CA2194282A1 - Optical random access memory having multiple state data spots for extended storage capacity - Google Patents

Abstract

An optical memory (100) in which data is stored in an optical data layer (190) capable of selectively altering light such as by changeable transmissivity or polarization. The data is illuminated by controllable light sources (330) and an array of multi-surface imaging lenslets (210) project the image onto a common array of light sensors (270). Data is organized into a plurality of regions or patches (called pages) in which each page contains a field of data spots that encode multiple states or levels of data by the amount of transmissivity or polarization of that spot.

Description

~096/02056 r~
~ 2 1 94282 OPTICAL RANDOM ACC~88 MEMORY
UAVING MULTIPLE 8TAT~ DATA 8POTS
FOR ~ 8TORAG~ CAPACIT~

Rl~l K~;ltl~llNI1 OF TE/E 1NV~L~11UN
The invention cnn~rn~ method and apparatus of optically storing and retrieving ma3s digital data stored as light altering characteristics on an optical material and providing fast random access retrieval.
Optical memories of the type having large amounts of digital data stored by light modifying characteristics of a film or thin layer of material and A~cP~8~d by optical addressing without - hAni~Al v~ L have ~een pLu~osed but have not resulted in wide spread commercial application. The interest in such optical recording and retrieval te~hnolo~y is due to its projected ~Ap~hility of faster retrieval of large amounts of data compared to that of existing ele~L~ Lical --hAni such as optical discs, and ~-gnetir storage such as tape and magnetic disc, all of which require relative motion of the storage medium.
For example, in the case of optical disc memories, it is necessary to spin the record and move a read head radially to W096l02056 2 1 9 4 2 8 2 P ~

retrieve the data, which is output in serial fashion. The serial a~c~sing of data gPnerRlly requires transfer to a buffer or solid state random access memory of a data yluces~ù
in order to r - '-te high speed data addre3sing and other data operation3 of modern computer3. Other storage devices such as solid state ROM and R~M can provide the relatively high access speeds that are sought, but the cost, size, rnd heat dissipation of such devices when G~pRn~ to relatively large data capacities limit their applications.
r ~lP~ of efforts to provide the relatively large capacity storage and fast access of an optical memory of the type that is the subject of this invention are disclosed in the patent literature such as U.S. Patent 3,806,643 for PHOTOGR~PHIC RECORDS OF DIGITAL TNT~nRM~TON AND PLAYBACR
SYSTEMS TNrTlTDTNr~ OPTICAL sr~NNRR~ and U.S. Patent 3,885,094 for OPTICAL SCANNER, both by James T. Ru3sell; U. S. Patent 3,898,005 for a HIGH DENSITY OPTICAL MEMORY MEANS EMPLOYING A
MULTIPLB ~ENS ARRAY; U. S. Patent No. 3,996,570 for OPTICAL
MASS MEMORY; U. S. Patent No. 3,656,120 for READ-ONLY MEMORY;
U. S. Patent No. 3,676,864 for OPTICAL MEMORY APPARATUS; U. S.
Patent No. 3,899,778 for MEANS EMPLOYING A MULTIPLE LENS ARRAY
FOR READING FROM A HIGH DENSITY OPTICAL STORAGE; U. S. Patent No. 3,765,749 for OPTICAL MEMORY STORAGE AND RETRIEVAL SYSTEM;
and U. S. Patent No. 4,663,738 for HIGH DENSITY BLOCK ~RTRN~Rn SOLID STATE OPTICAL MRM~RTR-~. While some of these 3ystems attempt to meet the above mentioned objectives of the present invention, they fall short in one or more respects.

W096/02056 2 1 9 4 2 8 2 = r~

For example, some of the systems ~roposed above have lens or other optical ~Llu~LuL~ not capable of providing the requisite resolution to retrieve useful data density. ~he optical resolution of the data image by these prior lens systems does not result in sllff;r;rnt data density and data rate to compete with other forms of memory. Although certain lens systems used in other fields such as miuLuscu~e objectives are theoretically capable of the needed resolutions, such lens combinations are totally unsuited for reading data stored in closely spaced data fields. Another difficulty rnrollnt~red with existing designs is the practical effect of temperature and other physical distllrhAnr~R of the -h~n;rAl relationRhir between the data film or layer, the lens A~= ' 1; ~R and the optical sensors that convert the optical data to electrical signals. For example, the thermal r~rAnRinn effectg of even moderate density optical memories of thig type can cause severe misregistration between the optical data image and the read out sensors. Similar ~;ff;rlllties are encountered in the required registration between the recording process and the subse~u~..L reading operation3. Intervening misregistration of the high density optical _ ~ R can cause Riqn;fir~nt data errors if not total 1088 of data.
Accordingly, it is an object of this invention to provide an optical mass memory having random Arc~cRihil;ty in a relatively compact size comparable to or even smaller than tape and compact disc storage -hAni~ and yet still serving data processing equipment in the same manner that solid state 2l 94282 W096/02056 1~~

random acces3 --es move data into and from the processor's data bus.

SUNMaRY OF T~IE lNVt;h ~ lUN
Data is stored in an optical data layer capable of selectively altering light such as by ~hAng~hlP or storable transmissivity, absorption, reflectivity, pnl~r;~tion~ and/or phase. In the case of a transmissive data layer, data bits are stored as relatively trAn~p~rent spots on a thin layer of material 3uch as photographic film and are i~ min~ted by controllable light sources. An array of imaging lenslets project an optically enlarged image of the ;llnminAted data onto a fixed array of light sensors. The layer of data is organized into a plurality of regions or patches (called pages) and, by selective illumination of each data page, one of the lenslets will image the selected data page onto the array of sensors. Transmitted page data, in this c~e light pa3sed through the transparent bit locations on the data lnyer, strike different ones of the arrayed light 3ensors, thereby outputting a pattern of binary bits in the form of electrical data signals. By selectively and sequentially ~ lm;nPting different ones of the data regions (pages) on the data layer, corresp~n~ingly different data patterns are imaged by the coLle~y~..ding lenslets onto the same photosensor array, thereby Pn~hling many data pages to be multi rl PY~ at electro-optical speed onto the common phot~s~n~or array image plane.
~ i Ls of data storage and retrieval systems relatedto the present invention are disclosed in the above-referenced W096/020S6 21 94282 1 .1~ .

cQpr~nr7;ng ~7prl;~77-j~n SN 07~815,924 as read-only devices, write-only devices, and read/write devices. In accuLdance with the preferred ~ 7i L of the present invention, each ~ data location on the record, i.e., each data spot has three or more states and thus holds more than a single bit of information. The additional information, in the preferred '-'; L, is in the form of modulation of the light energy that reaches the sensor, such as by varied 7~r~n~ ivit absorption, reflectivity, or polarization.
The recording medium in the preferred form aB ~7.; ~c10~7 herein has a variable optical density, i.e., a variable amount of absorption of light passing through. Silver halide is one simple example of such a medium. When it is properly exposed and developed, it will have a range of "hl ~knr~88~ . The range can be con~;r7~red an analog value, or the range may be divided into several discrete steps or states in density to give a digital value (where digital is tertiary or larger multiples, not just binary).
An alternative is to record each data spot with a varied diameter. The amount of light energy that reaches the sen30r will depend on the area size of the spot. In addition, the energy will be modulated by the extent of diffraction by the relatively small spot compared to the wavelength of the ;lln~;n~tion. That is, when a data spot aperture becomes _ -r~hle to the wavelength, transmitted light will spread out. In the limit of a very small hole, the transmitted light will behave as though the hole is an isotropic point source.
This diffraction causes a reduction in the bit energy reaching W096/020~6 2 1 9 4 2 8 2 r~

the sensor simply because the edges of the light bundle are not caught by the first element of the lens system.
Another alternative i5 to record each data spot as a variable angle polarizer. In this case, the data source may be polarized and the polarizer at individual spots would vary between zero and 90 degrees relative to the source. A
variation is to u3e an llnpolAr;~ d source but place a general polarizer between the record and the sensor.
The~e and other feALuLes, objects, and advantages of the invention will become apparent to those skilled in the art from the following detailed description and App~n~d drawings.

.

BRIEF DES~n~ . OF THE DRAWINGS
FIGURE 1 is a section view in elevation of an optical memory having multiple state data spots for extended storage capacity in accordance with the preferred -'i L.
FIGURE 2 is a plan view showing schematically the layout of individually switchable light sources for illnminnting during reading the data region pages, each page containing the multiple state data spots.
FIGURE 3a is a plan view similar to FIGURE 2 showing the layout of a data layer in Arcnr~An~e with the preferred form of the invention and an associated FIGURE 3b is a greatly enlarged view of one of the hexagonal data regions for pages of the data layer illustrating schematically the or~hngnn~l 2~ ~rray of rows and columns of data spots in which the actual number of data spots may be greatly in excess of the density schematically shown in FIGURE 3b.

W096l02056 21 942 ~2 r~

FIGU~ 4 i8 a plan view similar to FIGU~ S 2 and 3a showing the array of one of the lenslet system elements, in this instance a refractive element in which the array of such refractive elements ifi in a hexagonal cell pattern that ~ Ls the h~Y~q~n~l data layer pattern shown in FIGU~
3a.
FIGV~ 5a is a plan view similar to Figure 4 of another of the lenslets elements, in this case being diffractive elements, and FIGU~ 5b is an enlarged view taken from one area of FIGUR~ 5a and shows the oV~rl ~rr; nq circular gratings or holographs of the diffractive elements making up this ~urface of the lenslet array.
FIGU~ 6 is a ~~gnif;~d fragment of the data layer illustrating data spots having variable optical (in this ca3e, variable tr~n~ ivity) and in which certain of the data spots are reference spots of known density for nor~~l;7;nq the _ _ ~nts during readout.
FIGU~ 7 is a greatly enlarged fr~3 Lary view of a data layer similar to FIGU~ 6 but in which the field of data spots are of variable area.
FIGU~ 8 is a greatly enlarged fragmentary view again of a data layer similar to FIGURES 6 and 7 but showing the individual data spots in the data field as variably pol~r;7~d.
FIGU~S 9a and 9b are electronic block ~i~3 of the readout control electronics associated with the optical memory of FIGU~ S 1 through 6.

Wo9~0205G A ~~
.

FIGURE 9c is a timing diagram showing the sequence of signals and operations assori~ted with the control electronics shown in the block ~i~gr~q of FIGURES 9a and 9b.
FIGURES 9d, 9e and 9f are schematic ~ r~C depicting the topology of a multi-bucket (or multi-well~ charge coupled device array that i8 in~oL~oL~ted into the data image sensor array of the optical memory of FIGURE 1 and which also appears schematically in the FIGURE 9a block diagram of the control electronics.
FIGURE 9g is a timing diagram showing the sequence of signal operations associated with the multi-bucket charge coupled solid state sensor of FIG~RES 9a and 9d through 9e.
FIGURE 10 ic a 8rh ~ir diagram of the recording control electronics for recording the data spots as multiple levels or states, such as density, area (size) and polarization.
FIGURE lla and the enlarged fragmentary view in FIGURE
llb are schematic diagrams of an alternative 'i ~ of the individually energizable light sources for use in the optical memory of FIGURE l.
FIGURE 12 is an enlarged fragmentary view similar to Figure llb showing a further alternative ~ of the individually energizable light sources.
FIGURE 13 is an enlarged fragmentary view similar to FIGURES llb and 12 showing still another alternative : 'lo~i- t of the energizable light sources.

W096l02056 2 T q 4 2 8 2 r~

nR~TTRn DESCRIP~ION
With reference to Figure 1, an optical random access memory 100 incorporating a preferred : -'i L of the multiple state data spots in accordance with the invention i5 shown to include a housing 110 of a regular polygon shape, in this instance in~ ing top and bottom walls llOa and llOb, respectively; opps~;ng side walls llOc and llOd, and a front wall llOf (cut away) and back wall llOe as ~i ~rl ~~e~ in related coppn~ing U.S. Application Serial No. 07/815,924 and international Application Serial No. PC~/US92/11356, the pl~hli~hP~ spe~if;~ations of which are incoL~oLated herein by reference. Although not critical to the present invention, housing 110 is substantially bisected into left and right ~h: ' ~ each of a gPn~rnlly cubical shape in which the right hand chamber contains the electronics and optical _ _ ~8 for reading data by projecting an imPge of each selected region or page of data spots onto a photosensor array ~;~p:~ ' adjacent the bottom wall llOb at the right hand side of the housing. The left hand chamber of housing 110 contains the electronics, light sources, and other optics that function to compose, project and record data images onto a blank data film or a recordable data layer positioned in the right hand chamber as dp~r;hpd more fully below.

;REAr~ UN~
~o more easily understand the construction and operation of the c ~;nPd write/read system shown in Figure 1, only those elements of the system used for reading data will be W096/020~6 ~ ~ 2 1 9 4 2 8 2 ~r~rr;hed first, then the write (or record) elements will be introduced and PYpl~;n~ Thus, with reference to the right hand side of the bisected house 110, there is provided adjacent the upper housing wall 110a, an array of light source drivers 130 formed in an integrated circuit and coupled by micro leads (not shown) to an array of solid state photoemitter elements serving as the read light sources 150.
~ight sources 150 are mounted on a circuit board or other integr~ted bL-u~Lul~l unit to fix the source3 in a closely packed dense light pattern. Immediately beneath and p~r~l 1 r~l to light sources 150, a unitary data/lens card bLLu~Lu-~ 170 i8 removably mounted for storing on a data layer 190 having multiple state information spots organized in an array of data pages or regions. A _ 1~ Lary lenslet array 210 i8 bonded to data layer 190 and has a plurality of lenslets ~;~pos~ in precise, fixed optical registration with each multi-spot data region or page. Unitary data/lens card bLLu~Lule 170 is fabricated as a bonded unit so that the array 210 of lenslets i8 fixed in spatial relation to the data layer 190 and so that the structure 170 iB readily removable as a unit from housing 110 of the optical memory 100 without disturbing this critical optical spatial relationship.
In the now preferred form of the data/lens card bL-u~LuLe 170, the lenslets 210 include one or more refractive and/or diffractive elements 210-1 and 210-2 shown separately in Figures 4 and 5a. Following the card structure 170, the rays of data images projected from data layer 190 and lens array 210 as a result of ;lln~;n~tion by light sources 150 are W096/02056 2 t 9 ~ 2 82 further redirected by field lens 250 having an ~u~LLuLe that ~ - C8PC the entire depth and width of the right hand chamber of housing 110. Hence, field lens 250 in this : '; L is of gPnPr~l1y rectangular 3hape about its perimeter and otherwise ha3 conventional sphPrirAl or plano optical surfaces.
Beneath the field lens 250 there i5 a cavity that allows for the optical cu..v~Ly~noe of the data image rays which in turn form the data image onto an upwardly facing common image plane of sensor array 270. An intervening beam splitter 310 functions during recording of data, as described below, but is essentially transparent to the projected read image between field lens 250 and sensor array 270. The data image projected onto array 270 in this preferred _ -'; L is in the general shape of a hexagon, or roughly circular, cnnfn~m;nrJ to the image generating data pages on layer 190, and the light sources and optics; however, the sensor array 270 itself may have a substantially rectangular or, in this case, square perimeter. Beneath sensor array 270 is located the sensor interface circuitry 290 which is preferably fabricated as an integrated or printed circuit wafer of similar ~h; rknPc8 and rectangular perimeter to complement and lie subadjacent sensor array 270 as shown.
Thus, in operation, a single page of data selected on data layer 190 by energizing a chosen cell of light sources 150, causes a data image to be generated that has roughly the shape of a hexagon and fills the image plane on the upper surface of sensor array 270.

W096/02056 2 ~ 9 4 2 8 2 The individual data spots within a single data page are here arranged as shown in Figures 3a and 3b in close packed fields and at densities that use to advantage high resolution optical films and other record media inr~ ing but not limited to photorhPmicAl films. To provide storage competitive with other types of memory, the data spots must be in a size range of 2.25 to 0.5 microns and a center-to-center spacing also in that range. II~Leuv~I, in accordance with the present invention, each data spot encodes three or more states and preferably from 4 to 64 states or levels by such plucesses as variable density, size or polarization, 80 that more than one binary bit of data is retrieved from each data point or spot.
Each data page is then formed by the amount of individual data spots that can be collected and grouped into a single hexagon cell as shown in Figures 3a and 3b and at the preferred density range of 2 x 107 - 4 x 108 spots per cm7, it has been found that about 1 x 106 spots of data per page (or region) is an advantageous quantity that results in the generation of a data image after r-gn;f;ration that can be reliably sensed by photosensitive elements of sensor array 270. In this case, the preferred ~ provides an optical m-gnification through the various lens as~: ' 1;P~ of approximately 20 to 30 times. Thus, ~snminrJ a ~gnifir~tion of 25, the spacing of the projected image elements on sensor array 270 is on the order of 25 microns and a hexagon cell consisting of a page of data will, in this ~ , contain one million data spots per page that are imaged onto a coLLe~ol-ding number of photosensitive elements in array 270, to yield data of at W096/02056 2 ~ 94282 least 4 times the number of spots or 4,000,000 bits of data per page.
The particular 3tructure and operation of the sensor - array 270 and various alternatives to the preferred -'i are described in greater detail below. For the present, however, it will be appreciated that each data spot causes a photosensitive element of sensor array 270 to detect the multiple states, or levels of photo 3ignals. Although different forms of data layer 190 may be employed, in the present preferred : -'i L data layer 190 i8 a variable density tr~n~miQsive mask or film (a fr-_ L of which is shown in Figure 6).
It will thus be seen that the read elementg and op~rat; nn of optical memory 100 provide for a~c~sing each of hundreds of pages of data having, for example, a minimum of 4,000,000 bits per page at 1 micron spot size. The read out of data from sensor array 270 i5 des~r;h~fi in greater detail in connection with Figures 9a, 9b and 9c, but briefly it is seen that by selecting a single data page on data layer 190 by energizing one cell of the light ~ources 150 an entire page of 4 x 106 bits is made available at the output interface circuitry 290 associated with sensor array 270 at speeds typical of ele~LL~ ~Lical switching, e.g., equal to or less than 100 n~nns~c.,~c. Data words that ~ake up different portions of the entire page may be addressed, such as a column or row of data on each page, or the entire page may be output.
- Each row or column of data within an ~rc~ page may contain as many as 4,000 data bits, hence making fast retrieval of W096/020s6 2 ~ ~ 4 2 ~ 2 . ~
.

~Y~ee~i ngly long bit worda of this magnitude within the rApAhil;ty of the optical memory 100. In terms of rli ~irn~, a 1,000,000 (106) spot page imaged on sensor array 270 will occupy a hexagon that would fill an area of 6.5cm2 or about 1 square inch. Similarly, at the above stated preferred density r~nge of 2 x 107 - 4 x 108 spots per cm2, an area of 6.5cm2 (about 1 square inch) contains as many as 640 patches or pages of data. In effect, the multiple pages of data bits are mult;plr~Yr~ onto the image plane at sensor array 270 by electronic ~witching of read light sources 150.
In the preferred ~ i L, each page 195 of data layer 190 contains the above spr~if;r~ number of data spots, such as 106 spots per page at a density of about 1 micron spacing.
Each spot 196 other than reference or fi~ ;Al spots that are discussed later, have a variable density that is depicted in Figure 6 by a LL _ L of page 195 illustrating the levels of different density by the cross-hatched circular data spots 196-1, 196-2, 196-3 -- 196-n.
The varied level or density of each data spot 196-1 through 196-n is recorded by grey scale recording tr~r~hniqllr~
discussed more fully below, or by photographic copying processes using photo~hr~;rAl devr~ L. Thus, for example each of data spots 196-1 through 196-n stores multiple levels or states, for example 4 to 64 individual states or levels, in order to L~ se.lL 2 to 6 binary bits of data. A four level density or trAnr~iR~ivity with adequate dead bands between levels would thus I~Iese.-L two states raised to the power of two, or four two-bit binary value~ of 00, 01, 10 and 11. Data W096/02056 21 94282 r~

apots having 8 separate levels of tr~nR~i~sivity would thus store 3 bits or two raised to the power of three binary values of 000, 001, 010, 100, 011, 111, 101 and 110. It is thus seen - that the amount of data stored and retrievable from data layer 190 is multiplied and enhanced greatly by the number of dGta levels or density levels that can be stored at each spot. In order to faithfully retrieve this data, the sensor and associated electronics must be able to differentiate effectively between the density levels, in this case transmissivity levels, of each data spot. An analysis of the detection pr~ccs~e~ as well as the uniformity of the data layer material and the varied density at each ~pot shows that each sensor element on array 270 must be able to measure the amount or level of data spot light energy to an ac~uLG~y of one part in 2~th, with a confidence of one part in 18, where n is equal to the number of bits to be r~yL~se.lLed by a single spot. Put in another way, the signal to rms noise of an ordinary single bit two-state system of maximum and minimum transmissivity, should be at least 25db or roughly 18:1 s/n ratio. For a multi-state or level spot system as in accordance with the present invention, the overall s/n ratio must be 2~th x 18:1. Thus for a system that stores at each spot a six bit value or 26 = 64 levels, the overall s/n ratio must 64 x 18:1 or 1152. While the individual dynamic range of available solid state sensor~ of the charge couple device (CCD) type is on the order of 2000:1, the variation in sensitivity between sensor element to sensor element is typically on the order of 5%. This inter-sensor effect is W096/02056 r ". ,~
2~ ~2~2 called photo response non-uniformity or PRNU. This means that the noise attributed to the sensitivity variation from sensor to sensor cre~tes a noise constraint that limits the range of ,the sensors to about a ratio 20:1. While this may be suitable for some applications in which the number of states to be stored and retrieved from each data spot is relatively ~mall, not more than 3 states, a preferred ' 'i L of the invention provides for calibrating and norr-l; 7ing the data layer to sensor readout of the data image to yield an ~pAn~A
practical dynamic range of the data layer to sensor operation and moreover greatly enhance the repeatability and r~l;Ahil;ty of the optical memory. While one alternative t~chn;que in accordance with the present invention would be to calibrate and trim out the individual sensor elements of array 270 to minimize the sensor to sensor differences in sensitivity, a preferred way is to configure the data layer and the electronics associated with the readout of the sensor array in such a manner that the sensor elements and the multiple state spots are both calibrated to a known level and then norr-l;z~d from sensor element to sensor element to achieve the needed signal to noise ratio for reliable data storage and retrieval.
Thus in the presently preferred ~ Al;hrAf;nn of the sensor array is achieved in part by selecting one or more pages in the data layer as a reference page that is recorded with all ls or maximum transmissivity at each of the data spots within that page. In the '~~; L ~;~nlosed, data layer l90 in Figure 3a is shown to have a centrally located page 195-rp as a reference page of all "ls". Prior to W096/02056 2 1 9 ~2 82 P~

reading data from the other pages of the overall data layer or chapter of pages, the ,ef~L...ce page 195-rp is ;llllm;nAted by its associated one of light sensors 155 causing the image of all 18" to activate the sensor array 270. ~sllm;ng perfect _ Ls, all of the sensor elements of array 270 would thus be ;llllminAted or energized to maximum L~onse, thereby estAhl; ~hing a reference or base level of sensor element sensitivity. The sensor output for each reference "1" is stored in a reference bucket provided as an All~; 1; Ary storage bucket or location in the CCD array as shown and ~ r; hed in connection with Figures 9d through 9f. Thereafter the variable level data is read out from a page 195 of chapter data layer 190 and the resulting charge or signal level shifted into a data signal bucket underlying the sensor elements and being available along with the reference value in the reference bucket for readout. From the storage buckets of the CCD array, the reference value and the data value are outputted into electronics that n~rr~l; 7e the sensor to sensor variations and produce a grey scale or variable level signal which is then analogue to digital (A/D) converted to yield the ~extended amount of data available on each page.
Thus with Ief~Le..ce to Figure 9a, sensor array 270 is associated with readout control electronics 290 to enable detection of the multiple states or levels of the sensed image resulting from the variable trAn~ sivity or grey scale spots 196-1, 196-2 -- 196-n of page 195 as shown in Figure 6. A
~ data request ;n~ ;ng page address and sub-page addressing, if required, is applied through an interface 290-1 that in W096/02056 2 l 9 4 2 ~ 2 turn is ;~a~P~ over a user bus to address buffer 290-2 of the interface electronics 290. The address data in huffer 290-2 supplies x and y decode addresses that select a data page and cause page or light source drivers 130 to energize the selected page light source. Initially during page readout, the light source driver r~rrP~pon~i ng to the reference page 195-RP (Figure 3a) is selected and through the sequence of operations ~P~rihPd helow, the reference "18' for the entire page are imaged onto sensor array 270 thereby causing reference values for each of the sen30rs of the array to be stored in one of the charge coupled device buckets.
In addition to the page addres6, buffer 290-2 i8 loaded with column and row ad~L~sse6. The column addresses for each page are monitored in a column counter 290-6 that, in conjuncticn with a start signal on lead 290-14, controls the phase and operation of sequence and timing circuit 290-7 that in turn provides the proper sequence and phase of timing signals to page light source drivers 130 and to output devices that include norr~ e and A/D converter 290-9, buffer register 290-10 and multiplexer 290-11. A clock 290-8 receives control signals from sequence and timing circuit 290-7 and produces a needed - 1 ~ of variously phased clock signals to sensor array 270 for shifting 3ignals into and out of charge couple storage buckets that underlie the sensor elements as PYp1~;n~ more fully in connection with Figures 9d through 9g.
Figure 9b shows the preferred construction of norr~ P
and A/D converter 290-9 and its connection to the rows of W096/020~6 r ~
~ 21 94282 charge coupled device sensor array 270. In this instance ~or ~; l;rity and clarity, an extension of a single row n is shown as element 270' in Figure 9b. The divisions of the - extension of row n of the LL _ ~ 270~ of the array are used to illustrate the organization and s~r~rafinn of the lef~.ence ~ignal in the form of a reference charge and a data level signal in the form of a data charge as these charges are shifted from the multiple storage buckets underlying sensor array 270 into nnrr-l;7e and A/D converter 290-9. For each row of sensed data in sensor array 270, a reference (value) charge and multi-state or level data (value) or charge are outputted into the nnrr-li7ation portion of the circuit shown in Figure 9b innlllAing a reference ~ lifi~r 290-9-1 and a corresponding data ~ 1ifirr 290-9-2 which couples the reference and data charge signals into a pair of clocked sample and hold circuits 290-9-3 and 290-9-4. These sample and hold circuits, when sequenced by clock 290-8 over a first of a group of timing leads 290-12, hold the reference and data charge signal for proc~cfiing by the norr~ n7 circuit including a fixed ; lifi~r 290-9-6 and a variable ; lifi~r 290-9-7, another sample and hold circuit 290-9-8 and a differential ; _lifiPr 290-9-10. ~he outputs of fixed and variable ; _lifirrs 290-9-6 and 290-g-7 are applied as the inputs to differential ; _lifirr 290-9-10 and are also applied as the reference and variable analog signal to A/D converter 290-9-12. Sequence signals from seqn~nre and timing circuit 290-7 ccntrol the timing of operations of the nnrr-li~e and A/D converter circuit 290-9 in accordance with the phase and W096/02056 2~ 94282 ~ . s ,~

timing diagram of Figure 9c ~i~c~sed below. In essence, the reference and charge values cause a di~erence value to be stored in s mple and hold 290-9-8 and when a nnrr-li2e ~ources control signal is applied, this value is held for adjusting variable ; lifiPr 290-9-7 to nnrr-li7e tbe source3 to the reference levels sampled when the center reference pages are flashed and re-~nrr-li7e the values whenever the pPrin~in full scale rnrr-li7~tinn data spot shows up in the page of data during readout as mentioned above in connection with Figure 6 as the leading spots for nnrr-1;7;ng 196-R and the re-~nrr-li7ing Bpotg 196-R placed in this instance at every eighth spot position in the field of data within each data page. A leading detect circuit 290-9-14 provides a detection of the first set of spots within the data field that represent reference levels and correspond to the first or leading spots 196-R in the left hand column of the data page 195 shown in Figure 6. ~his leading edge signal is fed back to sequence and timing circuit 290-7 to mark the start of a data readout sequence as ~P~r;hPd below in ccnnection with Figure 9c. A/D
converter 290-9-12 receives a convert timing signal over one of leads 290-12 as indicated from circuit 290-7 to effect the A/D conversion that is Lep ese.lted by the inputs to circuit 290-9-12 from the outputs of ~ rs 290-9-6 and 290-9-7.
The reference signal to the A/D es~hli~hP~ the .~fe~ ce current within the UBual - _-ri ~nn ladder of the A/D. The conversion of the data signal takes place as a fraction of the reference, rather than as a fraction of a fixed voltage or current level.

-I W096/02~56 2 1 94282 P ~

Thus with the data readout direction indicated by arrow 197 in Figure 6, a leading edge of an array of data spots are for quantifying and nnrr-l;7ing the sensor electronics and for - this purpose the3e leading spots 196 R are recorded with the greatest level of f rAn~i csivity which may be Le~L~s~nted by the value "1". In the direction of readout 197 there is also an additional set of quantifying and n~rrol;7ing reference spots 196-R for re-nnrr-1i7ing the sen80r readout as the data is retrieved from the array of spots along the page.
10With reference to Figure 9c, the sequence of op~r~ n~
of the circuits of Figures 9a and 9b is shown to - e with a data request timing signal that in turn initiates an address reference page source signal followed by a signal for pulsing the reference page source, that i~ reference page 195-rp shown in Figure 3a that contains the full scale or all "18". When the reference page is flashed by pulsing its light source, the resulting reference image on sensor 270 is stored by a timing signal that shifts reference charge into a reference bucket channel of the CCD network underlying the sensor elements of array 270. The system has now stored the needed reference level for all of the individual sensor elements of array 270.
At this time, and as shown in Figure 9c, the input data request causes the appropriate data page to be addLessed as indicated by a signal address data page source which in turn causes a signal to pulse the data page source. The multiple data levels or states from the data page spots cause the ~ensor elements to be energized to different levels. ~he iensed signal level is in turn temporarily stored and shifted W096l020~6 2 1 9~82 P l/u~
.

out of the sensor element of the array into a data output channel. Now a signal i8 applied to the sensor array 270 tor addressing the charge coupled device row and word, which in turn starts the charge coupled device clock 290-8. The resulting operation of clock 290-8 is ~Pc~rihpd more fully below in connection with Figures 9d through 9g and relate8 to the manner in which the sensed spot signal levels, both reference and data, are shifted from charge storage wells or buckets out of the array and into the input ~ ~li f; ~rs of nor~l; 7e and A/D conversion circuit 290-9. The next timing signal in Figure 9c is the feedback signal from the leading detect circuit 290-9-14 (Figure 9b) that reflects the pLes_l-ce of a full scale n~rr~ ing value in a first position of the row of image data to be outputted and which initiates a nnrr-1i7nti~n aequence in~;r~tPd by the signal entitled "Nnr~-l; 7e Sources~ ~S/H A) referring to the operation of sample and hold circuit 290-9-8 of Figure 9b and the setting of the gain of ~l;fiPr 2990-9-7 that cauge nnrr~li7nti~n of the individual reference spota interspersed on a data page with the earlier stored values from reference page 195-rp.
The timing and sPquPnre continues with the reading of the next 3pot, which will be a data spot COLL~ ;ng~ for example, to data spot 296-1 of the row of data spots 296-1, 2, 3, 4 -- n.
~he ensuing timing signal operates the A/D conversion as a convert pulse input to circuit 290-9-12 (Figure 9b) which is followed by signals causing the output of the A/D conversion to be stored in buffer register 290-10. Thereafter the sequence continues to read data spots in~ ing the re-WO 96/02056 2 ~ ~ ~ 2 8 2 norr~l i 7~f i o~ reference spots in this instance at every tenth location in a row until the end of a data row marked by a timing signal referred to as end of row (S), at which time a - stop clock signal is ~Lodu~ed and ~fter a delay, a timing signal issues indicating that all data are now available at buffer register 290-10 and mult;pleY~r 290-11.
It ~hould be understood that the sequence of lefeL~IIce source flash and data source flash may be interchanged. The only resulting change would be that the inputs to the 0 ~ ,1 i f i ~r8 290-9-1 and 290-9-2 would be interchanged.
~ow, with reference to Figures 9d-9g, to reduce contributions of photo L~ .se non-uniformity (PRNU), a sensor architecture is advanced which allows in each data spot for the dual use of the photo element. A reference pattern may be imposed on the sensor and the resulting charge stored in an interline transfer register, then the signal of interest is sampled and the two results (stored in separate parts of the output register) are clocked out sequentially.
The charge coupled device ~L u~LuL_ is shown in Figures 9d and 9e. Figure 9e shows the placement of the polysilicon layers. The Poly 100 electrode is driven high to create a depletion region for the integration of photo charge. The charge A~ ted under this electrode is eas3ed to a well or bucket under phase ~2 in the transfer channel by pulsing this E~hase high, as the integrating electrode (Poly 100) is forced }ow. The charge is then moved to the electrode under phase ~4 by pulsing phase ~3 high as ~2 goes low and then phase ~4 high as ~3 goes low (all with the Poly 100 electrode low). A

W096~2056 2 1 942 82 .

subsequent ~ nAtion of the sensor can place a 3econd sample of charge under the phase ~2 electrode. These two samples from the 3ame photo diode A~ h a mean3 by which a reference pattern can be used to correct for PRNU in the data image. Polysilicon layers 200 and 300 make up the two clock phases used to actuate the output transfer register.
The Poly 100 layer is closest to the sub3trate surface 50 it has sole control of the surface potential nn~rneAth it.
Figure 9e shows the under layers. For the preferred : '- ;- L, a surface channel device is used. Alternatively, a buried channel structure can be employed. There is a stopping implant 305 on both side6 of the interline trAn~f~r channel 315 for lateral charge con~A; L. The exception to this is under phase ~2, where there is no implant interdicting flow of charge from the photo 3ensitive portion of the data 3pots. This implant 305 is of the same sense as the substrate but of a higher cu..ce.lLL~tion, such as a P+ type, for a device which uses electron3 as signal. To force directionality on the charge transfer, with the minimum number of clock pha~es, a barrier implant is also ;n~ d. This is also the same 3ense doping as the substrate (P typeJ, and somewhat higher in concentration (though not as high as the channel stopping region). The existence of this implant forces the channel potential in this region to be lower than in the area without it, thus charge will move along in the direction away from the barrier implant as the electrode voltage is reduced. Figure 9f shows the layout of these CCD devices in a sensor array.

W096102056 Z 1 9 4 2 8 2 L~,llll... _ n~
~l Figure 9e illu3trates the timing of the 3ignal3 used to shift out the CCD reference and data signals.
Figure 10 showa a preferred configuration for the ~ r~cor~; ng electronics. To record data onto the patche3 or pages 195 (Figures 3a and 3b), a recording process similar to that disclosed in the a~uv~ - -Lioned cop~n~ing Arrlira~inn~
serial No. 07~815,924 and PCT/U592/11356, is employed using the page s ---~ consisting of recnr~ing light sources 330, imaging lens 350, light valves 370 and recording interface circuits 295 as shown in the left hand chamber of Figure 1 for forming and directing the recording image light onto the dnta layer 190 via beam splitter 310. The general functions and operation of the3e recording elements are ~; ~rl ose~ by L~fe~ ce to the ~rAw;ng~ and written spe~;f;~ation, incu,y~L~Led herein by reference, of the ab~ Lioned related cor~n~;ng patent applications. Briefly, the page to be L~colded is selected by addressing a ~p~c;f;~ r~C~r~;ng light source 330. The light from that source is imaged onto the selected patch by lens 350, modulated by light valves 370.
One spot at a time is recorded by opening a CULL~ -.A; ng ~valve and energizing the selected source at the required energy level.
For this purpose, the recording interface circuits 295 llre constructed as 3hown in Figure 10 to include a u3er interface buffer 295-2 for 3upplying addres3 information, ~n~ln~;ng a page addregg, for selecting a source driver - through buffers 295-2-1, 295-3 and 295-4. The 3pot location ~ithin a page is selected by an addres3 on 2g5-2-2, which is W096l020~6 2 1 9 ~ 2 ~ 2 decoded by x and y axis decoder3 295-11 and 295-12 to select a spPci f; ~ valve in 295-13. Several spots are recorded in 3equence, so this addre~s is for the first starting spot.
Data words loaded into buffer 295-2-3 co-function with the start address in buffer 295-2-2 to cause the light sources to be driven at certain multiple states or levels through a D/A
converter 295-8 and source _ ~?tor 295-10 controlling a recording light pulse generator 295-6 as indicated. The energy levels which the light sources are driven at ~PtPrminPC
the degree of transparency (hence trAn~ ;vity) Iec~Lded (Figure 6) of each data spot. In such case, the density is a function of the ~hP~inAl reactions that are mediated by direct photon-electron excitation yIocesses. Another such material would be photon bleached dye, preferably in a plastic binder.
Alternatively, the energy of the rPnnr~; ng light sources may burn different size holes in the d~ta layer record (Figure 7) as described below. A word count buffer 295-2-4 governs a timing and sequence generator 295-12 that sn~cPqsively pIoduces signals as indicated for selecting the next valve address, loading data registers, pulsing recording light ~ource, in~L Ling valve address, de~L~ Ling word count, looping through each suncP~sive recording sequence and producing a done signal when f i n; ~hed.
When a record having variable hole size i8 to be recorded (see Figure 7), a data record having a medium with a sharp energy thre3hold is used. Since the recording spot cross section is cone-shaped (close to gA-l~s;An), the resulting hole size burned into the record depends on the light energy .
W096/02056 P~~
~ ~1 9~Z8~
8~1rrl; Pd~ intensity and duration. To record the variable polarization spots in the record shown in Figure 8, a material such as a ~LL~3sed li~uid crystal may be used with variabl~
A pnl~ri7~h;1;ty between zero and 90 degrees, responsive to the selected energy level in the page o -a--.
The record r~t~r; Al could also respond to thermal energy, as delivered by an optical beam. Variable diameter data or spot holes can be recorded with thermal processes as well as high gamma photonh~mir~l systems. In the thermal case, photons are still the source of energy just by gross absorption p.ucesses. The materials such as dye-polymer, thin metal film, ; ~huus/crystalline phase change mixtureS (e.g., chalcogenide glass), or magneto-optic are suitable. The problem is that these materials are binary. The t atu is proportional to the write energy Snrpl; ~d, but when the temperature rises to the action point, i.e., the melting point, a different ~~n; ~ makes the actual change. It is surface tension in the case of dye-polymer or metal films, phase growth for the phase change materials, or a domain flip by an external field for magneto-optic. This se~nn~Ary effect is not controllable in the same sense as the photnnh~m;
The only effect that is controllable is location of the transition point, that is, the edge of the hole (or spot), thus making thermal responsive records suitable for variable hole size data, as is the example shown in Figure 7.
Figures lla, llb and Figure 12 show an alternative configuration of the light sources for ; ~l~ Ling a different quantitization and nnrr~l; 7 /tion process. As an 21 q4282 W096/02056 r~

alternative to the terhniqU~- for quantitizing and ~r~-1;7;n~
the 3ensor cells by flA~hinq a reference page of solid "ls" is to flood the sensor array prior to data readout with a uniform light. For this purpose, the sources shown in Figure lla, and in greater detail in the enlargement in Figure llb, have two color or wavelength light sourcea within each page or patch cell. One of the colors, e.g., C1 = ~1, is selected for optimum resolution and imaging by the particular lenslet and field lens optics and is energized to readout the individual data spots. This is called the "gQod" color light. Each source cell also ;n~ p~ "bad color, in this case L~yL~ ed by the color C2 = ~2, and these source devices are flashed at the bPginning of a data read seq~l~n~e to flood the sensor with ill nmi nntion of uniform light that enables the quantitization and n~r~l;7Ation circuitry ~Ps~r;he~ above to m; n;m; 7P sen80r to sengor variations in photo sensitivity.
Color C2, the "bad" color, is snff;~;pntly different from the lens design color C1 that an image is not formed. Thus the b~d'~ color is pulsed using the light source cells C2 (Figure llb) and the c~LL~ ;ng sensed level at the sensor array is shifted into holding wells or buckets as described above in connection with the pulsed center reference page of all "ls".
After the "bad" color C2 sources are pulsed, the data is read out by energizing the "good" light source by flA~h;ng the elements Cl of the source array. Since the correction light, or "bad" light, is coming from the very page from which data is to be read, any geometric biases due to angle or lens Pff;~iPn~y differences are automatically , ~~ted. While there are various alternatives to thf r~ of the "bad"
light sources C2, a preferred approach is shown in Figure llb in which the sources Cl and C2 are ~;Ag~nAlly distributed within the source cells as indicated, hence four sources are used--two Cls and two C2s--alternated at 90~ from each other.
Two of these cells, the pair of Cls or the pair of C2s, are flashed at any given time. This tends to even out the effects of the asymmetric arrAn~; L of the sources. It may be ~e~irAhl~ in i ~ ing this : ' ~i t to use a diffuser to 3mooth out spatial noise, e~pP~iAlly when LEDs are used for the sources. A holographic, diffractive ~iffn~r is preferred over ground glass diffusers. The holographic element more ~f f i ci~ntly directs light rays in a pseudo-random way, rather than scattering and hence causing losses of the source light when a ground glass diffuser is used.
An alternative to the diagonal or 90~ distribution of the ,good and bad light 30urce elements shown in Figure llb is to place the good light sources Cl at the center of each of the h~ y, ' cells as shown in Figure 12 and make these sources ~omewhat larger than the "bad" source elements C2 which are distributed on the divisions between source pages as indicated. Thus each light source cell 155'' as shown in Figure 12 is primarily made up of the good source C1. ~o flash the "bad" light sources, the distribution of sources C2 around each cell are energized at one time for any particular l?age .
As yet another alternative, the opaque character of the medium may be c ~~~~ of a dye. The data spots have a w096~2056 2 1 9~282 ~ u~

variable amount of dye to provide the attenuation required by the data to be recorded. The source color selected for data rePdout, Cl, is variably ~h3~rhP~ by the dye, and cau3es an image of the data to form on the sensor array when Cl sources are flashed as ~i~CU~5~ before. The wavelength of the second source color, C2, is close enough to Cl that a good image is formed, but the wavelength is outside of the absorption band of the dye. When the reference source, C2, is flashed there i3 substantially no absorption and all "13" are imaged on the lû 3en30r, which are 3ubsequently 3hifted into the reference bucket3 as before. This alternative has the advantage that n~rroli7~tion is available for each individual data spot as well as the COL~ ; ng sensor. Further, the 3ame dye technique can be applied to the variable diameter spots (Figure 7). In this ca3e, the dye ha3 only two valve3, high absorption and no absorption, i.e., a hole. The reference C2 light will pass through all spot location~ and again all "15 will be imaged on the sensor.
With reference to Figure 13, a still further alternative ~ is shown for ; 1 ~ing the reference light source to quantitize and ~nrr-l; 7e the sensor array. In Figure 13, a group of source cells is shown to include a distribution of both polarized and unpolarized sources P and NP 3ub-page 30urce3. A pair of polarized P 30urce3 and a pair of unpolarized NP sources may be used as illu3trated, arranged ~;~;1 ~rly to the good and bad light 30urce3 of Figure llb, at diagonally oppo3ite or 90~ rotated position3 a3 depicted.
This type of light 30urce would be u3ed preferably with the W096l020~ 2 1 9 4 2 8 2 P~

variably polarized record shown in Figure 8. Alternatively, the reference source, instead of being non-pnl~ri 7e~, could be circularly pni Ari 7~d 50 as to be transparently passed through the variably polAri7~d data spots for uniform energization of the sensor elements. As in the case of the "good" and "bad"
light ; '-'i L shown in Figures lla and llb, and Figure 12, the '-'i- L of Figure 13 does not require a separate reference page as does the preferred 'i L de~rrih~
herein before. The readout process is the same as ~rrihe~
above using the dual well or dual bucket CCD sensor array.
This attenuation also has the advantage that each individual data spot and coLLe~nding sensor can be nnrr~li7~d.
While only particular ' ';- Ls have been ~; ~rl n~
herein, it will be readily apparent to persons skilled in the art that numerous changes and ~;f;rations can be made thereto, ;nrlll~;ng the use of equivalent means, devices, and method step3 without departing from the spirit of the invention.

Claims (6)

I claim:

1. An optical system comprising:
an optical data means for storing data as light altering characteristics and being organized into a plurality (P) of juxtaposed data regions each having capacity to store (N) data spots, in which at least certain ones of said data spots have ~3 light altering states to store >1 bit of data.
controllable light source means for selectively illuminating at least one of said separate data regions of said optical data means;
data imaging lens means having a plurality of juxtaposed lenslets each being shaped and arranged in such proximity to and in optical registration with a separate one of said juxtaposed data regions so that the image resolving power thereof is substantially uniform over the field of view of that data region to form an image thereof on a common image surface spaced from said data means and lens means;
sensor means having a plurality (S) of juxtaposed light sensor elements arranged at said image surface for sensing data as a light image corresponding to an illuminated data region, said sensor means including light state detection means for detecting at each of said sensors the ~3 light altering states of said certain ones of said data spots as imaged onto said common image surface.
data signal output means coupled to said sensor means for outputting data signals representing said >1 bit of data of each of said certain ones of said (N) data spots of an illuminated and imaged data region.

2. The optical data system of claim 1, further comprising sensor element normalization means for sequentially flashing certain ones of said controllable light source means prior to or after the selective illumination of said separate data regions for creating a sensor level reference, and sensor level output adjustor means for adjusting each sensor element signal level output as a function of said sensor level reference.

3. An optical data system comprising:
an optical data means for storing data as light altering characteristics having a density on the order of one data spot per square micron and being organized into a plurality of juxtaposed data regions each adapted to store a field of data spots each of which has 3 or more light altering states for storing more than 1 binary bit of data;
controllable light source means for selectively illuminating at least one of said data regions of said optical data means;
lens means having a plurality of juxtaposed lenslets configured to resolve objects on the order of one micron or less and each arranged proximate and in optical registration with a separate one of said juxtaposed data regions for forming an image of the field of data thereof on a common image surface spaced from said data means and lens means;

sensor means having a plurality of juxtaposed light sensor elements arranged at said image surface for sensing said data as a light image of an illuminated data region, and including means for discriminating between said 3 or more states of each of said data spots; and data signal output means coupled to said sensor means for outputting data signals representing said data of more than 1 binary bit per data spot of an illuminated and imaged data region.

4. An optical data system comprising:
an optical data record means for storing data as light altering characteristics and being organized into a plurality (P) of juxtaposed data regions each having capacity to store (N) data spots, in which at least certain ones of said data spots have ~3 light altering states to store >1 bit of data.
controllable light source means for selectively illuminating at least one of said separate data regions of said optical data means;
data recording means for photo-optically writing data to each of said (N) data spots to store thereat one of said light altering states;
data imaging lens means having a plurality of juxtaposed lenslets each being shaped and arranged in such proximity to and in optical registration with a separate one of said juxtaposed data regions so that the image resolving power thereof is substantially uniform over the field of view of that data region to form an image thereof on a common image surface spaced from said data means and lens means;
sensor means having a plurality (S) of juxtaposed light sensor elements arranged at said image surface for sensing data as a light image corresponding to an illuminated data region, said sensor means including light state detection means for detecting at each of said sensors the ~3 light altering states of said certain ones of said data spots as imaged onto said common image surface.
data signal output means coupled to said sensor means for outputting data signals representing said >1 bit of data of each of said certain ones of said (N) data spots of an illuminated and imaged data region.

5. The optical data system of claim 4, further comprising sensor element normalization means for sequentially flashing certain ones of said controllable light source means prior to or after the selective illumination of said separate data regions for creating a sensor level reference, and sensor level output adjustor means for adjusting each sensor element signal level output as a function of said sensor level reference.

6. An optical data recording system comprising an optical data means for storing data as light altering characteristics having a density on the order of one data spot per square micron and being organized into a plurality of juxtaposed data regions each adapted to store a field of data spots each of which has from 4 to 64 or more light altering states for storing 2 to 6 or more binary bits of data at each spot; and data writing means including controlled writing light source and controlled shutter means for writing said 4-64 states to each data spot.