`
`
`R. BALLINGALL et al., “Two-Dimensional Random Access Infrared Arrays,”
`
`IEE Advanced Infrared Detectors and Systems, No. 228, London, 1983
`
`(“BALLINGALL”)
`
`
`
`TRW Automotive U.S. LLC: EXHIBIT 1025
`PETITION FOR INTER PARTES REVIEW
`OF U.S. PATENT NUMBER 8,599,001
`IPR2015-00436
`
`
`
`1025-001
`
`
`
`Organising Committee
`
`P R Wallis (Chairman)
`Professor G S Hobson
`R J Home
`Dr K F Hulme
`D Lane
`
`W D Lawson
`P D Lomer
`M C Rowland
`G S Walton
`G F Weston*
`P D L Williams
`
`* representing the Institute of Physics
`
`Overseas Corresponding Members
`
`Professor C Corsi
`A Guilmain
`Dr M Hartl
`
`Published by the Institution of Electrical Engineers, London and New York, ISBN 0 86296285 1
`
`ISSN 05379989.
`
`This publication is copyright under the Berne Convention and the International Copyright Convention.
`All rights reserved. Apart from any copying under the UK. Copyright Act, 1956, part 1, section 7, whereby a single copy of an article may be
`supplied, under certain conditions, for the purposes of research or private study, by a library of a class prescribed by the UK. Board of Trade
`Regulations (Statutory Instruments, 1957 No 868), no part of this publication may be reproduced, stored in a retrieval system or transmitted in
`any form or by any means, without the prior permission of the copyright owners. Permission is, however, not required to copy abstracts of papers
`or articles on condition that a full reference to the source is shown.
`
`Multiple copying of the contents of the publication without permission is always illegal.
`
`Produced by MULTIPLEX techniques ltd, Orpington, Kent.
`
`© 1983 Thelnstitution of Electrical Engineers
`
`1025-002
`
`1025-002
`
`
`
`This material may be protected by Copyright law (Title 17 U.S. Code)
`
`6
`
`TWO DIMENSIONAL RANDOM ACCESS INFRARED ARRAYS
`
`H.A. Ballingall, I.D. Blenkinsop, D.J. Lees, I. Baker*, M.D. Jenner*, R. Lockett*
`
`Royal Signals & Radar Establishment, Malvern,
`
`UK
`
`* Mullard Limited, Southampton, UK
`
`INTRODUCTION
`
`Many future infra—red systems find it desira-
`ble to use two dimensional arrays of electro-
`nically scanned or "staring" arrays. Many
`schemeshavebeen suggested ranging from a
`monolithic approach using silicon or cadmium
`mercury telluride as both the detector and
`read—out,
`to a hybrid approach where the
`detector and read—out are made from the most
`suitable materials and then combined using an
`interface technology.
`The approach taken by
`most research groups has been based on a
`hybrid combination of CMT photovoltaic detec-
`torsinterfacedto silicon charge coupled
`in
`devices. This has been very successful
`the 3-Sunlband but so far there has been no
`published report of an 8-l2um array.
`
`An alternative to the CCD read—out, which has
`been developedduringthe past
`few years,
`is a
`co—ordinate addressed approach (1,2 and 3).
`This consists of a two dimensional array of
`CMT photovoltaic detectors connected on a one
`to one basis to silicon MOS switches (Fig. l).
`The IR diodes are read—out by addressing the
`appropriate row with a shift register.
`The
`output of each column is integrated on a
`In
`capacitor of an integrating amplifier.
`this way a line of IR detectors is electroni-
`cally scanned across the scene.
`In the
`8—lNum band this approach has a number of
`advantages over the CCD. These are briefly
`listed below.
`1.
`Lower RoA diodes can be used. This is
`because the input
`impedance of the amplifier
`is much lower than a CCD.
`
`Higher dynamic range is achieved. This
`2.
`is due to the absence of well size limita-
`tions that limit the CCD. Consequently the
`longer end of the 8—l3um band can be used.
`3.
`Removal of the background radiation
`pedestal is easy. This is performed before
`A/D conversion.
`
`Uniformity correction is easier.
`A.
`a simple linear correction.
`5.
`Simple high yield MOS
`6.
`If one element of the
`the remainder of the array
`
`technology is used.
`array is damaged
`continues to work.
`
`It is
`
`The combination of 1,1: and 5
`7.
`cheaper focal plane array.
`
`leads to a
`
`8.
`
`Random access capability.
`
`The principle disadvantage is that at present
`only one line equivalent performance is
`obtainable.
`It is for this reason that it is
`uncompetitive with CCD's in the 3-5pm band.
`
`Line Sampled Array (LSA)
`
`Before discussing random access arrays, a
`brief review of the line sampled array is
`useful.
`The construction of the array makes
`it ideal for detailed assessment and testing.
`Each diode can be addressed in turn for as
`long as desirable. Hence a detailed study
`
`The
`can be made on any particular diode.
`following important parameters can be meas-
`ured on a routine basis for every element
`in
`the array.
`a.
`Current—voltage characteristic of the IR
`diode (Fig 2 shows a 32 x 32 element plot and
`Fig.
`3 an expanded plot of one of the array
`diodes).
`b.
`Short circuit photocurrent using a 300K
`background (Fig.
`A shows a 32 x 32 element
`plot).
`c.
`Zero bias resistance derived from Fig. 2
`(Figure 5 shows a 32 x 32 element plot using
`the sensitive areas of the diodes(5)).
`d.
`Responsivity using 300K and 39AK black
`bodies (Fig 6 shows a 32 x 32 element plot).
`e.
`Spectral response (Fig.
`7 shows a 5 x 5
`element sample taken from a 32 x 32 plot) in
`this particular case showing an average cut-
`off wavelength of ll.Uum. Relative response
`per watt is plotted versus wavelength.
`
`The detector noise can be calculated from the
`short circuit photocurrent and the zero bias
`resistance
`
`2
`
`1
`
`n
`
`_
`_
`
`HKT
`j_.
`R0
`
`2
`+ 2qI (A /Hz)
`
`Based on the above measurements the D* (Ap,
`2n, SOOK) of each detector can be calculated
`(Fig. 8).
`
`some form
`It can be seen from Figures 2-8 that
`of structure exists in the array. This clear-
`ly manifests itself in increased short circuit
`photocurrent and reduced zero bias resistance.
`The array shown was chosen to illustrate the
`structure which can be seen radiating from a
`point just below the centre of the array. This
`type of structure has been reported before,
`particularly good examples are shown in (A).
`The structure results in low resistance diodes
`and correlates well with low angle grain
`boundaries in the CMT observed in the scanning
`electron microscope. This structure is of
`limited importance in linear arrays since the
`starting material may be chosen to be rela-
`tively free of defects. However, as the area
`involved in 2D arrays increases, this becomes
`more difficult. Except
`in the most serious
`cases the effects of this structure on the
`image can be removed by signal conditioning
`electronics.
`
`In order to give an easier appreciation of the
`current status of the LSA arrays a typical
`image is shown in Fig. 9. This image was
`produced after correction for non-uniformities
`and consists of two 32 x 32 images taken with
`the LSA and Joined.
`The total losses in the
`window and f/l lens were about AO% (ie T =O.6l
`The measured MRTD using a calibrated thermal
`target is O.l5K.
`
`Continued progress in materials preparation
`and diode fabrication has reduced the struc-
`ture.
`Figure l0 shows an example of the
`
`1025-003
`
`1025-003
`
`
`
`current arrays with much reduced structure.
`Both short circuit photocurrent and zero bias
`resistance are shown.
`
`Random Access Line Sam led Arra s
`
`RALSA)
`
`1 is replaced
`If the shift register in Fig.
`by a decoder the line sampled array can be
`made random access. This allows any part or
`parts of the array to be addressed independ-
`ently of the rest.
`The information from these
`selected detectors can be read more frequently.
`As a consequence the sensitivity of the
`randomly addressed array is increased by
`(N/R)
`, where N is the number of IR detectors
`in a column and R is the total number of IR
`detectors that are randomly addressed in the
`column direction.
`For example if 8 columns
`are selected from a 32 x 32 array a factor of
`X2 increase in sensitivity is achieved. This
`facility may be used to selectively enhance
`parts of the image on a frame to frame basis.
`
`Focal Plane Multiplexing
`
`At present both LSA and RALSA require one row
`of leads to be taken through the dewar inter-
`face.
`For a 6M x 6h array this results in 6H
`signal leads plus 19 addressing and power
`supply leads crossing the dewar interface.
`Some systems would prefer this number to be
`reduced. Depending on the requirements of
`dewar design and heat
`loading the amplifiers
`may be placed outside the dewar,
`inside the
`dewar at
`intermediate temperatures or on the
`focal plane at 77K.
`By fabricating silicon
`amplifying integrators on the focal plane at
`cryogenic or intermediate temperatures the
`outputs from the IR array may be multiplexed
`to one signal line that crosses the dewar
`interface.
`In order to achieve this, small
`low power,
`low noise amplifiers operating at
`or above 77K are needed. These are currently
`being researched.
`The requirements for low
`current noise,
`low l/f noise,
`low input off-
`set and bias current dictate that a J.F.E.T
`input circuit should be used.
`
`The main problem is the increase in transis-
`tor l/f noise at 77K. These devices may be
`used directly as integrators or as current
`preamplifiers followed by integrators.
`
`In either case the multiplexer can include
`sample and hold circuits to allow signal
`integration and read-out
`to occur simultan-
`eously.
`-
`
`Uniformity Correction Electronics
`
`A major problem with all infrared arrays that
`are directly coupled to the scene is that of
`achieving sufficient uniformity to avoid
`excessive fixed pattern noise.
`It is unreal-
`istic to expect
`the required uniformity,
`which is about O.l%,
`to be achieved with
`existing focal planes.
`It is doubtful if
`focal planes will ever achieve this degree of
`perfection. Uniformity correction electro-
`nics are therefore required. Using the co-
`ordinate addressed type of array described
`here,
`the major non-uniformity is due to the
`IR diode. Unlike the CCD approach,
`there is
`very little non—uniformity due to the inter-
`action between the IR diode and the read-out
`electronics (1). Hence the arrays described
`in this paper require only a simple correc-
`tion process.
`
`Because of the low contrast of the infrared
`scene and the non-uniformity of infrared
`arrays, 12 bit electronics is required to
`
`provide sufficient dynamic range to reduce
`the fixed pattern noise to acceptable levels.
`
`In the system employing integrators and sample
`and hold circuits (Fig. 11),
`the serial signal
`from the multiplexer has a dc pedestal sub-
`tracted from it.
`
`It is then converted to a 12 bit digital
`signal.
`In this laboratory demonstration
`system,
`timing and arithmetic of the correc-
`tion process is handled by a 16 bit micro-
`processor.
`
`The signal conditioning system performs back-
`ground subtraction and division operations on
`the input data using pairs of factors for
`each array element. This results in a
`corrected 8 bit video output, which in this
`case is displayed on a TV monitor, but which
`would normally pass directly to signal
`processing electronics.
`The appropriate
`correction factors are generated in an initial
`setting-up phase in which the array is ex-
`posed sequentially to uniform cold and warm
`targets.
`The resulting 12 bit numbers are
`stored in pairs in an area of random access
`memory.
`
`This system demonstrates the practicability
`of such a correction scheme, but is limited
`in speed by the performance of the micro-
`processor used. Dedicated circuitry should
`permit frame rates of several hundred Hertz
`to be achieved.
`
`Some idea of the degree to which the signal
`correction electronics is effective can be
`seen by comparing uncorrected and corrected
`images. Figure 12 shows an image before
`correction and Fig.
`9 after.
`It can be seen
`that very substantial improvement is achieved
`with relatively simple correction electronics,
`made from commercially available parts.
`
`Conclusion
`
`A relatively simple and cheap way of achiev-
`ing electronically scanned arrays in the
`8-lflum band has been demonstrated.
`The
`facility of random access permits flexible
`interactions between the signal processing
`electronics and the detector array.
`
`REFERENCES
`
`1.
`
`Ballingall R.A., Advanced Infrared
`Detector and Systems, 2§—30 Oct l§8l,
`D 70-75.
`
`2.
`
`Baker, I. et al Ibid, p76—8l.
`
`3. Ballingall, R.A., et al, Electronics
`18, No. 7, p2B5—287.
`Letters, Vol.
`
`M.
`
`Syllaious A.J. and Colombo L.,
`International Electron Devices Meetin
`l§82, 5.3
`
`5.
`
`Baker, I. et al. This conference.
`
`Copyright C) Controller, HMSO, London 1983
`
`1025-004
`
`1025-004
`
`
`
`1.4..23.7...,.J_.,
`
`
`
`..Jlvu.Z.J.yU
`
`uJ;,,.JJ:9
`
`J..J.u..r_J.U.J_w.TJ_..f
`
`OUTPUT
`
`COLUMN SCAN
`
`Fig.2 /V characteristics of a 32x32 array
`Hor. scale +/— 0.7V
`
`Ver. scale +/-2,uA
`
`Fig.1
`
`Schematic of co—ord/‘note addressed array
`
`Fig.3
`
`Expanded /—V plot
`
`92nA
`
`280nA
`
`1'g.4
`
`_
`.
`Short circu/t photocurrent.
`
`1025-005
`
`1025-005
`
`
`
`
`
`
`
`
`yaw?
`
`2.8 /(Ohms
`
`88 k0hms
`
`Rig.5
`
`Zero bias resistance
`
`0.34;:/w
`
`3.2A/W
`
`1'g.6 Responsivity (500K)
`
`Fig. 7
`
`Spectral response
`
`Hor. scale 8- 14pm
`
`1.2x10
`
`Fig.3 D: cm»/f7Z-'7W (500K)
`
`1025-006
`
`1025-006
`
`
`
`
`
`
`
`
`
`u"
`-‘+.'r:«'::-
`-"E,-‘-‘c vi}!
`
`
`.«~.»_-,
`
`
`
`an-95k’
`
`
`
`
`-- 0 «. *: Q .
`-:,—.»l‘~ fl
`.“.a< in .- 2.
`
`" "»".‘9é+“z.»"*«';'*.‘-1-1:.
`
`
`.r:.~;.«:a-. f:=.’:’**-=‘.*»’:§,\*"'-’.-t.*«x.,.«
`V-.a*.".' 3' v_»_",..J-' *;-‘ ,,..~.‘.g9-‘,_"€I‘;‘ -J, V
`'«' ' ‘.2’. .
`.
`'
`‘L
`
`
`F’;-I9.;3.».:;~.§‘:;,,€-;¢%3_,’_"
`fr ¢
`
`it
`
`
`;.*»‘-‘.-'§+:_::« :31:-.-.*.n v.g.m«,.
`
`Fig.10
`
`Short circuit photocurrent
`and zero bias resistance.
`
`Diagram shows both grey
`
`scale and carpet plots.
`
`1O
`
`1025-007
`
`x.-_..
`
`. ___.-.J
`
`Fig. 9
`
`Thermal
`
`image of man on bicycle
`
`1025-007
`
`
`
`11
`
`CONTROLS
`
`Ill‘
`Illll 1lI|‘
`§ 12911" ADC -
`
`
`
`32-1
`ANALOG
`
`M’ PLEX
`PEDESTAL
`CPU. BOARD
`
`
`SUBTRACT
`
` RA‘-SA 32 CHANNELS
`(278002-
`
`16 BIT)
`
`
`SIH
`
`I'll
`
`
`
`
`DEWAR (77K)
`
`
`+ SYNC.
`
`VIDEO
`CLOCK
`
`S
`
`LOOK — UP
`RAM
`
`§kx16
`
`CONTROLS
`
`DAC
`
`VIDEO
`
`F1'g.11
`
`Schematic of uniformity correct/‘on electronics
`
`Fig.1.? Uncorrected thermal image
`
`1025-008
`
`1025-008