DE1421762C - Device for dividing a glass plate - Google Patents

Description

The invention relates to a device for opening such a device is much simpler share a glass plate, which randomly scattered constructed as z. B. electronic calculating machines, Has errors, in error-free programmed law, with which the problem can basically also be solved

angular parts of a given size. · Let.

It has already been proposed that glass plates 5 Appropriately, a device is provided, to be divided in such a way that in the manufacture which prevents the selection of lengths, the one the same in all of the cooling of a glass remaining length of glass from the new guide band following manufacturing phases, the error edge of which left to the end of the glass plate, the continuously according to the progress of manufacture, not uniformly established by the combination of the profestones and the values obtained in this way can be divided into programmed length. be applied to according to the dimensions of the The invention is in the following on hand schema cutting A cutting program of topical drawings of exemplary embodiments in more detail , which is explained in the course of the ongoing production. In the drawings shows Position of the glass ribbon by means of an automatic Fig. 1 shows the stages of a according to the invention Cutting machines is carried out. This pre-15 logical two-step process for staking out blow gives no teaching in which way the glass plates,

division is to be carried out, but only concerns F i g. 2 is an example of a leading part of a

the detection, recording and automatic glass plate on which a special fault course locally

Marking the mistakes. loan is given,

The invention is based on the object of providing a 20 Fig. 2A table for the example

Device of the type outlined at the beginning of the error scheme to be created, which eliminates the faulty square

fen, with which an optimal utilization of the units of FIG. 2 with regard to the local position

joinable defect-free areas of glass plates with defined,

relatively simple means is possible. 2B shows a table of the, for example, pro-This object is achieved according to the invention by means of 35 programmed sizes with degree and value, which are to be cut out of the one device for arranging the different glass panels of FIG. 2, programmed sizes in width groups with the same- a table which shows the experimental length dimensions in a given development of the plate according to FIG. 2 illustrates the result; by a device for reading home choice of a particular group of cutting commands Shaped mation relating to the presence of glass 30 measured the sizes of the programmed 2B fehlem adjacent in each row to run along _ER-,.;

stretching examination zones, which extend transversely Flg. 2D, 2E and 2F graphic representations

extend through the glass and have a length, measured by the experimental divisions, which the table of the

from the leading edge of the glass plate, Fig. 2C represents

which of the length dimension of the investigated pro- 35 Fig. 3 a graphical representation of a general Group corresponds to; by means of a device solution of a division problem, device for displaying width fields of the glass plate, FIG. 4 shows a block diagram of an area division computer intended for a glass switch that contains the faults in the sub-factory Exclude length dimension, where .this is more the subject of the further FIGS. 5 to 25, The device responds to the error reading device; FIG. 5 is a block diagram of the computer time; by a computing device, which on dial gauge,

the device and the device responds, for Fig. 6 a circuit diagram showing the electron

Determine the closest possible combinations of tube circuits of the computer clock according to FIG. 5

the latitude created by the facility,

dimensions, which in the error-free strips 45 Fig. 6A is a circuit diagram of an auxiliary time measuring circuit,

of each examined glass plate longitudinal section Fig. 7 is a circuit diagram of the stepping system for

fit in; by a device to add the computer experiment step switch,

add an amplitude value signal, which is shown in FIG. 8, a circuit diagram of the step voltage and

relative value of each programmed right-hand memory circuit used by the step switch represents angular pane size; operated by an input 50 according to Fig. 7,

direction for summing the value signals of the grain Fig. 9 is a representation of the plug-in switchboards of the

combination of the computer determined by the arithmetic unit for the Z-S size program,

Width dimensions for the length studied; Fig. 10 is an illustration of the plug-in switchboards

by a device for actuating the set-up computer value program and the associated valuation, by a different length dimension 55 voltage source,

to be checked with associated width groups, after Fig. 11 is a circuit diagram of the rear switching dic Summicrcineinrichtung a sum for a sub-panel connection between the main Z connector Length dimension created; through a switchboard and the A 'plug-in switchboard for installation Z of FIG. 9 left to store the total word signal to represent several for each length and for selecting the length- 60 Z program circuits,

dimension of the sample for which the total signal Fig. 12 is a schematic representation of a

is greatest; through a device for the storage system and a computer input

separate the selected length from the leading edge synchronizing circuit, which can be

the glass plate to create a new filling edge for use together with the computer according to FIG. 4 to create the following compliance regulations; and turn 65,

by a device for controlling the splitting. Fig. 13 is a circuit diagram of a paper tape error

ciiK'r each opened lung in width dimensions query and scanning system for mirror and window

the preferred Hrdlc which belongs to this length. glass qualities,

Fig. 14 is a broken partial representation of the FIG. 12 to illustrate the mechanical arrangement of the reading fingers with reference to FIG Tape failure recording,

15 shows a 5-division matrix of the one in FIG. 4 shown computer, with the highest-ranking 5-cell block partially in detail and the rest five 5-cell blocks are shown in block form, 15A is a circuit diagram of the six banks of the Universal S-cell switching relays, their contact pairs are in the matrix of Fig. 16,

Figure 16 is a circuit diagram of the input automation circuits for the S matrix of FIG. 15, which contains the values used by the error reading system of FIG. 14 controlled Contain voltage distribution circuits, as well as a representation of the unit residue eliminator circuits,

17 is a circuit diagram of the value register reading relays the S matrix according to FIG. 16,

Fig. 18 is a circuit diagram of an S-value-per-Z-stage adding circuit each Z step with a block diagram of a value comparison circuit,

FIG. 19 is a circuit diagram of the electron tube circuits included in the block diagram in FIG are shown

F i g. 20 is a block diagram of a Z accumulator that is converted from the decimal-to-binary converter according to FIG Fig. 11 is actuated.

Fig. 21 is a circuit diagram of the electronic pulse generator and one of the counter stages of the accumulator according to Fig. 20,

22 is a circuit diagram of a binary-decimal

Converter for the accumulator according to FIG. 20, F i g. 23 a circuit diagram of the paper tape brake

and diode starting circuitry provided by the converter of FIG. 22 is operated,

F i g. 24 a computing device, the busbars of which from the binary-decimal converter according to F i g. 22 are fed,

F i g. 25 an S-marker circuit which was used during the selected Z-computer attempt is operated,

F i g. Fig. 26 is a semi-schematic view of a cutting system for a glass factory incorporating the Calculator according to FIG. 4 to 25 contains, and

F i g. Fig. 27 explains the electrical symbols.

Since a computing system is not easily suitable for representation on circuit diagram drawings, certain agreements are made to make things easier. For example, switches for which no actuating solenoid or other actuating device can be found in the drawing are denoted by the symbol (a). Normally open and normally closed switches, which are operated by a solenoid winding shown in the drawing, are marked with symbols (b) and (c), respectively. Relay windings for operating such switches are shown as circles surrounded by a reference number according to symbol (d) . Capacitors are shown with the symbol (e).

I. Apparatus according to the invention A. Division of areas

The division of areas of plate or sheet material involves what can be referred to as two-cut logic. For a generalized Representation is referred to Fig. 1, which relates to flat glass, such as. B. plate or sheet material, relates from which error-free, prescribed sizes are to be cut.

There are three basic cutting conditions for cutting flat glass as sheet or sheet material to meet.

The board should be divided into rectangles with planned or programmed dimensions in contrast to a method in which the defective area is only cut out of the board or always cut the largest available flawless piece of glass plate will. The planned sizes are much larger than the rectangular ones with errors contaminated remnants that have to be eliminated as scrap.

Second, every cut must have one entirely

Dimension of the relevant piece of glass run.

Third, the partitioning must be selective so that the trash has a smallest surface area in accordance with the choice of sizes into which the flawless glass can be divided. Since the Defect severity influences the classification, a glass area can be free of defects with reference to a quality, however, it cannot be flawless with respect to any other quality.

For example, a sheet of glass can be a 6.35 mm thick plate that runs across as a continuous band a width of 3.05 m at a speed of approximately 5.08 m / min. A across running cut through the glass ribbon width is the first cut of each work step and is here as denotes the Z-cut ', and a group of second cuts, here called S-cuts, slit the cut Z-piece into smaller widths and eliminate the errors. This is the so-called Two-cut logic.

With reference to FIG. 1, the information required to divide the glass consists of the location information the leading front edge of the glass, in the details of the location of the imperfections and in the Program of sizes to be cut,

Various devices can be used to determine the location of the imperfections: and expediently in the form of a visual or photoelectric control of the glass plate of the to entire area to be considered before the first cut is made. This information can can be viewed as being stored directly in and on the glass, but can be more easily seen by addition be held by markers or marking devices. Preferably, however the error is stored secondarily, e.g. B. on tape or other recording device, which contains the location of each error square with respect to the leading edge and the side edge.

In terms of sizes per se, the lengths and widths of the different ones to be cut are Rectangles divided so that the corresponding width dimensions are grouped that have a common Have length dimension. All length dimensions are grouped in this way, where each is associated with one or more width dimensions. The need for the various Length-by-width sizes comes in a two part Invoice determined. First, a ranking or priority order is established for the appropriate Width dimensions that are any length dimension

5 6

are assigned, established, and then a programming is done with a view to

Numerical value developed for each quantity so that values can be set up for daily or weekly production,

from size groups collected and compared and an establishment census is expedient of

can be. Sizes made that cut, but not

In accordance with the two-cut logic, 5 are out of order so that the program

the attempts speak for the first cut which can be revised if the quantity requirements

different programmed lengths are measured.

from the front edge of the glass. Thus, for the actual time at which the cutting ereaches programmed length a trial length which follows can be delayed while the work glass plate considered, for which either the entire io sequence is determined. So a record for that Band width or parts of this for splitting as subsequent cutting of successive ones error-free fields, depending on the number of the selected lengths (with the accompanying adapted Locations of the imperfections, which are the widths) are available, each chosen affect the quality level to be considered. Each length represents the leading edge for the next. Defect is contained in an error strip, so that 15 The Z timing logic also reacts to request the programmed length is considered. longer than a basic procedural cut

Next, the programmed widths are in a Z-point in time multiple system. A larger number of options under the same Z-joinable widths according to an assigned rank and size and the same value program is fitted for this special programmed length. The scrap that mismatched parts 20 receive by having the highest aggregate value of the defect-free widths as well as the defect stripes at settlement Z in a given length with respect to themselves is selected by such a choice of the leading edge. Each additional number and type of programmed widths, which compound Z is a given number of the available error-free widths best programmed Z-values in a particular order, are reduced to a minimum. For each pro- 25 with respect to the leading edge, and its value is the programmed length and corresponding, thus subject to change by the order in which tentatively used widths are arranged the values of its proportional Z-sizes. The programmed quantities other than the value requirements stored in terms of information storage, which these trial lengths or in such a system are very high in which the cross-section is assigned. 30 substantial numbers of sizes from ZS loading

The stages aimed at fitting the programmed ones, and those specifically described here,

Widths in available error-free widths and the preferred system works on the Z point in time

The following values are saved for each pro logic *

programmed length repeated before any. An example of a void covered Cutting decision is made. A 35 glass ribbon is shown in FIG. To be useful Number of first test cuts across the width Relationship between the vacancy positions and made of the glass, making the cut for the required, missing sizes, the best fit is achieved, d. H. the cut, a basic dimension unit is established for which for which the value of the fitted programmed all length and width conditions whole multitudes is highest, is chosen. After the 4 ° are. With a belt width of 3.15 m, as in The selected section determination is made, Fig. 2, a unit of 5.08 cm was selected, The commands for cutting these are programmed giving the glass an effective width dimension Length and then for cutting the fitted of 62 units (5-dimension). Natural latitudes this programmed length. According to it, it would be the one size fits all, in case the pro can such a large number of programmed lengths 45 programmed sizes for cutting so small or test cross-sections are considered, and intervals such as 25.4 mm should change with 25.4 mm it depends on the specific errors and which one is chosen.

Cutting program depends on whether the first chosen If you use the slab or table as a mosaic Looking at the smallest or largest or one of them 50.8 mm squares, the errors in between of the programmed lengths are recorded. 50 place through the Z and ^ coordinates for the

After the local position of the new front square in which they are located, locally determined. The

Edge is systematically determined in this way, glass is back in units from the front

the process steps are repeated. This is how edge is measured along the Z dimension, and the

any programmed length in terms of the new 5 dimensions will be from one of the side edges

front edge viewed again. The flaws 55 from measured. As shown, the relate

with respect to the front edge there are of course coordinates on the outer boundaries of the squares

changes, and the available widths become themselves again. The starting boundary of the first

definitely. After re-fitting the program Z or 5 strips, it is regarded as zero. Therewith

mated widths into the available widths for each is a square Z 6.522 by Z-unit dimension

programmicrte length and storage of the value of the 60 lines 21 and 22 limited.

Sizes that can be cut will be the length as the second cross-section in the repeated through Z and 5 coordinates selected in the table according to the procedure. The latitudes for Fig. 2A are localized. The defective areas of this second cut are cut or assumed to be arbitrary, since the actual slit, and the fitting process is again with material production, their number and position in relation to the newly created front edge cannot be predicted. The flaws g _{C] e} it _e t. can be microscopic in size and much

ner than a unit square, or they can intersect multiple mosaic units, such as. B. at the defect shown d. However, in any case, each defect is considered to occupy a square. Likewise, since each square must be examined either in whole or in part according to the nature of the defects present in order to determine defect classification, the smallest unit size may be effected by the limitations of the defect detection device. Although bulbous edges on the sides of the belt are continuous, elongated imperfections, such foreseeable defects are eliminated simply by folding or trimming.

The table of Figure 2B illustrates a simplified program of sizes to be cut, listing eight sizes which can be reduced to three different length or Z dimensions, each of which has an associated group of two or three S dimensions. Thus, for any given Z-cut along the length of the sheet of glass, there may be a choice for the S-cuts. In the table of FIG. 2B, the larger S in a group of S dimensions for a particular Z ranks above each smaller S. Furthermore, each size Z · S is assigned a value which, in this case, corresponds to the 5 dimension of that size.

With this information supplied with regard to the desired sizes and the distribution of errors the usable areas are defined by the two-step logic (as shown in the table according to FIG. 2 C is specified). As a result, each Z dimension (35, 20, and 9) forms the basis of a tentative first cut, measured from the front edge of the glass. Then every flaw will measured from the reference side edge of the glass, marked as a strip the width of one unit, with certain available widths remaining.

When fitting the 5 dimensions into an available width, the division with the smallest waste is determined by using the highest-ranking 5 dimensions wherever a choice is made. If an exact division is not available, the available width is reduced by 1 and the fitting is tried again. This process is repeated until a perfect split is achieved or the error-free width is used up. The S-fits for the three Z-tests listed in FIG. 2C are shown graphically in FIGS. 2D, 2E and 2F. The crossed area is the flawed strip, and the single hatched areas are the broken glass with the flawed strip, if the Z-cut is chosen. Sufficiently large pieces of waste can be saved if necessary by setting out special or non-programmed sizes. By summing the assigned values of the quantities for each Z trial, the choice of Z by the highest storage is determined. In this case, Z is chosen to be 20, since the value is 58, so that there are only flaw strips of two units and two additional strips of one unit of non-fitting scrap glass.

It should be noted that in the example shown, a smallest scrap area per Z-unit length logic is determined by making the value proportional to S. If the value were proportional to Z · S, Z would have been selected in this example with 35 despite the larger amount of rejects. If in the 5-fit the edge is not in the order of the 5-dimension, the links of the division in the S-dimension are interrupted in favor of the smaller 5-values and, of course, cut. In the course of the production of a day, a week or a year, the reject becomes by definition

ίο minimal, since the programmed values are in the programmed Okay have been cut.

If the tape is divided into separate plates of e.g. B. 5.47 m or 90 units remaining excess lengths are preferably avoided. This is achieved by always a Z is chosen which is part of an exact fit of the various Z dimensions in the total length is. Choosing Z with 20 in the first cut leaves 70 units in the two Can fit Z lengths of 35.

Usually, by including a large number of Z's in the program, one remains large number of length options open. The number of tentative z-scores that matched the 5-ver search splits and are subject to evaluation is naturally enlarged. However, the hassle of additional choices is made by theirs still facilitated systematic solution. If the choice is likely due to less or only large programmed Z-numbers are to be limited, a small remainder can be added to the total length be tolerated.

The starting point of the two-dimensional division is a fundamental method of the one-dimensional or length division. Such a division can generally be in the form of groups of integers are expressed whose sum is exactly equal to a given integer. Splitting a large sum into selected smaller integers is like resolving a large number into prime factors, often easier to do than explain. In the formation of prime factors, the Simplification tables are used, and the results are valid regardless of whether they are determined by explicitly expressed logical steps or by arithmetic trial and error procedures are.

With regard to the discovered usefulness of the defined concept of partitioning, two types of solution are of particular interest. One type can be called the general solution, in which all possible divisions of A are recorded and evaluated by the integer a _t. The other type can be referred to as the selective solution, in which conditions for the uniqueness of the solution are imposed by _assigning rank or priority order to the various given integer a {.

General splitting is used in the Z-splitting method discussed earlier. The information required indicates whether a given a _{t is} or is not present in any of A's possible partitions (ie, whether its n _t is different from zero).

In Fig. 3 the possible fits of any or all of the integer α, values 3, 5, 8 and 12 into an integer A with the value 20 are shown graphically. This A is by means of a

109 617/18

9 ίο

ordered group of evenly spaced par system 78 localized, with the. Unitary

allele horizontal lines 0,1,2,3 ... 20 dimension of the coordinates in this case 50.8 mm

shown. The divisions are by vertical amounts. The sizes Z ■ S of the glass can be a

Link chains shown, the corresponding length minimum of 9 x 6 units up to the full

have spacing lines, the number of which can be shown - 5 plate size of 90 x 62 units. It will

differs according to 3, 5, 8 or 12. The one glass production line speed in that

Choice of any values from 3, 5, 8 or 12 as a proximity of 5.08 m / min, taken into account,

The first Z-step out of a Z-length of 20 is The decision-making takes place in the computer

thus acceptable. Operation sufficiently fast to make successive

In Fig. 3 also shows the availability of various reuse of the same computer elements with different integer α, for each of all subsequent implementation of the Z and 5 test divisions of the options. In this way, for each of the programmed Z-lengths, A will be borrowed equal to 17 if an α, of 3 is the first choice. It will encompass the many steps and all integer α remain, still as second arithmetic operations available for solving options, since terms 3, 5, 8 and 12 15 distribution problems depend on line 17 by means of general purposes. If the second determined electronic digital computer necessary is chosen O ₁ 12, a link which extends from are eliminated and sufficient speed is extended to line 17, then this is maintained for speed, albeit this relatively long third choice only a, of 5 available, since only one link is sam. The full list of reliable electrical solenoid relays of the usual division is thus 3, 12, 5, and the other possible types used. The final decision will be split easily from the table - after 20 attempts within a few seconds. reached, during which the glass moves from the transverse

It should be noted that F i g. 3 not all cutting station moved forward a distance has divisions for the integer A less than 20 35 that is less than the smallest Z size. The all-illustrative, but only such divisions, the common arrangement and the general system of the by subtraction from one or more of the computers described here can also be achieved. Thus, no reliable apportionment solutions end up, instead of calculating 30 every link on line 16 despite the fact that the interstage problem that appears is re-divided by two numerical values 8. However, the extensive memory because 16 is not a division of 20. and reading conditions that match the possible number

Although the general divisions shown in Fig. 3 are assigned to the divisions contained herein,

Solution the acceptability or future options - avoided.

freedom with respect to any selected a _t shows In Fig. 4 is a block diagram of the computer contains them which make the components of the above-described 5-dividing indefinitely computer as well as their functional correlation would be multiple choices, which system shown 35. In the mentioned ranking, in which and work sequence illustrated,
For example, a larger number always precedes, a computer clock 135 and the test step, the divisions with smaller numbers of switches 135 according to FIG. 4 steer sideways and rank highest integer a _; eliminated. If one considers the various operations during each of the highest-ranking a _t as the β, -member, which on his computer work period. The clock delivers rows at the upper time end with the line 20 of the parallel set of controlled pulses, which with the error recorder F i g. 3 is connected, the highest a _{t is} chosen from the voltage device, the glass conveyor belt and the local any of those available in an all-location of the glass plates on the conveyor belt as well as their common division of 20. This 45 consecutive leading edges, if is 12, with a next A of 8 remaining, cutting decisions are made, synchro for which the general partitions are 8, 5-3 or nized.

3-5 are. According to the rules, the step switch timed by the clock closes 8

another, so that automatically one makes a series of 20 attempts, after which one

Solution for n ₁ (a _l ) is obtained, which is 1, and is repeated for 5 ° selected as a further attempt.

/ I ₂ (fl ₂ ), which is 8. During this 21st attempt, the calculation

In the following, a computer is described, with repeated calculations for the selected Z-trial length, which uses distribution matrices. The results are stored and / or as a computer includes for glass Z · S or surface area commands are issued, and the device is divided back Z to a point in time described above, in order to set the computer period for the next intermediate logic. The computer size and 20 attempts with respect to a new advance or complexity will initiate to a large extent the leading edge. In this example, the number of choices that must be made in succession to the twenty electrical channels and the speed are energized. In ^6o voltage source for the specific sizes of values to the example shown, the positions and Gesamtanforde- provides is prepared to each of the approximately program associated with the performance of a glass factory with 20 attempts. The channel corresponding to the 20 trial Z-steps, each of which corresponds to the choice chosen Z-trial, is energized by the Verder size Z, six S sizes, one margin for each S, pace switch during the 21st attempt, two Passage quality grades (inferior quality) and the S-value of the clock, runs to the test step depending on the rank or size. The faulty switch and from this to the Z program switch are indicated by the rectangular coordinate table 137. The selected experiment becomes the step

11 12

switch supplied on the dashed line to the pro-fed back for each test step, so that the correct channel during the programmed S-lengths in the error-free clamping 21. or command step can be excited. The wide ones to fit in for the Z-length are different Connections from the clock and taking into account the level of quality are available. Step switch to elements of the other computer 5 This arithmetic unit has six hierarchical defi-component circuits are the nth cell blocks in the block diagram, around the six S sizes per Z F i g. 4, not shown for the sake of clarity, will provide and provide selective detachment for each however, in the detailed description of the construction fault-free busbar span. elememschaltungen explained. Input circuits set the size program

The Z-program switchboard 137 and the S-pro-Z-io and the request information of each step in the program switch 138 together define commands for switching and activating the computing size aspect of the computer requesting device. The size scheme for every program. A value is each programmed cell block is determined for each test step of the size Z · S in a value program switching device signals that are assigned by the 5-per-Z program 139 . This information can be output together with the switching device 138. The choice of quality or degree also completes the requirement of each step. The twenty test channels 5-busbar spans excited one after the other by the step signals from the error interrogation circuit 141 the switches 136 , to which the 5-arithmetic unit will apply voltage sources to the selected individual size and value values by programming the computer on device voltage distributors 143. A unit residue eliminator circuit 144 connected as lines. an addition of the ^ computing device programs the

In this way, the Z-program switching voltage source connections to the fault-free sampler 137 reconnects the Z-connections with the test rail spans, which are generated by the screaming channels in the desired test order. The tendency to reduce the excited span is not S-per-Z program switching device 138 connects 35 are divisible.

six 5-size circuits with each programmed A totalizing device 145 fits every 5-size

ten S. As long as there is a programmed Z variable in a division element reading of the 5 arithmetic unit 142 possible division of the available length of the split to the value voltage for this 5 quantity contained by the plate to be allocated, the step of the device 139 is supplied and adds the switch voltage to the value voltages for each step with the Z variable pro 30. As the groups programmed programmed 5-size circuits. Each of the 5 division elements and the values generated during one of the programmed S is assigned to a specific value test step in relation to the voltage level. Values are thus produced for the Z variable via the Z program switching device, the programmed variables Z · S , which are valid if every value voltage sum is a Z · erring every Z attempt is recorded. 35 Total value for a tentative Z step be-The other part of the computer input program is plus the leading edge. The summing device the error information. A failure detection and further selects the highest value sum and returns storage system 140 contains the failure position - the apportionment solution accompanying this.
coordinates for each error unit square. This means 146 for storing the total information must be related to the leading 40 value signal for each length and for selecting the or leading edge of the lens and for each new length dimension of the sample for which the total leading edge is set by each signal is largest, only records the step, selected Z-cut or a saved cutting edge while the highest value stress sum is made from the decision. Regardless of which summing device is obtained. The Z-size programmed during which memory system is used reads the error 45 of this attempt, the interrogation circuit 141 the S-coordinates of the incorrectly selected Z. This value is taken as a cutting point, which in every Z-attempt as from the previous - the decision is recorded and also measured at the edge of the glass. The Z cutting tool 147 supplied. The 5-division query signal is sent by the step switch via the information is not retained, but is supplied as Z-program switchboard 137. 50 21st step of the experiment, the corresponding high value

A particularly useful and simple Anchende Z test channel re-energized.
Okay to program more than one During the first twenty steps, the

Usable quality is achieved in that arithmetic facility 142 automatically feeds a preselected circuit and fault position information with the 5-Pro during each attempt, from separate fault memory recorder circuits accompanying the selected Z. The reading signals are queried, each of which for the recorded quality does not correlate with the value assignments that define areas that are error-free for this. but are sent to an S-marking system 148. In this way there is a memory recording switched. Here the 5 divisions are recorded as a cutting circuit for mirror quality as defects, all input instructions for the selected Z length are recorded, squares that are not relevant for mirror quality after making the Z cut, and one more Fault memory S-cut information used to record the Z-length circuit gives as flaws in the slits.

usually a smaller number of squares, which for Thus the working period is ended, and the

Window quality are not acceptable. next starts with a new front edge. As

The two parts of the computer input program, 65 determined by the two-cut logic, are assigned to each step of the experiment, ie (1) all waste is recorded in the 5-cuts. The request program and (2) the error information- According to the definition, the total waste, ie the Hon, is sent to a selective arithmetic unit 142 in excess of the error unit squares

13 14

Glass to a minimum for the entire amount of glass. In the first circuit, triode 162 is normally reduced, that for any given circuit is blocked and triode 163 is conductive. In the mated requirement and the deficiency taken into account, the second circuit is that presented by the anode to the point distribution for cutting. Triode 162 driven triode 164 blocked and the With relatively short glass panels, 5 triode 165 conductive. The first multivibrator is set to a noticeable proportion of the length of the remaining end parts with a pulse duration of 30 milliseconds. As a further part of the definition of the smallest drop following the termination of a generator pulse, the selected Z-lengths are summed up in one input and one winding of a timing relay 168 via a direction 149 which excites the selection of cathode amplifier stage 169. The second multi-length prevents the remaining length of the glass io vibrator is set to a pulse duration of 20 milli from the new leading edge to the end of the glass seconds left on the termination of the plate, which is not evenly followed by a pulse from the first multivibrator and The combination of the programmed length is divisible over a winding of the second timing relay 166. The computing device has energized numbered collecting cathode amplifier stage 167,
rails corresponding to the length to be considered 15 According to FIG. 6 has circuit 154 on top and selectively energized. The cell blocks have a conductive triode 170 in each computing period and bind busbar spans that are the same as the corresponding non-conductive triode 171. When a corresponding Z program variable is applied, time-controlled reading of all Z variables derived from the pulse generator in any " possible signal, the line states of tubes 170, 20 and 171 are reversed. In this case, a capacitor switches when the output signals are fed back to the Z-program closed contact pair 181 of the first auxiliary timer 137, control relay 181 (FIG. 6A) becomes a capacitor 172 in programmed Z-sizes, which are not in the grid circles of the two tubes at a negative Z-splitting solution, from the Z-program bias voltage, and a diode 173 and a switched off. After the Z is selected, 25 needs stand 174 form a shunt for the condenser no further test for the fitting of the Z-variable to earth contact pair 166 a opens to be made. For continuous tape or, the negative charge of the capacitor 172 is discharged via another system in which the entire plate length of the diode 173 is discharged. If the pair of contacts closes again a few very long times with respect to any programmed milliseconds later, a nega-Z value arises, the value of the mismatched 30 tive charge pulse on the resistor 174, so that end pieces are meaningless. the operating conditions of the tubes are reversed. As in the block diagram of FIG. 5 illustrated the. «Odd and even relay stages 155 and is, starts and stops a clock control circuit 151 for 156 each generating an output pulse from the arithmetic clock 135 a pulse exciter 152 with 15 milliseconds. As in this particular constant frequency example illustrated by the basic time measuring element 35, the anode of the tube 171 is that of the clock. A series of relay stages 153 form time lattices of the first of the odd-stage multivibrator switching devices for actuating intermediate tubes 175 connected to a positive pulse in relays which control the voltage application sequence. The output of the second odd-level multi-one bistable multivibrator 155 is generated by the vibrator tube 176. The pulse generator is operated in the same way via the intermediate relay stage. 40 the anode of the tube 170 to the grid of the tube It is used to alternately coupled odd and even stage multivibrator 156 to energize even relay stages 155 and 156 in response to a positive output pulse at the anode of the generator pulses' the contact pairs of the second tube 178 of the odd step multivibrator to generate odd and even stepping relays 157 and. The corresponding odd and 158 are alternately energized in order to operate the step 45 even multivibrator circuits by cathode sequence switches 136. If a single amplifier 179-180 is used coupled to the even input step switches, odd and even step switches can drive the multivibrator and separate odd and even relays 157 and 158, respectively. The relays need to be omitted and replaced by a single non-relay stage for a step duration of 120 milliseconds. 50 to be actuated, an actuation of

Details of the clock circuit are shown in FIG. 6 to 15 milliseconds is used,

indicates. The pulse generator 152 is a freely running The timing auxiliary switch shown in Fig. 6A

Multivibrator with electron tubes, which uses the main contact pairs of the sensitive

zcitmeßelemcnt forms and in this case so an interim control relay 166 and 168 which are shown in FIG. 6th

is set to show a 55 every 120 milliseconds.

5-milisecond impulse delivers. For example, the winding of a first closed contact pair 161a of a clock start relay auxiliary relay 181 in the auxiliary circuit through the 161 in clock control 151 provides a negative closing of the open contact pair 166a of the first bias voltage for the grid of triode 160 so that timing relays are energized. The pair of contacts 181A of the multivibrator is blocked until the relay 161 60 Relay 181 is expediently energized in the input and its contacts 161a are opened. of circuit 154 of Fig. 6 and is a This is the case when the void recorder shows the various means by which timing can be introduced to the correct position for interrogation and delays since the reading at the S -Distribution matrix of the contact pair 181a has only moved several milliseconds after the computer. In another way, the 65 opens when the contact pair 166a closes. A start signal from the Z-tool or another normally open contact pair 181 is discharged to the station b at the glass-production line route checking the stored value signal in the staling. ^do g ^after f i g. 18 used.

15 16

A second auxiliary relay 182 is switched off by a control pair c of relay 186-26, which is in series with the voltage source by closing the open relay winding of clock start relay 151, contact pair 168 α of the second intermediate relay, no further pulses are received from the Stepping rain. The winding of the relay 182 is also taken up via switch actuation relays,
a diode 183 is connected to the pair of relay contacts 166 α . In this way the auxiliary relay 182 will periodically be required for auxiliary functions. That energizes first for approximately 30 milliseconds if the clock the step switch is designed to close relay 168, and for further approximately such additional time conditions without altering the 20 milliseconds when the relay contacts meet installed facility.
166 α to be closed. io step switch circuits, which are made by contact pairs of the

A device 136 operates as a 26-stage time-stage relay 186-1 to 186-23 are actuated

controlled sequence switch with step relay, which is shown in FIG. 8 shown. One that is in the upper part of the

get excited each other and stay excited when the F i g. 8 is the source of the graded

Stepping process continues. The additional trial voltages and uses those normally

Relay contact pairs facilitate the completion of 15 closed pairs c (as contact pairs 186-1 c to

of the twenty-first or chosen Z-step 186-2Ic) and the normally open pairs d

and prepare the computer for the next working (as contact pairs 186-1 to 186-21 d d). This ver

period before. sequentially bind channels 192-1 to 192-20 and

As in Fig. 7, an arrangement is a selected Z trial relay 193 with a step of 26 step relays 186-1 through 186-26 in succession 20 switching voltage source 194. The stepper switch actuated by a source 187 when pairs of contacts are with stepper pawls 1-20 on one of the odd ones and even stepper switch switch panel (Fig. 9) wired, switching relays 157 and 158 the corresponding odd The positive terminal of a grounded voltage and even numbered step relay in sequence source 194 (Fig. 8) is on contact pairs 186-1 d to excite. 25 186-2Id connected. The series connection will

The step relays 186-1 to 186-26 are connected to mercury from one end terminal of each open contact pair d

silver-coated relays with several groups. To the common terminal the next higher

switching contact pairs. _ numbered group of contact pairs made.

The step switch connection for the first 25 relays When the step switch relay 186-1 in

in Fig. 7 is equal. The end clamps of the 30 F i g. 7 applies voltage to the stepper rails

closed contact pairs α every odd number 192-1 via & m now closed contact pair

Step relays are connected to the positive switching rail 186 d-1 and ■ another pair of contacts 186 c-2.

189 connected while the end terminals of the laid. When stage relay 186-2 is energized, open contact pair b of any odd numbered stu, stepper switch trial rail 187-2 will be level relay with positive latch rail 188 35 if energized. However, the stepper rail will be connected. The contact pairs α and b of the straight 187-1 by opening the contact pair 186 c-2 numbered step relays are with the negative straight currentless. Thus, the busbars are each numerous switching and locking bars 191 or progressively for a step of 120 milli-

190 connected. The other terminal of every relay second energizes when the step switch is on winding is with the common terminal of the 40 working period.

Contact pair α and b of the previous stage In F i g. 9 plug-in switchboards are connected to the program relay. Thus, each step relay winding is shown with a group of Z · ^ sizes in the intervening terminal of its own desired Z trial order. Above in switching pair α and b and the common terminal F i g. 9 shows the step-by-step switchboard 196 of the switch pair of the preceding step relay, which switches terminals 197 with the numbering 1, which has opposite polarity. to 20, to which the corresponding voltage

The locking step rails 192-1 to 192-20 according to FIG. 1 remain in the first 25 steps. 8 ruled 188 and 189 of the odd and even-numbered ones. A second patch panel 198 energizes numerous interlocking circuits. Therefore, if a group carries terminals 199 which are first briefly energized on the busbar 189, the 5 ° subsequently numbered from 9 to 90 will be. The stepped relay 186-1 from the first represent the possible Z-lengths in units represent. Source 187 over his closed contact α ER Z-sizes from 82 to 89 units for glass plates stimulates this winding and then closing its open contact 90 Units length are not programmed, as Paares b interlocked. In the next step, if residues smaller than the smallest Z of 9 units are energized for the switching rail 191, the step relay 55 would belong to waste.

186-2 energized and locked, so that also one end

the winding of the next relay is marked with 186-3 control panel for approved Z values,

is locked. . also carries a complete group of Z-sizes

Relay 186-26 is used to reset terminals 202, numbered 9 through 90. six

Step relay at the end of a computing period. Its 60 separate S-size plug-in switchboards 203-1 to 203-6

Winding is with the positive side of the voltage carry corresponding groups 204-1 to 204-6 of

source 187 via a closed reset contact- 57 terminals. Each terminal group is "numbered,

pair 1576 of the stage relay 189 connected. That is about the possible 5 standard sizes from 6 to 62

the other end of the winding is shown with the voltage. Plugs 205 connect each terminal 202

source or with the locking bar 190 and 65 on the control panel 201 with one clamp of each

connected to the shift rail 191. S size switchboard 203-1 through 203-6 for a full

Since the calculator clock itself by opening a program. With the single switchboard 201 for each other normally closed contact-left Z values is any group of six Z-to-5

17 18

Plugs 205 connected to a single Z terminal 202 on Z panel rails. The other wicked switch The end of each coil is grounded so that the relays

As further shown in FIG. 9 is a Z of 216, the size number of which is that for the first twenty

50 for the first test step with a plug corresponds to steps programmed Z-sizes, after-

200 from terminal no. 1 on the stepper switch 5 to each other from the step voltage source 194 (Fig. 8)

control panel 196 to the Z size terminal to be energized.

No. 50 on the main Z plug-in switchboard 198. In Fig. 11, the mirror glass-window glass-

elects. Accompanying S sizes 55, 53, 52, 51, 45 and 9 quality selectors 217-9 through 217-90 shown. Everyone

are in this order by plug 205 between- Switch connects an approved Z-rail of the

see the terminal of no. 80 on the control panel 201 io same Z size number in series with a diode,

for permitted Z values and the terminals of the 218 with one winding end of a selector relay 219

Numbers 55, 53, 52, 51, 45 and 9 in the S-plug for a quality. The other end of the winding is

Switchboards 204-1, 204-2, 204-3, 204-4, 204-5 or grounded, so that the relay is switched on during each test

204-6 classified. Step of a Z-magnitude of η is excited, for which

The value program which the size program 15 of the selector switch 217-n is closed. While 139 is shown in Fig. 10, where a group of such trial Z-steps for which the entvon six 5-value plug-in switchboards 207-1 to 207-6 speaking selector switches 214-n are open, this represents is shown. Each has a group of 57 terminals, Z query relay 216-n a record reading those from 6 to 62 as in the case of the 5 size for an error in window glass quality. Plug-in switchboards 203-1 to 203-6 according to FIG. 9 nume- ao When switch 214-n is closed, a are ured. A value voltage switchboard 209, the record reading for an error pertaining to Also shown in Fig. 10, clamps 210 has mirror glass quality manufactured by the circuits, numbered from 1 to 75. These represent those determined by the control pairs of the selector relay 219 Values that the programmed will be controlled.

5-sizes by means of connectors 211 between the 5-value 35 As shown in Fig. 12, forms a paper tape

switchboard and its value voltage switchboard to 220 the fault memory record. It will be with

are ordered. fed at a speed that

A value voltage potentiometer 212 has taps in relation to the glass conveyor belt speed,

213, which are arranged in the same resistance intervals and defective positions are punched

net are seventy-five even increments 30 of holes 221 entered in the tape. Included

To define voltage levels. The taps are when the tape is fed from a feed roller 222

the value voltage terminals 210 connected to pulled ,. tJie has a switching mechanism 223 that

Establish tension on the clamps that sweeps the paper one Z-unit scale at a time

Correspond to the value of the terminal. If that triggers an error input. A sprocket drive 224

Potentiometer parallel to a voltage source 214 35 for the paper is driven by a motor 225

is connected, the voltage increase is ben. The tape is in a precise location to

each tap on the order of V2 volts. Reading stopped when an electrical coupling

A value program for the first step of the in and brake 226 is energized by a control signal ¹ . F i g. The Z x 5 size program shown in FIG. There 40 hole punches 227G and 227 M for windows are the ordered 5 sizes 55, 53, 51, or mirror glass quality. A leading edge hole 45 and 9 is provided on clamps 212 which the punch 228 establishes a reference hole (close to the same value as the 5 size. It becomes the side of the paper tape) around the leading edge of each a reference when estimating the likelihood - To mark the glass plate. Reading mechanisms borrowed garbage in a given requirement 45 229 G and 229 M, which are supplied directly under the drive program. All 5 sizes and values located in sprocket 224 scan the fault positions associated with a Z step are scanned for when mirror or window glass quality circuits through the punched holes are further selected. - between transverse Z and 5-conductor panels

The value voltages of the plug-in switchboards on opposite sides of the paper tape

Fig. 10 are not directly connected to the Z x 5 sizes - 5 ". A laterally movable tension pulley

Plug-in switchboards according to FIG. 9 connected, but 230 between the input and reading devices

the assignment of each value to a Z · 5 variable is used for the paper tension.

is achieved in the value accumulator circuits. As further shown in FIG. 12, a

The control stage of a device 141 for the longitudinal group 231 of contact fingers 232 for reading out information relating to the existing edge is numbered from 90 to 0. A conductive be of glass flaws in each row of adjacent rod 233 on the opposite longitudinally extending examination zone from the fingers, which side of the paper will decrease from each finger extending across the glass, is touched in Fig. 11 by the Z number in turn when shown a hole. The left row of numbered clamps 9 in the front edge during the belt movement to 90 is the group of clamps 199 which ^advances the main 6o. The paper tape fingers are optionally Z-switchboard 198, and the right, similar row to a potential above ground potential is the group of terminals 202 of the switchboard 201 for the Z values described in connection with FIG. 22 in FIG. 9. The conductors, which benen Z-accumulator brought. The winding of one of the terminals of the same number can connect the paper brake relay 234 is connected between the conductive of the normally closed switch lower rod 233 and ground. This relay has to be broke. Contact pairs in the excitation circle for the paper

A terminal end of the windings of a group brake and other circuits. When the finger is 90

Interrogation relay 216-9 to 216-90 is energized with the main, for example, to indicate that the

If the full length of a sheet of glass is to be cut, the tape will stop when its hole in the front edge reaches the finger 232-90, and the tape parts containing the mosaic patterns for mirror glass and window glass quality will be read with the reading mechanisms 229 M and 229 G aligned.

The electrical circuit of the reading mechanism 229 M and 229 G for reading faults

tape moved to produce a new leading edge position, as is ensured by opening the paper brake relay contact pair 234 α.

The error relays 246-1 to 246-62 are energized, 5 in order to read the S-position of all defects occurring in a certain Z-length. Each relay winding is normally connected to the S-bar 236 G of the same number of the window glass quality reading device via a normally closed contact relating to mirror and window glass quality is connected in io pair 219 a of the quality selector relay 219. F i g. 13 shown. The Fault Reader 229 G The same end of each window glass grade fault relay winding 246-1 has a numbered group up 246-62 with the S-rail 236 M being the mirror of 62 conductors 236 G that align with the direction of the quality reader via a normally paper tape movement . A second group of open contact pair 219 b of relay 219 is connected, conductive rails 238 G transverse to the paper 15, if an attempt step disposed for mirrored glass quality band movement, is from 1 reading to 90 has been programmed, the Wählnumeriert be. If the paper tape is energized between the relay 219 and the error relay 219, around the group of S-rails 236G and the Z-rails S-positions of all unit squares in this ver-238 G, contact is made between search-Z-length indicate which intersections of Z and S through 20 do not meet the requirements of the mirror glass quality.
Defect punches are manufactured that eliminate glass defects. Since mirror glass quality is higher than quality

is of the same dimensions Z · S with respect to window glass quality, contains the mirror glass reading correspond to the same front edge. The Z rails for a given Z plus all window glass defects 238 G are further faults by means of a flexible conductor 239 G. Contact pairs of the error relay 246-1 bound. If a positive control voltage 240 at 35 to 246-62 are distributed in the S-matrix voltage common conductor 239 is connected, are shown in Fig. 16 and in the S marker circuits S-rails, for which the imperfections are determined, according to FIG. 24 used.

One after the other to lie above ground potential - The reading of the value during each test step

of potential. is by closing contact pairs b the

A mechanical arrangement for making 30 relays 253-6 through 253-62 in each of the six S-blocks electrical contact via the flaw punch is indicated. Un ^ the various shown in this way in Fig. 14, where the spaced-apart indicated size divisions collect and with S-rails 236 G placed behind the paper tape 220- the special tension which are arranged for the detected S-size. The Z-rails 238 G before the paper is plugged in to interrelate, each carrying a number of spaced-apart 35 large numbers of circuits is required. A hostile, resilient contact fingers 241, which are aligned with the S-rails corresponding to the shaft solution for the equipment requirements. The Z-rail is effected in the manner shown in FIG. can be moved over a small angle of rotation in order to move it. Since the smallest S-span for a smallest finger to lift out of touch with the paper, S-size is 6, contact pairs b and α are the one without the raised Z-rail of the other 40 relays 253-62 through 253-57 in parallel for each of the six to separate. Cell blocks as the first register, 253-56 through 253-51

In Fig. 13 are Z-Schienen238G and 238 M with second than the register 253-50 and so on up to 253-45 connected as the corresponding numbers third register by dashed lines to the tenth registers that both rails overall to indicate switched, which is made up of the contact pairs of 253-8, when the Z-actuator 242 of the 45 253-7 and 253-6 consists.

same number goes dead. Closing a pair of contacts in each

To read off special Z values η , all registers must be displayed by the register finger splints 238 G and 238 M from the Z-Num- of parallel-connected pairs of contacts in the tensioner η or smaller in the reading position through the excitation circuit with a register value relay winding and spander corresponding Solenoids 242 are lowered. 50 power source is inserted. As in Fig. 2, one terminal of each solenoid is connected, one end of each register is grounded, via which one is grounded, and a voltage source 243 is applied via a positive voltage to a terminal of all value relays, normally closed contact pair 234 a of the. For the ten registers for the first paper brake relay 234 (FIG. 12) with a sole cell block of the S ₁ group, the value relays noide input terminal 235 are connected. The input 55 261-1 through 261-10 is used. Register for the remaining terminal 235 is at the low-numbered end Blocks S ₂ , S ₃ , S ₄ , S _? , S ₆ are correspondingly with 262-1 normally closed contact pairs 216-1 α
to 216-90α of the interrogation control relay according to FIG. 11
connected.

When each Z-rail solenoid 242 is connected to chain 60 directly under the pair of query contacts with the same number

to 262-10, 263-1 to 263-10, 264-1 to 264-10, 265-1 to 265-10 and 266-1 to 266-10, respectively. One single voltage source 267 is used for all 60 relays. A second simplification in the value store switching is in the value adding circuitry shown in FIG. There tension rails 270-1 to 270-6 carry the value tension for everyone of the corresponding S-Blocks 1 to 6 independently

When connected, all solenoids 242-1 through 242-90 are energized if there are no interrogation relays lower Number than 216-90 are aroused. The chain is through

Actuation of an interrogation relay with certain 65 of which S during a certain Ver Z number interrupted, all solenoids 242 programmed with search steps. This becomes easy in higher Z number to de-energize. All finger splints are lifted while the paper

the first S ₁ -BIoCk achieved, for example, that each S-size value terminal 208-1 on the

Sj-value panel 207-1 of FIG. 10 is connected via a normally open contact pair b with the same S number of the relays 254-6 to 254-62. Since only one 5 ¹ J size is programmed for each Z trial step, only one of size relays 254-6 through 254-62 is energized and a single value voltage occurs on value voltage rail 270-1. The size relay contact pairs b for the other 5-blocks of different rank connect in the same way

connects to ground via a resistance-capacitance network 285, 286.

The negative pulse developed across ground diode 284 as the capacitor discharges, 5 is applied to the grid circle of the triode 287 of the three resistor-coupled amplifier stages, 'forming amplifier 276 of FIG. A second resistor 288 is shunted to that Series combination of resistor 283 and each group of 57 S-value pawls 208-2 to 208-6 io grounding diode 284. If capacitor 282 is selected the corresponding value voltage rails have been loaded from an earlier step, 270-2 to 270-6. no pulse is sent to the grid of triode 287

The value voltages assigned to each register are given, if not the negative amplitude of the an are summed in an actuation amplifier 271. the voltage recorded at terminal 274 is greater Its input network includes ten resistors 15 than the voltage that was previously applied to the capacitor 272-1 to 272-10 same size. One end each has loaded.

Resistor is connected to the amplifier input terminal. At the end of a calculation period, the capacitor

and the other end of each register is connected to either sator 282 to earth through a normally open to earth or to one of the six Wertsammei- contact pairs 186-26 α of the 26th stage relay in the rail, as through the contact pairs α so the step switch circuit of FIG . 7 discharged. and b the value register relay according to FIG. 17. The memory shown in FIG. 8 for maximum

is true. Z-value assigns the actuation of storage volume

Each register resistor 272-1 to 272-10 of the relais280 (Fig. 18) corresponds to the special test step input network of the adding amplifier v / ird. The number displayed in this way is neither connected to earth nor with a certain 25 highest step is fed as a one-value voltage for conversion. During a step, the twenty-twentieth or selected Z-step with no more than a 5-value voltage is applied to any of the registers previously programmed for this experiment, the adding function completely storing quantities Z * 8.

in the resistor network and in the amplifier As in FIG. 8 is one winding end

he follows. Whether or not different value voltages are applied to each of the 20 V / er storage relays 291-1 to 291-20 with different of the excited register resistances to earth and the other with one of the corresponding ones depends on the distribution solution. Storage rails 292-1 through 292-20 connected. In the block diagram part according to FIG. 18, the sometimes closed contact pairs α and the entire standard comparison system for selecting the highly open contact pairs b of each storage relay value during the 20 test steps are shown. 35 are used in an interlocking circuit, the at the amplifier output terminal 274 to the memory relay highest number and its increasing total voltage is first applied to an assigned memory rail by a voltage capacitive comparator 275, which is a source for the calculation of the 21st stage 293 to pulse during the first step of the experiment emits, excite. Contacts α and b of the lock for which some value is obtained. An amplifier 40 circuits are adjacent their respective 276, a pulse shaper 277, a multivibrator 278, memory relay windings 290-1 through 290-20 in FIG. 8 and a cathode amplifier 279 and shown. A chain of normally closed strengthens the pulse that extends a value storage relay 280 contact pairs α of the storage relay from being excited. During subsequent attempts, no pulse is emitted from an emergency terminal 294 via contact pairs 290-1 α, the comparator 275 if the voltage is not greater than that on 45 290-2a, 290-3a ... 290-2Oa and then via one the normally closed pair of contacts 186-23 c he was previously charged. The end of the relay 186-23 for the 23rd stage (Fig. 7) to be diverted for the valid information is the last or the highest positive terminal of the stage source 293. The experimental step for which a voltage is applied to the Storage rail 291-1 stage relay for maximum value is energized, with 5 ° from another source, the storage relay will not save the value sums themselves. 291-1 energized and locked by source 293. At the

A particularly useful embodiment of the energizing of a higher number storage rail Comparison circuit and in particular the locks its corresponding storage relay itself Value comparator 275 is shown in FIG. One and opens the chain to any storage relay negative output voltage of amplifier 271 occurs dropping 55 lower number. at output terminal 274. This voltage The stepping contact pairs open and close

is between the ends of a series combination regardless of the state of relay 280. However, a diode 231, a capacitor 282, one is only a storage rail 292-1 to 292-20, resistor 283 and a grounding diode 284 when energized when the pair of contacts 280 α closes . If a normally open contact ^{pair 6o closes again} later during the series of 20 attempts, contact pair 280α, 181b of auxiliary relay 181 shown in FIG. 6a, closes again, the contact pair 280α is released. The diode 281 prevents the speaking storage rail from being excited, and the storage storage capacitor 282, which is connected to the negative relay latching circuit, lets the previously excited potential of the VeisUi.kerausgangs be charged, dropping into relays, while the storage rail and its charging circuit are discharged back. To remedy 65 relays with a higher number are locked. of switching transient processes is included in the charging circuit. At the end of the 20 test steps, only one

also a normally closed pair of contacts of the storage rails 292-1 to 292-20 via the 181c of the same auxiliary relay, which the charging line Speicherrclaisverriegelungskette remain energized. If

23 24

no value for any test steps input contact pair closes. When the amplifier

has been stored, none of the storage relays is energized, 311 conducts, changes a stage with triodes 312, 313, and the locking chain of contacts α of which initially the tube 312 is non-conductive and

Storage relay 291-1 to 291-20 leaves the emergency - the tube 313 was conductive, its operating state,

terminal 294 at the potential of source 293. 5 An output cathode amplifier stage with a

When the 21st stage relay 193 in Fig. 8 is closed- triode 314 energizes the accumulator output relay sen, normally open contact pairs 304-1 to 304-7 connect, if a pulse in this counter 193 a, the corresponding storage rails 292-1 to is received. Transfer pulses are sent as input 292-20 with the step switch rails 192-1 to output pulses from the. previous level of 192-20. The voltage of the source 293 is thus obtained at 10 circuit, and transfer pulses from the stepping terminal on the stepping plug-in circuit of the counter stage are transmitted to a multi-switch panel in Fig. 9 corresponding to the Z stage over-vibrator, which apply triodes 315 and 316, during which the highest saved points.
Value was received. If a counter stage has a trigger pulse of

To add up the Z-quantities when they are from 15 the pulse generator and one from the previous

a plate are cut, and the front receives the step, a transmission pulse is emitted.

Glass edge when interrogating the error record The accumulator output relay 304-1 to 304-7

Placing anew during each calculation period responds relatively slowly and changes the

the selected Z states do not exist during each calculation period if two pulses per stage are received

summed up. 20 genes. At the end of the counting process, open at 90

A Z accumulator system is shown in FIG. 20, a pair of contacts of a reset relay 332a, in which the binary Z output terminals 300-1 up to a negative bias voltage from the normal 300-7 according to FIG. 11, correspondingly with accumulator mode To remove non-conductive amplifier 314 , input relays 301-1 to 301-7 are connected. There the flip-flop circuit returns to its stable anist one end of each input relay winding is earthed, so the initial state, and the accumulator output that various combinations of the input relay relays are de-energized for the purpose of re-counting up to 90,
from the voltage source of the 21st step as the although the collected binary count is excited by the physical equivalent of the chosen Z-decimal length tending or non-conductive state of the corresponding. Seven binary counter stages 302-1 until the output relay 304-1 to 304-7 is displayed, 302-7 represent successive powers 30. It is more expediently represented in decimal form, from 2. For simplicity, the seven counters are the binary-to-digit converter According to FIG. 22, 302-1 to 302-7 graduate one after the other as counters "1", illustratively · the conversion is denoted by using "2", "4", "8", "16", "32" and "64", respectively . of contact pairs α and b of each of the accumulator-each counter is connected to a pulse generator 303 via output relays 304-1 to 304-7 using the normally open contact pair 301-1 α to 35 of the usual tree-like shape.
301-7 α of its corresponding input relay. A positive voltage terminal 321 of the convertible. Transmission pulses from each counter stage to lers according to FIG. 22 applies a voltage to the next higher counter level in the municipality are denoted by "carry" the same terminal of the contact pairs α and b of the return. There is no transmission pulse from the control relay 305-7 . If the reset relay is not energized last stage 302-7 back to the first counter 40, the accumulated count is less 302-1, used when the counter is stopped at 90 than 64, and when the relay is actuated, the count is and is not up to its full count of 128 worth 64 or greater. Each of these contact pairs will be used. Output binary accumulator relays 304-1 in sequence with the common terminal of one to 304-7 of each counter stage collect the Z-lengths, contact pairs α and b of the reset relay 305-6 when they are cut. A reset device 45 connected to the presence or absence of a device 305 is connected to each counter stage to display count 32 value. This row grows with the count value after a total of the last digit of the output relay 304-1 90 is reached to the accumulator for the Z-Wah- displayed binary number, which for a full extension len of the next 90 units - Prepare the plate. 128 contact pairs controls.

In FIG. 21, the circuit of a pulse generator 50 is shown. The end of the converter at the end of the chain 303 and one of the seven counters used here - the normally closed contact pairs α of the stages 302 . The pulse generator is a seven accumulator output relay as a connection multivibrator with triodes 307, 308, which are designated by a terminal 90 and corresponds to the cathode amplifier triode 309 , which is customary in Z-counting. The value zero. The other connections are in the downward generator is numbered to zero for its single pulse output during the 55 direction. When the accumulator of the 21st step received by the closing of a contact no pulses, delivers the input clamping couple 186-22 d of 22 stages in the relay step voltage at terminal 90 with respect to earth. The circuit according to FIG. 7 switched on. (The gen- erally excited terminal is incremented downwards, since the rator start pulse can also be selected later by a con- each Z variable during a calculation period is clock pair of the 23rd step depending on the time- ⁶ °.

delay can be controlled laterally, which in the case of an accumulator reset relay 235, whose winding response of the binary input relay is available the hole on the aisle rail 310 is supplied. Leading edge of the paper tape fingers 232-0 er-Each counter stage 302 in Fig. 21 is a bistable 65 ranges. Its pair of contacts 235a in the counter stage circuit which comes from a reference position at the beginning 302 (Fig. 21) is then closed to actuate the battery of a new count. To reset a triode modulator so that all accumulator amplifiers 311 become conductive when the accumulator output relays 304-1 to 304-7 are de-energized.

The output terminals 323 of the transducer 320 provide the signal for shifting the leading edge of the fault record, which is queried when a new leading edge of the lens is set, from any selected Z-size with respect to the previous leading edge. The leading edge fingers 232-9 to 232-90 of the finger share 231 in FIG. 12 are connected to corresponding terminals 9-90 of the transducer of FIG. 22. When the fault finger 232-90 is at the infeed end of the finger share , a leading edge hole in the paper tape must pass first under the finger 232-90 and then under the subsequent fingers. When the accumulator is reset to apply positive voltage to finger number 90, no circuit will be completed to leading edge contact rod 233 and thence to ground through the winding of paper brake control relay 234 until the paper has advanced enough to have a leading edge hole put under finger no. 90.

As further shown in FIG. 23, the solenoid control winding of paper brake 226 is in series with a normally open pair of contacts 234a of the paper brake control relay in parallel with a control voltage source. As long as the brake relay 234 is therefore energized, the brake is released and the paper tape advances until a leading edge hole passes under the finger 232-90.

During the next calculation period, the brake remains energized for the first 20 attempts. After the selected Z is collected, however, the input voltage is switched to a lower terminal. The paper brake relay 234 is de-energized again, the brake 226 is released, and the paper continues to move until the leading edge hole comes under the next finger, which has a potential to earth. The paper tape is then brought to a standstill by the brake, with a new leading edge being made for error reading.

The results of the S-splitting solution in the twenty-first or selected Z-trial are applied to the S-cutting readout circuit of FIG. 24 related. A positive voltage source 331 is g with the 5-Schneideableserelais 332-1 to 332-61 via the normally open contact b of the step 186-22 186-22 relay according to F i. 7 connected. The control network includes pairs of contacts for reading relays 259-1 through 259-62 ft in Fig. 16 and pairs of contacts for fault relays 246-1 through 246-62 in Fig. 13. The cut read relays provide information for cutting the glass. Spans of the glass that are smaller than the smallest programmed size are shredded into pieces of waste.

The cutting information is stored until the Z cut is made and an appropriate one The way to store the information is to mark the glass itself with the 5 cutting positions.

A row of boards or plates made of ground and polished flat glass is placed horizontally on rollers a main conveyor belt to a glass cutting device moves, from which the glass plates are separated transversely in Z-lengths by means of a glass cutting device will. In the production line, the Z-lengths are shifted parallel to scan lines, where suitable glass cutting devices, here referred to as S-Schneidwerkzcuge, are arranged are to make longitudinal cuts to divide the Z-lengths into smaller programmed sizes.

The glass plate can for example be a 6.35 mm thick plate with a length of 4.57 m and a width of 3.15 m or as a continuous belt of that width at one speed of approximately 5.08 m / min. To Carrying out the Z-cuts, the 5 marks are made on the individual Z-lengths of the glass, In order to determine the location of the longitudinal cuts, the cutting slots follow the edges of the programmed 5-widths of usable glass, between which the defective parts and the one by the Cutting process are excluded waste. This is the so-called two-cut logic.

As a preliminary step in performing the apportionment calculations and decisions will be made at a control station at a front of the glass production the Z-cutting tool and the S-marking device checked on the assembly line, both internal and surface defects of the glass being determined will. Information regarding these paying agents is sent to the computer. The calculator leads * 5 the division calculations based on the error information in time with the glass production. The flawed glass plate provides a primary record of the flaws contained therein. With this facility, the void information is available on request for calculation purposes .to make a secondary record of the faults, "in the glass in a fault storage device made, the recording taking the form of a perforated paper tape, such as this is shown in FIG. If the error information when requested from the secondary record is delivered and the sizes that have been previously set up are available, the usable areas become cut out piece by piece by using the two-cut logic is tracked and the defective parts of the glass are excluded.

In order to produce the necessary information regarding the defects occurring in the glass, the glass is first checked by hand for the presence of defects of such strength that its quality influence. The dimensions of these are indicated by colored pencil markings on the surface of the Locally marked glass plate. For useful interrelationship between the vacancy positions and certain A dimension unit is set up for the required sizes of the glass, for all length and width requirements are whole multiples. For a panel that is 3.15 m wide, a unit of 50.8 mm is selected giving the glass a width dimension of 82 units (5 dimension). The Glass is measured in units back from the front edge along the length of the Z dimension, hence a sheet of glass is 90 units in length, and is the width or S dimension of 62 units is measured between the side edges. In this way the glass and the locally determined flaws measured on the glass using a matrix of 50.8-minute squares.

By means of markings, the information regarding faults is stored directly on the glass, which is thus a primary record. The secondary record of the fault location information contains the dimensional position of each

Squares of defects in relation to the front edge and the side edge. The facility includes the Division of the glass plate into the desired sizes, which are graded in terms of quality. To do this too separate secondary records are made on paper tape in the fault storage device as shown in FIG. 12 to provide separate sources of defect location information To create window and mirror glass quality.

It becomes a © of the separated secondary records made to obtain information on the position of total defects while the Another secondary record is made to provide information related to the fault location to get mirror glass quality. This is shown schematically in FIG. 12 shown.

The information regarding the failure location information (F i g. 12) on the continuously moving paper tape recorded by punching holes in the tape by means of the punching units 227 and M 227 G. The 50.8 mm square glass is transferred to the paper tape, shrinking in the process to allow the application of a narrow tape so that a flaw in a 50.8 mm mosaic square is on the glass is represented by a 3.18 mm long and 2.39 wide rectangular defect punch hole on the tape.

In this way, the information regarding the position and severity of defects on an entire Glass plate or a predetermined length of the glass ribbon in easily evaluable for the computer Form to be made available. While the second recording is in time with the movement of the Glass is produced on the conveyor belt, the computer can use the information gathered regarding of the given glass lengths in order to systematically fit all programmed widths and allow lengths into the available defect-free areas.

At setup, a photosensitive defect detection device (Fig. 26) at control station 414 checks the glass plate on the conveyor belt for the presence of defect markings as it moves along the conveyor belt past the bridge on which the defect detection device is located. This provides information with regard to the location of the flaws in the form of coordinate positions with respect to the front edge on a side edge of the glass plate. To this end, the photosensitive defect device 62 includes photosensitive elements 416 in two rows of 31 elements each, each element 416 being positioned to face a ribbon of glass the width of a mosaic or unit square as it is the bridge happens. Each photosensitive element therefore responds to a defect which is present in a unit square at a certain distance S from an edge of the glass plate. A signal change from a photosensitive element caused by the passage of a flaw mark is used to control the fault storage device. The operation of the punch mechanism 227 M, 227 G of the error memory device is controlled in accordance with signals caused by errors which are received by the photosensitive device via information feed channels.

The punching mechanism of each error storage unit (Fig. 12) contains 62 selectively operable punches. The elements 416 of the photosensitive defect detection assembly are arranged in the same manner as the punches in the punching mechanism. Each photosensitive element 416 is mounted to cover a 50.8 mm unit of des

ίο the glass is directed and thus responds to a defect, which only appears in the square at a certain S coordinate. A signal change from the photosensitive element caused by a flaw mark becomes Control of the corresponding punch used in such a way that it makes a perforation in the mosaic of the paper tape recording the defects.

With reference to FIG. 4 and 26 there is the information which is necessary to carry out the calculations for the division of the glass, in the Position of the leading edge of the glass, in the position of the flaws, in the strength of the flaw, d. H. Mirror or window glass quality, and in the program of the sizes to be cut.

If this error information is displayed on the parts of the If the paper tape is available in the error storage device, the information can be stored on the various Tape parts are given in the calculator to get data regarding the strength of the defect as well as the To supply error existentials. The computer picks up the data and sticks through the size programming the required sizes depend on the defect-free areas of the glass.

The size programming is the quantitative requirement description, which includes both the absolute sizes as well as the relative requirement for them. Regarding the sizes themselves are the various rectangles to be cut are arranged so that they have the corresponding width dimensions group as parts of a common length dimension. The requirement for the various Length by width sizes is preferred given in a two-part breakdown. Thus a numerical value is produced for each size, so that values from size groups can be collected and compared. The in F i g. 26 with Block labeled "Request Program" indicates the size program and priority order are required to operate the calculator.

The defect location information obtained from the photoelectric defect detection device via the Conveyor belt is derived, is electrically transferred to a control room, which the fault storage devices and which is preferably arranged in the vicinity of the glass production line. the Computer is preferably also arranged in the control room and electrically connected to the 5 marking device and the Z-cutter connected so that the command decisions are electrically connected these components of the facility can be transferred. The one in Fig. 26 with "buffer memory" designated block is a facility, which is also located in the control room and for the temporary Storage of fault position information is used. The buffer storage device is provided because

29 30

interrogated with the light-sensitive error detection device bands in the error storage units for error information with regard to the presence of information. This is not obtained in the present faults in a 50.8mm tessellation or unitary device by coordinating the operation of the square of the glass before the calculator and fault storage device is passed the full length of the unit square from the light leading edge finger 232 (Fig. 12) achieved. Thus, the sensitive element that is located above the unit when the error information is located at different squares on the bridge, is viewed or from the paper tapes containing the degree of thickness are scanned. Each photosensitive element detects the input section into the reading section of a very small area of the glass, moving along the error storage device, passing through the selective center line of the unit square. Hence, 10 excitation of the fingers 232 and coincidence between buffering or temporary storage and a leading edge hole the tape is applied to set the defect information to read first when the part that is to be sampled until the full unit square defect data for a full , uncut glass plate has been. contains, entirely within the reading device

To make the separate records, 15 lies. If so, the computation period becomes like that in FIG. 12 is started according to the error in the calculator. Trial Z-lengths to the degree of strength is preferably a device prior to paper tape recording to distinguish between the degree of severity of defects Defect data in an order determined by the requirement, so that separate mirror glass and program-specific order are considered.
Window glass failure records in the failure 20 The failure information of the device after storage device can be made. 12 is supplied to the computer in the form of the S-Ko-A device for the production of strength degree ordinates of information occurring in the test lengths contains separate channels for over-errors. The electrical circuitry of stripping flaw information from each reading mechanism for reading mirror and photoelectric elements 416 of the device for window glass flaws is shown in FIG. 12 shown. The reading mechanism shown in the buffer memory device having a circuit shown in FIG the error numbers. Certain rails of this kind are caused by defects on the glass. 30 Λ 'rails are located on an above ground

The block marked with "error reading" is potential lying at the potential and are to be

di '; Device for reading the error information ner connected, se that they are the S coordinate of the

from the recording of the fault memory device, feed fault position.

how this is further carried out by the calculator, d? the required fault position

In the 35 information, the, ^ - positions of the flaws in a

Fault memory device according to Fip. 12 are determined, `` ηten Z-length is, the calculator does not work on one

separated reading mechanisms 229 M, 229 G vor'ian limited special form of the faulty storage device,

as a result of this, the error position information can be stored in

to remove glass or window glass defects. Likewise in any suitable manner, e.g. B. by electrical

Way includes the error reading in the Fig. 12 40 actuated devices for magnetic retention

Mechanism shown separate window glass and ten of the. Data stored on a magnetic tape

Mirror glass quality flaw punch that will be the tape.

punch holes in places that are marked by error signals.

that have been mentioned by the buffer storage device. Each plate consists of

can be obtained here. In addition, there is also a front 45 6.35 mm thick flat glass and is 3.15 m wide and

Edge punch 228 provided to make a hole (4.57 m long. The flaws of the entire plate,

near one side of the paper tape) to a mosaic of 50.8 mm unit squares

Ask to be allocated to the front edge of each glass plate, in terms of the possible

mark. Consider size cuts that are made

Furthermore, the device of FIG. 12 contains 50 could before the first Z-cut is selected. The one set of leading edge fingers 232-1, entire remainder of the panel is re-232-90 that are selectively connected to the computer before the second Z-cut is selected and in cooperation with the leading edge and process is repeated until the plate is full hole in the tape the movement of the paper tape is constantly divided. The maximum time to meet by the reading section with the functioning of the 55 each Z-Schnfttentscheid is coordinated by the conveyor computer. Limited by a belt speed shown in FIG. 23 and the lowest permissible control device for the brake and clutch Z-cut. Thus, if the conveyor belt 226 associated with the belt feed motor 225 is net at a speed of over 5.08 m / min, the belt is moved in dependence on the co-, an admissible smallest Z-cut of incidence between the selected leading edge 60 45.72 cm about 3 seconds time for each decision finger232 and the leading edge hole forwards conduction. That is, because work is done continuously, parts of the belt must be allowed to be read, and because the calculation is an assembly line process. a complete calculation step is ended and

To the computer taking into account a command decision during the 3 seconds

Assembly line time conditions for misinformation to be reached, during which the front 45.72 cm

and despite the fact that they are taken into account on a glass plate under the Z-cutting tool.

Then follow the trial Z-lengths before choosing the slide. In the case of strip glass, after a command

To enable Z, selected parts of the paper decision is made, the entire program of the

I 421

Z-lengths viewed from the new front edge and the best fit of programmed Z · S-sizes selected.

Following each calculation period during which the computer looks at the request program of trial Z lengths and selects trial Z values that form Z x 5 sizes that give the preferred fit, the Z cutting tool, as shown in F i G. 26 is actuated when commanded by the computer to carry out the successive Z cuts xo. These commands are transmitted to the Z-cutter via a cable. In order to perform the 5 longitudinal cuts, on which the calculations for choosing the Z-lengths depend, without affecting the flow process on the conveyor belt, the separated Z-lengths of the glass for the S-cuts are marked and then shifted in parallel to side sections, on which have suitable slot cutting facilities. In Fig. 26, a preferred device is provided to apply the markings on the pieces of glass at the required locations. These 5-markings are used to control the cutting course of the cutting tools on each side path with a suitable cutting marking display device or another similar device, which is arranged immediately in front of the ^ -cutting device against the direction of movement. To facilitate S-mark detection, the markings are made with suitable marking devices, shown here as chalk pens 418 and preferably containing a fluorescent substance that is excited by an ultraviolet light source and emits visible light as it passes through the cut-mark fixture.

As further shown in Fig. 26, the.! »Marking device is carried by a fixed bridge over the conveyor belt 402 and is connected to the computer by a cable so that command signals can be transmitted. There, the bridge carries 63 individually actuatable markers for making markings on the glass, which define the local positions for the S-cuts. These markers contain a swiveling pencil or crayon holder.

In the present case, a solenoid is provided for unlocking each holder and individually connected to the computer so that the command to make a 5 mark in a special place and get the command by unlocking it and lowering the markers onto the glass can be done.

Claims (2)

Patent claims: A device for dividing a glass plate, which has randomly scattered defects, into defect-free programmed rectangular parts of a given size, characterized by a device (137, 138) for arranging the differently programmed sizes in groups of widths with the same longitudinal dimensions in a given sequence; by means (141) for reading information relating to the presence of glass defects in each row of adjacent longitudinally extending inspection zones which extend transversely through the glass and have a length, measured from the leading edge of the glass plate, which corresponds to the length dimension of the inspected corresponds to programmed group; means (143, 144) for displaying widths of the sheet of glass excluding the flawed zones of the length dimension being examined, said means being responsive to said flaw reading means (141) ; by computing means (142), responsive to the means (137, 138) and the means (143, 144) , for determining the closest possible combinations of the width dimensions created by the means (137, 138) , which are incorporated into the defect-free strips of a fit into each longitudinal section of the glass plate examined; by means (139) for adding an amplitude value signal representing the relative value of each programmed rectangular glass sheet size; by means (145) for summing the value signals of the combination of the width dimensions determined by the arithmetic unit (142) for the examined length; by means (136) for operating the means (137, 138) to examine a different length dimension with associated width groups after the summing means (145) has created a sum for an examined length dimension; means (146) for storing the total value signal for each length and for selecting the length dimension of the sample for which the total signal is greatest; means (147) for severing the selected length from the leading edge of the sheet of glass to create a new leading edge for subsequent partitioning determinations; and by means (148) for controlling the division of each divided length into width dimensions of the preferred width associated with that length. 2. Apparatus according to claim 1, characterized by a device (149,150) which prevents the selection of lengths which leave a remaining length of glass from the new leading edge to the end of the glass plate which cannot be divided uniformly by a combination of the programmed length. In addition 7 sheets of drawings