Ulllted States Patent
`
`[19]
`
`[11] Patent Number:
`
`5,940,182
`
`Lepper, Jr. et al.
`
`[451 Date of Patent:
`
`*Aug. 17, 1999
`
`US005940182A
`
`[54] OPTICAL FILTER FOR SPECTROSCOPIC
`]\/IEASURENIENT AND METHOD OF
`PRODUCING THE OPTICAL FILTER
`
`4,346,992
`4,901,728
`4,957,371
`5,200,855
`
`8/1982 Schwartz ................................. .. 350/73
`
`2/1990 Hutchison .......
`. 128/633
`. 356/419
`9/1990 Pellicon et al.
`4/1993 Meredith, Jr. et al.
`. 359/588
`
`.
`
`[75]
`
`Inventors, James M_ Lepper Jr.’ Trabuco
`Mow Km
`M15510“ V1619 bmh Of Ca“
`[73] Assignee: Masimo Corporation, Irvine, Calif.
`[*] Notice:
`This patent issued on a continued pros-
`ecution application filed under 37 CFR
`.
`.
`1.53(d), and is subject to the twenty year
`patent
`term provisions of 35 U.S.C.
`154(a)(2).
`
`[21] App]. No‘: 09/088,397
`.
`[22]
`Filed:
`Jun. 1, 1998
`
`_
`_
`Related U-S- Apphcatlon Data
`.
`.
`.
`.
`_
`[62] Division of application No. 08/486,798, Jun. 7, 1995, Pat.
`No‘ 1760910‘
`[51]
`G01N 21/25; (30213 5/22
`1111;, C]_5 N
`
`[52]
`U_g_ C]_ ......... N
`. 356/416; 356/419; 359/838
`[58] Field Of Search ................................... .. 356/416, 419,
`356/73; 250/226, 227, 166; 350/166; 372/99;
`35 9/888—890
`
`References Cited
`Us. PATENT DOCUMENTS
`2,708,389
`5/1955 Kavana h
`g
`3,442,572
`5/1969 Illsley et al.
`3,771,857
`11/1973 Thomasson et al
`Ifyludel
`------
`,
`,
`,
`omasson e a
`356/186
`3,929,398 12/1975 Bales .............
`350/317
`3,981,568
`9/1976 Bartolomei
`356/189
`4,054,389
`10/1977 Owen ...... ..
`4,187,475
`2/1980 Wieder ................................. .. 331/945
`
`
`
`~-
`
`88/112
`350/166
`350/166
`
`[56]
`
`
`
`............ ..
`5/1993 Cole et al..
`. 128/633
`5,209,231
`2/13::§:s::::“;d:,~~1~~~
`~~,1:,::/:3:
`9:12:31
`5/1995 Barshad 5: al.
`.. ...
`.. ... 356/300
`5:416:579
`.
`1.
`.................. ..
`’
`5
`.
`6/1998 Lepper’ Jr et a
`356/416
`760910
`FOREIGN PATENT DOCUMENTS
`466403
`1/1992 European Pat‘ Off‘ ‘
`3724852
`1/1988 Germany .
`61_035681
`2/1986
`Japan
`'
`OTHER PUBLICATIONS
`
`Yang, Shumel, “Circular, V./ariable, Broad—Bandpass Filters
`With Induced Transmission at 200—1100nm , Applied
`Optics, Vol. 32, No. 25, Sep. 1993, pp. 4836-4842.
`Squire, J.R., “An Instrument for Measuring the Quantity of
`8
`Y8
`Blood and its De ree of OX enation in the Web of the
`Hand”, Clinical Science, Vol. 4, pp. 331-339, 1940.
`
`Primary Examiner—Frank G. Font
`Assistant Examiner—Michae1 P. Stafira
`Attorney, Agent, or Firm—Knobbe, Martens, Olson & Bear
`[57]
`ABSTRACT
`
`An optical filter used in applications involving spectroscopic
`measurements is fabricated by depositing layers of optical
`coatings onto a substrate. The layers are deposited so as to
`have a substantially constant thickness in a first direction
`along the surface of the substrate, and a gradually increasing
`thickness along a direction perpendicular to the first direc-
`.
`tion. The structure of the optical filter allows for large scale
`production of the filter so that costs in producing the filter
`are greatly reduced. The filter may be used in a Variety of
`-
`-
`-
`-
`-
`-
`-
`,
`-
`g[:I;1;i°:]E:§;§C1ud1:g’.butnotjlgmfiimChemlcal analws’
`mm“ Ormg’ an
`e
`6'
`
`11 Claims, 8 Drawing Sheets
`
`
`
`ASML 1227
`ASML 1227
`
`

`
`U.S. Patent
`
`Aug. 17,1999
`
`Sheet 1 of 8
`
`5,940,182
`
`
`
`F737
`
`(P/?/0/P /4/«=77
`
`

`
`tHEtaPQMU
`
`Aug. 17,1999
`
`Sheet 2 of 8
`
`5,940,182
`
`/70
`
`777
`
`Z20
`
`

`
` U.S.Patent
`
`OHoo4<z<
`
`4<:oa
`
`mmHmm>zoo
`
`4<o:ao
`
`mokompmo
`
`WVN
`
`4<zom4<:o:_
`
`mommmooma
`
`Aug.17,1999
`
`HDQHDO
`
`Sheet3of8
`
`5,940,182
`
`
`
`

`
`U.S. Patent
`
`Aug. 17,1999
`
`Sheet 4 of 8
`
`5,940,182
`
`Z
`
`//._\= 850nm
`
`9Z
`
`f,100%
`
`2 (
`
`50%
`
`2<
`
`CK
`I'-
`
`100%
`
`50%
`
`Z
`
`Q Q
`
`2 Q<
`
`:
`CI
`F-
`
`
`
`0
`
`32
`
`64
`
`95
`
`128
`
`150
`
`192
`
`224
`
`256
`
`DISC ROTATION
`
`)\_= 115Dnm
`
`0
`
`32
`
`64
`
`95
`
`128
`
`150
`
`192
`
`224
`
`255
`
`Z
`
`DISC ROTATION
`
`}\¥=1350nm
`
`Q E
`
`£100%
`
`E 3<
`
`F-
`
`05
`
`50%
`
`0
`
`32
`
`64
`
`96
`
`128
`
`160
`
`192
`
`224
`
`256
`
`DISC ROTATION
`
`f
`
`“*1
`
`f‘°1fl’*2 '
`
`’°2’\1
`:
`
`fpzk-2
`
`'
`
`'
`
`-
`
`'
`
`_
`
`f
`
`£01‘/“n
`
`go 9
`
`/-76,
`
`417
`
`

`
`U.S. Patent
`
`Aug. 17,1999
`
`Sheet 5 of8
`
`5,940,182
`
`100%
`
`TRANSMISSION
`
`50%
`
`O
`
`32
`
`64
`
`96
`
`128
`
`160
`
`192
`
`224
`
`256
`
`DISC ROTATION
`
`F/615
`
`

`
`U.S. Patent
`
`Aug. 17,1999
`
`Sheet 6 of 8
`
`5,940,182
`
`.300
`
`BEGIN
`
`J05
`
`.3’/O
`
`572
`
`J/5
`
`
`
`HOUSEKEEPING
`AND
`SELF TESTING
`
`
`TRANSMIT LIGHT
`THROUGH FILTER
`
`TO OBTAIN Io
`
`INSERT
`TEST MEDIUM
`
`SAMPLE DETECTED
`LIGHT AT VARIOUS
`WAVELENGTHS
`
`52.3
`
`DIVIDE BY 10
`TO NORMALIZE
`
`-355
`
`J25
`
`CONSTRUCT SIGNAL
`INTENSITY VECTOR
`
`HLTER
`CHARACTERBTICS
`
`MATRIX
`
`55;
`
`
`
`
`
`
`
`TAKE INVERSE
`TRANSFORM OF
`FILTER MATRIX
`
`
`.330
`MULTIPLY SIGNAL
`VECTOR BY
`INVERSE VECTOR
`TO OBTAIN OPTICAL
`DENSITY VECTOR
`
`
`
`
`
`
`
`
`
`
`
`550
`
`END
`
`F/G16’
`
`

`
`U.S. Patent
`
`Aug. 17,1999
`
`Sheet 7 of 8
`
`5,940,182
`
`‘W
`
`
`
`
`ILLUMINATE FILTER
`AT FIRST
`POSITION OVER "n"
`WAVELENGTHS TO OBTAIN
`SPECTRAL ABSORBTION
`COEFFICIENTS
`
`
`
`
`
`
`
`8.30
`
`505/
`
`555
`
`
`
`ILLUMINATE FILTER
`AT SECOND
`POSITION OVER "n"
`WAVELENGTHS TO OBTAIN
`SPECTRAL ABSORBTION
`COEFFICIENTS
`
`
`
`
`
`
`
`
`
`
`
`
`ILLUMINATE FILTER
`
`
`
`
`
`
`AT "Mth" (LAST)
`POSITION OVER "n"
`WAVELENGTHS To OBTAIN
`SPECTRAL ABSORBTION
`COEFFICIENTS
`
`
`
`540
`
`845
`
`
`
`
`CONSTRUCT MATRIX FROM
`SPECTRAL ABSORBTION
`COEFFICIENTS OBTAINED
`FOR EACH ROTATIONAL
`POSITION OF FILTER
`
`
`
`
`
`
`RETURN
`
`/-7637
`
`

`
`U
`
`9mM,g
`
`S9
`
`2O01,
`
`m%QR
`
` Mvfih.QQ\He23:%Q3mmmxQN\r\\\mi&\M..\
`
`
`|uII||II.||N/.\,.HAI,T.AK.3...mKlNWH1’!Qa9.......
`
`
`
`
`M.Exp.EfimhxWN\wt..):
`
`3%9%
`
`
`
`
`
` 00.f.._NM2mEmz<E...K3L:a>z_mawn:is$55.:2am
`
`
`
`
`
`:KE&mfmfiam
`
`
`
`

`
`1
`OPTICAL FILTER FOR SPECTROSCOPIC
`MEASUREMENT AND METHOD OF
`PRODUCING THE OPTICAL FILTER
`
`This is a Divisional of U.S. application Ser. No. 08/486,
`798 now U.S. Pat. No. 5,760,910 filed Jun. 7, 1995.
`BACKGROUND OF THE INVENTION
`
`1. Field of the Invention
`
`The present invention relates to optical filters which are
`used in applications where spectroscopic measurements are
`used to determine the properties of substances such as
`chemicals and other substances.
`
`5,940,182
`
`2
`filter are formed in a pattern so that rotation of the optical
`disk results in the transmission of selected optical bands. In
`many previous applications involving precise spectroscopic
`measurement, optical filters have been designed with very
`high tolerances. Furthermore, the methods for manufactur-
`ing such filters have often precluded the possibility of
`manufacturing the filters by mass production. Thus, even
`optical filters of this kind may be prohibitively expensive to
`fabricate.
`
`SUMMARY OF THE INVENTION
`
`10
`
`The present invention provides a rotating dichroic filter
`for spectroscopic measurement wherein the cost of the filter
`is approximately 100 times less than conventional rotating
`dichroic filters. This is accomplished by first relaxing the
`specifications of the filter and compensating for the relax-
`ation of filter specifications through more intensive signal
`processing steps. In addition, the filter is constructed in a
`manner which allows for easier production. The filtcr con-
`structed in accordance with the present invention allows
`from 10 to 100 times as much light to pass while maintaining
`the necessary precision through signal processing.
`One aspect of the present invention involves a method of
`manufacturing an optical filter. The method involves a
`number of steps. An optical substrate is provided having a
`top surface and a bottom surface, and layers of optical
`coating are deposited on the top surface such that the layers
`vary in thickness across the top of the substrate in a first
`direction. The thickness of the layers is substantially con-
`stant in a second direction substantially perpendicular to the
`first direction.
`In one embodiment,
`the method further
`involves creating a mounting hole in the center of the
`substrate. In addition, an opaque strip along at least a portion
`of the substrate is deposited in one embodiment.
`Another aspect of the present
`invention involves an
`optical filter. The optical filter has a substrate having a top
`surface and a bottom surface. The filter also has a plurality
`of optical coatings deposited on the top surface of the
`substrate such that the coatings vary in thickness in a first
`direction across the top surface. The coatings are substan-
`tially constant in thickness across the top surface in a second
`direction substantially perpendicular to the first direction.
`Another aspect of the present invention comprises an
`optical filter having a generally a generally circular sub-
`strate. Layers of optical coatings deposited on the substrate
`provide a non-imaging interferometer wherein approxi-
`mately one-half of light incident upon the coatings passes
`through the coatings over the entire surface of the substrate.
`Yet another aspect of the present invention involves an
`optical filter. A substrate having a top surface and a bottom
`surface has a plurality of layers of optical coatings varying
`in thickness in a first direction across the substrate. The
`
`layers provide optical transmission characteristics for the
`optical filter to provide an optical filter which transmits more
`than one wavelength through the filter at all locations across
`the surface of the filter.
`
`BRIEF DESCRIPTION OF THE DRAWINGS
`
`FIG. 1 depicts an exemplary dichroic filter as constructed
`by conventional methods.
`FIG. 2 depicts schematically the general method used in
`accordance with the present invention to manufacture a
`rotational optical filter.
`FIG. 3 depicts the dichroic filter of the present invention
`depicted in FIG. 2 in a blood glucose monitoring applica-
`tion.
`
`15
`
`2. Description of the Related Art
`Optical filters are well known in applications involving
`spectroscopic measurement. Spectroscopic measurement is
`used to determine the properties and chemical composition
`of various substances in a sample based upon the optical
`characteristics of the sample. In a typical spectroscopic ,
`measurement, light (in the visible and non-visible range) is
`used to illuminate the sample over multiple frequency
`spectra. More than one optical frequency (wavelength) is
`used to more precisely determine the optical characteristics
`of the sample and also to subtract out interference. In some
`applications, the light reflected from the sample is detected,
`while in other applications light transmitted through the
`sample is detected to determine the optical characteristics of
`the sample. In addition, a combination of the transmission
`through the sample and the reflections from the filter may be
`employed.
`The detected light is usually quantified to provide an
`indication of the “frequency response” of the sample at each
`of the frequency spectra. As is well known in the art, each
`substance has definable optical properties determined by the
`frequencies at which the substance reflects and absorbs light.
`Thus, the optical characteristics of a given substance may be
`quantified (e.g., plotted as intensity of reflected or transmit-
`ted light versus frequency) to provide an indication of the
`optical characteristics of that substance. Since different
`substances typically have distinct optical characteristics,
`quantified measurements of the optical properties of a
`sample containing several substances can serve as the basis
`for distinguishing among or making other measurements
`relating to the several substances within a sample. Precise
`measurements of the reflected or transmitted light can be
`used to determine the precise concentration of the various
`substances within a sample.
`Some present spectroscopic measurement systems use
`multiple light emitting diodes (LEDs) or laser sources to
`provide light at
`the desired wavelengths. However, very
`expensive, high precision wavelength light sources must be
`employed in order to manufacture such a system with the
`necessary wavelength accuracy for each of the sources.
`One alternative method of generating light at multiple
`frequencies involves rotating an optical filter between the
`sample to be measured and a broadband light source. Cur-
`rent optical spectroscopic devices, as identified by the inven-
`tor for use in the present invention, often require expensive
`custom-made filters which are used to generate a pattern of
`optical signals to be transmitted. One such filter, commonly
`known as a dichroic filter, comprises a rotating optically
`coated disk which includes regions of varying optical thick-
`ncss. As the wheel spins, light from the broadband light
`source passes through different portions of the wheel so that
`light of various frequencies are passed by the filter to
`illuminate the sample. That is, the regions on the dichroic
`
`40
`
`45
`
`60
`
`65
`
`

`
`5,940,182
`
`3
`FIGS. 4A—4C depict in graph form the optical transmis-
`sion characteristics for an exemplary dichroic filter over
`dilferent degrees of rotation in accordance with the present
`invention.
`
`FIG. 4D illustrates a matrix used to specify the optical
`characteristics of an exemplary dichroic filter in accordance
`with the present invention.
`FIG. 5 depicts in graph form the optical transmission
`characteristics of an exemplary conventional dichroic filter
`over dilferent degrees of rotation in accordance with the
`present invention.
`FIG. 6 depicts a general flow chart of the signal process-
`ing operations which are used to compensate for the lower
`optical tolerances of the filter of the present invention.
`FIG. 7 illustrates a flow chart which sets forth the general
`steps of obtaining the optical characteristics matrix of FIG.
`4D.
`
`10
`
`15
`
`FIG. 8 represents a functional block diagram of the
`general steps of using the filter of the present invention in .
`conjunction with signal processing to accommodate impre-
`cision.
`
`DETAILED DESCRIPTION OF THE
`INVENTION
`
`FIG. 1 shows an exemplary dichroic filter fabricated
`according to conventional methods. Previous methods
`employed to fabricate such optical filters typically involved
`laying out a circular substrate and then selectively increasing
`the coating thicknesses on the surface of the circular sub-
`strate as the substrate is rotated with uniform speed.
`Such a filter 150 is depicted in FIG. 1 as having coating
`layers 152, 154, 156, etc., of increasing thicknesses to form
`a spiral configuration as the filter 150 is rotated. Of course,
`it should be understood that the coating thicknesses depicted
`in FIG. 1 are exaggerated for ease of illustration. This
`method of optical coating is carried around substantially the
`entire circumference of the circular substrate so that as the
`coated substrate revolves, the thickness of the optical coat-
`ing grows throughout the entire revolution and then sud-
`denly drops back from the thickest coating to the thinnest
`coating at the end of one revolution.
`It has been found, however, that such methods of optical
`coating require high precision and are extremely costly.
`Furthermore, manufacturing these filters is typically carried
`out one-by-one, since production methods do 11ot allow for
`laying out several disks on a single sheet for mass produc-
`tion purposes.
`In addition, conventional filters of the type depicted in
`FIG. 1 generally have many layers (e.g., 100 or more layers
`is common). The number of layers in conventional filters are
`provided to provide very precise pass bands (for a bandpass
`filter). FIG. 5 depicts an exemplary transmission character-
`istic for a conventional rotational dichroic filter versus
`
`degrees of rotation for a selected wavelength. As illustrated
`in FIG. 5, the pass band of the filter is very precise for the
`selected wavelength, generally without side-lobes, and also
`provides essentially zero transmission outside the pass band.
`A very high number of layers is required to obtain a filter
`with this near ideal precision. It should be understood, that
`this very narrow passband is in dilferent rotational positions
`for different wavelengths. In other words, a conventional
`dichroic filter can be characterized as a monochrometer
`which passes a different wavelength at different rotational
`positions.
`Creating each layer is expensive due to the continuous
`rotational variation from thin to thicker. Thus, when many
`
`4
`
`layers are created (e.g., 100 or more for good precision),
`such conventional filters are very costly.
`In accordance with the present invention, a dichroic filter
`is disclosed which dilfers significantly from conventional
`dichroic filters. FIG. 2 depicts a filter 120 along with the
`steps followed in the method of producing a filter in accor-
`dance with the teachings of the present invention.
`The dichroic filter according to the present invention is
`made in a novel manner in which the multiple optical
`coatings are created on a substrate to form a wedge-like
`substrate. For a rotational filter, the substrate is then cut to
`form a rotational disk filter.
`
`In addition, according to one aspect of the present
`invention, the dichroic filter has fewer layers than conven-
`tional filters. This provides for less precision in the trans-
`mission characteristic of the filter. FIGS. 4A—4C depict the
`optical transmission characteristics for selected wavelengths
`of an exemplary rotational filter made in accordance with the
`present invention having only 17 optical coating layers. As
`illustrated in FIGS. 4A—4C, the transmission characteristic is
`not as precise as the transmission characteristic of the filter
`represented in FIG. 5. As depicted in FIGS. 4A—4C, the
`dichroic filter of the present invention has a several pass-
`bands for each wavelength depicted. In addition, outside the
`pass-bands,
`the transmission does not fall completely to
`zero, as with the conventional precision filters. The reduced
`precision in the passbands is due to the reduced number of
`layers in the filter. It should be understood, that the reduced
`precision explained above is not limited to rotational dich-
`roic filters, but could also be advantageous with dichroic
`filters that are vibrated (e.g., through oscillation or the like),
`and for any other optical
`filter which conventionally
`involves high precision in the pass-bands. The decreased
`precision of the filter of the present invention is accommo-
`dated with signal processing as filrther explained below to
`obtain the required precision. In this manner, the cost of the
`filter can be reduced.
`
`40
`
`45
`
`When both aspects of the filter in accordance with the
`present invention are used (layering process and reduced
`number of layers), the resulting filter is much less expensive
`to construct than conventional dichroic filters. However, it
`should be noted that using either aspect of reducing cost is
`advantageous in itself. For instance, a conventional rota-
`tional filter could be fabricated with far fewer layers, but
`using conventional layering techniques such that the filter
`increases in thickness through the entire revolution of the
`filter. Alternatively,
`the method of fabrication disclosed
`herein could be used to form a rotational filter with conven-
`
`tional precision (e.g., many layers) at reduced manufactur-
`ing costs due to the improved manufacturing method.
`In the method which reduces the cost of layering the
`optical filter, a fiat substrate 110 (FIG. 2) is coated with
`optical coatings of increasing thickness to form a wedge-
`shaped coated layer 111. It should be noted that for purposes
`of clearly illustrating the present invention, the thickness of
`the optical coating 111 has been exaggerated, and in prac-
`tical applications the thickness of the optical layer 111 varies
`from roughly 1.66 micrometers to about 3.33 micrometers,
`with an average thickness of about 2.35 micrometers. It
`should also be understood that these thicknesses are approxi-
`mate and may vary depending upon the index of refraction
`of the layer materials. Therefore, in accordance with one
`aspect of the present invention, the optical coatings which
`define the filter are applied across a substrate rather than
`continually applying coatings circumferentially, thus, sig-
`nificantly reducing the cost of the filter. The filter at this
`
`60
`
`65
`
`

`
`5,940,182
`
`5
`point provides a dichroic filter which could be used in
`oscillating filter type applications.
`For a rotational filter, once the optical layers 111 have
`been applied to the substrate 111, a cylindrical portion 112
`is cut from the wedge-shaped slab formed by the optical
`layer 111 together with the substrate 110. A cylindrical
`aperture is then formed in the center of the cylindrical
`portion 112 to form a mounting hole. In certain applications,
`it is desirable to form an optically opaque strip such as a
`brass strip 122 over a portion of the optical filter disk 120.
`The brass strip provides a zero-transmission reference por-
`tion of the disc 120 which may be helpful for noise cancel-
`lation in certain signal processing applications.
`The above description provides ease of illustration for
`understanding one aspect of the present invention. However,
`it should be understood that the method may, in practice,
`involve first cutting the substrate into a disk. Thereafter, the
`optical coatings are applied onto the disk as though the disk
`were still square so that the excess falls onto the platform
`(not shown) supporting the disk within the vacuum tank. In
`this manner the wedge is formed on the surface of the disk
`120 as shown in FIG. 10.
`It will be understood that the disk 120 does not continu-
`
`ally increase in thickness through the entire circumference
`of the wheel, but increases in thickness and then decreases
`in thickness. However, both halves of the circumference can
`be utilized as further described below.
`
`In addition to the reduced manufacturing cost of the filter
`described above, in accordance with a further aspect of the
`present
`invention, a minimal number of optical coating
`layers are deposited. In one preferred embodiment, only 17
`layers are necessary to obtain the desired resolution.
`Although reducing the number of layers results in less
`precise filters, such imperfections can be accommodated in
`digital signal processing steps. For example, as explained
`above, conventional dichroic filters typically pass a single
`frequency band at a time (FIG. 5), while the filter of the
`preferred embodiment may allow for multiple bands to pass,
`since this is accounted for, and can be compensated through
`signal processing.
`the resolution typically
`It should be noted here that
`necessary for applications involving more expensive inter-
`ferometers or monochrometers is typically not necessary for
`analyzing liquids. However, additional layers can be added
`at greater spacing intervals in order to increase resolution of
`the filter.
`
`COMPENSATING DIGITAL SIGNAL
`PROCESSING
`
`As briefly set forth above, the imprecision of a filter made
`in accordance with the present invention having a minimal
`number of optical coatings can be accommodated through
`signal processing.
`FIG. 6 is a data flow diagram which details the method
`used to compensate for the imprecision of the filter made in
`accordance with the present
`invention.
`It should be
`understood, however, that prior to run-time, initialization is
`performed.
`
`PRE-RUN-TIME INITIALIZATION
`
`The initialization is performed at the factory or other time
`prior to use. In general, a filter characteristics matrix is
`constructed, as described in greater detail below with ref-
`erence to FIG. 7. The filter characteristics matrix represents
`the transmission characteristics of the dichroic filter 120 at
`
`10
`
`15
`
`40
`
`45
`
`60
`
`65
`
`6
`different portions of the filter 120 and for various wave-
`lengths of light. The filter characteristics matrix is used in
`order to extract portions of the electrical signal generated by
`a detector which are due simply to the optical attenuation
`caused by the filter 120. In other words, by knowing the filter
`characteristics, the impression of the filter can be accounted
`for.
`two-dimensional
`The filter characteristic matrix is a
`matrix. The filter characteristic matrix includes one column
`
`for each wavelength of light which is characterized and one
`row for each position (rotational in the present invention) of
`the filter 120, at which characterization (of the filter
`characteristic) is performed. Thus, in one embodiment, the
`filter characteristic matrix includes 16 columns and 256
`rows when 16 wavelengths are characterized and 256 posi-
`tions of the filter 120 are defined. It should be understood
`here that it is not necessary that 16 different wavelengths be
`used;
`the use of additional wavelengths is particularly
`advantageous for increasing the signal-to-noise ratio. Since
`about half of the incident light is transmitted through the
`filter at each position of the filter, the same wavelength is
`detected multiple times (although in a unique combination
`with other wavelengths each time) so that the overall signal
`intensity is from 10 to 100 times the intensity of any single
`wavelength and much higher than the noise floor. This is
`commonly referred to as Felgate’s advantage. In this manner
`the spectral
`response of the entire filter 120 over the
`expected measured wavelengths is completely character-
`ized. The method employed to construct the filter charac-
`teristics matrix is described in detail below with reference to
`FIG. 7.
`
`Dl:'RIVAl‘ION OF THE FIL'l‘l:lR
`CHARACTERISTIC MATRIX
`
`FIGS. 4A—4D, together with FIG. 7, illustrate in greater
`detail, the method employed to obtain the filter characteristic
`matrix. The derivation routine is illustrated in FIG. 7 and
`starts with a begin block 800.
`The activity blocks 830-845, together with FIGS. 4A—4D,
`illustrate the method used in accordance with the present
`invention to construct the filter characteristics matrix. The
`
`filter 120 reflects and transmits optical radiation in different
`proportions for different wavelengths at different places on
`the filter disk 120. This is clearly illustrated in FIG. 4A—4C,
`wherein FIG. 4A represents the optical transmission of light
`at a wavelength of 850 nanometers plotted versus each of a
`possible 25 6 disk rotational positions (for one embodiment).
`As shown in FIG. 4A, when the disk 120 is in the initial
`starting position (i.e., q)=0 where 0 represents the rotational
`position of the filter 120), the transmission of light at 850
`nanometers is approximately 10% through the filter 120,
`while when the disk 120 is rotated so that (|>=32, the optical
`transmission of light at 850 nanometers through the filter
`120 is approximately 25%. Again, between the disk rota-
`tional positions of <|)=l28 to ¢=l60, the transmission of light
`at 850 nanometers wavelength through the filter 120 is
`approximately 75%. Thus,
`the optical
`transmission for
`}t=850 nanometers is entirely characterized over 256 rota-
`tional positions of the disk filter 120, as depicted in FIG. 4A.
`FIG. 4B depicts the optical transmission characteristics of
`light at 1,150 nanometers over the same 256 rotational
`positions of the disk 120. Similarly, FIG. 4C depicts a plot
`of the optical
`transmission of light at 1,350 nanometers
`through the disk filter 120 at each of the 256 rotational
`positions of the disk 120. In one actual embodiment of the
`invention,
`the optical
`transmission characteristics of the
`
`

`
`5,940,182
`
`7
`filter 120 are described for 256 rotational positions at each
`of 16 wavelengths between 850 nanometers and 1,400
`nanometers.
`
`Thus, from these measurements, a filter characteristic
`matrix may be constructed, as shown in FIG. 4D. The filter
`characteristic matrix designated in FIG. 4D as F(<|),
`7»)
`includes 256 rows and 16 columns. Each column of the filter
`characteristic matrix comprises the spectral transmission
`characteristics of the disk 120 at each of the 256 rotational
`
`10
`
`15
`
`positions of the disk 120 for the selected wavelength for that
`column.
`the filter characteristic matrix
`to construct
`In orcer
`depicted in FIG. 4D, the filter 120 is illuminated at a first
`rotationa position over each of the 16 wavelengths to obtain
`spectral
`transmission coefficients for each of the 16
`wavelengths, as indicated within an activity block 830. Once
`the spectral transmission coefficients have been determined
`for the
`irst rotational position as indicated within the
`activity block 830,
`the filter is illuminated at a second
`rotationa position (i.e., q)=1) over the 16 selected wave-
`lengths to obtain spectral transmission coefficients for the
`second rotational position, as represented in an activity
`block 835. This method is carried on for each of the possible
`rotationa positions of the disk 120 until, as indicated within
`an activity block 840, the filter is illuminated at the “mth,” '
`or last, rotational position (i.e., position 256) of the disk filter
`120 over the 16 selected wavelengths to obtain the spectral
`transmission coefficients for the last rotational position. In
`one preferred embodiment, where a stepper motor is used,
`the rotational positions will be precise from revolution to
`revolution of the disk 120. Of course, a computer disc motor
`with salient poles and run at a constant speed could be used
`provided that phase dithers are minimized to less than one
`part in 256.
`Once spectral transmission coefficients have been deter-
`mincd for all 16 wavelengths of all 256 rotational positions
`of the disk 120,
`the filter characteristics matrix is
`constructed, as indicated within an activity block 845. The
`matrix defined by column and row where columns represent
`coeffluents and row represents the wavelength by putting
`coefficients. Once the filter characteristics matrix is
`constructed,
`the system has the necessary constraints for
`processing.
`It should be understood that derivation of a filter charac-
`
`40
`
`45
`
`teristic matrix has been described for purposes of the rota-
`tional filter 120. However, an oscillating filter, or any filter
`with defined positions on the filter such as Fabry-Perot type
`filters and even fixed filters such as those used in CCD
`
`applications can also be characterized in accordance with the
`discussion above.
`
`RUN-TIME PROCESSING
`
`Discussion of the overall processing in accordance with
`the present invention in order to account for impression of
`the filter through the use of the filter characterization matrix
`is made with reference to FIGS. 3, 7 and 8.
`FIG. 3 illustrates the use of the filter 120 in a system for
`monitoring blood constituents. FIG. 6 illustrates a general
`flow diagram for the steps of accounting for the imprecision
`in the filter to obtain the characteristics of a medium under
`
`test. FIG. 8 illustrates a general functional diagram of the
`process of accounting for filter imprecision thro11gh signal
`processing. As dcpictcd in FIG. 6, the start of processing is
`represented in a begin block 300. First, housekeeping and
`self-testing procedures are performed, as represented in an
`activity block 305. Briefly, housekeeping and self testing
`
`60
`
`65
`
`8
`involves boot operations and conventional initialization a
`self testing. For example, the system first determines if there
`is a sufficient signal intensity to take an accurate reading.
`After housekeeping and self testing is completed, the light
`source 110 (FIGS. 3 and 8) is activated to transmit light 115
`through the filter 120, as represented in an activity block
`310. Initially, the light source 110 is activated while no test
`medium 131 is interposed between the filter 120 and the
`detector 140. Thus, the light which is detected by a detector
`140 (FIG. 3) represents a baseline light intensity (Ia) which
`can be used as a test to insure that a bulb which is too dim
`
`is not inserted as a replacement bulb for
`or too bright
`example. In one embodiment, a lens 117 (FIG. 8) can be
`provided between the light source and the filter 120 to
`provide focused light 115 on the filter 120.
`Once the initial baseline light intensity constant has been
`determined, the medium 131 under test is inserted as indi-
`cated in an activity block 312.
`As indicated within an activity block 315, the light which
`is incident upon the detector 140 is converted to an electrical
`signal and this signal is amplified in a pre-amp (not shown),
`filtered with the band pass filter (not shown), and sampled by
`an analog-to-digital converter 142. Since the filter 120 is
`rotating (at approximately 78.125 revolutions per second in
`one actual embodiment, although other rotational rates could
`be advantageous as called for by the particular application),
`samples of the electrical signal output by the detector 140
`are indicative of the light
`intensity detected at various
`rotational positions of the filter 120. In one advantageous
`embodiment, one complete rotation (i.e., 360°) of the filter
`120 corresponds to 512 digital samples. That is, 512 samples
`are taken within the period corresponding to one revolution
`of the filter 120. Thus, for example, if the filter 120 rotates
`at 78.125 revolutions per second, then 512 samples will be
`taken within approximately 1/78th of a second, so that the
`sampling rate of the analog-to-digital converter 142 will be
`approximately 40,000 samples per second.
`As described,
`the filter 120 constructed in accordance
`with the present invention includes rcdundant rcgions within
`an en