US2925586A

US2925586A - Method of, and apparatus for, electronically interpreting a pattern code

Info

Publication number: US2925586A
Application number: US360540A
Authority: US
Inventors: Levy Maurice Moise
Original assignee: Individual
Current assignee: Individual
Priority date: 1953-04-29
Filing date: 1953-06-09
Publication date: 1960-02-16
Anticipated expiration: 1977-02-16

Description

Feb. 16, 1960 M. M. LEVY METHOD oF, AND APPARATUS FOR, ELECTRONICALLY INTERPRETING A PATTERN CODE 6 Sheets-Sheet 1 Filed June 9. 1953 Feb. 16, 1960 M. M. LEVY METHOD OF', AND APPARATUS FOR, ELECTRONICALLY INTERPRETING A PATTERN CODE 6 Sheets-Sheet 2 Filed June 9. 1953 v -5IN Feb. 16, 1960 M. M. LEVY METHOD oF, AND APPARATUS RoR, RLECTRONICALLY TNTERPRETING A PATTERN CODE 6 Sheets-Sheet 5 Filed June 9. 1953 Feb. 16, 1960 M. M. LEVY 2,925,586

METHOD 0R, AND APPARATUS PoR, RLRCTRONICALLY INTERPRRTING A PATTERN CODE OUT PBT INPUT OUTPUT MARA-MRR Feb. 16,1960 M. M. LEVY 2,925,586

METHOD OF, AND APPARATUS FOR, ELECTRONICALLY INTERPRETING A PATTERN CODE Filed June 9. 1953 6 Sheets-Sheet 5 @Mmmm u HVLVLVL /N VIA/77E Feb. 16, 1960 M. M. LEVY METHOD OF, AND APPARATUS FOR, ELECTRONICALLY INTERPRETING A PATTERN CODE 6 Sheets-Sheet 6 Filed June 9. 1953 NSG. OUTPUT Pos. mrtg-W LLA States Patent O METHOD OF, AND APPARATUS FOR, ELECTRONI- CALLY INTERPRETING A PATTERN CODEv Maurice Mose Levy, Ottawa, Ontario, Canada, assignor to -Her Magesty the Queen in lthe Right of Canada as represented by the Postmaster General Thisinvention relates to codes and apparatus for electrically identifying articles having imposed thereon a pattern code, and more particularly to such methods and apparatus suitable for use in identifying letter mail for purposesA of sortation. i

It has long been considered desirable to have automatic systems for sorting mail. There are a number of mechanical systems designed to assist in the sorting of mail. However, these are merely mechanical aids to the actual handling of the mail, the actual sorting of which is ,still being effected by human beings.

The real diiiiculty in producing an automatic mail sortation system is the replacement of the human eye, i.e. the provision of a device which is able to read and reproduce in usable form any one of the untold thousands off"fi addressed letters which pass through any largey post o ce. y t x The first stepneeded to overcomethis diiiiculty is the provision of a uniform code sufiiciently compact to be imposed upon an envelope in a position so as not to be obscured by lthe hand-writing on the envelope, and yet which is capable of conveying virtually any address in the world. The need of a code which could convey so many countless combinations immediately eliminated single line codes such4 as havetbeen used in the past in other coding systems. It has been mainly for this reason that previous sensing devices have been unsuitable for use in mail sorting systems. Another and most important reason Why prior sensing devices have not been suitable for this field is that they have been designed merely. to compare codes against a predetermined standard and to produce a yesorno output according to whether the codes scanned thereby do conform with the predetermined standard or not. This, of course, is quite unsuitable for the present field where it is necessary to know exactly wha-t is contained' in any oneV of innumerable codes. i

Having decided that single line codes were unsuitable due Yto the limited number of combinations conveyed thereby, a multi-lined, or what may be called a two-dimensional, pattern code was evolved. Insuch a code, information is not only conveyed by the position'ofmarks, in any one rowof the codebut itis also conveyed by the interrelationship of code marks in any number v of several associated rowsg4 Thus a code frame-work` of a given number of rows and a given number of columns,

vas will be described. later in detail, was chosen.V In such a' code frame-work, information could-,be conveyed by the placing of marks inappropriate positions Within such a framework. 'y p t TheV next requirement is the lprovision of a sensing devicecapable of reading such a patternV code and reproducing` in nsable' form exactly the information conveyed in any one such code, i.e. in the form of electric signals. To fdo this it was necessary to add vto the codeframework a certain number of marksin a given pattern extraneous of the frame-work itself. These latter marks are referred to in this specification as synchronization` 2,925,586 Patented Feb. 16,1960' marks as opposed to those marks which convey information and which have been termed classification marks. For purposes of clarity it may be mentioned that the term code marks shall be construed as being generic to both classification marks and synchronizationv marks.

Of course, having provided a device which is capable of reading and converting into the electric signalsrthe information conveyed in a pattern code capable ofexpressing virtually any postal address inthe world,I such electric signals can then be adapted to actuate suitable equipment for affecting the actual mechanical sortation of mail.

According to the present invention, a method of identifying articles for purposes of subsequent sortation comprises imposing upon an article a pattern code in `the form of rows and columns of classification marks and then scanning said pattern code with an electric sensing device to detect the individual classification marks and produce corresponding electric signals, separately identifying each such signal and applying it to the appropriate terminal of a group of terminals each corresponding to one position in the framework of said pattern code.

In addition to the classification marks there are spaced about the frame-work of a pattern code a number of synchronization marks. It is the signals excited by such synchronization marks which are used `to enable identification of those signals excited by the classification marks.

A more specific form of the present method consists framework of rows and columns, a first synchronization 1 mark being disposed before a first row of said frame-work,v a second synchronization mark being disposed an equal distance before a second row of said `frame-work not adjacent said first row, and a series of further synchronization marks disposed at the same end of, equidistant from, and in line with each column of said frame-work, said further synchronization marks thus forming a row parallel with the rows of said frame-work, at least said first and said second synchronization marks being of different configuration than said classification marks, simultaneouslyV scanning each row of said frame-work together with said frstand said second synchronization marks and said Vrow of further synchronization marks with an electric sensing device, applying the signals excited in said sensing devices by said first and said second synchronization marks to two synchronization signal selectors adapted one to select the signal excited by said first synchronization` mark and one to select the signal excited by said second synchronization mark, applying the signals selected by said selectors to a single electronic gate, the presence of a signal atthe output of said single gate indicatinglcorrect alignment of said pattern code, applyin'gthesignals excited in said sensing device bysaid I series'of further synchronization marks to a number of multi-vibrators corresponding one to each column of said frame-work andforming a cascade circuit, applying the signal selected by one ofsaid selectors to a starter inv electrical connection with the first multi-vibrator in said cascade circuit, applying the signal which appears at the output of each of said multi-vibrators to a separate group of electronic gates, the number of electronic gates in each of said groups 'corresponding to lthe number of rows in said frame-work, and applying the signals excited in said sensing devices by the classifica-tion marks in any one 1 row of said frame-work to the electronic gate in .each

of, said groups of electronic gates corresponding to that particular row.

The apparatus for effecting thls method may comprise an electric sensing device, a group of terminals one ooranew@ responding ton each position'in .the frame-work of a pattern code, and means for separately identifying each of the signals excited in said sensing device by said classication marks and applying each such signal to the appropriate terminal of, said group of terminals,

. It is, highly 'desirable that means.. should be provided tgindicate when a pattern code being scanned is incorrectlypositioned in relation to the scanning device thus gi. .e ilse.. t9 incorrect reading 0f the code.

111.111,@ acgrrlpanyng drawings which illustrate the rreferred( embodiments of the present invention:

Fig. lpfsvh'ovvsl a patterncode,

Fig. 2 shnvsY a blockV diagram of an electronic device @treading a Code, as; shown in Fig. 1,

- Eig.: 3 ShQWS a. mvdiflcaticrl of Fis- 1, i

Fig. 4 shows a block diagram of an electronic device for reading., a Code asVV shown. in Fig. 3,

i ,gf 5. ShQWS, a further modication f Fig. 1,

Y Fig.Y 6 s hoyvs blockl diagrarnof; an electronic device Yfor readies. a 904e. as. Shown in, Fig, 5,

7 shows a circuit of a synchronization pulse selcctor as shown by numerals 7 8- and 10k of Fig. 2,

"lis- 8 Shows. a Pulse diagram,

Eig,l 9:'schowspoa circuit of anY electronic gate,

Fig.'A 1i)` Y'shows-a lCircuit 0f a pulse width synchronization puise selector,

Fig. 141E shows a circuit of a synchronization pulse selector, as. shownby numeral 39 of Fig. 6,

Fig. 1.2 shows a voltage control circuit, and

Fig: h3,l shows a cold-cathode tube cascade circuit.

.For the purpose of, describing the preferred embodiments of the present invention, pattern codes will be considered, as being scanned: by the variousY sensing devices employed, In practice, however, the codes are passed before the sensing devices.

'Iihereareat least twovbasic methods of scanning pat- 'fmwdea Orlewhere. thecode is scanned by asingle en alongits length, at, a` constant or at a varying SRQdfef, Sanlinsfby. rows, and4 one 'Where a large 1111.,.lgr1 0f- SCflllills, Sweeps', arefmade at right angles t0 thendirectionin which the code is moving at a constant speedin1 relationv to the sensingdevice, i.e. by columns. Many welll-gknown,'types` off sensing devices may be employed tofscan pattern codes,y according to thev present ri so to facilitate the. following description an arbitrary choice of., sensing devicesV has been` made to ate scanning,Y by rows. and scanning by columns. 'i hoices arefa b'atteryof photo-electricr cells in the lf sc,anningbyfrows,'and a yinglspotcathode ray tube inconjun'ction with'r a photo-electric cell `in the case ofi,'scanningbyV columns. l

Itis tobe understood that this arbitrary choice does not in any way restrict the. present, invention which may beuadapt'edfor use Withvarious code imposing mediums, ie;blaclcand-l white,k color, fluorescent, magnetic powderk and'ph` spho`rescent codes, It `will be understood, of course;y that the, sensing device` wouldV have to be mod-iedaccording to the medium in which/the code is irnpose Fluorescent codes would require illumination by ultraviolet light andphosphorescentfcodes would require the i absence of visible light before they would excite the pulses in photoelectric cells, while a code imposed in magnetic powder would require the use of magnetic headsv rather-'than photo-.electric cells,-

Referring to Fig.. 1,'classicationmarks l aregrouped intovthree blocks each ofA ve rows R, and five columns C.' The rows R representthe following numbers: 'Rl=l;

Rza-72K; R s`=4; R4i=s 'and 125:16.

Any number upto and including 3l can thus bel represented in any vone` column C -by combinations o f 'classi- -catinmarks l, i.e. the classitication rnark lin,v row RZa'nd thefclassication markr lin row R5 bothY in columnv C, represent'the number 18. Similarlyicombina# tions of classification marks l inany one columnI 'C1 '4 Y 3=C and so on, i.e. the class' cation marks 1 -in row R2, row R3. and row R5., all in column C8 represent the letter V.

Other types of code may be used having other than 5 rows; however this 5 row type of code has proved preferable. The number of columns used Vmay vary and other forms of grouping the classification marks 1 into` blocks may be employed according to the grouping most suitable for any particular application. The grouping shown in Fig. l is, for example, suitable for an addressv where the street number is given in one block, the street or an abbreviation of same in another andthe town and state or province in the remaining block.

Synchronization marks S are provided, to enable determination of the position of anyclassication mark in the code, to ensure correct alignment of the pattern c0de` when being scanned and to enable all blocks together or any desired block of the code individually to be read. A pattern code as shown inv Fig. l is scanned from left to right at a constant speed by a batteryV of photo-electric cells`2-6 asv shown in Fig. 2. The photo-cells 2,-6y are`r so, arranged thatwhen theV code being scanned is correctly positioned, the photo-cells 2-6 correspond each to one- `9. If the code substantially correctly aligned when scanning commences, the selected synchronization pulses will be simultaneously applied tothe gate 9 and a pulse will thus. appear at the output of the gate 9.

VIl? a pulsev does` not appear at the output of the gate 9then. thecode isl not correctly positioned. Thus if anl electronic code-interpretation device according oto thei present invention isfusedf in rconjunction with aknown electricallyroperated:sorting system, i-.e. an Il BeM. sorter,

suchasystemcouldzbe set to-divert all codes into'a rejectbox unless a pulse appeared-at the output of the gate-9.

The pulseslexcited` in photo-cell 2 are also applied to a further lsynchronization pulseselector 10 throughr one contacts of a three-way switchl 10A. `Theselector 10 selects: the larger. pulses excited by the synchronization mark S1Y andrthislpulse isapplied to a delay line 11. Tapping points 12 to 17 inclusive are provided along the" length of "the delay` line 114 and since the code lin this? methodv is being. scanned( at a constant speed, the delay of-.the .delay line 11y issuch that the synchronization pulse. applied to. thedelaydine 11 will reach the tapping point 12 at the precise tn omentthat any classification marks. incolumn ClI excite pulses in therphoto-cells 2 tol ziinclusive. Similarly tapping points13 to 16 corretronicfgates corresponding toonefblock of the code shown in Fig.'1. The photo-cells 2-6 arev connected each one tofthe row of gatesvlfsinthe pattern :ofkgatesvl'which c corresponds tothe row Rofjthe code which the particular 'c photo-Celli@ warming, 4Thus eachgate. 18Y ls` electrically connectedto one of the tapping points 12-16 andto one ofwthe photofcellsuZf-.

Th.- synhwaizatgnnllse ,applied-to,V thedelayl line' 1,1.` ffmlhe slestgrllwlllthisbeapplled to all "the gates? 18 corresponding, to.V one'jcolumn orv thejcodel atfthe' Wsaine`vr time as .impulsesI are e itelA4 in the photo-cells` .l

2 w16 .inclusive by arly,"Classicationmarks which. might apurar` in ythat particular. silmn. Chf; the. Code Since. theseirnnulsea ,ifY any, ,arillltntanrpusly appliedtolie,

gates 18 corresponding to the rows R of thecode, and since a pulsewill appear at the outlet of a gate 18 if two pulses are applied to it simultaneously, a pulse will appearat the output of any gate 18 corresponding to the intersection of a column C and a row R of the code at which a classification mark 1 appears.

For example, the pulses excited in

photo cells

3 and 6 by the classification marks 1 inrow R2 and row R5 of the code both falling in column Cl will be applied to the corresponding gates 18'connected to tapping point 12 at the same instant that the synchronization pulse selected by selector appears at the tapping point 12 and is applied to the column of gates 18 in electrical connection therewith. Thus a pulse will appear at the output of the gate 18 which is in electrical connection with both photo-cell 3 and tapping point 12 and at the output of the gate 18 which is in electrical connection with both photo celll 6 and tapping point 12.

A precaution against a code being incorrectly read because of being scanned at an incorrect speed is provided by synchronization mark S3. The synchronization pulse applied from selector 10 to the delay line 11 should reach the tapping point 17 by the time the synchronization mark S3, which is of the same size as the classification marks 1, is scanned by, and excites an impulse in photo cell 2. If the code is being scanned at the correct speed the pulse excited in photo cell 2 will be applied to the electronic gate 19 at the same moment the pulse at the tapping point 17 is applied thereto, and a pulse will appear at the output of gate 19. This may be adapted to prevent the code being scanned from being rejected in the manner similar to that described in connection with gate 9. This is also a further check on the alignment of the code since if the code if sufiiciently out of alignment to give anA incorrect reading, a pulse would not be excited in photo-cell 2. 'This indication of a code having been incorrectly read may be used, if desired, to actuate the apparatus designed to reject the article bearing the incorrectly read code. l y Y It will be noted that yby means of the switch 10A, any single block of the code may be selected for scanning. If, for example, photo-cell 3 were in electrical connection with selector 10 through the switch 10A, no pulse would be applied to the delay line 11 except that excited by synchronization pulse S4 while the scanning of the second block of the code is in progress with the result only t pulse applied from the counter 24,v when caused to"` those classification marks appearing in the second block ofthe code would give rise to the appearance of pulses at the outputs of the corresponding gate 18.

Similarly the third and last block of the code may be read by the electricalconnection of photo-cell 4 to the selector 10 through the switch 10A. Here the delayed pulse is that excited by the synchronization mark S5.

At the end of the second and third blocks of the code in Fig.V 1,k synchronization marks S6 and S7 appear in row R1. These are to ensure that the code is scanned at the correct speed and in correct alignment as was the case with synchronization mark S3.

Having now a pattern of electrical pulses, or rather of terminals at which such pulses appear, corresponding to the occurrence of classification marks 1 of the code being scanned, these pulses may be used to actuate sorting systems. f Y

The scanning method described 'above `requires that thev speed of scanning'be constant. Since it may be difficult 'to maintain such a constant speed of scanning, the following method of scanning by rowsv which does'not rely on' a constant speed of scanning may be employed Referring to Fig. 3, a pattern code is shownV vwhich is suitable for scanning by rows at non-constant speeds.

For the sake of simplicity this code is shown as a single block codev having an initial synchronization pulse S8 and a row of synchronization marks-Sl0AS14 at the top of' the block, onehalignedwith each columnC. If it were desired to break down the code-.into va number of tion marks S4 and S5 in Fig. 1. Alternatively, synchronization marks of different configuration could be disposed one at the beginning of each block of the code all in one row, the one photo-cell scanning that row being in electrical'connection with as many synchronization pulse selectors as there are blocks, each selector selecting a synchronization pulse of a specific configuration.

In Fig. 4 the photo-cells 2 to 6 are aligned each to scan a particular row R of the code shown in Fig. 3. An additional photo-cell 20 is provided to scan the row of synchronization marks S10-S14. As in the case of Fig. 2, the photo-cells 2 to 6 are in electrical connection each with a row of electronic gates 18 forming a part of a pattern of gates 18 corresponding to the rows R and columns C of the code. l

Photo-

cells

2 and 6 are electrically connected to

synchronization selectors

21 and 22 which select the pulses excited in photo-

cells

2 and 6 by synchronization marks S8 and S9. Here again if the code is substantially correctly aligned a pulse will appear simultaneously at the outputs of

selectors

21 and 22 which when applied to an electronic gate 23 will produce a pulse at the output of gate 23 which may be used, if desired, to prevent rejection of the article bearing the code being scanned.

The pulses excited by the synchronization marks S10 to S14 in photocell 20 are applied to a row of cold cathode tube counters 24 to 28 inclusive, one for each column of the code while the pulse excited in photocell 2 by synchronization mark S8 and selected by the selector 21 is applied to a starter 29. The pulse appearing at the outputV of the starter 29 is applied to the counter 24 preparing it for striking.

When the synchronization mark S10 excites a pulse in photo-cell 20 this pulse is instantaneously applied to the counter 24 which strikes and applies a pulse to the first column of electronic gate 18 and to the counter 25 preparing it for striking when the pulse excited in photo-r cellf20 by synchronization pulseSll is..applied to it.

Since the synchronization mark S10 is aligned with the leading edge of column VCl of the codein Fig. 3, the

strike by the pulse excited in photo-cell 20 by the synchronization mark S10, to the first column of gates 18 will coincide witlrthe application of any pulses excited' in photo-cells 2 to l6 by'any classification marks l in column C1 of the code, to the rows of gates 18. Thus the pulse'excited by the classification mark l in row 1 and column 1 of the code in Fig. 3 will be applied from photocell 2 to the first row of the gates 18 during the time the pulse" from the counter 24 is applied to the first column of the gates 18. A pulse will then appear at the output of that gate 18 in electrical connection with both, the counter 24 and photo-cell 2.

The successive synchronization marks S10 to S14 inclusive strike in turn each of the counters 24 to 28 respectively, As each counter 24 to 28 is struck it prepares the next counter 24 tor 28 for striking, i.e. counter'24 when struck will prepare counter 25 for striking which when struck itself will in turn prepare counter 26 for striking. Conversely as each counter 24 to 28 is struck` by the application of a pulse excited by a synchronization mark S10 to Sl4,*it cuts off the counterimmediately before it, i.e. counter 26 will cut off counter 25 which itself, when struck, cut off counter 24. It will be noted at the outputs of the counters 24 to 28 will be ofthe same duration as the time taken for any one column to pass be..Y

fore photo-cellsy 20 and 2 to 6. Thus the'speed-with which the code as shown in Fig. 3 is scanned, within the limits of the'mechanical means provided for movingthe code in relation to the photo-'cells 20 and 2 tov 6, has no effect oirl assumant .the readies of.; the @de 'lhs pulse an. marine a1. ,the cmp-1t @fthe counter 28.A is applied to: andA resets starter.; 29. Suclra row of counters, 24. to 2.8 will'betmore fully described later in. this specification.

As anadditional precaution against incorrect reading ofY acode due to. incorrect alignment, thev pulse appearing at the output ofV the last. counter in: the cascade circuit,l counter Z8, may also be; used. in conjunction with the pulse atV the output of gate 23 to prevent rejection of the article: bearingthe? code beingA scanned.

Various, adaptations of the above methods'w-hich fall with-in the: scope of the present invention are envisaged. Amon-g these are methods whereby synchronization marks are placed above, or below, the columns in two rows, those aligned with the odd numbered columns being in one Yrow while those inline witht'ne even numberedV columns being in a second row. Alternatively the synchronization marks S could a-ll be inthe same row, as shown-.inl Fig. 3l but 'alternately large and small, or by alternately different combinatio-ns;r of a number of smaller marks.

It should be understood that the main characteristic of :1t-'synchronization mark is` that the pulse or pulses it excites. b e,electronica-ilyY distinguishable from the pulses excited. by c lassiiicationy marks and may therefore be of a number ofconfgurations. These variations would allow the codedmessage to be divided into two parts, one part being,representedy by the odd numbered columns, the second part by the. even numbered columns- Thus, where alternately large. andy small synchronizationl pulses are used either part' ofthe code could be interpreted without thel other by. using only the pulses excited by one row of` synchronization. pulses in the former case or by the use of a synchronization pulse selector selecting only synchronization pulses of the desired width.

AIn this method the image of the spot'excited on the screen` of a cathode ray tube 29A of Fig. 6 is suitably focused on a; code as shown in Fig. 5 and the reflection of,or V emissionexcitedy by theimage is'suitably focused on a photo-cell@ 30Fig-t y6. Asit is desired that the columns of the code be rapidly scanned,ia sawtooth wavemay be applied; to they cathodeiray tube .29A thus causing the image ofthe spot continuously to travel down the columns and y, back. astheacode moves before the' tube 29A at right: angles to the direction of movement ofthe image.

.It maybe mentioned hereI that if the iiying spot type sensing device isused in conjunction with the methods earlier v described :of scanning ,byfrows a stepped wave-form vvoltage may be applied tothe cathode'ray'tube causing the image tomove successivelyffrom one. column to the next iny they same row, pausing momentarily in each column.

The secondof the two broad methods of scanning de.

fnedfearlier in this. specilicationis that whereby the code is scanned by columns at a rapid rate as the code is movedv fromright to` left, or viceversa, in relation to the sensing. device, whichhas in this case beenarbitrarily chosen as. aflying` spotcathode rayvtube in conjunction with a photo-electric cell. Y

A', patternicode suitableforV scanning by columns is showniinFigi, 5; The synchronization pulses Sare narrosvin-.relation to the` direetionofscanning, i.e. by columns, with-,ther exception of.; synchronization pulses S16 and S157 whichv are: larger in size to the classification marks'4 1.'

'In-A thiszmethod: the letter moves.` beforezthe'sensing device `fromyright to left so thatthescanning commences ati-the.lefttzhandside, offthe code. The first pulses excited ,;inthe .Phot0-'cellf3o0,- see Fig. 6, are thoseexcited by synehronization; marks Sl 8, andi Sl 9; The photo-cell 30 is in ,electrical connection-with',aseries of .electronicgates 31 to 36.inclusive and to va narrow pulsesynchronization l Mector .37. The ,narrowpulses excitedby synchr0--Y :fearless 1.8 and Sl 9.7,areselected-.by selector 37 and,

annledsfaaaelectrcni. sate-.38 and. teamone Sanctum*V nizationiliulse selector 39. which, combinesY the. two. excited` by.:synchronization:-marks-S18 and- S19v in any oney scan into one. sharp. high pulse. f If the code is corY rectly alignedf,l this combined pulse will. be of Su'lciently high voltage to actuate selector 3-9 and'to cause` a pulse.- to. appearv at` the output thereof. This pulse Vis applied to a voltage control 40. andi may also be used` topreventV Thevoltagecontrol40, when actuatedby a pulse from,

selector 39, applies a voltagefto the gate 38 which is maintained. throughout the scanning of the code. Thus.

as scanning continues; lthe pulses excitedi by synchro-V niza-tion` pulsesSZO to=S34 andi selected by the selector 31 are applied to, and pass through the gate 38a These.A

pulses. are thenapplied through; a, pulse. Widener 41- to. a high frequency delay line. 42.

The speed withwhich the flyingf spot of the cathode ray tube 29A1travels down the columns is constant. Thus the rate atY which the tive rows R are traversed, andy the ratte at which pulses` arel excited' by any.4 given classication mark 1 are; constant. The delayl of the line 42-is such, therefore, tha-t tapping points 43 to 4 7' inclusive. correspond tothey rows R1 to |R5 while the tapping point 48 corresponds to the row in which. the synchronization marks S16l and. S17vr appear; For example, the-pulseexcited by synchronization mark S25 in any single scanY of the column C6 will appear at the tapping point45 a-t the` same instant that a pulse is excited by the classification. mark l at the-intersection of columnC and row` R1 during thesamescan of column-C6.

The tapping pointsV 43. to: 48 aretin electrical connection` each with one of the series ofV electronic Vgates 31 to. 36-

inclusiye `to-{al-l ofl which 'the-pulses excited in the photov cellg30fareapplicd. v

Let us here assume that in the time it takes for thewidth` or one classication mark 1-. to'w pass beforeI the photocell 30,-,t'he columnl Cl inwhich the classification mark 11 appears is scanned live times; Referring back to the aboveexample regardingithe appearance ofga pulseat tapping point 43- coinciding with, the pulse excited by the.

'n classification mark in' column C6,4 row'Rl, such a coincidence of pulses willoccur five, times while the code moves the widthV of a classilication mark` Iin relation to photocell 30. Therefore since both the tapping. points 43 to. 48 inclusive and photo-cell'30 are in electrical connection with the gates 31 toorespectively, whenever a classificationrnark 1 `appears in a row R or a synchronizationVY mark S16 or S17 appearsin the case of tapping'point 48I and gate 36, ltive short pulses will., appear in rapid suc-v cessionI at the-outputtofthe gates 31 to 36 inclusive-.correspondingtozthe row'in which the classiiication mark lA or synchronizationmiarks S16 and S17 appeared.

v Eaclrone of ythe gates 31i to y36 inclusive is in e1ectrcal connection with a low pass lter49 to 542at-theoutput o of which a single wide .pulse appears whenever a series of nar-row'pulses, e.g. live, are applied thereto from the o Asis wellknown, arlow. pass iilter. has,

the effect of widening and delaying each pulse passed through it. The single wide pulse is thus the resultoffthe additioneto'one anotherwofthe individualpulses-that have been widenediin thelowpassfilter so'as to` overlap'oney another.Y 0f these low passf.flters, 49-to 53 inclusives-areelectrically c onnectedeach -tofarow of `electronic :gatesfv l'j'asginpreyious. methods.v Thenumber of columns in i the l pattern of. gates; 18 in., this; casev corresponds to\ the number. of. colurrmsl C in the code.

9 each corresponding to one ofthe synchronization marks S20 and S21 and so on, are applied to a delay line 56, the delay of which -is such as to delay the pulses by threequarters of the time delay between the synchronization marks S20 and S21 and so on in passing before the photocell 30.

The delayed pulses at the output of the delay line 56 .are then applied to a short pulse generator 57. For every delayed pulse applied to the short pulse generator 57, two short pulses appear at the output thereof, one positive-going and one negative-going. Since only one of these latter two pulses is desired, the negative-going pulse may be suitably suppressed. The positive-going pulse appearing at the output ofthe generator 57 is then applied to a -row of cold cathode tube electronic counters or multi-vibrators M1 to M15 inclusive in cascade circuits, each row corresponding to one of the columns C in one block of the code as shown in Fig. 5.

The pulse at the output of the voltage control 40 and applied to the gate 38 is also applied .to a starter 58 which immediately prepares the multi-'vibrator Ml for striking. VIt will be remembered that the voltage at the output of the: voltage control 40 is initiated by the combining of the pulses excited by the synchronization marks S18 and S19. Thus the multi-vibrator M1 is prepared for striking by the-pulse applied from the starter 58 immediately upon thecommencement of scanning.

Due to the delay of the delay line 56, the first, or positive-going, pulse ofthe two vshort pulses generated by the generator 57 from the pulses originally excited in photocell 30 by the synchronization mark S18 will be applied to mul-ti-vibrator M1 at the time the scanning of column C1 commences, the second of the two pulses having been suppressed since it is unwanted. This rst short pulse, or for the sake of clari-ty the S18 positive-going short pulse causesthe multi-vibrator Ml to strike and a voltage willlappear at the output thereof. It should here be mentioned again that cold cathode tubes are used in the multi-.vibrator cascades Ml to M15 inclusive which will be considered in detail later. The voltage appearing at the output of Ml will continue until the positive-going short pulse originating from the pulses excited -by synchroniz'ation mark S20 appears at the output of the generator 57. This S20 positive going short pulse is applied to the row of multi-Jvibratorsand will strike M2,4 M1 having already been struck. As soon as M2 begins to conduct, the anode voltage of multi-vibrator Ml drops and the lvoltage appearing at its output is cut 01T. It will lbe seen that s ince the first and positive-going pulse generated by the generator 57 -appears at the output of the generator V57 coincidental with the time the scanning of a particular column commences, the `output voltage of the multi-vibrators M1 to M15' will have a duration each corresponding to and coinciding with the time taken to scan one column C of thecode. The outputvoltage of the multi-vibrator M1 is -applied to the rst column of gates `18in the pattern of gates 18 previously described 42 and to'tapping point 48. The series of pulses excited bythe synchronization mark S24 and widened by pulse Widener will reach the tapping point 48 and be applied to the gate 36 at the same time as a series of pulses excited by the synchronization mark S16 are applied to I the gate 36. Thus a series of pulses will appear at the output of the gate 36 and will be converted into one long pulse by the low pass lter 54. This long pulse is applied through a pulse diterentiator 59 to a starter 60 which then prepares multi-vibrator M6 which corresponds to column C6 ofthe code, the rst column ofA the second block of the code, for striking.

Multi-vibrators M6 to M10 inclusive, corresponding tocolumns C6 to C10 of the code, are prepared for striking,. struck and cut off in the same manner as multi-vibrators: Ml to M5 inclusive and the output pulses thereof areapplied to the columns of gates 18 corresponding to the columns C6 to C10 of the code. Similarly multi-vibrator M10 when struck,` applies a pulse to, and resets starter 60.

-The synchronization marks S16 and S17 are, however', not essential, their use being simply to enable separate blocks of the code to be scanned rather than the entire code. In describing the actuation of the row of multivibrators M11 to M15 which correspond to columns C11 to C15 of the code it will be assumed that no synchronization mark S17 is used. In such a lcase multivibrator M10 will also be unnecessary. Instead of the output pulse from multi-vibrator M10 being applied to multi-vibrator M10' it is applied directly to multi-vibrator M11. The cascade of multivibrators M11 to M15 thus strike and cut off as did multi-vibrator M1 to M5 f and M6 to M10, the output pulses being applied to the appropriate gates 18. `Here a multi-vibrator M15' is used to supply a pulse to reset the starter 60.

Thus it will be seen that `in each of the methods described a series `of pulses.k are obtained at terminals in a pattern which exactly corresponds with the pattern in which the lclassication marks are imposed on the code being scanned. These pulses can be adapted for use in many ways such as to operate the earlier mentioned sorting systems, to select codes having certain specific data on them as with a cataloguing system, or to actuate memory devices. e

The details of the circuits will now be considered. Referring to Fig. 7 a synchronization pulse selector of the type shown bythe

numerals

7,8 and 10 in Fig. `2 is illustrated. Y

As pulses are excited in photo-cell 2 of Fig. 2, for exi ample, they are applied across resistance 61 of an unterminated delay line 62 comprising inductances 63 and condensers 64. Y,

In the case of a white surface ywithblack synchronizationand classiiication marks beting scanned, these pulses would be positive and it will be assumed that this is the case.V The sign of the pulses if excited by a uorescent and shownin Figs. 2 and-4, thusthegates 1'8 correspond- `f ing to Vcolumn'Cl ko f .the code are ready yto openfshould any classiiication marks appear in that column and cause ythe ,pulsesjto appear at thel output of one-of the low pass:`

lters 49 to 53 inclusive.

The outputv voltage of thefrnulti-fvibrator M1 is alsoA pulse generated by the pulse excited by synchronization.

pulse S244 is applied thereto and the output pulse thereof is applied to and resets the starter 58.`

Nownreferring 4back to the high-frequency delayv line code, thus being negative, could simply be reversed by applyingthe' negative pulses to the grid of a triode and applying the positive output pulses appearing at the anode l' of the triode tothe delay line 62.

A pulse diagramrepresenting the input of the delay line 62 -is shown at (a) in Fig. 8.' The pulses travel lalong thedelay line 62 and are then reilected back along thedelayrline 62 lunchanged in sign but delayed. A pulse diagram of the reected pulses is shown at (b) of Fig. 8. The input pulses and the reccted pulses appear combined at the4 input of the delay line 62. A pulse diagram representing the input plus reflection of the delay line 62 is shown at (c) of Fig. 8. It will be noted that the delay of the delay line 62 is such that only theA wider pulses excited by the larger synchronization marks Y actually overlap to producea high voltage pulse. In the case of the apparatus as shown in Fig. 2, this is possible sincel the speed of scanning is constant.

The.. original input pulses and the reilected, delayed.V

11;` Y pulsesgcombnedare. appliedI from the. input-.of :the id elay' line62 through condenser 65 to the gridI 66;of a.triode 67. A-grid `

bias resistance

6,8 i s connected between thesgrid 66. and ground. The cathode 69 offthe triode-67 is grounded through cathoderesistance 70 and by-passl condenser 71. A positive HT. voltage is applied to the cathode 69,' of the triode 67 of such a-value that `only ka positive pulse of the voltage of the high pulsev created by the overlapping of the original wide'pulse, excited by the synchronization mark Sl of Fig. l, and its delayed reflection willbe suflicient to overcomethe positive bias of the cathode 69 and cause it to conduct thus producing a negative pulse on the anode 72 of the triode. 67. This negative pulse corresponds to thepulse originally excited iniphoto-cell 2 of Fig. 2 by the synchronization mark'Sl ofI Fig. 1. If desired, ofcourse, the sign of this pulse may be changed by applyingitto'the grid of aisecond triode and tapping off the positive pulse which would appear in the anodef circuit of this secondtriode.

A number of diiferent types of electronic gatestmay'be employed inthe lthreepreferred embodiments of the present invention.

The-preferred form of gate shownin Fig. V9, may be used in all instances in any of the circuits illustrated-in Figs. 2, 4 and 6. Here again the apparatus illustrated in Fig. 2 will be used as an example. rhe pulse appearing at tappingpoint 12 of the delay line 11 in Fig. 2'is1-applie'd to the cathode 73 of diode 74 in Fig. l0. The

, anode 75 of the diode 7'4 is electrically connected to a the platelresistance 78 whenl pulses.are-simultaneously appliedtoboth` thecat-hodes y713 and 79. Therefore the pulseexcitedV by a synchronization mark appearing at tapping :point 12 of` delayline 111 in Fig; `2 and-applied tothe cathode 73 of' diode 741 wil'lnot cause apulse'to appear across resistance 78, Fig. lI unless aNpulse excited in-photo-cell 2- by va classification mark .is 'simul taneouvsly4 appliedtothe cathode 79 ofdiode 77;

Fig.-l'0 yillustrates. the detailed'circuit of a pulse width Y synchronization pulse 37`Yin Fig. 6. v

Let us, Afor example, considerthepulsesexcidf in selector. as shown by thenumerl photo-cell30 of Fig. 6 during one scan of column` of Fig. 5v. HereV a narrowpulse will firstfbeexcited by the synchronizationV mark S27 whichwill='be followed -by wide pulses excited bythe four classi'ilatit'ni marks ll in columnrCS. A diagram of pulses excitedgin thephoto.- cell 30. is shown aty(df) of;y Fig. 8. i

These pulses are all applied tothe grid8 0"'of:1a

(this may be a triodeorvpentode) througha condenser 82 and across a grid bias resistance 83'. The anopder84 Thetinitial inputA voltage on the grid 80, i.e.the;begin. l

ning. of. thefnarr'ow.. synchronizationV pulsei-,excited-.by synchronization mark S27,`isshownatl (e), in.-Fig."8. setsup; an oscillation intheltunedl circuitfof inductance plate resistance 8,8.,see ('f) of Fig. 3.

The; secondl input on` thetgrid l is 'a voltage:drop,.i. e. the Vend' of the f narrowsynchronizationA pulse excited `by synchronization mark S27, is `shown at (G) of Fig. 8. Here again a damped oscillation is set up in the tuned circuit Yof inductance;` 86 and the condenser 87, delayed by the -width of thefnarrow synchronization pulse and ofreverse signv to the initial oscillation, see (h) of Figf8. Therefore the combined initial and secondinputs as shown `at (i) of Fig. 8 will produce a combined oscillatory output, as shown atv (j) of Fig. S. that thegrst swing of the-secondoscillation will` combine with the second swing of the initial oscillation, both being positive, to.y produce an extra high voltageswing. Asmentioned before, thetuned circuit must be tuned so The output of the vtuned circuit of inductanee 86 and s condenser 87 is then appliedV tothe grid 89 ofaftriode- 90` having a positive bias on its'catho'de- 91. .As` is the case of the'trio'de 67 of Fig. 7'the value of this positive bias must beI such that only the positive voltage offthe combined positive swing ofthe initial and second oscillations asshown at (i) of Fig. 8 will cause the triode 90 to conduct thus producing a negative pulseon the'anode 92 of the triode 90. fThis pulse may then be converted to a positive pulse by'applying it to the grid-sof a'further triode and tapping off the` positive pulse appearing at'the anode of this further triode. The numerals'9r3, 94` and 95' showt-the cathode resistance, the by-pass condenser and the plates resistance respectivelyiof the trio,de9,0.

Referring now to Fig. 11, the details of the'synchro- Vnization pulse selector 39 of Fig. 6 is shown. 'Thisselec` tor operates on similar linesto those shown at numerals 7 8 andltl of Fig. 2 and shown in detail in Fig. 2. Since the speed at which the filying spot'of thel cathodejray tuber29a of Fig. 6,'t'raverse the columns C of a code is constant, the,synchronization marks S18v and S119 of the code'shown in Fig. 5 will' excite'pulses ,in photo-cell30 of Fig. 6 at; regular intervals.` Therefore an unterminated delay line`96isbuilt up of inductances 97' and condenser 98having a *totalV delay of 'half' 'thedelayf'between the appearance'of the pulse excited by theisynchronization mark S128 and that excited by synchronization markfSl'9 inthe same scan. Thus as'thesepulses appear at the outputof selector 37 ofFig. 6,.they` are, assuming. them to be positive, applied directly to the input resistancej99 of' the delay line 96,. If theinitial pulsesare Ilgative they may bek applied to'the ,grid of a triode and thepositive pulses'a'p'pearing in the kplateA circuit` ofthe triodje 'be appliedglto resistanceKA VThe delayv line.A 96T" being yun- ,trerminat'zedL thefpillsesy are reilected.,y backtothefinput pwithout a .change of sign and' since the rdelay ofthe delay lli'ne`96 is half ofthe delay betweenthe occurrence `of two pulses Vin anyone scan of the synchronizationpulses S18 vand S19, the pulseexcited by markS 18 will ,be

reflectedtback to the input resistance 99r coincidentaljwith 'the'application ofthe pulse excited bythe mark S19 thereto. Thusa series of'narrow positive pulses, see ('k)` Fig. Y8, applied to .the delayline 96 willV appear at'they input-'of thedelayfline 96 as a series ofA narrow pulses,V every second, pulse being of` twice the positive voltage `of' the others,"see (l)i of Fig.j`8.

The -input'of the delay line 96'is'then-appliedthrough condenser 100,-to the grid101rof'a triode 102. Thegrid 101, being grounded'- through `resistance v103, andlthecathode 104 of the triode V102 beinggrounded through It will be seenl Y the' cathode 104 being positivelyrbiased soY that only a positive voltage ofthe value of the larger positive pulses applied to the grid 101 will cause the triode 102 to conduct thus causing a positive pulse to appear at the cathode 104 of the triode 102 and a negative pulse to appear in the plate circuit 107 of the triode 102. It will be seen that ifthe code, Fig. 5, being scanned is incorrectly aligned, the delay between the narrow synchronization pulse being applied to the delay line 96 will not correspondto the delay ofthe delay line 96. Therefore a pulse will only appear in the plate circuit 107 or at the cathode 104 of triode 102 when the code is in the correct position.

When a negative pulseiappears at the `anode output of the synchronization pulse selector 39 of Fig. 6, it is applieduto voltage control `40. The details of voltage control `40 areshown in Fig. 12.

The negative pulses are first applied to the anode 108 of a diode 109, the cathode 110 of which is connected to ground through a condenser 111 and to the grid 112 of a cold lcathode triode 113. As the pulses appear at the anode 108 of the diode 109 a positive charge is built up on the plate of condenser 111 remote from ground. This positive charge is applied to the grid of the triode 113 the anode 114 of which is connected to a positive H.T. source through resistance 115 and will tend to strike that tube. Since the code is being scanned by column however, and since the pulses will be excited by the synchronization marks S18 andY S 19, Fig. 5, several times, for example five, as early mentioned, iive pulses will appear at the output of the selector 39, Fig. 6, if the code ciently charge the condenser 111 that the voltage applied'V to'the grid 112 strike the triode 113.

Once thegtriodei'113 Lis struck, a positive voltage will appear at the cathode 116 of the triode 113 which will continue indefinitely. This voltage is applied to the gate 38,-Fig. 6, and will-'thus allow any synchronization pulse applied directly from the selector 37, Fig. 6, tol pass through the gate 38.

Referring again to Fig. 5, it will be seen that an addi-V tional ,synchronization marl: is provided at S35. .The

pulses which this markV S35 excites'in the photo-cell 30, Fig. 6, are applied to thegridv 117 of a second cold cathode `triode 118 causing it to strike. The anode 119 of the triode 118 is in parallel with the'anode 114 of triode 113 Vthuswhen the triode 118 strikes anegative-going pulse appears in the combined anode circuits 114 and 119 asia result of which the triode 113 will be cut off. 'Y

The triode 113 will then remain quiescent until the scanningrof another code commences.

Referring to Fig. 13, a cascade of cold cathode tubes as shown by the numerals 24 to28 in Fig. 4 and by the numerals M1 to M15 in Fig. 6. A starter pulse is applied to the grid 120 of'a starter tube '121 the cathode 122 of which`is vgrounded through resistance 123l and by-pass condenser 124. The anode 125 of the tube 121 is connected to a positive l-H.T. source through plate resistance 126.

The cathode 122 of the tube 121 is electrically connected through a resistance 127 to the grid 128 of a second cold kcathode tube 129. The anode 130 of the tube 129 is connected to plate` resistance 126 in parallel with the anode 125 of the tube 121, the cathode 131 thereof being grounded through resistance 132 and by-pass condenser 133. Again the cathode 131 of the tube 129 is electrically connected through a resistance 134 to the grid of the next tube and so on until tube 135 is reached which is connected as with all the other tubes, its anode 136 to the plate resistance 126 and its cathode 137 to ground through resistance 138 and by-pass condenser 139k and its grid 140 to the cathode of the preceding tube. f

In operation the pulse, let us assume'it to be a selected :pulse excited by a synchronization mark as in Fig'. 4, 'is

applied to the grid'1`20 of the starter tube 121. A positive-going voltage appears at the cathode 122 of the tube 121 and is applied to the grid 128 of the tube 129. The

cold cathode tubes used should be such thatthe'positive- The voltage of this pulse in addition to the voltage applied to the grid 128 from the cathode of the. starter tube 121 is sufficient to strike they tube 129.

When the tube 129 strikes, a negative-going voltage appears at the anode Aand henceV at the anode 125 of tube 121 thus cutting off the starter tube 1121.

This whole processis repeated, a further cold cathode tube being struck each time a new pulse is excited in the photo-cell 20 of Fig. 4 by another synchronization mark, each tube in turn preparing the next and cutting off the one previous, until the last tube ofthe cascade, tube is struck. This last tube 135 will cut olf the preceding tube and the positive-going voltage at the cathode 137 thereof is applied back to the grid 120 of the starter tube 121 thus preparing it for striking by the next selected pulse either to begin the scanning of a new block of the code or to commence-the scanning of the next code.` The pulses utilized to open the gates 18 of Figs. 4 and 6 are those appearing at the cathodes of all the tubes in the cascade except that of the starter tube 121.

Other components ofthe block diagrams such as low pass filters, short pulse generators and pulse wideners are well-known units in the artY and need not be describedin detail.

Itwill be understood that the form of synchronization marks used may vary greatly and that the type of code used may also vary. Similarly a number of sensing devices may be employed. The present invention is mainly concerned with an elfective system of using pulses excited by synchronization marks to enable the pulses excited by classification marks to be isolated and identified in relation to the positionof the classification marks in the code being read. f

While no specific mention has been made of what use the pulses appearing at the outputs of the gates 18 in the various figures maybe put, these pulses are identiliable and'may be used in Vnumerous manners by one skilled in theart. A use for these pulses specific to theV sorting of mail will be the subject of another application.

I claim:

1. An Vapparatus for detecting and reproducing in the form of electric signals information conveyed lin a patframe-work of rows and columns and synchronization marks of generally similar configuration to the classification marks disposed one at the same end of, equidistant from, and aligned with each column of rsaid frame-work to form a row substantially parallel with the rows of said frame-work together with an initial synchronization mark of different configuration than said classification marks disposed before and in alignment with a row of said frame-work, comprising a number of sensing devices so disposed as tov scan each one row of said pattern code including all the rows of said frame-Work and said row of synchronization marks, acascade of multi-vibrators one corresponding to each column in said code in electrical;v connection with the sensing devicen alignn're'ntl4 with;- said 1 row off synchronization-L marks, aI synchronization pulse selector in electrical connection with the sensa...

said frame-work being in electricall connection'7 witht the` electronic .gate electrically. connectedltoeach-y multi-vibra-V tor and-^corresponding to\.the,row beingscann'ed by that-1 particular sensing device.

2. An apparatus asI claimedV inclaimi 1 in which said initial synchronization mark is disposedbeforeA and aligned with'- a first rowof'the frame-work, asecondiinitial synchronizationmarkof adifferenti configuration-- than the classification marks is: disposed an equaly distance before and in alignment with a second'rowofsaidframe-y work not adjacent saidffirst row, affurtlierL synchronization: pulse selector is providedin electrical;y connection` wlth the sensing device scanning the said-secondi row of saidframe-work, anda single electronic gate :isfconnected to .both.saidisynchronization.pulse selectors, said separate Single gate being adapted to emit a. signal when' the pulses selected by both said synchronization pulse selec tors are applied to said single gate simultaneously, thel presence of a signal at the output of saidy single gate indicating correct alignment of said'patterncode.-

3. Ina method of identifying'articles.forpurposesfofiy sortation, comprising imposing a pattern code'upon' an article, said code including classilicationmarks disposed. within a vframework ofzrows and columns, atrleastf one starting mark disposed at one'. end of.` said framework Y and atleast one finishingl mark disposed at` the-"other end of said framework;` scanning saidcl'assification.marksto detect the individual classification.marks-andfproduce corresponding classification electric signals; separately identifying each such signalfor subsequentsortation off said article; the improvement of a verification procedure: comprising carrying out said-scanning step-'so as` toy gencrateI a vstarting signal corresponding to said starting mark. and a iinishing signal .corresponding-1to1 saidfnishing mark, in addition to saidz classification signals;

calculating fromV `the time relationship of at least' said' starting and finishing signalsthe distance-'between said starting and finishing marks;v comparing this'- calculatedf distance with a datum distancetofdetect any discrepancy therebetween, and generating Va rejection signal for cancellation of sortation in accordancewithsaid classification signals whenever such a discrepancy'is-detected.

4. The method of claim 3 wherein said; article' is.

moved relatively to the scanning meansand-said scan- 6. The method of claim 3, wherein said step,.of'.ca1`'V culating4 tiie'distance` betweensaid' starting and' finish- Y ingl marksY includes the step of; calculating from-i said' classsiii'cation signals'the number' of-I columns`- of f classi#- fication markssituated between said starting and iin; ishi'ng, marks. j

7i. The method'of claim 3, wherein-said step of ca1- culat'ingl the distancel between said@ startingv and` finish ing'- marks includes thestep; of determining the mean rate=offscanningof said code' and relating this ratek toVv the time-intervalbetween-Said starting and-finishing signals.

8. A method of' identifyingl articles' for purposes: of;

sortation, comprisingimposing upon an article a pattern' cod'e in the form ofrowsand columns of classification marks', scanning said' codeinv a mannerv such as to` scanA eachsuch mark aV plurality Aof` times and produce'aL plurality of spaced electric:4 pulsesy corresponding to suc*- cessive portions of-each said mark, separating the pulses correspondingv to-each mark from-the pulses correspondving to 'other-marks and widening anddelaying each' off said separated pulses to such extent that said: pulses` overlap` one another, and addingusaidV separated'A pulses' togetherl to produceV ay unitaryy signal representing eachsuch-mark;

9. Apparatus yfor detecting and reproducing in they form of electric: signals' informationA conveyed in a pattern code on an article, said code-having marks disposed within a framework of-rows andlcolumns, said apparatus comprising sensing means forv detecting saidl marks; means for moving said'article and said sensing-means to cause said sensing means-toscan said codeinY a-mana nere such as to seaneachof saidy marks aplurality of timesto forma plurality ofindividual-pulses eachl cor responding toa` successive-portion' of-v said mark, meansv for separating the pulses produced by afselected mark.y

and alow-pass filter for widening, delayingandadding said` separated pulses` to-produce a signal? representingsaid` mark.

References Citedzinethe file of this* patent:

UNITED-STATES PATENTSV 1,217,092v Hopp, Feb. 20, 19:17v 2,224,646 Friedman et.al., Dec. 10, 1940r 2,358,051 Broido Sept. 12, 1944r 2,513,112 Shepherd June 27, 1,950 2,575,034A Tyler. .Nov. 13, 1951. 2,580,741 Dickinsonv Jan.v l, 1952. 2,592,202 Smith Apr. 8, 1952 r2,593,206 Short .Apr. 15,'1952 2,609,828 Doust -s Sept'. 9, 1952v 2,612,994 Woodland et al. Oct.- 7, 1952 2,615,992 Flory et al. Oct.` 28 1952I 2,618,386 Samain Nov. 18,1952 2,675,170 Sebesta Apr. 13,- 1954.v 2,677,815 Brustman --.May 4, 1954v 2,731,200 Koelsch Jan. 17, 195,6:

OTHER REFERENCES Rectifier Networks for Multiposition Switching,

Brown and Rochester, June 1949, pp. 139-147, Proceed-i ings of the VI.R.E.