program then returns to Junction 1.

If keyboard character R, P, S, D, A and T have not been depressed, control passes to the IF CHAR=C branching step. If keyboard character C has been depressed, the program enters the RESET TIME STAMP FLAG step wherein the Time Stamp Flag flip-flop is reset. The program then returns to Junction 1.

If the Flag Console flip-flop has been set but none of the above keyboard characters are depressed, the program enters the PRINT ERROR step in which the keyboard is caused to print an error indication, such as the word "Error," to indicate that the Flag Console flip-flop had been set to indicate that information was available at the keyboard but that an invalid key had in fact been depressed.

At any time during execution of the Console program, Interrupt signal 26 may be received at the microcomputer input/output section. The Interrupt signal 26 is generated at the beginning of an inspection run and at specific intervals therein by the circuit shown in FIG. 4. Sensor 10 and activate control 12 provide signals 14 and 16 to the input of signal processing circuit 18. Signal processing circuit 18 processes signals 14 and 16 and provides external data to microcomputer 22 in the form of signals 20. The microcomputer enters the Interrupt program in response to the signals 20 as will be described in detail hereinbelow.

The activate control 12 generates signal 16 in response to the rotation of a main shaft which drives the bottle detecting mechanism described in U.S. Pat. No. 3,557,950. Activate control 12 includes a sensor-arming cam fixed on the main shaft, a microswitch and a relay, such as the combination described in U.S. Pat. No. 3,557,950 at column 3, lines 16-26. The activate control output 16 indicates that an inspection run is beginning.

Sensor 10 generates signal 14 in response to the rotation of the bottle under inspection as described in application Ser. No. 467,702. For the distribution of timing marks 100 on the bottle bottom shown in FIG. 1, the signal 14 will take the form of a series of pulses as shown in FIG. 3. These pulses are shaped by a conventional pulse shaping circuit 30. The output of the pulse shaping circuit is fed to optical isolator 32 and then to noise immunity circuit 34. Optical isolator 32 may be any optical isolator circuit known in the art suitable for providing the appropriate voltage levels for operation of the digital circuitry in signal processing circuit 18. For example, optical isolator 32 may be a Monsanto MCT-2 isolator circuit. Noise immunity circuit 34 is shown in greater detail in FIG. 5. The circuit includes a one-shot circuit 36 and NAND gate 38. One-shot circuit 36 may be a positive-edge triggered one-shot such as a Texas Instruments 74123. The output of the one-shot is taken off its complementary output terminal and is fed to NAND gate 38. The input of one-shot circuit 36 is fed through the RC circuit 40 to the second input of NAND gate 38. The noise immunity circuit 34 cuts off the rising edge of the input signal to one shot circuit 36 and produces an inverted pulse of slightly shorter duration at the output of NAND gate 38. RC circuit 40 ensures that the NAND gate 38 is insensitive to noise pulses which may trigger one-shot circuit 36.

The output of noise immunity circuit 34 is fed to NAND gate 42 which, along with one-shot circuit 44, supplies the Interrupt signal 26. The output is also fed to driver circuit 46 which may be any suitable transistor driver circuit known in the art. Driver circuit 46 supplies the Data signal 24.

The signal 16 generated by activate control 12 will take the form shown in FIG. 3. The signal is fed to optical isolator circuit 48 and noise immunity circuit 50, which are identical to circuits 32 and 34, respectively. The output of noise immunity circuit 50 is fed to one-shot circuit 52 which may be a negative-edge triggered one-shot such as a Texas Instruments 74123. The output of the one shot is taken off its complementary output terminal and is fed to NAND gate 42. The output of NAND gate 42, therefore, will include a pulse corresponding to the leading edge of signal 16 and a series of pulses corresponding to signal 14 pulses. This output is fed to one shot circuit 54 which may be a positive-edge triggered one shot such as a Texas Instruments 74123. The complementary output of the one shot is the Interrupt signal 26.

The output of noise immunity circuit 50 is also fed to driver circuit 56 which is identical to driver circuit 46. Driver circuit 56 supplies the Active signal 28.

In view of the above, it is seen that the microcomputer 22 will receive information as to the signal 14 generated by sensor 10 by means of Data signal 24; it will receive information as to the signal 16 generated by activate control 12 by means of Activate signal 28; and it will receive information as to the sequence in which the Activate signal 28 and Data signal 24 are generated by means of Interrupt signal 26.

Referring to the flow diagram for the Interrupt program, see FIG. 7A, the system enters the START terminal of the Interrupt program when an Interrupt signal 26 is generated at the input/output section of the microcomputer 22. The particular Console program instruction being executed by the computer when the Interrupt signal is generated is completed and the program counter in the central processing unit is saved at the next instruction. Additionally, a call is made for the Interrupt program. In the preferred embodiment herein, the Interrupt program is stored at a predetermined location in the microcomputer memory. Alternatively, the Interrupt program may be stored in a time shared computer memory associated with teletype 23 or it may be stored in a paper tape or other storage medium. These operations are indicated generally by the START terminal in the Interrupt program, FIG. 7 A.

The program then enters the STORE STATUS step wherein the central processor unit registers are saved at the Console program instruction just completed. Thus, if the Console program had been interrupted during the PRINT MOLD FILE step, FIG. 6, the status of the registers would be saved in the microcomputer memory so that execution may be re-started at this step.

Control is then passed to the READ INPUT PORT step wherein the input port to the input/output section at which the Activate signal appears, hereinafter the "Activate input" is inspected. The program then enters the ACTIVATE ON branching step. If no signal is present at the Activate input, this indicates that the Interrupt signal was due to noise for the output of sensor 10 while the system was not ready for an inspection run. Accordingly, the program enters the RESTORE STATUS terminal wherein the central processor unit registers are returned to their status at the last Console program instruction. Control then passes to the RETURN TO CONSOLE step wherein the system returns to the next Console program instruction to be executed.

If a signal is present at the Activate input, this indicates that the Interrupt signal 26 was valid, that is, that an inspection run is taking place. Accordingly, the program enters Junction 2, FIGS. 7A and 7B. Control then passes to the FLAG ACTIVATE step, FIG. 7B, wherein the Activate Flag flip-flop in the central processor unit is inspected. As will be described hereinafter, the Activate Flag flip-flop is set after the INITIALIZATION routine in Branch I is entered. If the flip-flop is set, the program branches to Branch II and control passes to the DATA ON terminal wherein the microcomputer inspects the input port to the input/output section at which the Data signal 24 appears, hereinafter the "data input."

If Data signal 24 is not present at the Data input, this indicates that the sensor is malfunctioning or that no bottle is present during the inspection run. Accordingly, the program jumps to the RESTORE STATUS step which restores the central processor unit registers to their status in the Console program as saved in the microcomputer memory. The system then returns to the next instruction in the Console program. The system will return to the Interrupt program upon receiving a new Interrupt signal at the input/output section. If, however, a signal is present at the Data input the program enters the NO. OF DATA SIGNALS = 24 branching step. In the preferred embodiment, the maximum number of timing marks 100 on the bottom of the bottle is 16. Thus, there can be two timing marks in each of the S and T sectors and one timing mark in each of the twelve intermediate sectors, see FIG. 1. Assuming that a bottle is rotated 11/2 times during the inspection run, the maximum number of Data signals per inspection run is therefore 24. For each Data signal pulse generated during the inspection run, a pulse will appear at the input port to the input/output section to which the Interrupt signal 26 is fed, hereinafter the "Interrupt input," since the Activate signal 28 is on throughout the run, see FIG. 3. Accordingly, a count of the number of Interrupt signal pulses in a run, following the DATA ON step, is identical to a count of the number of Data signal pulses during the same run. The numbers of Interrupt signal pulses received at the input/output section is counted by the Interrupt counter in the central processing unit.

If the bottle has not yet been rotated 11/2 times, the number of Data signal pulses will be less than 24. Accordingly, the program will branch to the INCREMENT INTERRUPT COUNTER BY ONE step. For the first Interrupt signal pulse (corresponding to the first Data signal pulse) the Interrupt counter is incremented from zero to one, for the second Interrupt signal pulse the counter is incremented from one to two, and so forth.

Control then passes to the TIME STAMP DATA SIGNAL LEADING EDGE terminal wherein the clock count at the leading edge of Data signal pulse is stored in the microcomputer memory. The program then enters the READ INPUT PORT step wherein the Data input is inspected to determine whether the Data signal pulse is still present. Control then passes to the DATA ON branching step. If the Data signal pulse is still present, the program enters a loop inspecting the Data input until the DATA ON step determines that the Data signal pulse has ended. The program then enters the TIME STAMP DATA SIGNAL TRAILING EDGE terminal wherein the clock count at the trailing edge of the Data signal pulse is stored in the microcomputer memory.

The system then enters the RESTORE STATUS step and returns to the Activate Loop, Branch I. The system repeats the operations described upon the occurrence of another Interrupt signal during the inspection run.

If the Activate Flag flip-flop has been set but no Data signal is present at the Data input or if the maximum number (for example, 24) of Data signal pulses has been received (indicating that all useful data has been collected), the program skips to the RESTORE STATUS step and then returns to the Activate Loop, Branch I. The next Interrupt signal takes the system back into Branch II wherein the Data input will again be inspected.

The program will not enter Branch II if the Activate Flag flip-flop is not set. After the Activate signal first goes on, the Activate Flag flip-flop remains reset, in its condition immediately after the end of the preceding run, and the system enters an INITIALIZATION routine in which the Interrupt counter in the central processing unit and the Mold Number Display 25, FIG. 2, will be cleared and the Activate Flag flip-flop set. Thereafter, the system enters the TIME STAMP ACTIVATE SIGNAL LEADING EDGE step in Branch I. In the TIME STAMP ACTIVATE SIGNAL LEADING EDGE step, the clock count at the leading edge of the Activate signal is stored in the microcomputer memory. The program then enters the READ INPUT PORT step wherein the Activate input is inspected for the Activate signal. Control then passes to the ACTIVATE ON branching step. If the Activate signal is still present at the input/output section, the system enters a loop in which the READ INPUT PORT and ACTIVATE ON steps are repeated until the next Interrupt signal is received or the Activate signal ends.

When the next Interrupt signal is received, the system leaves the Activate Loop and returns to the START terminal of the Interrupt program, see FIG. 7A, and enters Branch II of FIG. 7B.

Once the Activate signal ends, the system enters the STATUS ON branching step wherein it is determined whether a Status signal 61, FIG. 2, is present at the Status input to the input/output section of the microcomputer. The presence of a Status signal indicates that the bottle under inspection should be rejected. Referring to FIG. 4, the Status signal 61 is generated by signal processing circuit 18 in response to Reject signal 58. Signal 58 is generated in response to the detection of a flaw in the bottle under inspection by a set of sensors and relays as described in U.S. Pat. No. 3,557,950 at column 3, lines 72-end and column 4, lines 1-51. The Reject signal is fed to optical isolator circuit 60, noise immunity circuit 62 and driver circuit 64 which are identical to circuits 32, 34 and 46, respectively. The output of driver circuit 64 is the Status signal 61.

If a Status signal is present at the input/output section of the microcomputer, the Status Flag flip-flop in the central processor unit is set in the SET STATUS FLAG step, FIG. 7B. Control then passes to the RESET ACTIVATE FLAG step wherein the Activate Flag flip-flop is reset. Thereafter, the program will not enter Branch II until the FLAG ACTIVATE and INITIALIZATION routine terminals are passed through. After the Activate Flag flip-flop is reset, control passes to the TIME STAMP ACTIVATE SIGNAL TRAILING EDGE step wherein the clock count at the trailing edge of the Activate signal is stored in the microcomputer memory. The system then enters Junction 3.

FIGS. 7A and 7B represent the Time Stamp portion of the Interrupt program. FIGS. 7C-7F represent the Computation portion of the program. Thus, Junction 3, FIG. 7C, marks the start of the computation of the mold number of the bottle currently inspected. As previously mentioned, Junction 3 is not entered until the Activate signal goes off. The trailing edge of the Activate signal, therefore, marks the end of the inspection run as far as data gathering is concerned. The computation of the bottle mold number is based on the data gathered while the Activate signal was on for a one-half revolution of the bottle.

As shown in FIG. 7C, the Computation portion of the program begins by inspecting the contents of the Interrupt counter in the IF NO INTERRUPTS < 2 step. The number of Interrupt signal pulses counted in Branch II, FIG. 7B, is determined by the number of timing marks on the bottle bottom. In the preferred embodiment described herein, for a one-half revolution of the bottle the least number of timing marks 100 in the bottle bottom is three, corresponding to two timing marks in the S or T sector and one timing mark in an intermediate sector, see FIG. 1. The timing marks in the S or T sector are either spaced to provide two close Interrupt signal pulses or a single long Interrupt signal pulse, as compared to the briefer pulses separated by longer inter-pulse spaces which are generated by the timing marks in the intermediate sectors. Thus, assuming that the digital number 0 is not used to identify a bottle, the least number of Interrupt signal pulses logged into the Interrupt counter will be 2, corresponding to one long pulse from the S or T sector and one short pulse from an intermediate sector.

If the number of Interrupt signals counted is < 2, this indicates that the bottle has been incorrectly marked or, if correctly marked, incorrectly read. Accordingly, the program enters Junction II, FIGS. 7C and 7G. Therafter, the time stamp information is cleared from the input area of the microcomputer memory, in the CLEAR MEMORY step, FIG. 7G, and the registers are restored to their status at the last executed instruction in the Console program in the RESTORE STATUS step. The system then returns to the Console program.

If, however, the number of Interrupt signals counted is .gtoreq. 2, indicating that a mold number can be identified based on the data gathered, the program enters the DISCARD 20% OF RUN step, FIG. 7C. It has been determined that, due to non-uniforities in the Interrupt signals arising from the varying speed of rotation of the bottle when it is first being rotated for inspection, roughly 20% of the data at the beginning of the inspection run should be ignored. Thus, all Interrupt signals falling within the first 20% of the run are ignored in the DISCARD 20% OF RUN step. For example, if the leading edge of the Activate signal occurs at the 97th clock pulse and the trailing edge at the 749th, then all Interrupt signals having leading edges which occur at time stamps before 1/5 (749-97) = 130 are ignored. The data stored in the microcomputer memory corresponding to the first 20% of the run will not be used in computing the mold number.

If desired, the non-uniformities in the Interrupt signals due to the varying speed of rotation of the bottle may be compensated for by synchronizing the clock pulses to the speed of rotation of the bottle instead of using a free-running clock. Thus, an encoder 31 (shown in dashed lines in FIG. 2) can be used to generate a train of clock pulses for time stamping the Interrupt signals in synchronism with the rate of rotation of the bottles. Encoder 31 comprises a magnetic pickup and pulse genrator for sensing the rate or rotation of the motor drive shaft which drives the carriages in the bottle detecting mechanism described in U.S. Pat. No. 3,557,950. Accordingly, the pulse repetition rate of the clock pulses generated by encoder 31 will vary in synchronism with the bottle speed, thereby compensating for the non-uniformities in the Interrupt signals cuased by non-uniform speed of rotation of the bottle at the start of an inspection run.

Following the DISCARD 20% OF RUN step, control passes to the COMPUTE PULSE WIDTHS, INTER-PULSE SPACINGS AND CENTERS OF PULSES step. In this step, the remaining data in the microcomputer memory is processed to provide the width of each Interrupt signal pulse, the spacing between Interrupt signal pulses and the center of each Interrupt signal pulse. For example, in Table 1 below it is assumed that the following data is stored in the memory for 1/2 a rotation of the bottle during a run, where the numbers represent time stamps:

Based on experience, any pulse width .ltoreq. 2 time stamps can safely be ignored as a spurious signal or a valid signal which was cut-off by the trailing edge of the Activate signal.

After the above computations are made and stored in the working area of the memory, control passes to the SELECT NARROWEST PULSE step wherein the narrowest Interrupt signal pulse, the eighth in the above example, is selected. The program then enters the SUM PULSES < 3 .times. NARROWEST wherein all pulses less than 3 times as wide as the narrowest pulse are added. In the above example, this computation would reduce to adding all pulses < 3 .times. 6 = 18 counts wide:

Control then passes to the COMPUTE AVERAGE PULSE WIDTH step wherein the average width of the Interrupt pulses is computed and stored in the working area of the memory. In the above example, the computation would be:

The program then enters the SEEK S and T SECTORS step wherein the Interrupt pulses corresponding to the S and T sectors are selected based on the inter-pulse spacings data. In the preferred embodiment herein, inter-pulse spacings are compared to 2% of the difference between the trailing and leading edges of the Activate signal. If a spacing is less than this figure, it indicates the spacing between two timing marks in an S or T sector rather than the spacing between consecutive timing marks in the intermediate sectors. Thus, in the above example, 2% of the Activate signal is 7.04 counts or clock pulses. The third and fourth and the seventh and eighth Interrupt pulses are the only ones having inter-pulse spacings < 7.04 counts. These pulse pairs, therefore, mark the S and T sectors of the bottle, FIG. 1.

Control then passes to the NUMBER branching step, FIG. 7C. In this step, the number of Interrupt pulse pairs selected in the SEEK S AND T SECTORS step determines which branch the program will follow. If the number of pulse pairs selected is one, meaning that only one S or T sector has so far been located, the program enters Junction 5, FIG. 7D; if the number is none, meaning that no S or T sector has so far been located, the program enters Junction 4, FIG. 7E; if the number is two-the ideal condition meaning that both an S and T sector have been located--the program enters Junction 7, FIG. 7F. Thus, Junctions 4 and 5 represent the non-ideal conditions wherein only one or no S or T sectors could be identified in the portion of the program following Junction 3, FIG. 7C. The non-ideal conditions result from a too close spacing of the timing marks in the S or T sector. Thus, the speed at which the bottle is rotated determines a lower limit on the distance separating the two timing marks in the S or T sector. If the marks are too close, the sensor 10 will to be able to differentiate the adjacent edges of the pulses produced by the two timing marks. In other words, there will be no "zero crossing" between the two pulses and they will appear to be joined as a single long pulse. If one of the S and T sectors contains timing marks so spaced then only one sector will be selected in the SEEK S AND T SECTORS step, FIG. 7C. If both sectors, however, contain such timing marks, then neither S or T sector will be selected in the aforementioned step.

If only one pulse pair is found in the SEEK S AND T SECTORS step, FIG. 7C, the program enters Junction 5 and the SEEK SECOND SECTOR A step, FIG. 7D. In this step, a pulse following the pulse pair already identified, FIG. C, having a width > 1.5 times the average pulse width previously computed is selected as the other S or T sector. The factor of 1.5, like the 20% factor referred to above, is chosen from experience. Other factors may also be empirically derived based on the characteristics of the timing marks, the bottle speed and so forth.

The program then enters the FOUND branching step. If the second S or T sector following the one already selected cannot be found, control passes to the SEEK FOUND SECTOR B step, FIG. 7D, wherein a pulse preceding the pulse pair already identified in the SEEK S AND T SECTORS step, FIG. 7C, and having a width > 1.5 times the average pulse width previously computed is selected as the other S or T sector. Control then passes to the second FOUND step in the flow diagram. If the second S or T sector preceding the one already selected cannot be found in the SEEK SECOND SECTOR B step, the program enters Junction 6. Thus, at Junction 6 the other S or T sector still remains to be identified.

If the pulse is found in the SEEK SECOND SECTOR A step, control passes from the FOUND step to the TREAT AS END SECTOR step wherein the pulse is temporarily stored in the working area of the microcomputer memory as the sector which marks the end of the digital number on the bottle bottom. If, on the other hand, the pulse is found in the SEEK SECOND SECTOR B step, control passes from the FOUND step to the TREAT AS START SECTOR step wherein the pulse is temporarily stored in the working area of the microcomputer memory as the sector which marks the start of the digital numbers on the bottle bottom.

If a pair of Interrupt pulses having an inter-pulse spacing of < 2% of the Activate signal time cannot be found in the SEEK S AND T SECTORS step, FIG. 7C, the program enters Junction 4 as previously mentioned. From Junction 4, the program passes to the SEEK FIRST SECTOR AGAIN step, FIG. 7E. In this step the program looks for a single long Interrupt pulse rather than the Interrupt pulse pair described. Thus, a pulse is selected which has a width > 1.5 times the average pulse width. This pulse represents an S or T sector. Control then passes to the NUMBER branching step. If no such pulse is located in the SEEK FIRST SECTOR AGAIN step the run is considered a failure and the program enters the ASSIGN MOLD NUMBER = 77 step and proceeds to Junction 9. The number 77 is arbitrarily chosen to indicate that the bottle under inspection could not be read. If instead a pulse is located in the SEEK FIRST SECTOR AGAIN step the program enters Junction 6. If two such pulses were located, the program enters Junction 7 as if the ideal condition had been satisfied as in the SEEK S AND T SECTORS step, FIG. 7C.

Assuming that the program has been able to locate both S and T sectors, control passes through Junction 7, FIGS. 7C-7E, to the CALCULATE SLOPE terminal, FIG. 7F, wherein the number of degrees of rotation of the bottle per clock count is calculated by computing the center lines of each S and T sector Interrupt pulse or pulse pair, subtrating the lesser center line from the greater, and dividing the result into 180.degree.. The slope is stored in the working area of the microcomputer memory for use in the next inspection run, if so required, as will be described more fully below. In the above example, the center lines of the S and T sector Interrupt pulse pairs would be:

And the slope would be:

The program then enters Junction 8 and the CALCULATE MOLD NO. step in which the first center line, above, is subtracted from the center lines of the Interrupt pulses due to the timing marks in the intermediate sectors following, seriatim, and the result is multiplied by the computed slope and compared to the number of degrees per sector. The number of degrees per sector is a predetermined number stored in the input area of the microcomputer memory. This fixes the location of each Interrupt pulse in the intermediate sectors of the bottle bottom. Depending on the sector location, the pulse is weighted and decoded into its decimal equivalent. Thus, in the above example, the series of subtractions reduces to:

Since, in the preferred embodiment herein, the number of sectors per semi-circle of the bottom bottle is seven (six intermediate sectors in which the identifying number appears in binary and one S or T sector), the number of degrees per sector is (180.degree./7) = 25.71. Accordingly, it is found that the fifth Interrupt pulse lies in the second intermediate sector on the bottle bottom and the sixth interrupt pulse in the fifth intermediate sector. The intermediate sectors are weighted in binary code as shown in Table 2 below:

The weights are stored in the input area of the microcomputer memory. In the above example, then, the mold number is identified as 2.sup.1 + 2.sup.4 = 18 in decimal notation.

After the mold number is calculated, the program enters the MOLD NO. = 0 branching step. If the computed mold number is not 0, the program jumps to Junction 9 and the DISPLAY MOLD NO. step wherein the mold number is displayed in the Mold No. Display 25, FIG. 2, which may be any suitable digital display compatible with the microcomputer described herein. If, on the other hand, the mold number identified is equal to zero, a forbidden mold number in the preferred embodiment, the program passes to the ASSIGN MOLD NO. = 77 step wherein the mold number is arbitrarily designated as 77 to indicate that the inspection run was a failure. The number 77 is then displayed in the DISPLAY MOLD NO. step.

Referring back to FIGS. 7C-7E, if only one of the S and T sectors could be identified from the data gathered in the current inspection run, the program enters Junction 6 and the USE SLOPE FROM LAST RUN terminal, FIG. 7E, wherein the slope calculated in the last inspection run and stored in the microcomputer memory is retrieved for use in calculating the current bottle mold number. In particular, in this step the slope computed for the last run is divided into 170.degree. to find the range of time stamps within which all data with respect to the intermediate sectors must occur in a 1/2 rotation of the bottle. As already explained, the span from the beginning of an S sector to the beginning of a T sector, including the six intermediate sectors, is 180.degree.. In the preferred embodiment herein, each sector spans 25.71.degree.. Thus, a span of 170.degree. from the center line of the Interrupt pulse or pulse pair corresponding to an S or T sector will include all Interrupt pulses corresponding to the six intermediate sectors in a semi-circle on the bottle bottom. In the above example, then, all data from which the mold number can be identified is located within (170.degree./0.65) = 261.5 clock counts or time stamps. The computed range of time stamps is stored in the working area of the microcomputer memory.

After executing the USE SLOPE FROM LAST RUN step, the program enters the ARRANGE INTERRUPTS terminal wherein the Interrupt pulses corresponding to the intermediate sectors are re-arranged, based on the symmetrical placement of sectors on the bottle bottom, so that the mold number can later be calculated in the CALCULATE MOLD NO. step previously described. In the ARRANGE INTERRUPTS terminal, the following operations are executed:

1. The number of time stamps corresponding to 170.degree. is added to the center line of the only S or T sector known - this serves as an upper limit in re-locating all Interrupt pulses arising from intermediate sector timing marks preceding the S or T sector Interrupt pulse or pulse pair;

2. the number of time stamps corresponding to 180.degree. is added to the center lines of all Interrupt pulses arising from intermediate sector timing marks preceding the S or T sector pulse or pulse pair; and

3. Interrupt pulses arising from intermediate sector timing marks following the S or T sector pulse or pulse pair but not prior to the earliest re-arranged Interrupt pulse are ignored. For example, assume in Table 3 the following time stamps history for 1/2 rotation of the bottle:

The only known S or T sector is indicated by the pulse extending over 170 - 190 counts or time stamps since this is greater than 1.5 times the average pulse width which is 10 counts. There are three Interrupt pulses preceding this sector and one following it.

Using the previously calculated slope of 0.65.degree./count 170.degree. of rotation is equivalent to 255 counts. Thus, all Interrupt pulses corresponding to the intermediate sectors of interest will lie within 180 counts (center line of S or T sector shown in brackets in Table 3) to 435 counts. The Interrupt pulses preceding the S or T sector pulse are brought forward by adding 270 counts (the equivalent of 180.degree. based on a slope of 0.65.degree./count to their center lines. The re-arranged time stamp history of the Interrupt pulse is shown in Table 4 below:

The bracket number pair represents the known S or T sector and the bracketed number [435] represents the upper limit of the Interrupt pulses. These counts are then stored in the working area of the microcomputer memory.

The program then enters Junction 8 and the CALCULATE MOLD NO. step, FIG. 7F. The center line of the known S or T sector pulse or pulse pair is subtracted from the center line of each Interrupt pulse as re-arranged, the result is multiplied by the slope calculated in the previous run, and each Interrupt pulse is fixed at a particular intermediate sector on the bottle bottom, all as previously described. If the mold number identified is valid, that is, not zero, then the number is displayed digitally in the DISPLAY MOLD NO. step.

Control then passes to Junction 10 and the FLAG REJECT step, FIG. 7G, where the Reject Flag flip-flop in the central processor unit is insepected. The Reject Flag flip-flop is set by Status signal 61, FIG. 4, to indicate that the bottle currently being inspected must be rejected. The Status signal is generated by the operator by depressing the appropriate key on the teletype keyboard 23, FIG. 2. If the Reject Flag flip-flop has been set by the Status signal, the program bypasses the DISPLAY LIGHT AND EJECTOR step, and enters the ACCUMULATE REJECT COUNT BY MOLD NO. step. The Status signal activates an Ejector mechanism 29 for ejecting the bottle under inspection such as the ejector mechanism described in U.S. Pat. No. 3,557,950 at column 4, lines 11-end and column 4, lines 1-21.

Thereafter, control passes to the ACCUMULATE REJECT COUNT BY MOLD NO. step, FIG. 7G, wherein the mold number of the rejected bottle is stored in the microcomputer memory and the Reject Counter in the central processor unit is incremented by one count. The Reject Counter counts the number of rejected bottles for a series of inspection runs.

Control then passes to the FLAG TIME STAMP steps wherein the Time Stamp Flag flip-flop in the central processor unit is inspected. As previously explained, the Time Stamp Flag flip-flop indicates that a time stamp report is requested by the operator. The flip-flop is set by depressing keyboard character T, FIG. 6. If the flip-flop has been set, the program enters the STORE TIME STAMPS AND MOLD NO. IN OUTPUT AREA step. In this step, the time stamps or clock counts for the leading and trailing edges of the Interrupt pulses, as well as the leading and trailing edges of the Activate signal, together with the computed mold number, are brought to the output area of the microcomputer memory for use in printing out the time stamp report further on in the program.

The system then enters the HAVE X BOTTLES PASSED branching step where X can be any predetermined number of bottles, hence inspection runs, for which a report is to be printed. If the predetermined number of bottles has passed through the inspection area, the program passes to the PRINT AND CLEAR step. In this step the information stored in the memory output area, namely the time stamps and mold numbers for X inspection runs, is printed out and the memory working area is cleared. Control then passes to Junction 11 and the CLEAR INPUT AREA step wherein the memory input area is cleared.

The system then enters the RESTORE STATUS step. In this step the centralprocessor unit registers are restored to their status at the last executed instruction in the Console program. The system then returns to the next instruction of the Console program.

If the Time Stamp Flag flip-flop had not been set as of the FLAG TIME STAMP step, that is, if no report had been requested at that time, or alternatively if the required number of bottles, X, had not been inspected as of the HAVE X BOTTLES PASSED step, then the program skips to Junction 11, FIG. 7G, and no report is printed. Thereafter, the input area of the memory is cleared, the registers are restored and the Console program is re-entered.

If the Reject Flag flip-flop had not been set as of the FLAG REJECT step, the program enters the COMPARE COMPUTED AND STORED MOLD NOS. branching step. If the computed mold number is identical to a stored mold number in the updated memory file, the program enters the DISPLAY LIGHT AND EJECTOR step wherein a DISPLAY LIGHT 27 and Ejector Mechanism 29 are activated by the input/output section of the microcomputer, see FIG. 2.

The program then enters the ACCUMULATE REJECTED COUNT BY MOLD NO. step and the sequence of operations previously described in respect to the Reject Flag flip-flop being set is executed. If, however, the computed mold number is not identical to any of the stored mold numbers, the program enters the ACCUMULATE ACCEPT COUNT BY MOLD NO. step in which the Accept counter in the central processor unit is incremented by one count and the computed mold number is stored. The Accept counter counts the number of accepted bottles for a series of inspection runs. Thereafter, control passes to the FLAG TIME STAMP step and operation proceeds as previously described.

A program, written in PL/M and Assembler languages, for controlling the operation of the microcomputer 22 disclosed herein for on-line identification of the mold source of a bottle and storage of a running log of the identity and status of each bottle in an entire procession of bottles is shown below. The program represents one suitable set of instructions for performing the operations disclosed in the flow diagrams in FIGS. 6 and 7A-7G as indicated by the marginal headings. It should be understood that the program may be modified by persons skilled in the program art to effect additional or related operations by microcomputer 22 in identifying a bottle and providing a running log of the information computed within the scope of the present invention.

The microcomputer utilized in the present invention is an Intel MCS-80 microcomputer. Other programmable computers may, however, be substituted for the Intel MCS-80. In addition, other programming languages may be used for the programs disclosed herein without departing from the scope of the invention. ##SPC1## ##SPC2## ##SPC3##

The present invention may be embodied in other specific forms without departing from the spirit or essential attributes thereof and, accordingly, reference should be made to the appended claims, rather than to the foregoing specification as indicating the scope of the invention.