Document jmyBL1OL2vEQZZ7DG8NYJaoxN

1 -----------,-- ' Vi'>! mlismliimii' -1 t WM&fc^..y...r.H....W... '1 IjKpB V !'l -(i!i-! i \ S1UDY OF MI/I ALL 1C COLOR 'IOLLRAKCLS \ 'vJLU l-j. R. II. VIMJ NG - ] 0/] 5/bS -PHILADELPHIA ,`, K. V. i* <Vr p i , i* 1*1 J^Wmw^za'rh i ... ffiWiU . I. DU PONT DE NEEMC -;a};CdMPANT|;:: FABRICS & FINIS e IIr t ^n t : ' -4*03. RESTRICTED - COMPANY CONFIDENTIAL REPORT NO. PE-65-11 COPY NO.: / DUPONT V' Mar s h al l l abo r at o r y l ibr ar y 1 E. I. DU PONT DE NEMOURS COMPANY, INC. FABRICS FINISHES DEPARTMENT PRODUCTION DIVISION DIVISIONAL PROCESS ENGINEERING GROUP COLOR OPERATIONS SECTION ifUM COPT ii Ofteyum A STUDY GF METALLIC COLOR TOLERANCES BY: R. H. VINING DATE ISSUED: OCTOBER IS, 1965 PROJECT: P-58, INSTRUMENTAL COLOR TECHNOLOGY 4Pi?2?/966 ATTRACT An analysis of the factors influencing metallic color control and customer acceptance patterns are presented together with a summary of the results of the 1964-1965 model years. Based cm the data analysed, instrumental color tolerances are proposed for metallic colors. DUP030002751 A SUM OF METALLIC COLOR TOLERANCES TABLE OF OMENTS PAGE I, SMiMSf Introduction ..................................... ........... Results Conclusions ............................................ .......... SecODSTndations .................. Continuing Action ............................ AcloicwledgEffiiit ............................ Restrictions .......................... 1 1 5 4 5 5 6 II. DISCUSSION . Background ................ Visual Appraisal Procedures .................. Repeatability and Reproducibility ............ Color Tolerance ................................................... .. Batek-t~Eatch Ur&fbrn&ty ............................ . Custor Acceptance Tolerance................. 1 7 3 10 12 13 III. APPENDIX 1. Tucker Prograa - .1966 Colors 2. General References 3. Distribution list DUP030002752 A STUDS' OF METALLIC COLOR TOXSRANCES I - SIM4ART 1MHJCTION The successful application of color technology to the routine control of color is largely dependent on defining mathematical tolerance zones in a manner such that they define the acceptability or non-acceptability of a color match to a customer when the match is viewed visually. If metamerism is not present, there should be no conflict between visual and instrumental appraisal of a color difference. The term "color measurement" brings to the minds of most people thoughts of certain specific types of iistruments. This, however, is only correct in part. The visual examination and comparison of colored materials constitute a measure ment which is qualitatively the same as the most elaborate and complex instru mental method. The success of any color measurement procedure requires a stan dardised procedure for the preparation of the sample which is both repeatable (by the same person in the same laboratory! and reproducible (by different people in different laboratories at different times). Should visual rod instrumental assessment procedures fail to agree, then it must be determined whether the criteria of repeatable and reproducible sample preparation techniques have.been fully met by both systems. With respect to metallic colors, specifically "Lectte*" tetcmtive colors, there is strong evidence to support.the belief that such is not the case with current visual 'color meroureraent procedures based on conventional sample preparation techniques and that the instrumental approach using the. sparger concept Is more reliable. Production for six colors ammifactured during 1965 at tour plant locations has been monitored by visual and instraaental methods. Data reported and conclusions drawn' in this report are based on production batches which were visually approved using conventional panel spray procedures versus a visual oymmWM&Hy' sprayed standard, and .then subsequently resprayed via the sparger procedure and measured using Coloraasters specially modified tor both normal and high angle metallic color measureraent, RESULTS ' The results obtained from this study are- suwrized below.: 1. The., spray variable inherent in the preparation of conventional metallic panels.is of sufficient magnitude to violate the fundamental criteria of repeatability rod reproducibility and effectively "masks" color shifts existing between batch sad standard. The actual, as compared with the apparent, batch-to-batch variations between conventionally sprayed metallies are .not .accurately defined by existing procedures nor can throe actual variations be considered equal to those tradition ally reserved tor products requiring "exacting matches." The actual batch-to-batch variability existing between metallic colors as defined by sparger spray and instrumental measurement is several orders of magnitude greater than that applied to Automotive solid colors,- yet on a visual, basis, color tolerance, limits for metallic colors are more restrictive. The data summarized in the body of this report clearly define toe .actual batch-to-batch variation that has existed regardless of the immlzEEmFing location through toe application of control predicated a conventionally sprayed panels' and visual color tolerance limits. DUP030002753 2 2. A significant and not fully recognised factor in multi-plant production is the color deviation existing between production for any given time period for plants operating independently trader the visual appraisal methods. Under existing approval procedures where individual preferences can significantly influence the nebulous approval zone for any given product, color deviation between in dividual plants can be of sufficient magnitude that it probably could not be masked by the normal spray manipulation inherent in Automotive production plants. 3. Based cm data compiled during this study, there is no justification for determining or maintaining a visual offset predicated on a customer's implied or imagined preference for a given color or color family, since all evidence shows that none truly exists, assuming no basic formulation weakness. All plants, regardless of whether they are attesting to maintain a visual offset for a given color, have shipped batches over a period of one or two months that completely "bracket" the standard. Batches shipped during a given time interval, would, at another time period, be rejected visually due to unacceptable color deviations from the visual offset. 4. Evaluation of the data compiled for the product codes studied show the overall variability that existed on a Divisional basis determined from sparger panels prepared from conventionally approved batches {see Table I), DIVISIONAL PRODUCT COLOR VARIABILITY TABLE I Product 867-97212 867-972IS 867-97216 867-97217 867-97219 867-97223 "Head-In" Color JL 6a ~sr l.S 0.7 0.8 2.2 1.8 1.5 2.0 0.6 1.5 2.4 2.2 1.2 2.5 1.5 0.8 1.6 1.4 0.8 "Side-Tone" Color -nr 6a --w 1.3 1.0 1.6 2.5 1,5 1.2 1.5 1.6 2.3 3,3 1.4 1.6 0.6 1.7 0.9 0.8 1.2 0.7 5, There is .no evidence to indicate that there is a preferred "side-tone" color match for a metallic standard. It has been assured that problems will result if metallic colors are shipped exhibiting a light, milky side-tone and/or are less saturated than standard. It is very evident from this acrasaulated data that the customer has accepted and used such batches through the 1965 model year for some of the colors evaluated. In other instances, one customer plant has received production that has been consistently on the dark, saturated side of standard, a second customer plant has received and used the same basic color on the light, less saturated side, yet no color ossplaints have been reported. It is apparent that in many instances troublesome batches have been a case .of '^wsh-ado-about'-nothing," and the principal criteria is to maintain a reasonable level of uniformity between shipments. DUP030002754 3 OQNCIiUSIONS the conclusions drawn from the data presented in this report are delineated below: 1, Visual appraisal techniques, though valid and accurate at any one time and location, are subject to psychophysical factors which result in. a transient color preference for a 'hatch" to one side or another of the standard. The net effect of these factors indicates that at any one time different manufacturing locations could produce the same product on the opposite side of standard or in a given day close color matches can be erroneously evaluated as being on opposite sides of the standard. The elimination of the highly variable con ventional spray system and its replacement with the more precise sparger system will not eliminate the human variability factor. 2. The color appraisal, based a visual viewing and conventional spray techniques, is subject to many variables that are not readily cun'* trollable. There is strong evidence to support the contention that these appraisal techniques do not result in tee quality of match or match reproducibility credited to the overall system. Success of tee visual system has been predicated on each plant attempting to minimise baich-to-batch variation at its location and a major factor in achieving success steam fro the feet that the customer for the most part restricts his purchase of any given product to cm Du Pont location. If the customer were to change his procurement habits and acquire metallic finishes m a random basis from any of our four producing plants, it is highly unlikely that under tee existing procedures we would .escape formal color complaints. The variables inherent in application of visual color appraisal techniques are such that further refinements to the overall system are not readily obtainable, therefore, the increas ing demands.by ..industry for more uniform and exacting color matches are not likely to be achieved by visual techniques site . 3,. Color approval .based on instrumental 'techniques will not replace visual evaluation of a color. The eye will and must main the final arbitrator of a. color, match, being used primarily as the check against human error in instrumental measurement. . Application of instrasental techniques to color control will make possible better interplant uniformity and bateh-to-batch uniformity unattainable by current visual appraisal techniques, 4. The successful application of instrumental techniques to metallic color control is dependent on proper sparger spray procedures, minimising human variability and applying a procedure teat achieves the criteria of repeatability and reproducibility. 5. The successful application of instrwental color techniques to color control is predicated on having defined meaningful color tolerances. Color tolerances to be meaningful must be based m a customer's history pattern of acceptances for individual colors and/or color families. Acceptance pattern data are available ter a large group of metallic finishes. These patterns are based on a sampling of 1964 model year colors and the 1965 model year colors reported in the body of this report 6. The practice of establishing "offsets" for metallic colors should be dis continued since date show that they have no effect on tee overall accept ability pattern. DUP030002755 4 SUMMATIONS The following recoamiendations result from this study: 1. Metallic batches be sprayed via sparger techniques and ins&rumentally measured, matched and approved with an after-the-fact visual check to eliminate possible human measuring or calculating error, 2. The following instrumental color tolerances are recommended for the light to medium metallic colors - Table II. TABLE II RECOMMENDED COLOR TOLERANCES "(ngIt'"to Wdiiii'MeeiHcsJ Head-in Color Side-Tone Color ncoxi--. -- AL 0.40 Aa 0.30 Ab * 0.30 0.30 0.30 0.36 E 0.58 0,52 REPRESENTATIVE COLOR RANGE 867-97591 867-97841 867-97847 867-97692 867-97223 867-97217 867-97593 867-97344 867-97846 867-97219 867-97589 3. Dark metallic colors should be approved visually to the sparger standard and instrumental measurements obtained. Feedback data . must be forwarded to the Color Operations Section and numerical tolerances will be established as rapidly as sufficient feedback data are available to predicate meaningful tolerances. 4, The following overall procedures are rectapended for batch shading and approval: a. All panels be sprayed Via sparger following 1M-349-P. b. All panels to be measured via Model V or V-A Colormaster and feedback data using FF-7746 provided to C.O.S. c. Batches shaded.to within the above tolerances (Table II) should be considered acceptable regardless of minor visual color deviations. It should be borne in mind fSaFIth improved reproducibility of sparger spray techniques can be coupled with broader visual tolerances and still achieve greater bateh-to-bateh and plsat-to-plant reproducibility. ' DUP030002756 -S* 4* If doubt exists as to the visual validity of an instrumental watch held to within the above tolerances the sample may be sprayed conventionally at the viscosity extremes allowed by the conventional spray procedure and examined relative to the conventional standard. If the color of these two panels bracks* the standard the desired color match has been achieved. It must be remembered that the inherent variability of the conventional spray (see Tables III and IV) is such that even under the above circumstances a misleading representation of the color may be obtained, e. tore must be exercised by personnel visually examining a color match to assure elimination of personal preferences for a given color, since data on the 1965 colors clearly show that the customer has accepted and used without comment, production batches that have completely bracketed the color standard. The numerical values for these batches are in excess of the ones proposed as tolerances for color control purposes for 1966 colors. CCMXNUING ACTION 1. All plant locations must report to the Color Operations Section the numerical values at which each batch was approved, indicating whether the approval was based on numerical values, on visual assess* went of the sparger panel versus the sparger standard, or on a con ventional sprayout without direct reference to the sparger panels. Color tolerances at best are wily guide lines for operation and though carefully established based on available data, are to sane extent subject to the desires of the customer. The Color Operations Section should be advised by any plant if there are informal or formal color complaints lodged against computer-controlled batches shipped to the customer. Close and continuing liaison will be necessary in order to establish and maintain the full validity of color tolerances. 2. Accmilation and analysis of feedback data on "dark" and saturated metallic colors should be continued in order to establish valid tolerances for this color range. 3. Colors exhibiting inadequate formulation characteristics in the Selection of pigmentation or aluminum may necessitate either re formulation or their Divisional approval to sane numerically offset standard in order to allow Production to manufacture within an acceptable range of variability. A potential example of the problem is 867-97223 (see Figure H). ACfflCTiEMeir This study has been made possible only through, the cooperation of the various manufacturing plants who have faithfully forwarded panels, feedback data, etc, to C.O.S. DUP030002757 -6 - RESTRICTIONS This report is Restricted - Company Confidential and must be limited to personnel familiar with the Instrumental Shading System. PREPARED BY: 'vf*. J; XT37TOTO' COLOR OPERAH i SECTION APPROVED BY: ammt DIVISIONAL PROCESS ENGINEERING SUPERVISOR RHVj IL DUP030002758 II - DISCUSSION Rarely, if ever, does the manufacture of a batch achieve a perfect color match relative to a given standard. This is especially true in the field of metallic finishes where a compromise match between head-in and side-tone color must usually be accepted. Industry has, by necessity, relied cm visual appraisal techniques by highly trained aid competent specialists who have spent years developing their ability to correlate visual color differences with differences that they feel are acceptable to a customer. Color tolerances can vary for each color and can vary for each customer; consequently, the person responsible for deciding whether a match is adequate must rely heavily on a "color sense" that can best be described as intuitive. The success of the visual appraisal system is amply demonstrated by the -volume of business enjoyed by. the Finishes Division in the highly competitive Automotive field. To achieve this degree of success, the visual appraisal system has been subject to constant refinements over the years and it would appear that no sig nificant improvements in. the. overall system can be anticipated in the future. The trend in industry, however, is toward ever-decreasing color tolerances result ing from an increasingly stringent market reqmrammt. Additionally, the Finishes Division will, in the 1966 model year, depart from a manufacturing procedure that has in the past resulted in production from one Finishes plant being more or less restricted to a given Automotive plant. During the 1966 model year, production .*for several of the large-moving metallic finishes' will be manufactured simul taneously at the Flint and the Toledo Plants, and, in turn be shipped to a cornea Automotive plant. This procedure will result in a direct comparison by the customer of batch-to-batch variability inherent within a plant and more importantly, plant-to-plant uniformity* This is a situation not normally encountered in the Automotive segment of our business. Experience in other segments of our business has clearly shown that the problems of batch-to-batch uniformity, though signi ficant, may'be minor to the overall problem of attaining and maintaining an acceptable level of plant-to-plant mifbradty. - VISUAL APPRAISAL PROCBDDRES in order to make's -visual assessment of a color, it- is first necessary to prepare - a representative sample, usually a panel of the batch for comparison versus the standard panel. All panel -preparation procedures are subject to an inherent variability. Iba basic criteria of panel repeatability and reproducibility must be met as the visual examination and' comparison of colored materials is analogous to the most complex and exacting analytical procedure performed in a laboratory. In assessing the quality of the match, panel position, color of the surround, and lighting are equally critical. The observer, that is, the parson making the decision regarding the quality of the match, is the remaining factor involved in . the visual assessment process. Light from the source illuminates the object. A part of this light is reflected either unchanged or spectrally modified by the absorption characteristics of the panel into the eye of the Observer. From the response of the eye the reaction is conveyed via the optic nerve to the brain tr-\ where the sensation of color is evoked, The observer then decides whether the ' ) color is or is not within some intuitive specification or tolerance. As an aid to arriving at a decision, use is sometimes made of limit standards. The usual procedure is to .assign the last approved batch as a limit standard and tbs approval is given if the current batch falls between the limit .standard and the official standard. DUP030002759 3 Visual appraisal techniques at first glance would appear to be thoroughly reliable assuming the basic conditions outlined above are adhered to. In the assessment of colors, one sometimes encounters variables and transient effects which make positive conclusions very difficult, variables which appear to shift either with time or with relative position. Physical influences such as secondary light sources may often be contributing factors. The quality of the light for visual color comparisons is too often unstandardiaad and may be more or less changed by reflections from sources all too commonly ignored or unsuspected. For example, extraneous laboratory lighting nay influence the characteristics of the light under which the match is made and/or afternoon sunlight entering the laboratory can change the light spectral characteristics. The color response of the two eyes is not necessarily identical and a color may appear to fluctuate slightly in shade as the response of one eye or the other dominates when viewing the panel. This variable is "retinal rivalry" and the shifting response is linked with a shifting of attention, conscious or otherwise, from one eye to the other. The sensitivity of the eye for spectral hues changes with intensity of illumination, and from high to lew illumination intensities the shift in response is toward colors of shorter wavelengths; e.g., greens appear bluer. This is the "PURINJE EFFECT" and is a property of all normal eyes. The intent or mood of the observer can also influence what he sees and frequently a color panel can be rejected in the morning if the observer is under some form of pressure, and yet later be considered to be acceptable due to a change in environmental conditions. Personal fatigue can influence both perceived color differences and attitudes toward acceptance. The factors influencing visual perception and color interpretation for one observer apply to all observers. Mo two observers can be considered identical, as each will react differently through varying stimuli. The human eye and brain are unsurpassed as a null detector but with a barely perceptible difference be tween the standard and the sample, there my be considerable difference of opinion amongst trained observers as to the nature of the color difference and the overall significance relative to a given customer's requirements. REPEATABILITY AMD REPRODUCIBILITY Under the most optimum viewing conditiong with highly trained observers, the quality and uniformity of successive matches are dependent upon a panel being representative of the color of the batch being assessed* As with all analytical procedures, all facets of the process must be reproducible. With visual appraisal procedures, the criteria of repeatability and reproducibility has been assumed but rant proven to be firmly established. Repeatability can be defined as the ability of one operator to repeat a given operation successively and obtain the same end results. Reproducibility is the ability of several operators at a given location and/or separate locations to perform the same operation using samples from a common source and achieve the same results as the original operator. There is ample .evidence,to show that the fundamental precepts of repeatability and reproducibility are not achieved with metallic-type finishes, nor is the basic criteria of repeatability with a function as fundamental as the preparation of conventional standards or batch panels achieved. DUP030002760 -9 E. A. Miller, as a part of the development program leading to the sparger spray procedure, undertook a separate study to determine the variability existing between conventionally sprayed metallic standards. His work indicated that the repeatability and reproducibility of these standards had declined markedly as a result of major quality changes that affected the sprayability of the finish. The data were developed for the medium Turquoise color family and encompassed standards for three model years. Data indicated that the 1963 color standards (867-line quality) differed sufficiently from on another that if one panel were selected as the prime standard, many of the standards would haw to be rejected if the normal visual color tolerances used in batch assessment were maintained. The standard panels were selected at random and measured for "head-in" color using a wide aperture Colormaster, "Side-tone" color data were reported in terms of two-tone as measured on the J-5 Two-Tone meter. The variability found between panels within the sets is reported in Table III. TABUS in STANDARD PANEL VARIABILITY Model Year 1961 1962 1963 Quality 887 887 867 ah 0.60 0.80 2.00 aa 0.20 O.SO 1.20 Ab 0.20 0.50 1.20 an 3 5 12 Sixteen "O.K." color, panels for 867-95183 were in turn measured versus a single standard panel. Similarly retain samples for ten O.K. colors were resprayed and measured against the same standard panel. The variability existing between the panels is shown in Table IV. BATCH-PANEL VARIABILITY TABLE IV 867-95183 aL as ah an "OK" Color Panels 2.20 1.50 1.50 Retain Resprays 2.00 1.10 0.75 12 6.0 The "O.K." panels were sprayed at different times, whereas, the retains were sprayed at the .same time by the same operator using the same gun, tins minimizing the in herent variability of the conventional spray. In the preparation of sparger standards, 14 panels are prepared per deck and successive "decks" are prepared until an adequate number of standards is available. All panels are measured for the 6 value using a Model V-A Golormaster equipped with the dual angle measuring attachment. In addition, five panels selected at randan from each deck are measured for G, R, B values at both positions. These data are summarized in Table V.- Through the us of the Carrousel unit, the basic DUP030002761 - 10 requirement of repeatability and reproducibility for standards preparation has been met, A similar condition exists for the single panel production sparger wait, these standards, though developed originally for instrianental control purposes, can also be used equally as successfully for visual color assessment purposes. the level of repeatability and reproducibility which has been achieved with the sparger spray procedure may be assessed fro the variability existing between standards prepared on the "Carrousel'' standards unit for 1966 model year colors as shown in Table V. TABLE V SPARSER PANEL VARIABILITY - 1966 COLORS Product Code 867-97217 867-97589 - 867*97.593". 867-97223 867-97219 867-97844 867-97841 867-97591 ALp 0.04 0.36 0.22 0.09 0.13 0.11 0.23 0.13 AaF Sbp 0,10 0.10 0.20 0.06 0.10 0.35 0.17 0.05 0.06 0.21 0.12 0.06 0,07 0.16 0.16 0.03 0.06 0.06 0.34 0.11 0.07 0.02 0.20 0.12 m 0.33 0.07 0.32 0.23 0.17 0.20 0.09 0.14 AbH 0.10 0.06 0.09 0.19 0.05 0.07 0.12 0.04 COLOR TOLERANCE A color tolerance will seldom be equal to the acceptance .pattern established by a customer. In establishing this color tolerance, due regard should be paid to the acceptance pattern, since this pattern is established by actual shipments acceptable to. the customer and will include the occasional "maverick'' that somehow slips through all control laboratories. Acceptance patterns do serve to indicate whether the proposed tolerance is realistic and the difference existing between the assigned tolerance and the acceptance pattern can be con sidered as a safety factor between a complaint/no complaint situation. Under visual assessment conditions, a color tolerance must exist in the observer's mind and is a function of his experience, the tolerance will vary and is, to a large extent, dependent on the color preference of the observer. The observer may favor the light green side of standard, the dark yellow side of .standard, etc,, and to him any batch not falling inside his personal preference color dimension is unacceptable. Color memory is notoriously poor, the observer is himself subject f many factors, and as a result, his preference range will shift at frequent intervals and will often .reverse,' accepting today what a week ago was somewhat objectionable. Thoroughly trained observers can differ amongst themselves with respect to color preference, consequently, it is not uscanppa to have at any time diametrically opposed color preferences for the same color at different .wmfacturiag- locations (Figures IV - XI). limit standards bracking the standard color could be used as DUP030002762 - II a guide in visual assessment. Limit standards would define an acceptable color tolerance in all dimensions and if the batch fell anywhere within the parameters defined by the standards the batch would be approved. Divisionally, limit stan dards are not employed nor feasible and reference is usually made to the last batch panel and the standard panel. A "fan-out" of passing batch panels would be more truly representative of limit standards, bearing in mind that this would constitute the acceptance pattern for the color. This procedure is seldom followed due to the demonstrably poor repeatabiiity/reproducibility of conven tional metallic spray procedures. For illustrative purposes, let us assume that we have a basket ball, a volley ball and a tennis ball. Let us also assume that the volume encompassed by the basket ball is equivalent to the color tolerance for a given product; i.e., any batch falling within the basket ball would be acceptable to the customer. The three-dimensional characteristics of a color and/or color tolerance are shown in Figure I. The repeatability/reproducibility factor applicable for conventional metallic spray procedures can be represented by the volume of the volley ball; i.e., the true color location of the batch will be somewhere within this volume. Simlarly7"tK repeatability/reproducibility factor for the sparger spray procedure can be considered equivalent to that of the tennis bail. If the volley ball were placed inside the basket ball, the overall difference between the two is extremely small (see Figure II) and represents the visual tolerance that must be maintained to insure that all batches are within the tolerance (basket ball) acceptable to the customer; whereas, if the tennis ball were placed inside the basket ball, the overall difference is considerably larger, repre senting the larger visual tolerance that can be used with the sparger system to achieve the same end tolerance (see Figure III). With a system where the repeatability/reproducibility factor is very poor (large) the allowable deviation permitted for a color must be small, where the repeatability/rsproducibility factor is small, the allowable color deviation can be much larger. In actual practice we have recognized that conventional spray procedures restrict the color difference between a batch and standard to a barely perceptible visual difference if we are to remain within the acceptable color tolerance zona. With a more reproducible spray procedure such as the sparger spray procedure, the color acceptance zone can be mush larger visually and still be acceptable. This zone can be ascertained by preparing sparger panels from samples taken from batches that have been considered acceptable based on the conventional spray procedure and customer usage. A color tolerance zone can then be determined from these data. The batch-fo-b&teh uniformity of the material shipped to the customer will remain within the parameters defined by his acceptance pattern based m past experience, and, from a manufacturing viewpointj the degree of color variability that can be permitted is quantitatively defined. Color tolerances directly or indirectly will reflect the degree of confidence the observer has in the control system. A color tolerance can be made so restrictive that the ability to compete effectively is unduly hampered by unrealistic costs. By the same token, unrealistically broad tolerances cannot be assigned without the risk of customer complaint. Color tolerances will vary for different product lines, different end uses, and within a product line may vary with the lightness and saturation Characteristics of the colors* In addition to visual color tolerances it is possible and desirable to assign numerical color tolerances for a given color. If these numerical tolerances are properly detetmnsd, all batches falling within the tolerance parameter should be visually acceptable. The eye wsf be the .final judge of a color match and the eye DUP030002763 FIGURE I Customer Acceptance Tolerance (Basket Ball) Comrentiooal Spregr Variability (Volley Ball) *;}fh/:Mhli-, . i'i;;n,t ,h> Sparger Spray Variability {Tennis Ball) DIMESSIOHAL CHAMCTEBISSICS OF COLOR TOLERMCE b DUP030002764 FIGURE 11 DMHSSOSAL OF COLOR TOLHSMCI COOTEmOML SPRAY PROCEDURES DUP030002765 DIMEHSIOHAI* CRAMC2T1RI8TICS OF COLOR TOLERANCE SPARGER SPRAY PROCEDURE FIGURE III A, &L DUP030002766 * 12 and the tetrusent must agme as to the description of the cote deviation existing between the hatch aid standard. If disagreement exists between the eye and the roteisneter as to the gaitefe of the cote dotation aid the tetreasestal fates are choked aid found to be correct, the visual inter pretation m&y he in error due to factors discussed trader 1'Visual .Appraisal," HKFBKSWm, the data reported below are based era metallic "Incite*" production for four plants for the .1865 modal year unless otherwise specified.. All batches were approved visually based on conwatiarml spray procedures with a sparger-type panel prepared after the visual approval was given.- In many eases the sparger panels were measure both at the plant preparing the panels, and at the Cote Operations Action. The C,0,S. has on file cote data for over 500 oefcaUlp , batches., including mtems&iate shading data for '1964, 1965 and 1966 colors, These data are not appended to this report sine their inclusion would be superfluous. BAim-TO-MlUi W1TOTO Production batches that were visually approved following the convmtimal spray procedure were, after the fact, sprayed via the sparger technique. Sparger prepared panels for consecutive batches approved via visual procedures were requested .from each plant location in order to feteraiio whether a plant offset was required as a prerequisite t successful instnaisental shading. The data indicate that when the variability inherent in conventional spray is eliminated, the b&txh-toHbatds unifoimty is broader than generally realised, the cote differences existing between the batches when sprayed via sparger technique are, in many cases, of sufficient magnitude to result in a visual rejection .of the sparger psiel versus the sparger standard if the mm visual tolerawes used for approving visual spray were imposed, fee Figures IV and V showing consecutive batches for the four locations plotted in numerical sequence. These data not only Indicate greater differences than anticipated between b&tdk-te-tratch production, but also indicate a lack of plant-to-pleat uniformity. Note in Figure V (Aa, Ab chart) the epafent grouping of each plantSs prediction, .indicating that each plant approved the mm product cod with different raphasis era the preferred cote. The different grouping of each plant9-s production in dicates the visual presence of a preference, factor 'awl is mt indicative of a tewl&tiiig weakness. The- data d indicate that a .Divisional problem exists with 061-91223 with respect to side-ton color control. Note that 14 out of 16 batches reported were approved on the light, illy, less glaroow side of stendani, indicating the need for a brighter, coarser metilllc flate if the batch is to consistently match th standard. The data reported for 867-97217 indicate a plant pstew for ta* control with one plant, represented by the green numerals, approving production that .is considerably redder In CGssparisea to the Divisional spprm?al pattern. Note also that tins plant represented by the red .numerals fevered the yellow side stan dard, with the remaining two plants randomly distributed with respect to ysllw-bteness . With respect to side-tone cote, the problem of lightness control is not as prmmmcM md m equal rasher of .batches were approved either lighter or darker than standard* This would Indicate greater fomslating leeway with 867-87217 than with 867-97223, (*) Du Pont ifeglstered Trademark DUP030002767 DUP030002768 DUP030002769 13 CUSTOMER ACCEPTANCE TOLERANCES Data for six product codes; 867-97212 867-97215 867-97216 867-97217 867-97219 867-97223 Figure VI Figure VII Figure VIII Figure IX Figure X Figure XI have been plotted for two or more plant locations. Bata are for consecutive batches although they are not numbered as in Figures XV and V. Color tolerances, although specific for each color within the color solid, can be generalised for the six colors studied in this report. The customer acceptance pattern indicates that a lightness deviation equal to two units in the Modified Adams1 Coordinate System is acceptable. The data also indicate that the customer has accepted production, showing a greater deviation in the green-rad axis than in the yellow-blue axis. The acceptance range in the aa direction approximates one unit, in the Ah direction 0.8 units and these tolerances are shown on the transparent overlay atoompanying each chart. Customer tolerances for "head-in" color can therefore be generalised to a deviation of: Aa * + 0.50 Ab * + 0.40 equivalent to an E approximating 1,2 KBS units, Note that in most cases, batches have been shipped to the customer that exceed the above color tolerances, in side-tone color exhibits approximately the same color deviation as does the head-in color. With the possible exception of 867-97219, the customer has not objected to receiving prediction that is on the light, millcy, less glamorous side of stan dard. In effect, the "two-tone" characteristics of a color are not as critical as generally believed. Color tolerances for side-tone color can therefore be generalized to: Ah 0,75 6a m * 0,49 6b 88 * 0,30 or an E equal to 0.9 MBS mats around the mid-point of an offset centroid. Again note that the transparent overlay showing these tolerances does not include all batches, indicating a tolerance in excess of that stated above is acceptable. COLOR TOLERANCES - CONTROL Based on the 1.965 model year data, light and medium metallic colors controlled instnateatally to: AL ** *_ 0.40 6a * 0.30 Ab * * O.S) should be visually acceptable to the customer. . Colors controlled to the above values would exhibit color difference (E value) of 0,58 MBS units contrasted to an E value of 1,20 MBS units established during 1965 model year. DUP030002770 o - 14 Side-tone should similarly be controlled to: &h * + A*-* Ab * .30 . 30 .30 or equivalent to an E of 0.52 units versus the 0.9 HBS units established for 1965. The overlays show the proposed instrumental color tolerances for the six colors involved (dashed lines) and many of the 1965 shipments would not be considered acceptable based on these new tolerances. There is sufficient formulating latitude to place the colors within these tolerance tones assuming that no visual preferences exist. These same tolerances should be applicable to the 1966 colors that fall in the same color family and are of the same lightness range. f\ ') DUP030002771 -Ab q v+ i a aanoii i DUP030002772 If- * ---- 1 1 s %* * -----------1 c ----------1** --------------- T--]BT * 'if "i........ ....... ,u kw *w ------------ LkJ. d afc~4 X B *-< \ xK * o a ** r" -----11 h .......r*. H K* Uj l__? ie* sc_ X c*- *-' cO** GO o O - f' -ri w T x f----- l---- ... <-i % !\ >-> - < y; * K& *1 O H !* t i< i o H nr i 'a L * 4 --------------- ------ x--a-------- * C m r- * ! -- 1 c 1 t -L * iX x* _______i < --___,,,yi. ....JX ** Fionas t o /% DUP030002773 FIGURE T i l l DUP030002774 FIG URE IX - 6b nili i i> DUP030002775 -A t -Ab X 8SBX* DUP030002776 f ig u r e XI 1r &*> IV " S__A _____ -fib O H 1:1 \ *+q o *< >* c r~ si i * 1 *< L_ *. , o w J i 15. D 4 4+ o H >^4pwi + 1* K r pf * l * .J o dL at o A <I3 j DUP030002777 APPENDIX 1 TUCKER PROGRAM - 1966 COLORS At the outset of the 1966 model year, the Tucker Plant, based on their successful 1965 model year program, proposed shading the light to median metallic colors instrmeatally to tolerances of: AL * + 0,50 Aa * + 0.30 Ab * 0.30 E * ~ 0.52 for head-in and side-tone color. The C.O.S. agreed that the proposed tolerances appeared to be realistic based on a preliminary examination of approval data available for the 1964-1965 colors, and were in favor of the Tucker proposal. It was agreed that all batches would be shaded to the above values unless it became apparent that product formulation problems precluded the attainment of the limits. This formula limitation proved to be the case for 867-97223 (Figure XII) as would be expected based on 1965 experience (Figure XI). It was further agreed that all batches would be checked -visually versus the sparger standard and all colors, excluding carryover colors, would also be sprayed conventionally and checked visually versus the conventional standard as a further precautionary measure. It became apparent that the restriction placed on AL for head-in color was too stringent, and based on a more thorough review of 1965 data has been increased to the basic limits proposed in this report. The Tucker Program has demonstrated that the proposed tolerances of: Bead-In Side-Tone AL Aa 0.40 0.30 0.30 AL 0.30 Aa * + 0.30 Ab 0.30 are feasible. Batches shaded to the above tolerances are good visual matches and are acceptable to the customer. The overall accuracy of the .system, based m the use of the Model V-A, is more than adequate and repeatable and reproducible results can be consistently 'obtained.. To date some 39 batches, involving 10 - 1966 colors have been shaded successfully to color tolerances approximating the proposed tolerances without adverse effect on batch cycle time. The overall hits per batch average is 2.5, however, included in this total are a high number of aliment batches. Twenty-three batches since the initial alignment have averaged 1,5 hits for approval. DUP030002778 APPENDIX 1 (Cant'd) The product codes included in the lltcker Program and the number of batches shaded to date are shown in Table VI . TUCKER PLANT - 1966 COLORS TABLE VI Product Code 867-97213 867-97217 867-97219 867-97223 867-97224 867-97561 867-97589 867-97591 867-97592 867-97593 Batches (as o 10/11/65) 1 6 4 6 2 4 3 4 1 8 Data, though limited, has been plotted for five colors. The colors are; 867-97223 867-97217 867-97591 867-97593 867-97561 Figure XII Figure XIII Figure XIV Figure XV Figure XVI The tolerances, as proposed in the subject report, are indicated in red. DUP030002779 DUP030002780 DUP030002781 DUP030002782 I DUP030002783 T1 DUP030002784 O GENERAL REFERENCES: APPENDIX 2 Visual Perception 1. An Introduction to Color - Evans - John Wiley Sons 2. The Practical Art o Color Matching - Peacock American cyanamid Cmpany 3. Color and Instrumentation - E. Rhael - Sandos, Inc, 4. Color Measurement with Eye and Other Instruments N, Saltman - ''Dye Stuffs" Vbl, 45, No. 1 5. Encyclopedia Britannica - Vol. 6 6. Encyclopedia Americana - Vol*s. 7, 21, 22 7. Color Vision - National Association for the Prevention of Blindness, Inc. ;> Instrumental Measurements 1, Color in Business Science Industry - Judd 4 Nyszecki 2* Hi Measurement of Ctolcar - W. D. Wright - John Wiley Sons, MuMUian, N. Y. 3 Preparation of "bucite*" Metallic Color Panels E. A, Miller - S/27/62 4, Instnaeental Shading of Metallic - R, H, Vining - S/7/65 5, Color InstMBentation in Production - R. Hansen Official Digest - Vol, 37, Mo, 487 . 6, Color Problems in the Graphic Arts - Inter-Society Color Council, Inc. O (&) Du Pont Registered Trademark DUP030002785 APPENDIX 3 cm NO. DISTRIBUTION LIST 1 W. F. H. MATTLAGE WXIMNGTON F. F.) IN J. W* NESTOR ) TORN 2 F. H. BEADDBS B. B. FURST n " ) IN ) TORN 3 C. F. KALB B, T. BREUBR " " ) IN ) TORN 4 C. W. THEOBALD J. .S, ALLEN " " ) IN } TORN 5 R, B, DAVIS t. i* 6 C. E. CORRELL TOIEK) 7 F. S. BAKER TOCXER 8 R, A. DE WALL CHICAGO 9 N, I, DRESS PHILADELPHIA 10 R. J. KNAKE PARLIN 11 P. 0. PEASLBf SOUTH SAN FRAN* 12 H. R. STINE FLINT 13 C, H. THOMPSON FORT MADISON 14 R. W. LADKRELL MARSHALL DEVELOPWNT LAB* 15 s* i. mmujm FLINT DEyELOPMENT LAB. 16 W. H. TOOLE 17 W. E. CRMMSR J. G. HALL HHLAIBLPeiA m mm o f c a n am - Mo n t r eal ) in " " " - AJAX ) TORN 18 R. H. VXNXNG 19 W. H. EDWARDS 20 PROCESS MNEERING GROUP FILE DUP030002786