EPA/600/2-85/107
                                            September 1985
              TRANSFER  EFFICIENCY OF
      IMPROPERLY MAINTAINED OR OPERATED SPRAY
      PAINTING EQUIPMENT  SENSITIVITY  STUDIES
                        by


                  K. C. Kennedy


               CENT EC Corporation
              Reston. Virginia  22090
         Contract Number  68-03-1721, Task 1
                EPA Project Officer:

                Charles  H.  Darvin
Air and  Energy Engineering Research  Laboratory
       Hazardous Air Technology Branch
      Research Triangle Park, NC  27711
      AIR AND ENERGY ENGINEERING RESEARCH LABORATORY
           OFFICE OF RESEARCH  AND DEVELOPMENT
          U.S. ENVIRONMENTAL PROTECTION  AGENCY
            RESEARCH TRIANGLE  PARK, NC 27711
                     EPA Headquarters Library

-------
                                  TECHNICAL REPORT DATA
                           jP'.au nod Imuniciumt on tht rtrrnt btfort compltfint)
         NO
   EPA/600/2-35/107
I TITLE AND SUBTITLE
Transfer Efficiency of Improperly Maintained or
 Operated Spray Painting Equipment Sensitivity
 Studies
                                                         i. REC
                                                         B. REPORT DATE
                                                          September 1985
                                                         >. PERFORMING ORGANIZATION CODE
 T AUTMOR(S)
  K.C.  Kennedy
                                                       I. PERFORMING ORGANIZATION REPORT NO
  i PERFORMING ORGANIZATION NAME ANC AOCRESS
  Centec Corporation
  11260 Roger Bacon Drive
  Reston. Virginia  22090-5281
                                                       10 PROGRAM ELEMENT NO
                                                       11. CONTRACT/GRANT NO.
                                                         68-03-1721. Task 1
 13 SPONSORING AGENCY NAME AND ADDRESS
  EPA, Office of Research and Development
  Air and Energy Engineering Research Laboratory
  Research Triangle Park, NC  27711
                                                       13. TYPE OP REPORT AND PERIOD COVERED
                                                        Task Final; 6/83 - 6/85
                                                       14. SPONSORING AGENCY CODE
                                                        EPA/600/13
  6 •^""•"TAnviioTM AEERL project officer is Charles H.  Darvin. Mail Drop 54. 919/
  541-7633.
  6. ABSTRACT _.	'	,
           The report gives results of an investigation of the  impact of common indus-
  trial operating and  maintenance practices on the efficiency of spray painting sys-
  tems. The investigation included independent research, as well as assistance from
  both representatives of the spray painting equipment manufacturing industry and
  users of spray painting equipment.  The results  indicate strong directional  response
  in painting efficiency to certain common painting practices.,'	
                              KEY WORDS AND DOCUMENT ANALYSIS
                 DESCRIPTORS
 Pollution
 Spray Painting
 Maintenance
 Operations
                                           b IDENTIPIEnS/OPEN ENDED TERMS
                                          Pollution Control
                                          Stationary Sources.
                                          Transfer Efficiency
                                                                       COS
-------
                      NOTICE

This document has been reviewed in accordance with
U.S. Environmental Protection Agency policy and
approved for publication.  Mention of trade names
or commercial products does not constitute endorse-
ment or recommendation for use.
                        11

-------
                             FOREWORD


     When energy and material resources are extracted, proc-
essed, converted, and used, the related pollutional impacts on
the environment and even on health often require that new and
increasingly more efficient pollution control methods be used.
The Air and Energy Engineering Research Division (AEERD) at
Research Triangle Park, North Carolina, assists in developing
and demonstrating new and improved methodologies that will
meet these needs both effectively and economically.

     The research described herein was undertaken to address
how spray painting transfer efficiency is affected by operating
and maintenance parameters.  Air pollution impacts, energy,
and materials resource conservation are affected by loss of paint
and solvenc in poorly operated or maintained spray painting
facilities.

     Pour major types of spray painting equipment were tested
to determine their sensitivity to certain preselected operating
or maintenance parameters.

     This is the first published research into a very expensive
industrial and environmental problem.
                              ABSTRACT


      This report is submitted in fulfillment of Contract
 Number 68-03-1721, Task 1.  It describes sensitivity studies
 conducted on four types of spray systems to determine the
 effects of improper operations or maintenance on trtnsfer
 efficiency.  A Draft Standard Transfer Efficiency Kethod was
 used for the test program.  Three different target configura-
 tions were painted for each spray system.

      Test results show the strong effect proper selection of
 spray conditions has on transfer efficiency.  The particular
 3evel of response for specific factors varies frou spray system
 to spray system, and from target configuration to target con-
 figuration.  Case-specific regressions were developed for each
 spray system and target type.  These are presented and discussed
 in the report.
                                 ill

-------
                            CONTENTS
Foreword	ill
Abstract	ill
Figures	vii
Tables	viii
Abbreviations and Symbols 	   x
Acknowledgment  	  xi

  1.  Introduction  	   1
  2.  Conclusions   	   3
  3.  Mass Flow Rate Comparison   	   4
  4.  Air Atomized Conventional Spray Equipment 	   7
          Equipment Description 	   7
          Operating and Maintenance Variables 	   8
          Experimental Design 	  10
          AAC Test Performance	13
          Test Results  	 ...........  17
          Statistical Analysis  	  17
          AAC Test Conclusions	25
  5.  Air Atomized Electrostatic Spray Equipment  	  30
          Equipment Description 	  30
          Operating and Maintenance Variables 	  31
          AAE Test Performance	38
          Test Results	40
          Statistical Analysis	40
          AAE Conclusions	44
  6.  Airless Conventional Spray Equipment  	  50
          Equipment Description 	  50
          Operating and Maintenance Variables 	  50
          Experimental Design 	  52
          Test Performance	52
          Test Results	56
          Statistical Analysis  	  56
          ALC Conclusions	59
  7.  Airless Electrostatic Spray Equipment 	  64
          Equipment Description 	  64
          Operating zi,d Maintenance Variables	65
          Experimental Design 	  ....  65
          ALE Test Performance	70
          Test Results	73
          Statistical Analysis  	  75
          ALE Conclusions	77

-------
                       CONTENTS  (Continued)


  3.  Comparison of Targets	83
          Background	83
          Comparison of Variables Identified as Significant  83
          Worth Assessment of Three Target Configurations .  85


Appendices


  A.  Draft Standard Method  for Spray Fainting Transfer
      Efficiency Operations  and Maintenance Testing ....  87
  B.  Quality Assurance/Quality Control Plan Sensitivity
      Studies on the Effects on Transfer Efficiency of
      Improperly Maintained  or Operated Spray Painting
      Equipment	114
  C.  AAC Test Equipment and Paint Specifications	136
  D.  AAE Test Equipment and Paint Specifications	139
  E.  ALC Test Equipment and Paint Specifications	142
  F.  ALE Test Equipment and Paint Specifications	145
  G.  Glossary of Statistical Terms 	 148
                                vl

-------
                           FIGURES

Number                                                Page
  1   Air atomized conventional air cap
      (frontal view) 	  16
  2   Air atomized electrostatic air cap
      (frontal view)	33
  3   Air atomized electrostatic electrode position
      test levels	34
  4   Airless paint spraying system	51
  5   Airless electrostatic air cap showing
      test levels for electrode position	72
  A-1   Target configurations; for air atomized
        conventional and electrostatic spray guns. . .  90
  A-2   Target configuration for high speed bell ...  SI
  A-3   Set-up for paint supply equipment and platform
        scales	97
  A-4   Permissible methods for measuring conveyor
        speed	*	101
  A-5   Vertical cylinder wrapping technique 	 105
  A-6   Flat panel foil attachment technique	106
  B-1   Project organization as related to corporate
        structure	120
  B-2   Data validation responsibilities 	 127
                                vii

-------
                            TABLES

Number                                                Page

  1   Mass Flow Comparison Data	   5
  2   Operating and Maintenance Variables for AAC
      Spray Equipment	   9
  3   Experimental Variables Selected for Testing
      AAC Spray Systems	10
  4   AAC Experimental Design 	  11
  5   Levels of Operating and Maintenance Variables
      Tested on AAC Spray Painting Equipment  ....  15
  6   Air Atomized Conventional Test Results  ....  18
  7   AAC F-Statisties and Associated Probability . .  21
  8   AAC-FP Comparison of Predicted Versus Actual
      Transfer Efficiencies 	  26
  9   AAC-VC Comparison of Predicted Versus Actual
      Transfer Efficiencies 	  27
 1G   AAC-Graco Comparison of Predicted Versus
      Actual Transfer Efficiencies  	  28
 11   Operating and Maintenance Variables for AAE
      Spray Equipment	32
 12   Operating and Maintenance Variables for AAE
      Spray Painting Equipment  	  36
 13   AAE Experimental Design	37
 14   Levels of Operatinq and Maintenance Variables
      Tested on AAE Spray Painting Equipment  ....  39
 IS   AAE Test Results	41
 16   AAE F-Statistics and Associated Probibility . .  46
 17   AAC-FP Comparison of Predicted Versi s Actual
      Transfer Efficiencies 	  47
 i8   AAC-VC Comparison of Predicted Versus Actual
      Transfer Efficiencies 	  48
 19   AAC-Graco Comparison of Predicted Versus
      Actual Transfer Efficiencies  	  49
 20   Levels of Operating and Maintenance Variables
      Tested on ALC Spray Painting Equipment	53
 21   ALC experimental Design	54
 22   Order of Performance of ALC Test Runs	55
 23   ALC Test Results	57
 24   ALC F-Statistics and Associated Probabilities .  59
 25   ALC-FP Comparison of Predicted Versus
      Actual Transfer Efficiencies  	  ...  60
                              vlll

-------
                    TABLES (Continued)

Number                                                Page

 26   ALC-VC Comparison of Predicted Versus
      Actual Transfer Efficiencies  	  61
 27   ALC-Graco Comparison of Predicted Versus
      Actual Transfer Efficiencies  	  62
 28   Operating and Maintenance Variables for
      ALE Spray Equipment	66
 29   Experimental Variables Selected for Testing
      ALE Spray Equipment	68
 30   ALE Experimental Design	69
 31   Levels of Operating and Maintenance Variables
      Tested on ALE Spray Painting Equipment  ....  71
 32   ALE Test Results	74
 33   ALE P-Statistics and Associated Probabilities .  78
 34   ALE-PP Comparison of Predicted Versus
      Actual Transfer Efficiencies  	  79
 35   ALE-VC Comparison of Predicted Versus
      Actual Transfer Efficiencies  	  80
 ';6   ALE-Graco Comparison of Predicted Versus
      Actual Transfer Efficiencies  	  81
 37   Comparison of Significant Factors Identified
      by Three Target Configurations  	  84
 38   Worth Assessment Model Comparing Target
      Configurations  	  86

 A-1   Nomenclature for Spray Painting Transfer
       Efficiency Tests   	 104
 B-1   Spray Painting Transfer Efficiency Precision,
       Accuracy and Completeness Objective 	 122

 B-2   Performance Audit  Requirements  	 122
                                  ix

-------
            LIST OF ABBREVIATIONS AND UNIT CONVERSIONS

ABBREVIATIONS
ASTM
AAC
AAE
ALC
ALB
EPA
Pan air -
PP
PSIG
O&N
QA/QC
TE
VC
VOC
  American Society  for Testing and Materials
  air atomized  conventional paint spray equipment
  air atomized  electrostatic paint spray equipment
  airless conventional spray equipment
  airless electrostatic spray equipment
  United States Environmental Protection Agency
  shaping air or horn air
— flat panel (target configuration)
— pounds per square inch, Ib/in  , gauge
- operating and maintenance
- quality assurance/quality control
- transfer efficiency
- vertical cylinder (target configuration)
- volatile organic compounds
UNIT CONVERSIONS

To go from

°C
cm
9
kg
Jcg/L
kPa
L
m
m
m/s
mVs
rps
s
                 To

                 °P
                 in
                 Ib
                 Ib
                 Ib/gal
                 psig
                 gal
                 ft
                 mils
                 ft/min (fpm)
                 ftVmin
                 rpm
                 min
Multiply by
 ,8°C
           32
2.54
0.0022
2.204
8.328
0.145 kPa -14.7
0.264
3.281
   3.937 x
 196.85
211R.8
   0.017
  60.0
        10

-------
                           ACKNOWLEDGEMENT

     The contributions of Graco, Incorporated are gratefully
acknowledged.  Graco donated laboratory facilities, test equip-
ment and supplies, and provided technical support for the tests
described herein.

     The knowledge and experience of     Ray Myers, Professor
of Statistics at Virginia Polytechnic Institute and State
University and author of "Response Surface Methodology" and
"Probability and Statistics for Engineers and Scientists," was
invaluable to this effort.      Myers devoloptd the experimen-
tal design and evaluated the test data using the Statistical
Analysis System.      Myers1 contributions as a statistical
consultant are grateful)y acknowledged.
                                 xi

-------
                            SECTION 1

                           INTRODUCTION


     This test program was initiated to develop information
about how spray painting transfer efficiency (TE)* is affected
by operating and maintenance variables.  Four basic types of
spray equipment were selected for the test program: air atomized
conventional (AAC), air atomized electrostatic (AAE), airless
conventional (ALC), and airless electrostatic (ALE).

     Operating and maintenance (O&M) variables were developed
for each of these equipment types through a literature search,
by industry contacts, and through manufacturers of spray equip-
ment.  Over thirty separate variables were identified.  Based on
an evaluation of the possible effect of each variable on TE (for
each type of equipment) and on the ability to simulate the
variable in a laboratory, the most significant variables were
selected for testing in this program.  Up to 7 variables were
selected for testing on a single equipment type.

     An experimental design was developed to address selected
operating and maintenance variables for each type of equipment.
In sach case the design consisted of a fractional factorial
design augmented by a "star" design and a set of replicates.
The process of identification of operating and maintenance
variables, and of developing appropriate experimental designs is
detailed in "Subtask Report: Sensitivity Studies on the Effects
en Transfer Efficiency of Improperly Maintained or Operated
Spray Painting Equipment."  Levels for testing each variable
were developed on-site prior to testing.

     Once the test program was well defined, CENTEC began
contacting companies with well equipped spray painting labora-
tories to locate a qualified test site.  The electrostatics
laboratory at Grace, Incorporated in Minneapolis, Minnesota, was
qualified and willing to participate in the test program.  Graco
provided the laboratory, spray equipment, technicians, and some
other materials  for  testing.
 *TE  is  the  amount of paint solids  deposited  on  a  target
 divided  by the amount of  paint solids  sprayed  at the  target
 multiplied by 100 percent.

-------
     The test program was conducted in February 1984.  Each
equipment type was the subject of a single experiment consisting
of up to 34 test runs.  Each experiment lasted one week, for a
total of 4 weeks of testing.  Sections 4, 5, 6, and 7 of this
report describe the performance and results of each experiment.
Section 2 summaries the overall conclusions from the test
program and accompanying test results.

     As described in the Draft Standard Test Method (Ap-
pendix A), all tests took place with two target types, flat
panel (FP) and vertical cylinder (VC).  Graco has, for their own
purposes, developed a transfer efficiency determination method
utilizing a different target design.  In all of the testing
described in the report, the "SPA" targets (Standard Test Method
Targets) were first painted at a given set of conditions,
followed by painting the Graco target set under the same condi-
tions.  Thus all transfer efficiency results in this report are
reported according to flat panel, vertical cylinder, and Graco
target results.

-------
                            SECTION 2

                           CONCLUSIONS


EFFECTS OF OPERATING AND MAINTENANCE VARIABLES ON TE

     AAC transfer efficiency was most strongly affected by
restricted air lines.  This effect was pronounced over all
three target types tested, and should be considered the most
prominent OSM variable tested for this type of spray system.
Fan air adjustments had a strong effect for two of three target
types, and should be considered a major effect as well.  Re-
stricted paint lines and booth air rates had significant
effects, although not as strong as restricted air lines or
fan air.

     AAR tranfer efficiency was effected by the highest
number of variables.  The most prominent effect was voltage,
followed by restricted air lines and restricted paint lines.
Booth air, gun cleanliness, fan air, and electrode position
also had significant effects.  These effects were not con-
sistent across all target configurations; the VC and Graco
target configurations were much more sensitive to AAE test
variations than FP targets.  The FP target restricts the
ability of electrostatic spray to wrap around and increase TE.

     ALC transfer efficiency was overwhelmingly affected by tip
erosion.  Restricted paint lines were found significant for
Graco targets, but the effect of tip erosion was overriding
in all cases.

     ALB transfer efficiency effects are similar to both the
AAE and ALC systems:  voltage and electrode position had the
largest effects, but effects of other factors were contingent
on target configuration.


OTHER CONCLUSIONS

     The Graco target  configuration was  found to be the most
sensitive target design  for detecting O&M effects.  Transfer
efficiency regressions had  the  tightest  fit for Graco  results,
and the target configuration was  the most comfortable  to use
experimentally.   Thus,  the Graco  target  represents  the most
desirable target  design  tested  to  date.   It is recommended
as the  standard  target  for  all  future TE tests.

-------
                            SECTION 3

                    MASS FLOW RATE COMPARISON


     The original test method called for determining the paint
mass flow rate using platform scales and a stopwatch.  (Refer to
Appendix A.)  The paint supply pot rested on platform scales,
and readings were taken as t^e paint flow was initiated and
stopped for each test run.  & stopwatch was used to time the
interval between scale readings.  Mass flow rate was determined
by dividing the total weight difference by the elapsed time.
While this method had proven satisfactory in determining paint
mass flow for TE testing, no airless spray equipment had been
tested.  Airless pumps create vibration problems in using
platform scales.  Several sources recommended using a mass
flow meter to determine paint flow rate.  These sources main-
tained that a mass flow meter would be simpler to use, easier to
read, and more precise than the platform scale/stopwatch
method.  \ mass flow comparison test was designed to evaluate
the benefits and any CA/QC implications of these two measurement
techniques.

     The QA/QC plan for the TE test program specified require-
ments for determining paint mass flow in terms of weight and
time measurements.  To ensure that the mass flow meter met
these requirements, a test was set up to directly compare
methods.  The platform scale was set up, calibrated, and zeroed.
The paint supply pot was placed on the scales.  The paint flow
was routed through the mass flow meter, which was also zeroed
and calibrated.  TE test runs were simulated by spraying atom-
ized paint into an empty spray booth using AAC spray equipment.
A test run time of 16 seconds was selected, similar to the time
for earlier runs at TE testing sites elsewhere.

     Seven test runs were performed.  In each run mass flow
meter and platform scale/stopwatch readings were taken simul-
taneously. The experimental results are presented in Table 1.
It is readily apparent from Table 1 that the standard deviation
of the mass flow meter data was significantly lower than for the
original mass flow determination method, while the average mass
flow rates were virtually identical.

     Table 1 presents the results of seven experiments
performed to compare flow rates as determined by the
the mass flow meter and by the platform scale/stopwatch

-------
method.  The  use of the platform scale and stopwatch  is
described  in  the Draft Standard Test Method and has been
used in all testing to date. The use of the mass  flow meter
would offer certain simplifications in the proposed test
procedure.  The experiments listed in Table 1 were undertaken,
then, to determine if the two flow rate measurement methods  gave
substantially equivalent results in order to justify  the  sub-
sequent use of the mass flow meter.


                   TABLE 1.  MASS FLOW COMPARISON  DATA


       Experiment        Platform scale 6 stopwatch         Mass  flow
        number           	method, g/s	       meter, g/s

          1                   10.16                       9.99

          2                   10.05                       9.99

          3                   10.09                       9.99

          4                    9.96                       9.99

          5                   10.04                       9.99

          6                    9.74                       9.91

          7                    9.93                       9.93

        Mean                   10.00                       9.97

    Standard Deviation           0.14                       0.03
      It  was determined that the new method  would be accepted
if  it provided readings within the accuracy specifications
set  for  the flow rate determination, +2 percent.  The standard
deviation of the flow rate determination  by scale and
stopwatch had been estimated to be 0.1 g/s.   At  a flow rate
of  10 g/s,  then, a maximum acceptable difference of 0.2 g/s
was  set, or a ratio of acceptable difference to  standard
deviation of 2.0.

-------
     The risk of falsely accepting the mass flow meter as
meeting these criteria, the g-risk, was set at  0.05.  That
is, no more than a 5 percent risk was desired that the sample
would be judged to have come from an acceptable population
when it really came from an unacceptable r/opulation.  The
ct-risk, or the risk that the two methods night be judged dif-
ferent when they actually are equivalent, was set at 0.1
(double-sided test).

     The required number of observations to control the o and
0-risks to these levels under the stated conditions is seven.
The significance of the difference between the means was then
determined by performing a t-test at the 0.1 level.

     The t-statistic as determined for the data of Table 1 is
0.458 with 12 degrees of freedom.  Since the value of t is well
below the critical value at the 0.1 level, it may be stated that
the two methods do not differ by more than 0.2 g/s at the stated
levels of risk.

-------
                            SECTION 4

            AIR ATOMIZED CONVENTIONAL SPRAY EQUIPMENT
EQUIPMENT DESCRIPTION*

     A Graco Model 800 manual air spray gun was selected for AAC
testing.  The Model 800 gun was considered typical of pressure-
fed external mix air spray equipment.  The spray gun was equipped
with a 021-10^806 Graco air cap, and a 1.2 nun fluid tip.  Paint
flow was manually initiated by opening a valve on the paint
supply line.  Paint flow was measured by a daily-calibrated
Micromotion mass flow meter (see Section 3).

     A standard Graco black enamel was selected as the test
paint. The paint averaged about 28 weight percent solids
during AAC tests and was adjusted to 29 seconds {12 Shell
cup) viscosity.  A 16 L  (4 gal) batch of paint was mixed for
AAC testing.  One batch was sufficient to complete AAC testing.
Paint was mixed and stored in a 20 L  (5 gal) Graco Model 210-393
pressure tank, which was kept in a temperature-controlled
booth.  The paint pressure tank, stirrer, regulators, viscosity
measurement equipment, and some supply lines were kept in the
temperature-controlled booth at 25 C + 1 *C throughout TE
testing.  All paint supply lines  (spanning 6-8m) were Insulated.

     AAC tests were conducted in a Dynaprecipitator water wash
spray booth.  Air flow was in the direction of paint spray,
normal to the targets.  Air flow was  adjusted by opening or
closing a vent on the booth exhaust duct.  With the vent closed,
the booth air rate was 61 cm/a  (120 fpm)  at the plane of the
targets, and varied from 51 cm/s to 71 cm/s (100 fpm to 140 fpm)
across the booth  face during testing.  With the vent open, the
booth fan pulled  air directly from the room and through the
booth (instead cf. taking suction  from the booth alone), lowering
the booth aiv flow rate  to 36 cm/s  (70 fpm) and varied from
25 cm/s to 46 cm/s  (50 fpr= to 90  fpm) across the booth face
during testing.   Targets were kept  from being blown back from
the spray equipment by a polymer pipe  frame mounted between the
targets and  the water wash.
 *Air  atomized conventional (AAC) spray equipment  is  charac-
  terized by the use of air as the atomizing agent for  the
  paint  spray.

-------
     A variable-speed electric conveyer system (Reliance Electric
Company) was used to carry the targets in front of the spray
equipment.  All AAC runs were made with the conveyor set at 10.6
cm/s.  Very little fluctuation was observed in conveyor speed
during AAC testing.

     Foil weights were determined on Precise laboratory scales
accurate to 0.01 g.  Weight-percent-solids samples and dishes
were weighed on 0.0001 g accuracy laboratory scales.

     A forced air gas-fired oven was uued for curing weight
percent solids samples and painted foil TE samples.  TE samples
were mounted on a large rack for curing, while weight Percent
solids sample dishes were placed on a makeshift shelf for curing.
Both were cured at 148.9*C (300*P) for 20 minutes.  After
the first few TE samples showed signs of contamination (dirt
flecks in finish), the oven was cleaned out daily by vibrating
the walls and then vacuuming.

     A Micro Motion mass flow meter was used for paint mass flow
determinations.  Section 3 discusses the use of a mass flow
meter in comparison to the digital scales/stopwatch mass flow
determination method specified in the test procedure (Appen-
dix A).

     Medium temper 4x10~5tn (1.5 mil) thick, 15.24 cm (6 in)
wide aluminum alloy foil was used to cover VC and FP targets.
(Refer to Appendix A, Test Method.)  Medium temper 4 x 10  m
(1.5 mil) aluninum alloy foil 38.1 cm (15 in) wide was used to
cover Graco targets during testing.

     The test method in Appendix A was strictly adhered to for
AAC testing, as were the QA/QC requirements of Appendix B.
After each EPA test run was completed, a separate run was made
using Graco targets.  A summary of all AAC test equipment
specifications is presented in Appendix C.

OPERATING AND MAINTENANCE VARIABLES

     During an earlier phase of this project, industry represen-
tatives, consumers, and manufacturers identified 17 operating and
maintenance variables considered important in achieving optimum
TF for AAC equipment.  These variables are listed in Table 2.

     Five variables were selected for testing on the basis of
the number of tines it was identified by different sources, the
anticipated size of effect on TE, the ability to simulate it
within the prescribed test methodology, and finally, the limi-
tation of laboratory time.  The five selected teat variables
were:

     o  Restricted air lines
     o  Booth air rate
     o  Gun cleanliness
     o  Restricted paint lines
     o  Fan (or horn) air

-------
      TABLE 2.  OPERATING AND MAINTENANCE VARIABLES FOR
                AAC SPRAY EQUIPMENT*
                   Atomizing air
                   Booth air rate
                   Booth configuration
                   Cure schedule (time, temperature)
                   Paint discharge technique
                   Equipment design
                   Plash off
                   Gun cleanliness
                   Gun condition
                   Gun-to-target distance
                   Operator error
                   Paint mass flow rate
                   Paint characteristics
                   Restricted air supply
                   Restricted paint supply
                   Shaping air (fan air)
                   Target configuration

* As mentioned by industry sources contacted

-------
     Some of the variables could be quantitatively simulated
(for example by varying paint pressure), while others could only
be simulated qualitatively.

EXPERIMENTAI DESIGN

     An experimental design was developed to accommodate the
limitations of testing while addressing the effects of each
variable as completely as possible.

     The first restraint on experimental design was the avail-
ability of laboratory time: only about 30 test runs could
reasonably be completed during a week of testing.  The second
limitation was the number and type of simulation levels for each
variable.  Only two levels of linear air velocity (booth air
rate) were possible in the test laboratory, while three levels
of fan air (sometimes called horn air or shaping air) were
achievable, and five or more levels of the other variables could
be simulated.  Table 3 presents the type of variable (quantita-
tive/qualitative) and levels to be accommodated in the experi-
mental design.


    TABLE 3.  EXPERIMENTAI, VARIABLES SELECTED FOR TESTING
              AAC SPRAY SYSTEMS
  Factor                              Quant./    No. of
    ID    Factor description          qual.    test levels

    A     Restricted atomizing
            air lines                 Quant.        5

    B     Booth air rate (linear
            velocity)                 Quant.        2

    C     Gun cleanliness             Qual.         5

    D     Restricted paint lines      Quant.        5

    E     Fan air (sometimes
            called horn air or
            shaping air)              Qual.         3

     A central composite experimental design was selected as the
most thorough way to examine the effects of these factors with
the fewest number of test runs.  The experimental design is
characterized by combining a fractional factorial design portion
with a "star" portion, a uc; men ted by replicates.

     Table 4 presents the AAC experimental design.  In this
table, the abbreviations "a," "1," "0," "-1," and "-a" denote
                                10

-------
           TABLE 4.  AAC EXPERIMENTAL DESIGN
            Run Number

                1
                2
                3
                4
                5
                6
                7
                8
                9
               10
               1t
               12
               13
               14
               15
               16
               17
               )8
               19
               20
               21
               22
               23
               24
               25
               26
               27
               28
               29
               30
               -a
                a
                0
                0
                0
                0
                0
                0
                a
                a
                a
                a
                a
                a
                        FACTOR
                      B     C
                           -1
 0
 0
-a
 a
 0
 0
 0
 0
 a
 a
 a
 a
 a
 a
 0
 0
 0
 0
-a
 a
 0
 0
 a
 a
 a
 a
 a
 a
_E

 1
•1
-1
•1
-1
-t
-1
 1
-1
 1
 1
-1
 1
 1
 1
 1
 0
 0
 0
 0
 0
 0
Where:

     A
     B
     C
     D
     E
Restricted air  lines—test at  5  levels: a,1,0,-1,-a
Booth air rates—teat at  2 levels:   1,-1
Gun cleanliness—teat at  5 levels:   a,,1fO,-t,-a
Restricted paint lines—test at  5  levels:  a,1,0,-l,-a
Pan air—test at 3  levels:   1,0,-1
                                 11

-------
the level of each variable to be tested.
base level with a good spray pattern.  Level "-a" denotes the
poorest level of a variable to be tested.  The intermediate
levels "1," "0," and "-I" were determined along equal spacing
from "a" to r-a" for the particular variable.   Level "a"1 will
be different for each experimental design and for different
variables in the design.  It remains constant for a given vari-
able in a given design.

     Levels for AAC variables were determined in pretest
trials as described in the following subsection.

     The first 16 test runs in the experimental design (Table 4)
are the fractional factorial portion of the design.  When the
results of several factors are to be studied, a factorial design
it usually the most efficient method to use.*  The basic idea
of factorial design is to alter several aspects of a test at a
time, but in such a way that the effects of individual altera-
tions can be determined.  Fractional factorial designs sacrifice
some ability to test for interaction between factors but are
able to test for main effects very efficiently.

     Runs 17 through 24 in Table 4, are the "star" portion of
the experimental design.  This portion of the experiment tests
tbe effects of variables at the extremes of their range  (for the
system under test, at "a" and "-a").  The star design broadens
the range of information gathered in the test.  The star portion
of the design allows extra degrees of freedom in order to assess
lack of fit.

     The experimental design used in the AAC case involves a
central composite design for factors A, C, and D.  A classical
central composite design on all five variables was impossible
because of the necessity oi: using only 2 levels of variable B
and 3 levels of variable G.

     The last six runs of the test design are replicates.
Replicates ace provided at the base condition of all variables
to provide a measure of i:he test precision.
*Youde~n"r W. J. and Steiner, E. H. , Statistical Manual of  the
 Association of Official Analytical Chemists, Arlington,  VA,
 1932; and Davies, 0. L., Design  and Analysis of  Industrial
 Txperiments, Great Britain"^19T§! and Myers, Raymond H. ,
 Response Surface Methodology, 1976.
                                12

-------
AAC TEST PERFORMANCE

     AAC testing began on  February  8,  1984.   Equipment set up,
target assembly and hanging,  foil cutting  and preweighing, and
other preparatory activities  were completed  <-arlier in the week.
The morning of February  8  was spent mixing and adjusting test
paint to desired specifications.  (See Appendix C for paint
specifications.)  About  16 L  (4  gal) of paint was mixed in the
20 L (5 gal)  capacity paint pressure pot.  The paint was stored
inside a temperature-controlled  booth  at 25*C + 1"C.  Once
the paint was adjusted,  all equipment  and  line? were rechecked
for proper  installation  and freedom from obstruction.  The mass
flow meter was calibrated  and zeroed.   Haas  flow calibration was
double checked against unatomized paint capture a-.d found to be
within 0.4 percent of the  meter  reading, as  required.

     Preliminary paint weight-percent-solids determinations were
made using  the recommended ASTM  method, and  using Grace's own
technique.   (The ASTM method  is  included in  Appendix A.)  Basi-
cally, the  ASTM method required  sampling the paint and adding
0.5 g (+ 0.1  g) of the paint  sample to a dry preweighed 58 mm
aluminum sample dish.  Solvent (3 mL)  was  added to the sample
prior to curing to spread  the paint sample evenly in the dish.
The Graco method required  a 15 mL paint sample be taken and
spread out  by gravity into a  30.48  cm  (12  inch) preweighed
aluminum dish.  The results of the  two methods did not closely
agree, and  some unacceptable  variance  in the ASTM method was
also noted.   According to  the QA/QC plan,  TE data cannot be
accepted unless all component measurements meet precision
requirements.

     A number of weight-percent-solids determinations was made
to resolve  the differences.  In running these samples it was
discovered  that the ASTM-method aluminum dishes were coated with
an oily compound to keep them from  sticking  together in storage.
This coating  had to be burned off before the dish could be  ised
in weight-percent-solids determinations.  The weight of the oil
on the preweighed dish varied, causing the net weight percent
solids to vary as well.  It was also discovered that uneven
distribution  of the paint  (in either method) caused differences
in curing and consequently in weight percent solids.  The latter
problem manifested  itself  most frequently  in the Graco method,
and almost  none at all  in  the ASTM  method  when proper care was
taken to assure the dishes were level  during curing.  As detailed
in the TE test procedure,  the ASTM  method  was used for all EPA
weight-percent-solids determinations in this report.  Graco
amended their weight-percent-solids determination to follow ASTM
recommendations, but  continued to  take their weight-percent-
solids samples  from the  paint line  rather  than from the paint
pot as th^  TE test procedure  requires. Graco weight-perc«nt-
oolids samples continued to vary somewhat  from EPA values,
apparently  due  to  the  sampling technique or  position.   (The term
"EPA values"  as used  here  refers to the determinations nade
following  the Draft  Standard  Test  Method of  Appendix A.)

-------
     While one group of technicians was performing weight-per-
cent-solids determinations, a second group of technicians was
setting up the equipment at base levels of each variable.  Base
level ("a") was determined by Graco experience with the test
paint and spray painting system.  Selection of base level was
confirmed by checking the spray pattern at base level for a good
pattern.  No adjustments were required from Graco-recommended
base levels after checking the spray pattern.  Base levels were
thus determined as shown in Table 5.

     Deteriorated levels were selected by setting all variables,
except the subject variable (for each variable in turn), at
the base level.  The subject variable was altered until a
significantly worse spray pattern could be discerned.  The spray
pattern was checked by spraying onto a paper target for 5 or €
seconds then observing the resulting pattern.  Deteriorated
factor levels ("-a") thus determined are shown in Table 5.

     Intermediate factor levels were calculated to be evenly
spaced from the base level ("a") to the deteriorated level
("-a").  Intermediate levels are also shown in Table 5.

     Selection of levels for gun cleanliness were made by trial
and error pattern testing of air caps with different holes
plugged.  The resulting pattern of plugged holes for deterio-
rating levels of gun cleanliness is shown in Figure 1.  Gun
cleanliness must be considered qualitative because of the nature
of the progressively more plugged air cap.  Atomization and TE
may be affected as much or more by the geometry and design of
the plugged holes than by the total plugged area.

     Booth air rates were determined by the only two available
levels.  Neither level should be assumed to be an ideal level.
Base level was selected at the normal air flow level for the
booth, rather than the artificially lowered level.  Rates at
each level were measured using a hot wire anemometer.

     Level selection was completed on February 8, 1984.  AAC TE
testing began on February 9, 1984.  TE runs were made in a
randomized order based on the experimental design in Table 4.
During a single 16-hour experimental day 15 runs were made.
Three of the runs were thrown out because of incomplete data,
underspray, or losing paint from the targets because of dripping
or accidental contact with wet targets.  These three runs were
repeated at the end of the day.

     TE testing of AAC equipment was completed on February 10,
1984, after pertorming the remaining 15 runs.  One run  (Run 10)
was identified as an outlier by the QA/QC analysis (Refer to
Appendix B); it was repeated immediately.  Weight-percent-solids
samples were taken at the completion of testing as required by
the draft Standard Test Method.
                                 14

-------
             TABLE  5.   LEVELS  OF OPERATING AND  MAINTENANCE
                         VARIABLES—TESTED  ON AAC SPRAY  PAINTING
                         EQUIPMENT


                              Quant/    No.  of
	Factor	qual.     levels	Test levels	

A.   Restricted automizing
       air lines*             Quant.      5       a= 239.0 kPa (20 psig)
                                                 1= 218.6 kPa (17 psig)
                                                 0= 197.9 kPa (14 psig)
                                                -1= 177.2 kPa (11 psig)
                                                -a= 156.6 kPa (8 psig)


B.   Booth air rate +
       (linear velocity)      Quant.      2       1= 0.61 m/s (120 ft/min)
                                                -1= 0.36 m/s (70 ft/min)


C.   Gun cleanliness t        Qual.       5       See  Figure 1


D.   Restricted paint lines   Quant.      5       a= 180.7 kPa (11.5 psig) z
                                                 1- 170.3 kPa (10.0 psig)
                                                 0= 160.0 kPa (8.5 psig)
                                                -1- 149.7 kPa (7.0 psig)
                                                -a= 139.3 kPa (5.5 psig)


E.   Fan air I                Qual.       3       1s wide open
                                                 0= one turn shut
                                                -1= two turns shut
* Measured at the spray gun.
+ Actual booth air rates varied from 100 to 140 ft/min for level "+1"
  and 50 to 90 ft/min for level "-1."  Average air velocities are used in
  this table.
t Deteriorating gun cleaniness was simulated by blocking air cap holes
  as shown in Figure 1.
z Measured at control panel approximately 20 feet from spray gun.
f Fan air (sometimes called horn air or shaping air) was adjusted by
  setting the control knob on the gun wide open, then turning it the
  required number of turns towards the closed position.
                                     15

-------
          Level +ai all holes open
Level +1: 3 holes plugged
Level Oi 4 hcles plugged      Level -1: 6 holes plugged      Level -a: 8 holes plugged
  Figure 1.  Air  atomized conventional air  cap (frontal  view) showing
              selection of test  levels for gun cleanliness

-------
TEST RESULTS

     TE's  were calculated according to the Draft  Test  Standard
Method.  The  final AAC results are presented in Table  6.    Some
corrections were made to the original TE test data  because  of
mathematical  errors or incorrectly recorded weights.   These
corrections are reflected in Table 6.

STATISTICAL ANALYSIS

     Regression equations were developed to fit the TE data for
each target design.  The regressions developed for  AAC equipment
are based  on  the data in Table 6,  which were developed according
to the experimental design in Table 4.   Both qualitative and
quantitative  variables were coded  into the regression  analysis
according  to  the level rather than their numerical  value during
the test.  The coding procedure is a simple "centering" and
scaling of variable levels.   The variable levels  for the quanti-
tative variables were evenly spaced.   The following represent
the coded  or  "design units."

       Quantitative                  Coded level
      variable level            for regression equation
a
+1
0
-1
-a
2
1
0
-1
-2
     By using  these  coded values  for  levels, the regressions
become useful  for  either SI  (Le Systeme  International d'Unites)
or standard U.S. industrial  units.  Actual test values at different
levels than those  tested here may be  coded by interpolating
linearly according to  the above table.   Thus, a value exactly
halfway between  "-1" and "-a" would take on the numerical value
of -1.5 in the regression equation.

     Transformation  to design units is standard procedure when
one builds models  based on a planned  experiment involving
quantitative variables.  It  allows for interpretation of re-
gression equation  and  tests  to be in  terms of units that are
scale free and determined by the  region  of experimentation
selected by the  scientist.

     Qualitative variable levels  were coded in much the same
manner, assigning  either zero or  one  to  the level for each
variable.  The "0" level of  each  qualitative variable was
arbitrarily set  at zero for  all cases in this report.  For
                                 17

-------
TABLE 6.  AIR ATOMIZED CONVENTIONAL TEST RESULTS

                    Percent Transfer Efficiency
    Run number        Ft        VC      Graco

        1
        2
        3
        4
        5
        6
        7
        8
        9
        10
        11
        12
        13
        14
        15
        16
        17
        18
        19
        20
        21
        22
        23
        24
        25
        26
        27
        28
        29
        30
6^.5
85.9
83.8
•»7.1
73.8
75.8
77.8
57.4
82.0
58.2
60.0
85.3
61.0
68.7
70.7
65.0
78.1
63.5
66.7
64.8
59.5
73.5
80.5
65.7
58.3
59. 1
58.4
58.2
60.4
57.1
11.9
15.2
16.5
14. 0
13.7
13.0
12.3
11.0
15.7
10.7
9.6
16.4
10.7
13.6
13.5
12.6
15.1
11.7
12.2
11.5
11.2
13.7
15.7
11.7
10.3
11.2
9.5
11.2
11.3
10.5
27.8
39.9
41.7
36.3
33.2
34.6
38.5
25.8
38.2
25.0
26.1
38.4
29.0
31.4
32.3
29.5
38.9
30.1
30.3
29.5
28.7
33.2
35.9
28. i
25.1
26.2
26.2
27.9
28.3
25.7
                        18

-------
qualitative variables  tested at three levels, the value of the
variable associated with that particular level is denoted by
dummy variables xRd1 and *nd2» which take on the values:

     Factor     Experimental    Regression level

X
n
x
n
x_
level
+ 1

0

-1
«nd1
0

0

1
Znd2
1

0

0
     Qualitative effects can only be determined on a relative
basis.  For all of the regressions in this report, the "0"
experimental level has been designated the base level of
comparision for qualitative O&M factors.

     For qualitative variables at five levels, the assignment of
regression levels proceeds according to:

     Factor     Experimental      Regression level

X
n
x
n
x
n
x
n
x
level
a

+1

0

-1

-a
"nd1
0

0

0

0

1
*nd2
0

0

0

1

0
Xnd3 '
0

1

0

0

0

-------
Because x, and x- cannot be measured on a continuous quanti-
tative scale, they are termed qualitative variables.  The
various levels of such variables used in the experiment are
represented in the analysis of variance and the regression
analysis by "dummy variables.*  Thus, for gun cleanliness, x-,
and for fan air, x_, dummy variables were introduced.  For gan
cleanliness, dummy variables take on the following values:
     Level


      a

     4-1

      0

     -1

     -a
  Gun cleanliness
*3d1
 0

 0

 0

 0

 1
0

0

0

1

0
0

1

0

0

0
1

0

0

0

0
Gun condition


See Figure 1

See Figure 1

See Figure 1

See Figure 1

See Figure 1
Thus, for gun cleanliness at level *+1", all dummy variables
except x->d3 take on the value of zero; x,d, takes on the
value of one.
     In the case of fan air, dummy variables x,.d1 and x,.d2
take on the values shown in the table below.
                          Fan air
Level

 +1

  0

 -1
                        *5d1
         0

         0

         1
              *5d2
          1

          0
                            Gun setting
                 wide open

                 one turn shut

                 two turns shut
     The analysis of variance results  '.n  a comparison of  the
variance associated with each experimental variable  to  the
inherent error associated with repeat  observations.  This com-
parison is accomplished by forming the P  statistic,  the ratio for
each value of F has a probability associated with  It given  the
number of degrees of freedom in  the numerator  and  in the  denomi-
nator.  When the probability of  achieving a given  value of  F  by
chance is less than 0.05, the effect is said to  be significant
at the 0.05 level.  The F statistics and  associated  probability
for all factors found to be significant are presented in  Table  7.
(See Appendix G for a glossary of statistical  terms.)
                                 20

-------
       TABLE 7.  AAC F-STATISTICS (F) AND ASSOCIATED
                 PROBABILITY  (P)*
Effect

X1
  Plat  Panel
  F        P

140.89   .0001

 27.84   .033
Vertical  Cylinder
   F          P
 30.45

  4.49
.0027

.0376
X3d2
X4
X5d1
X5d2
X1X2
x2
X4
X2X5d1
x.x.d.
—
64.72
163.76
21.27
11.59
32. US
10.21
22.42
- —
.0005 8.66
.0000 16.88
.0058
.0192
.0024
.0241
.0520
—
.0321
.0093
-
-
-
-
-
                                                       Grace
  F      P

36.32  .0000

 8.57  .0022

 6.36  .0660

 5.33  .0120

53.17  .0000

28.14  .0001

 6.01  .0800


 7.66  .0034



 9.43  .0015
*F and P are dimensionless terms.  Refer to Appendix G for a
 definition of those terns.
                               21

-------
     In the case of all equipment types, a regression model was
postulated that applied for all target types.  The model considered
contained linear and sometimes quadratic effects for the con-
tinuous variables and dummy variables for the quatitative variables.
Certain interactions were put into the model on the basis of the
best engineering experience available.  All possible interactions
could noc be estimated due to limitation of time and resources.
In all cases, the experimental design was constructed to accommo-
date the model terms'.  Terms that were significant on the basis
of an F-test were retained and the final regression model is
reported for each target type.

     In the case of air atomized conventional, the following
model terms were considered:

     o  linear in x.

     o  quadratic in x^

     o  linear in x.

     o  dummy variables in x3

     o  linear in x.

     o  quadratic in x.

     o  dummy variables in xg

     o  interactions between x, and  x2

     o  interactions between x_ and  x5

     Regression models were constructed using mainframe SAS*
capabilities for e;-ch of the three target  types  including all  of
those affects found  to be significant at the 0.05  level.  The
resulting models and their associated R   (proportion of the
overall variance explained by the regression) ere  presented
below:

Flat Panel Target

     The regression model developed  for AAC  FP  is:

          TE - 68.95 - 3.12x1 -  0.98x2  +
 "•Statistical  Analysis System,  SAS Institute,  P.O.  Box  10066,
  Raleigh, NC  27605.
                                 22

-------
     The negative coefficient on x.  indicates a tendency for TB
to decrease with an  increase in restricted air levels.  However,
there is an important  interaction between the two factors as
evidenced by a positive and significant coefficient of x^j.
While this interaction does suggest  that TE continues to
decrease with an increase  in restricted air levels, the magni-
tude of that increase  depends on the booth air level.  The same
is true for x-, booth  air.  The mixed  coefficients on x4 and
X? indicate tnat TE  increases with an  increase in paint
line levels for the  low paint line levels, but the amount of
increase tapers off  as the paint line  levels become larger.
The heavy positive coefficient on x-d. suggests that TE
increases when "-1"  level  is used oft fan air.  The negative
coefficient on xsd,  suggests a decrease in TE  at the
•+1" level of fan air.

     Since both the  magnitude and direction of «-he effect of fan
air on TE are dramatically different at different- levels of fan
air, the operator must be  very careful to establish the appropri-
ate fan air level for  optimum TE.

     The proportion  of the overall variance explained by the
regression  (R  )  is 0.97.   This R  indicates a tight fit
of the regression model to experimental data.  The standard
deviation of replicate runs was  1.098, well within the targeted
2.0 standard deviation**  (expressed  in units of transfer effici-
ency ).

     The error  in the  regression model due  to lack of fit was
determined  to  be  insignificant  at the  0.05  level.  The P-
statistic for  lack of  fit  was  1.08  (probability-0.49).

Vertical Cylinder Target

     The regression  model  developed  for &AC VC is:

         TE-  11.99  - 0.91X,  -  0.28x2

            +  C.44x4  + 2.83X36,

     The P-statistics are shown in  Table  7.

     The negative coefficient  on x.  indicates  that TS decreases
with an  increase in  air  pressure.   The negative coefficient
on x0  indicates that TE  decreases with an increase  in booth
air  rates.  The positive  coefficient on x.  indicates  that  TE
 increases with increasing paint pressure.   The positive coef-
ficient on  fan air,  only  at level -1,  indicates there is


**CENTEC  Corporation,  "Development  of  Draft Stan-Sard  Test  Method
   for  Spray Painting Transfer  Efficiency,"  for USCPA  under Con-
   tract  68-03-1721,  Task 2.
                                 23

-------
something different about fan air at this level than at other
test levels.  Since fan air is a qualitative factor, it can only
be speculated that certain levels of fan air affect TE more than
other levels.  Reducing fan air from the "0" level improved
transfer efficiency, but increasing fan air did not signifi-
cantly degrade TE.

     The proportion of the overall variance explained by the
regression  (Rz) is 0.79.  This R  is lower than for the FP
target, probably due to the lack of overall variation in TE
level.  In the FP case, the total standard deviation for all
experimental data was 9.61, while it is only 2.030 in this case.
The FP target TE's were more strongly affected by experimental
variations and were thus easier to model.  The effect of experi-
mental variations on VC TE's is so small that it is almost below
the targeted 2.0 standard deviation for replicates.  Small
effects are difficult to tightly fit with renression models.

     The standard deviation of VC replicates was 0.706.  This
low standard deviation can also be attributed to the overall
insensitivity of the VC targets to experimental variables,

     The error in the regression model due to lack of fit is
insignificant:  F - 1.S3 (probability - 0.33).
TE - 32.22 - 1.61X, -1.24x2
Graco Target

            .22 - 1.6

      + 1.23x4 - 0.64xJ

      •f 0.67*^2 •*• 1.73x2x,d2

     The negative coefficient on x. indicates that TE decreases
with an increase in air pressure (restricted lines}.  The
negative coefficient on x, indicates a similar decrease in
TE with increasing booth air rates.  Cnce again, the positive
coefficient on the interaction between x. and x, indicates
that these trends are not constant but rather depend on the
level of the interacting variable.  For example, the negative
trend of TE with respect to air pressure Is not as pronounced
when booth air rate is high, according to the regression equation.
The positive coefficient on x3 (only at level d.) indicates
that TE is affected for one level of gun cleanliness. The effect
of gun cleanliness on TE at other experimental levels is insig-
nificant.

     Again, the positive coefficient on s. and negative sign on
the quadratic term suggests & nonlinear effect of paint pressure.
There is a positive slope on paint pressure until one goes beyond
the x »1.0 level, roughly.  At that point, the effect becomes
negative.
                                24

-------
     The strongly positive coefficient on x& at the d, level
contrasts dramatically with the strongly negative coefficient at
the d~ level.  This indicates that fan air can substantially
increase or decrease TB.  Again, very careful selection of fan
air levels is warranted by these results.

     Finally, two interactions are noted:  restricted air lines
and booth air rate interact to affect TB, and booth air rate
interacts with fan air  (only at the d2 level) to affect TE.

     The proportion of  the overall variance explained by the
regression (R >  is 0.97.  The standard deviation of repli-
cates is 1.261,  well below the target value.  The error due to
lack of fit is insignificant: P - 0.92 (P » 0.773).

    Tables 8, 9, and 10 present a comparison of predicted
values, based on the derived regression, with observed values
for the flat panel, vertical cylinder, and Graco targets res-
pectively.  The  residual  is the difference between predicted and
observed values.  The 95  percent confidence limits for the wean
give the upper and lower  bounds of the range within which the
mean of transfer efficiency (the "true regression") at
each experimental condition lies with 95 percent confidence.

AAC TEST CONCLUSIONS

     The regressions previously presented illustrate the
differences that target configuration can make in TB.  Even
with these differences, however, there are basic consistencies
between the results.  Three factors  (restricted atomizing air
lines, restricted paint lines, and fan air) were identified as
significant for  all tested target configurations.  A fourth
variable, booth  air, was  significant  for FP and Graco targets
and very nearly  significant for VC targets as well.  The con-
sistency of these results strongly implies that these four
factors have a critical influence on  TE regardless of target
configuration.

     Thus, selection and  maintenance  of appropriate atomizing
air pressure and paint  pressure should be given regular atten-
tion by the operator. Pan air rates  have o strongest influence
on TE across all target types, as demonstrated by highly
significant P-value and the large coefficient in each of the
AAC regressions.  Pan air levels are often set by individual
operators according to  their own judgment.  Por optimum TE,
plant management should determine optimum spray painting
conditions through  test runs and then specify those conditions
for the operator.

     Some reev.iluation  of booth air  rates may be justified  by
the test  reouits,  whicli indicate that the  lowest level of booth
air rate  should  be  selected  to maximize TB.  Care should be
taken  to  adhere  to  all  safety and environmental regulations,
                                 25

-------
          TABLE 8.  AAC-FF COMPARISON Of PREDICTED  VERSUS  ACTUAL TRANSFER EFFICIENCIES
to
en
Observation
number
1
2
3
4
5
6
7
6
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Observed
value
62.50
85.90
83.60
77. »0
73.80
75.80
77.30
57.40
82.00
58.20
60.00
85.30
61.00
68.70
70.70
65.00
78.10
63.50
66.70
64.80
59.50
73.50
80.50
65.70
58.30
59.10
58.40
58.20
60.40
57.10
Predicted
value
64.69
87.87
82.13
78.81
74.87
73.57
79.32
56.47
80.61
57.43
63.17
84.55
61.71
67.46
70.44
62. n
77.15
62.62
67.92
67.92
59.49
70.98
80.23
63.13
58.98
58.18
58.98
58.98
58.98
58.98
Residual
-2.19
-1.97
1.66
-1.71
-1.07
2.22
-1.52
0.92
1.38
0.76
-3.17
0.74
•-0.71
1.23
0.25
2.77
0.94
0.87
-1.22
-3.12
0.00
2.51
0.26
2.56
-0.68
0.11
-0.58
-0.78
1.41
-1.88
Lower 95% CL
for mean
62.47
85.48
79.73
76.54
72.47
71.20
77.16
54.51
78.22
55.20
60.93
82.13
59.35
65.09
68.19
60.75
74.25
59.72
65.64
65.64
56.30
68.22
78.3?
61.44
57.42
57.42
57.42
57.42
57.42
57.42
Upper 95% CL
for mean
66.92
90.27
84.52
81.07
77.26
75.94
81.47
58.44
83.01
59.65
65.42
86.98
64.08
69.82
72.68
63.69
80.05
65.52
70.20
70.20
62.67
73.73
82.12
64.82
60.55
60.55
60.55
60.55
60.55
60.55

-------
TABLE 9.  AAC-VC COMPARISON OF PREDICTED VERSOS ACTUAL TRANSFER EFFICIENCIES
Observation
number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Observed
value
11.90
15.20
16.50
14.60
13.70
13.00
12.30
11.00
15.70
10.70
9.60
16.40
10.70
13.60
13.50
12.60
15.10
11.70
12.20
11.50
11.20
13.70
15.70
11.70
10.30
11.20
9.50
11.20
11.30
••0.50
Predicted
value
12.73
16.44
15.56
15,01
13.74
13.19
14.07
10.35
14.62
10.91
11.79
15.89
12.17
13.05
13.61
11.23
14.08
10.44
11.70
11.70
11.38
13.14
14.54
11.70
10.76
10.76
10.76
10.76
10.76
10.76
Residual
-0.63
-1.24
0.93
-0.41
-0.04
-0.19
-1.77
0.64
1.07
-0.21
-2.19
0.50
-1.47
0.54
-0.11
1.36
1.01
1.25
0,49
-0.20
-0.18
0.55
1.15
-O.OO
-0.46
0.43
-1.26
0.53
0.53
-0.26
Lower 95% CL
for mean
11.93
15.44
14.72
14.09
12.74
12.22
13.25
9.50
13.65
10.03
11.06
14.92
11.28
12.20
12.73
10.70
13.08
9.50
11.06
11.08
10.38
12.20
13.77
11.08
10.05
10.05
10.05
10.05
10.05
10.05
Upper 95% CL
for mean
13.53
17.45
16.40
15.92
14.75
14.16
14.88
11.2i
15.59
11.79
12.52
16.86
13.07
13.91
14.49
11.77
15.07
11.38
12.33
12.33
12.37
14.08
15.32
12.33
11.48
11.48
11.48
11.48
11.48
11.48

-------
          TABLE  10.   AAC-GRACO COMPARISON OF PREDICTED VEKSUS ACTUAL TRANSFER EFFICIENCIES
to
oo
Observation
number
1
2
2
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
2*>
30
Observed
value
27.40
39.90
41.70
36.80
33.20
34.60
38.50
25.60
33.20
25.00
26.10
38.40
29.00
31.40
32.40
29.50
38.90
30.10
30.80
29.50
28.70
33.20
35.90
28.10
25.10
26.20
26.20
27.90
28.30
25.70
Predicted
value
23.03
41.58
40.55
35.32
34.57
34.86
35.89
25.81
38.45
24.90
25.94
39.20
29.11
33.14
31.91
29.68
38.0?
28.91
30.98
30.98
28.44
33.35
36.24
28.61
26.63
26.63
26.63
26.63
26.63
26.63
Residual
-0.63
-1.68
1.14
0.97
-1.37
-0.26
2. 60
-0.01
-0.25
0.09
0.15
-0.80
-0.11
1.25
0.38
-0.18
0.88
1.18
-0.68
-1.48
0.25
-0.15
-0.34
-0.51
-1.53
-0.43
-0.4J
l.>6
1.66
-0.93
Lower 95% CL
for mean
26.39
40.05
39.04
33.78
33.05
33.24
34.60
24.45
36.89
23.29
24.32
37.57
27.43
28.53
30.27
28.30
36.13
27.03
29.56
29.56
26.43
31.63
35.06
27.32
25.64
25.64
25.64
25.64
25.64
25.64
Upper 95% CL
for mean
29.67
«.11
42.06
36.86
36.10
36.47
37.18
27.17
40.02
26.52
27.55
40.82
30.79
31. 7S
33.55
31.07
39.88
30.78
32.41
32.41
30.45
35.08
37.43
29.91
27.62
27.62
27.62
27.62
27.62
27.62

-------
as well as providing for worker comfort when considering lower-
ing booth air rates.  The regressions  in the previous section
cen be used to make reasonable estimates of potential savings.
These savings should be weighed against all costs before a
change is made.

     The Graco and FP target configurations also identified
interactions between booth air rate and other variables.  These
interactions, while statistically significant, are not con-
sidered large enough to warrant direct practical attention.
It is recommended that the plant management emphasise selection
and maintenance of optimum levels for more critical variables.

     The Graco target configuration identified one level of gun
cleanliness as significantly affecting TE.  Since this finding
is not consistent across target J.ypes and ic relatively small
when it does appear, it is not considered critical to optimizing
TE.  This is not to say that gun cleanliness is unimportant to
the spray painter.  Gun cleanliness is one of the few O&M
factors universally stressed by gun manufacturers, spray paint-
ers, and other early participants in the test program.  Gun
cleanliness has a profound effect on paint finish, gun life, and
internal gun condition, which were not tested in this program.
Only the aspect of gun cleanliness tested during this experiment
is considered unimportant for AAC spray equipment.
                                 29

-------
                            SECTION 5

           AIR ATCHIZED ELECTROSTATIC SPRAY EQUIPMENT


EQUIPMENT DESCRIPTION

     A Graco Model AS-4000 manual electrostatic air spray gun
was was selected for AAE testing.  The Model AS-4000 gun was
considered typical of an external-mix, manual electrostatic
spray gun.  The sprav gun was equipped with a 177033 air cap,
776976 fluid tip, and a 215864 needle.  Paint flow was manually
initiated by opening a valve on the paint supply line.  The
spray gun was fixed in open position.

     Graco standard black enamel was selected as the test paint.
The paint averaged about 28.7 weight percent solids when cut to
30.4 seconds 
-------
OPERATING AMD MAINTENANCE VARIABLES

     As listed  in Table  11,  19 variables were identified
through interviews and literature  that may have potential
to exert an  important effect  in achieving optimum TE.  Seven
of the identified variables were selected for AAE testing
on the basis of:  (1) the number of  tiaies the variable was
identified for  AAE by different sources, (2) the ability to
simulate the variable within  the prescribed test methodology,
and  (3) the  limitation of laboratory time.  ?his prior
knowledge enabled us to  limit the  scope of TE experiments to
only variables  of particular  interest.

The  selected test variables were:

     o  Restricted atomizing  air lines

     o  Booth air rate  (linear velocity)

     o  Gun  cleanliness

     o  Restricted paint lines

     o  Fan  air (sometimes  called  horn  air or shaping air)

     o  Tip  voltage

     o  Electrode position

     Restricted atomizing air lines  can be simulated by de-
creasing  the pressure of the air supply to the  spray gun.  An
air  regulator was  used  for  reducing  the air  pressure to desired
levels.   Restricted  paint lines were simulated  by decreasing the
paint supply pressure  in a  similar manner.

     Booth  air  rate  (linear velocity) was  available at only  two
levels  at this  facility, 0.36 m/s  and 0.61 Vs  (70  ft/min  and
120  ft/min)  respectively.  Gun cleanliness was  simulated by
blocking  certain air holes  in the  air cap  in a  progressively
worse pattern  as shown  in Figure  2.   Fan  air was adjusted  by
using the adjustment knob on the  spray gun,  while voltage  sup-
plied to  the tip was adjusted at  the power  supply.   Electrode
position  was set manually as shown  in Figure 3.


EXPERIMENTAL DESIGN

      An experimental design was deve"  ed to accommodate  the
 limitations of testing  while addressing the  effects of  each
variable  as completely as possible.

      The first restraint on  experimental  design as  noted
 previously was the availability of  laboratory time: only
                                  31

-------
   TABLE  11.  OPERATING AND MAINTENANCE VARIABLES
              FOR AAE SPRAY EQUIPMENT*
                 Atomizing air
                 Booth air rate
                 Booth configuration
                 Conveyor speed
                 Cure schedule  (time, temperature)
                 Electrode position
                 Equipment design
                 Plash off
                 Gun cleanliness
                 Gun condition
                 Gun-to-target  distance
                 Operator  error
                 Paint discharge  technique
                 Paint mass  flow  rate
                 Paint characteristics
                 Restricted  air supply
                 Restricted  paint supply
                 Shaping air (fan air)
                 Target  configuration

*as mentioned by industry sources  contacted
                               32

-------
         Level +ai all holes open
Level +li 3 holes plugged
     Oi 4 holes plugged     Level -It 6 holes plugged      Level -ai 8 holes plugged
Figure  2.   Air atomized electrostatic air cap  (frontal view)  showing
            selection  of test levels  for gun cleanliness

-------
         FRONT VTEW
SIDE VIEW
                                V
                                  ^WH


                                Jb
                    LEVEL 1
                                7
                                Jb
                    LEVEL 0

                    LEVEL -1
Figure 3.  Air atomized electrostatic electrode

           position test levels
                      34

-------
about 30 test runs could  reasonably be completed during a
week of testing.  The  second  limitation was the number and
type of simulation levels for each variable.

     Table 12 presents the type of variable (quantitative/qual-
itative) and levels to be accommodated in the AAE experimental
design.

     A variation of a  central composite experimental design was
selected as the most thorough way to  examine the effects of these
factors with the fewest number of test runs and still allow for a
regression model to be constructed.   The experimental design is
characterized by combining a  fractional factorial design portion
with a "star" portion, augmented by replicates.  A slight variation
central composite experimental design was constructed for factors
A, C, D, E, F, and G.   Five levels were required for factors A,
C, D, and P but only three levels for factors E and G.  Thus the
design levels for the  star points could not be the same for all
variables.  In addition,  as in the AAC, the replicates were not
at the traditional center of  the design.  For pragmatic reasons
the replicates were taken at  the extremes in each variable.

     Table 13 presents the AAE experimental design.  In this
table, the abbreviations  "a," "1,"  "0," "-I," and "-a" denote  the
level of each factor  to be tested.  Level "a" denotes the base
level with a good spray pattern.  Level "a" is likely to be dif-
ferent for each  factor in each experiment.  It remains constant
for a given factor  in  a given experiment.   Level "-a" denotes
the poorest level of  a factor to be tested.  The intermediate
levels "1," "0,"  and  "-1" are determined  along equal spacing
from "a" to "-a"  for  the  particular factor.  Factor  levels for
AAE testing were  determined in pretest  trials as described in
the following subsection.

     The first  16 test runs in the  experimental design are the
fractional factorial  portion  of  the design.  When the effects  of
several factors  are  to be studied,  a  factorial design is usually
the most efficient  method to  use.*  The basic  idea of factorial
design  is  to  alter  several aspects  of a test at a time, but  in
such a way that  the  effects of individual  alterations can be
determined.   Fractional factorial  designs sacrifice  some ability
to  test  for  interaction between  factors but are able to test  for
main effects  very efficiently.
 *Youden. W. J. and Steiner, E. H., Statistical Manual of the
  Association of Official Analytical Chemists, Arlington, Va.,
  1982;  and Davies, O. L., Design and Analysis of Industrial
  Experiments, Great Britain,  1979.
                                 35

-------
TABLE 12.  OPERATING AND MAINTENANCE VARIABLES
           FOR AAE SPRAY PAINTING EQUIPMENT
          Variable
Quant/    No. of
qual.	levels
A.   Restricted atomizing
       air lines              Quant.
B.   Booth air rate
       (linear velocity)      Quant.      2
C.   Gun cleanliness          Qual.       5
D.   Restricted paint lines   Quant.      5
E.   Pan air (shaping air or
        horn *ix)             Qual.
P.   Voltage
G.   Electrode position
Quant.
Quel.
3


5


3
                          36

-------
          TABLE  13.   AAE EXPERIMENTAL DESIGN
Run number    A     B

    1
    2
    3
    4
    5
    6
    7
    8        -1
    9        -1
   10         1
   11        -1
   12         1
   i3         1
   14        -1
   15         1
   16        -1
   17        -a
   18         a
   19         0
   20         0
   21         0
   22         0
   ?.3         0
   24         0
   25         0
   26         0
   27         0
   28         n
   29
   30
   31
   32               1
   33               1
   34               1
                      Variable
                    C      D
                   -1
                    0
                    0
                    a
                   -a
                    0
                    0
                    0
                    0
                    0
                    0
                    0
                    0
                    a
                    a
                    a
                    a
                    a
                    a
 0
 0
 0
 0
-a
 a
 0
 0
 0
 0
 0
 0
 G
 0
 0
 0
 0
•1
 1
 0
 0
 0
 0
 1
 1
 1
 1
 1
 1
                                    -1
                                    -1
 0
 0
 0
 0
 0
 0
 0
 0
-a
 a
 C
 0
 0
 0
 0
 0
 0
 0
 0
 0
 0
 0
•1
 1
Where:
  A
  B
  C
  D
  E
  P
  G
Restricted air lines—test at 5 levels: a,1,0,-1,-a
Booth ait rates—test at 2 levels:   1,-1
Gun cleanliness—test at 5 levels:   a,1,0,-1,-a
Restricted paint lines—test at S levels:  a,1,0,-1,-a
Pan air—test at 3 levels:   1,0,-1
Voltage—teat at 5 levels:   a,1,0,-1,-a
Electrode position—test at  3 levels:  1,0,-1
                            37

-------
     Runs 17 through 28 in Table 13, are the "star" portion of
the experimental design.  This portion of the experiment tests
the effects of variables at the extremes of their range (for the
system under test, at "-a" and "a").  The star design broadens
the range of information gathered in the test.  The star portion
of the design allows extra degrees of freedom in order to assess
lack of fit.

     The final six runs o£ the experimental design are repli-
cates.  Replicates are provided at the base condition of all
variables to provide a measure of the precision of the test.


AAE TEST PERFORMANCE

     AAE testing was conducted from February  13 to February 17,
1984.  Spray equipment was set up on February 13 and initial
spray pattern was checked.  Some difficulty was encountered in
establishing a good spray pattern for base levels.  The fluid
tip and valve seats were replaced in the spray gun, and the
spray pattern improved.  Base levels ("a") for each variable
were established as described in the Test Method  (Appendix A).

     Deteriorated levels ("-a") were determined by setting all
factors except one at the base level, then decreasing the level
of the selected variable until a noticeably worse spray pattern
resulted.  Deteriorated levels of each variable were determined
in turn.  Intermediate  levels were calculated evenly between
•a" and "-a" for each quantative variable.  The final selection
of test levels is presented in Table 14.

     Deteriorated gun cleanliness levels were determined by
progressively plugging  more holes in the air  cap.  Final gun
cleanliness levels are  shown  in Figure 2.

     Electrode position was selected through  trial and error
spray pattern checks after alterations in electrode position
were made.  Selected electrode positions for  AAE  TE testing are
illustrated in Figure 3.

     The experimental design  in Table 13 was  followed.  Three
blocks of runs were made; all of the runs in  a block were of
the same electrode position.  Total randomization could not
be accommodated without introducing an unacceptable error in
trying to duplicate the desired electrode position.  Pretest
trials demonstrated the inability to assure consistent levels
ot electrode position in a totally random experiment.  (Spray
gun design caused straightening of  the electrode  whenever air
cap changes were made.)

     TE testing began February 14 after all documentation and
QA/QC measures were completed. Six  tests runs were completed.
On the second day of testing  17 runs were completed, with
                                 38

-------
     TABLE 14.   LEVELS OP OPERATING  AND MAINTENANCE VARIABLES
                  TESTED ON AAE  SPRAY  PAINTING EQUIPMENT


_ Variable     ____  Test  levels -

A.  Restricted atomizing air lines*           a= 293kPa  (20 psig)
                                              1« 218.6kPa (17 psig)
                                              0= 197.9kPa (14 psig)
                                             -1- 177.2kPa (11 psig)
                                             -a= 156.6kPa (8 psig)

B.  Booth  air rate +                                            ,  J  %
       (linear velocity)                       1- 0.61m/s (120 ft/min)
                                             -1- 0.36m/s (70 ft/min)

C.  Gun cleanliness t                         See Figure 2

D.  Restricted paint lines  z                 *• ISO.TkPa (15.5 psig)
                                              1- 170.3kPa (13.5 psig)
                                              0= ISO.OXPa (11.5 psig)
                                             -1- 149.7kPa (9.5  psig)
                                             -a- 139.3kPa (7.5  psig)
 E.   Pan air t                                 1- wide open
                                              0= 1 turn shut
                                             -1-» 2 turns shut

 P.   Voltage **                               »• '2 kV
                                              1- 63 W
                                              0- 54 KV
                                             -1- 45 kV
                                                 36 kV
 G.  Electrode position                        1" normal
      (See Figure 3)                           0- bent tt
                                              -1- clipped off
    "Measured at the spray gun.
    4Actual booth air rates varied from 100 to 140 fpm for level
     •+1" and 50 to 90 fpm for level "-1".   Average air velocities
     are used in this table.
    tDeteriorating gun cleaniness was simulated by blocking air cap holes
     as shown in Figure 2.
    zMeasured at control panel approximately 20 feet  from spray gun.
    tFan air (sometijnes called horn air or shaping air) was adjusted by
     setting the control knob on the gun wide open, then turning it the
     required nuntoer of turns towards the closed position.
   **Monitored at power supply.
   ttBent down and to the left.
                                       39

-------
the remainder finished on February 16, 1984.  All data were
gathered according to the requirements of the Test Procedure
(Appendix A) and QA/QC plan (Appendix B).

     All data satisfied the requirements of the outlier analysis;
no TE test runs had to be repeated.

TEST RESULTS

     TE's were calculated according to the test plan.  Final
results are presented in Table 15.  Seme corrections were made
to the original TE data when a QA scan identified several
unusual foil weights.  These foils were reweighed and the cor-
rect weights used to recalculate TE values.  These corrections
are reflected in Table 15.

STATISTICAL ANALYSIS

     Regressions are described for each target type in the same
manner as described previously for air atomized conventional
equipment. A discussion of how co uue the regression equations
is included in the AAC Statistical Analysis section.

     For AAE equipment, the variables are designated as follows:

                    x.arestricted atomizing air lines

                    x2-booth air rate (linear velocity)

                    x-agun cleanliness

                    x.orestricted paint lines

                    x=fan air (shaping air or horn air)

                    x -voltage
                     o
                    x."electrode position

                    TE°transfer efficiency

     Factors x.,x_ and x_ are qualitative variables and
therefore have duffimy variables associated with them.  The
regressions developed for each target type  follows.

     In the case of air atomized electrostatic- spray equipment,
engineering judgment suggested that the  following model terms,
including interactions, should be considered.

     o  Linear and quadratic in x.

     o  Linear in x_
                                 40

-------
     TABLE 15.  AAE TEST RESULTS
                      TE result
Run number        FP        VC      Graco

    1
    2
    3
    4
    5
    6
    7
    8
    9
   10
   11
   12
   13
   14
   15
   16
   17
   18
   19
   20
   21
   22
   23
   24
   25
   26
   27
   28
   29
   30
   31
   32
   33
   34
93.3
90.2
92.4
96.2
92.2
93.1
93.1
94.3
96.5
87.4
98.8
96.2
91.7
96.7
89.6
98.8
102.0
88.4
94.4
93.0
88.4
100.0
96.5
93.0
90.5
96.0
87.0
94.9
96.7
86. 8
92.9
94.0
98.6
95.4
33. 8
37.1
45.3
61.1
48.5
28.1
67.3
40.9
60.5
34.6
f.3.3
29.8
30.5
41.7
49. B
7. .4
50.4
42.0
47.4
45.0
44.5
56.5
3U9
29.6
28.6
44.8
36.3
52.7
76.4
77.3
79.6
75.4
79.8
73.9
56.2
60.2
60.2
72.2
62.8
49.5
77.7
57.6
70.1
54.4
72.8
57.3
52.6
62.3
61.5
78.7
67.4
60.8
67.9
64.0
65.5
70.5
56.7
56.7
54.9
70.0
54.0
69.1
78. 1
78.6
77.7
75.0
77.9
80.8
                        41

-------
     o  Dummy variables in x_
     o  Linear and quadratic in x.

     o  Dummy variables in x_

     o  Linear and quadratic in x,
                                 o
     o  Dummy variables in x_

     o  Interaction between x_ and x5


Flat Panel Target

      TE - 94.72 - 2.27x, + 1 . 56x .,
                        1        4

         + 1.22xg - 3.05x7d1

     All factors are significant only in linear form.  No quad-
ratic factors are significant for this gun type and target con-
figuration.  The negative coefficient on x. indicates a drop in
TE as air pressure increases.  Trie positive coefficient on x.
suggests that TE increases with increasing paint pressure.
Likewise, the positive coefficient on xg indicates that TE
increases with increased voltage.  The negative coefficient on
x-d. suggests a significant drop in TE, but only when the
electrode position is at level "-1."  (See Figure 3 to visualize
level "-1" compared to the other electrode positions.)

     The proportion of overall variance explained by the regres-
sion (R ) is 0.67.  This is a low R .   It is the result of a
lack of overall variance among test runs for this target con-
figuration.  The overall variance of all of the TE determina-
tions for AAE FP was  3.69, only about one and a half TE unit
above the targeted precision of 2.0.  When the overall variance
is low, it is difficult to tightly fit a regression model to
account for the small differences from run to run.

     The standard deviation of the replicates was 2.070, higher
than for most other cases, but very near the target value
of 2.O.*

     The error in the regression due to lack of fit was insig-
nificant with F = 1.42 (0.37 probability).
*CENTEC Corporation, "Development of Draft Standard Test Method
 for Spray Painting Transfer Efficiency," for USEPA under Con-
 tract No. 68-03-1721, Task 2.
                               42

-------
Vertical Cylinder Target

     The derived regression equation  for *AE VC is:

   TE = 41.34 - I.11x1 -  2.66x2 -  2

               * -  8.20x56, +  6.71x
     In this case,  several  factors  are significant, in both
linear and quadratic  form.  As  in the flat panel case, the
negative coefficient  on  x,  indicates that as air pressure
increases, TE drops.   Similarly, the negative coefficient on
Xj indicates that as  booth  air  rate increases TE decreases.
Tnis negative effect  is  moderated by the interaction of booth
air with x_, fan air.  The  positive coefficient of x-x-
indicates that  the  rate  of  change of TE with respect to booth
air depends on  the  prevailing level of fan air.  In particular,
this slope becomes  positive when the fan air is at the "wide
open" level.  Gun cleanliness,  x3,  exerts a significant effect
on TE only at the "+1" level.   (Figure 2 illustrates the different
levels of gun cleanliness.) Restricted paint lines are quadrati-
cally significant with a positive coefficient, indicating that
an increase in  paint  pressure also  increases TE.

     This may suggest that  there are interactions or special
effects on TE at certain levels of  fan air for this system.
The positive linear coefficient on  voltage indicates that as
voltage increases,  TE increases.  The negative quadratic
coefficient on  voltage indicates a  nonlinear effect for
voltage increases.  This negative coefficient moderates the
positive trend  for  the higher  levels of voltage.  Electrode
position was found  to be significant only at position  "+1" (shown
in Figure 3), but not for other electrode positions.

     The proportion of overall  variance explained by £he
regression  (R  ) is  0.92. This  is a relatively high R  , and
considered  indicative of a  good fit of the regression.

     The standard deviation of  replicate runs was 2.356, just
over the target standard deviation  of 2.0 for the procedure.

     The error  due  to lack  of  fit was statistically insignificant
at the 5 percent  significance  level (F = 2.66),

Graco Target

     The derived  regression for AAE testing using Graco targets
 is:
         TE = 66.78 - 0.89x1  - 1.14x2 -
            - 0.67x4 + 1.12x  -5.62x5d1  -
                                43

-------
     The directional effecto of x., x,, x- (at "+1" level),
x  (at "-1" level), and x, are the same aS for the vertical
cylinder AAE case.  Three new effects are identified for Graco
targets as compared to VC targets, as follows:

     o  Restricted paint lines have a small negative linear
        effect.

     o  Pan air is found significant at both the "-1" and "-H"
        levels-.  Both levels produce poorer transfer efficiency
        compared to the "0" level.

     o  Electrode position is found to have a significant
        effect at d1 (level "-1") and at d2 (level "-H").

     The proportion of overall variance explained by the re-
gression (R ) is 0.94.  This is considered a high value,
indicative of the good fit of the regression.  The standard de-
viation of replicate Graco test runs was 1.8606, well within the
2.0 limitation set by the test procedure.

     Table 16 presents the values of F and the associated
probability (P) for all variables and interactions found to be
significant.

     Tables 17, 18, and 19 present a comparison of predicted
values with observed for the flat panel, vertical cylinder, ?nd
Graco target respectively.

AAE CONCLUSIONS

     These regressions illustrate the differences target
configuration can make in TE.  Even with these differences,
however, there are  fundamental consistencies  among the
results.  Four variables  (restricted air lines, restricted
paint lines, voltage, and electrode position) are significant
for all target types.  Three other variables  (booth  air  rate,
gun cleanliness,  and fan air) are significant for VC and Graco
target configurations.  The consistency of these results across
target types strongly implies that all of the factors tested
for AAE spray equipment have an  important impact on  TE.

     The relative  importance of  each variable for a  certain
target configuration should be given individual consideration by
plant management.   It is recommended chat laboratory test  runs
be made with plant  paint and worXpiece targets  to determine
optimum combinations of factor levels  that result in acceptable
product finish.   The developed regressions should serve  as
guidelines  in  setting up  the  experimental design for site-specific
TE testing.  If such tests  are impractical,  the regressions may
                                 44

-------
serve as guidelines toward maximizing TE. Care mast be taken
when extrapolating the results for one spray system to another.
Previous test experience indicates that paint characteristics,
spray system characteristics, and target geometry can signifi-
cantly alter TE test results; however, the regressions may be
considered directionally sound for similar spray systems.
                                45

-------
 TABLE 16.  AAB F-STATISTICS (F)  AND ASSOCIATED  PROBABILITY  (P)*
            Flat   Panel
Effect
X1
X2
x3d3
X.
4
x«
F P
22.08 .0053
-
-
13.52 .0143

3.34 .034
x,a2
•J
            13.18   .065
Vertical  Cylinder
   F         P
  13.84   .0137
  34.58   .0020
  11.28   .020

  16.31   .0099

 159.6    .0000

  92.1    .0002
  16.62   .0076
  12.15   .0175
 7.68   .039
      Graco
     F      P
  5.71  .0038
  9.73  .0003
 13.85
  3.29  .0333
 11.93  .0000
  6.71  .0018
122.94  .0000
 18.77  .0001
 28.99  .0000
 10.61  .0002
  F and P are dimensionless.   Refer to Appendix  G for a definition
  of these terms.
                                46

-------
TABLE 17.  AAC-PP COMPARISON OF PREDICTED VERSUS ACTUAL TRANSFER EFFICIENCIES
Observation
number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Observed
value
93.30
90.20
92.40
96.20
92.20
93.10
93.10
94.30
96.50
87.40
98.80
96.20
91.70
96.70
89.60
98.80
102.00
88.40
94.40
93.00
88.40
100.00
96.50
93. 00
90.50
96.00
87.00
94.90
96.70
96.80
92.90
94.00
98.60
95.40
Predicted
value
91. IS
89.66
96.71
95.22
92.78
94.26
92.11
93.60
94.20
86.61
99.76
92.17
89.73
97.31
69.06
96.65
99.25
90.17
94.71
94.71
91.60
97.82
94.71
94.71
92.27
97.16
91.66
94.71
95.73
95.73
95.73
95.73
95.73
95.73
Residual
2.14
0.53
-4.31
0.97
-0.58
-1.16
0.98
0.69
2.29
0.78
-0.96
4.02
1.96
-0.61
0.53
2.14
2.74
-1.77
-0.31
-1.71
-3.20
2.17
1.78
-1.71
-1.77
-1.16
-4.66
0.18
0.96
1.06
-2.83
-1.73
2.86
-0.33
Lower 95% CL
for mean
89.40
87.76
94. 55
94.20
91.01
92.11
90.34
91.44
92.61
84.46
97.99
90.42
87.57
95.41
86.90
94.75
97.22
88.40
93.71
93.71
89.57
96.05
93.71
93.71
90.24
95.39
90.15
93.71
94.10
94.10
94.10
94.10
94.10
94.10
Upper 95% CL
for mean
92.90
91.57
98.86
96.24
94.54
96.42
93.87
95.75
95.80
88.77
101.52
93.92
91.88
99.22
91.21
98.55
101.28
91.94
95.71
95.71
93.63
99.59
95.71
95.71
94.30
98.92
93.17
95.71
97.35
97.35
97.35
97.35
97.35
97.35

-------
        TABLE  18.  AAC-VC  COMPARISON OP PREDICTED VERSUS ACTUAJ, TRANSFER EFFICIENCIES
oo
Observation
number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Observed
value
33.80
37.10
45.30
61.10
48.50
28.10
67.30
40.90
60.50
34.60
63.80
29.80
30.50
41.70
49.80
70.40
50.40
42.00
47.40
45.00
44.50
56.50
31.90
29.60
28.60
44.80
36.30
52.70
76.40
77.80
79.60
75.40
79.80
73.90
Predicted
value
33.49
40.57
51.46
65.82
54.08
36.38
62.94
37.96
56.30
34.16
65.16
35.74
31.27
42.79
49.24
68.04
43.88
39.44
38.68
38.68
51.52
51.52
26.34
J8.17
24.30
51.13
36.35
54.60
75.86
75.86
75.86
75.86
75.86
75.86
Residual
0.30
-3.47
-6.16
-4.72
-5.58
-8.28
4.35
2.93
4.19
0.43
-1.86
-5.94
-0.77
-1.09
0.55
2.35
6.51
2.55
8.71
6.31
-7.02
4.97
5.55
-6.57
4.29
-0.05
-0.05
-1.90
0.53
1.93
3.73
-0.46
3.93
-1.96
Lower 95% CL
for mean
27.12
34.27
45.20
60.99
47.38
30.24
56.20
31.56
49.76
28.36
58.31
29.62
24.60
36.34
42.93
62.31
38.37
32.53
33.85
33.85
44.22
44.22
19.89
32.80
15.39
31.73
31.73
48.17
71.47
71.47
71.47
71.47
71.47
71.47
Upper 95% CL
for mean
39.86
46.87
57.72
70.65
60.78
42.52
69.67
44.35
62.83
39.96
72.00
41.86
37.95
49.24
55.56
73.77
49.39
46.36
43.50
43.50
58.82
58.82
32.78
43.54
33.22
40.97
40.97
61.03
80.24
80.24
80.24
80.24
80.24
80.24

-------
TABLE 19.  AAC-GRACO COMPARISON OP PREDICTED VERSUS ACTUAL TRANSFER EFFICIENCIES
Observation
number
1
2
3
4
5
6
7
6
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Observed
value
56.20
60.20
60.20
72.20
62.80
49.50
77.70
57.60
70.10
54.40
72.80
57.30
52.60
62.30
61.50
78.70
67.40
60.80
67.90
64.00
65.50
70.50
56.70
56.70
54.90
70.00
54.00
69.10
78.10
78.60
77.70
75.00
77.90
80.80
Predicted
value
53.89
63.49
62.38
71.98
65.87
51.71
74.16
60.00
68.99
51.27
74.59
56.88
50.77
63.92
61.95
75.10
65.34
61.79
65.63
65.63
69.37
66.69
55.66
57.10
55.18
71.96
54.40
68.06
77.96
77.96
77.96
77.96
77.96
77.96
Residual
2.30
-3.29
-2.18
0.21
-3.07
-2.21
3.53
-2.40
1.10
3.12
-1.79
0.41
1.82
-1.62
-0.45
3.59
2.05
-0.99
2.26
-1.63
-3.87
3.80
1.03
-0.40
-0.28
-1.96
-0.40
1.03
0.13
0.63
-0.26
-2.96
-0.06
2.83
Lower 95% CL
for mean
51.25
60.43
59.26
69.97
62.85
48.65
71.03
56.94
65.92
48.22
71.47
54.10
47.63
60.86
58.82
72.13
62.62
58.65
62.08
62.08
65.13
63.13
51.90
53.33
52.45
68.82
50. 6b
64.29
75.92
75.92
75.92
75.92
75.92
75.92
Upper 95% CL
for mean
56.53
66.54
65.51
73.99
68.90
54.76
77.29
63.05
72.06
54.33
77.72
59.66
53.91
66.97
65.08
78.06
68.06
64.93
69.19
69.19
73.62
70.25
59.42
60.87
57.90
75. ',0
58.16
71.83
80.00
80.00
80.00
80.00
80.00
80.00

-------
                         SECTION 6

            AIRLESS CONVENTIONAL SPRAY EQUIPMENT
EQUIPMENT DESCRIPTION

     In airless spray painting, the paint flows from an orifice
at high pressure and breaks up into spray as it enters the at-
mosphere.  Typical paint line pressures are 6900 to 27600 kPa
(roughly 1000 to 4000 Ibs/in ).  Airless spraying avoids the
problem of turbulence caused by compressed air, which sometimes
prevents proper deposition of the paint on the workpiece.
Airless spray guns will atomize paint and permit application
into corners and recessed interior areas without the blow
back experienced with air spraying.

     Dirt or other small particles can obstruct the flow of
paint through the small orifice, which provides the atomization
in airless spray; therefore, special guns, pumps, hoses, etc.,
are required for airless spray.  Use of airless spray eliminates
the need for a hose from the compressor to the spray gun (See
Figure 4).

     Droplet sizes in airless spraying are larger than with
compressed air atomizing and consequently coatings applied by
airless spray are heavier and rougher.  Airless painting is used
widely to apply zinc primers and other highly pigmented paints
and is especially useful for large objectn.

OPERATING AND MAINTENANCE FACTORS

     Airless conventional spraying is an uncomplicated process
with few parameters involved.  The paint is supplied at high
pressure to the gun from which it  is expelled through a single
orifice.  The orifice is designed  to shape the spray.  Orifices
are designated by the diameter and half-width of the laydown
at 25.4 cm (10 in) target distance.  While plugging of one or
more holes in a conventional air spray cap is an operating and
maintenance problem, plugging of the single hole in an airless
cap, while possible, is such an obvious situation that the spray
gun operator always detects and corrects the problem before
proceeding.

     Erosion of the orifice with continued use does present a
maintenance problem for gun operation.  Obstructed paint supply
                                 50

-------
COMPRESSOR
PAINT
LINE
    Figure 4.   Airless paint spraying system
                            51

-------
lines leading to reduced pressure at the gun is also a c; »cern.
Finally, the flow of air in the vicinity of the target is of
interest.

     Test variables selected, then, were tip erosion, line plug-
ging, and varying booth air flow (See Table 20). The effect of
these variables on the spray painting operation were respectively
simulated by using orifices of progressively greater diameter,
by reducing the pressure of the paint at the gun, and by reduc-
ing the booth air flow.  Tip erosion was tested at three levels;
unused tip with 0.28 mm (0.011 in) diameter,  0.33 mm (0.013 in)
diameter, and 0.38 mm (0.015 in) diameter orifices.  Restricted
paint lines were tested at five levels:  9066.9 kPa (1300 Ibs/
in2), 8377.2 kPa (1200 Ihs/in2), n687 6kPa (11GO Ibs/in*),
6997.9 KPa (1000 Ibs/in ), and 6308.3 kPa (900 Ibs/in ).
Booth air rate- was simulated at two levels.

      ALC equipment specifications for the test are included
in Appendix '3.

EXPERIMENTAL DESIGN

     The experimental design for the airless conventional spray
is shown in Table 21.  It is discussed 
-------
     TABLE  20.   LEVELS OP OPERATING AND MAINTENANCE  VARIABLES
                   TESTED ON ALC SPRAY PAINTING  EQUIPMENT
                             Quant/   Ho. of
	Factor	qual.    levels	Test levels	

B.  Booth air rate*         Quant.     2      1- 0.61ra/s (120 ft/min)
       (linear velocity)                      -1- 0.36m/s (70 ft/rain)


C.  Tip erosion*-            Quant.     5t     a» 0.28 im (.011 in.) cap
                                              1» 0.28 mm (.011 in.) cap
                                              0= 0.33 mm (.013 in.) cap
                                             -1* 0.38 mm (.015 in.) cap
                                                 0.38 iim (.015 in.) cap
D.  Restricted paint lines  Quant.     5      a- 9J66.9 JcPa  (1300 psig)z
                                              1- 8377.2 JcPa  (1200 psig)
                                              0- 7687.6 kPa  (1100 psig)
                                             -1- 6997.9 kPa  (1000 psig)
                                                 6308.3 kPa  (900  psig)
    DUMMY                   Qual.      3      n/a
   •Actual booth air rates varied fran 100 to 140 ft/min  for level "+1" and
    50 to 90 ft/min for level "-1."  Average air velocities are used in this
    table.
   •K>un cleanliness w\s interpreted as "tip erosion"  for  this experiment.
    Progressively wider tip hole diameters were used  to simulate tip wear.
   tThe original experimental design called for five  levels; in practice
    we were orly able to simulate three levels.
   zMeasured at gun downstream of all paint filters.
                                     53

-------
          TABLE  21.   ATjC EXPERIMENTAL DESIGN
                Run  Number   B
                    1
                    2
                    3
                    4
                    5
                    6
                    7
                    8
                    9
                   10
                   11        -1
                   12         1
                   13
                   14
                   15
                   16
                   17
                   18
                   19
                   20
                   21
                   22
                   23
                   24
                   25
                   26
                   27         1
                   28         1
 Variable
 C      D
-1
-1
 1
-1
-1
 1
-1
-1
 1
 1
-1
 T
 1
 1
-1
 1
-a
 a
 0
 0
 0
 0
 a
 a
 a
 a
 a
 a
-1
•1
•1
 1
•1
-1
 1
.«
 1
-1
 1
 1
-1
 1
 1
 1
 0
 0
•a
 a
 0
 0
 a
 a
 a
 a
 a
 a
0
C
0
0
Where:
     B » Booth  air  rates—test at 2 levels-   1,-1
     C - Tip  erosion at 3 levels
     D « Restricted paint lines—test at 5 levels:   a,1,0,-1,-a
     Dummy  •  Dummy  variable not expected to affect  TE
                                 54

-------
TABLE 22.  ORDRR OF PERFORMANCE OP ALC "t'EST RUNS
                     23
                     21
                      5
                     25
                     29
                     11
                      2
                     16
                     15
                     28
                     22
                     10
                     17
                      7
                      6
                     13
                      9
                     19
                     26
                     12
                      8
                     27
                      14
                      1
                     24
                     18
                      3
                      4
                  (38  runs)
                        55

-------
to reduce some of the planned runs without sacrificing
information.  Runs 6, 11, 16, 23, 25, and 28 were dropped
from the design as shown in Table 21.
     In the case of airless conventional, the experimental
design allowed for estimation of regression terms of the
following type:
     o  Linear in x.
     o  Linear and quadratic in x2
     o  Linear and quadratic in x3
TEST RESULTS
     Tests were run and calculation:* performed in accordance
with the standard test method.  Values of transfer efficiency
obtained during testing are shown in Table 23.
STATISTICAL ANALYSIS
     Based on the TE test results, regression models were
developed to fit the data. Information on how to use these
regressions is presented in the AAC Statistical Analysis
section of this report.
     Variables are named in the regressions that follow according
to the table below:
                x.Bbooth air flow
                x2=tip erosion
                x3=restricted paint lines.
                TE=transfer efficiency
     Only those variables found to be significant have been
included in the final regression.  Tip erosion, x., is a
qualitative factor and therefore has dummy variables associated
with it.
Flat Panel Target
     The derived regression equation for ALC testing of flat
panel  targets  is:
            TE = 74.4 -  5.47x2  - 1.94x2
                                 56

-------
         TABLE 23.  ALC TEST RESULTS
                 Percent transfer efficiency
Run number         FP        VC      Graco

    1
    2
    3
    4
    5
    7
    8
    9
   10
   12
   13
   14
   15
   17
   18
   19
   20
   21
   22
   24
   26
   27
76.6
76.6
63.1
79.5
79.2
77.6
75.9
68.3
66.0
69.4
64.8
6C 5
80.5
77.5
65.5
75.7
74.8
76.4
70.7
67.7
68.3
70.2
13.0
13.6
10.4
13.7
13.8
12.9
>3.2
11.2
10.4
10.6
10.8
10.5
14.8
13.2
10.8
12.7
13.1
13.3
12.5
10.6
11.0
10.9
33.7
33.1
28.4
33.0
33.6
33.4
33.0
26.9
25.9
27.6
27.5
27.5
35.0
34.1
26.7
32.2
34.5
32.4
34.4
27.3
27.3
27.9
                         57

-------
    The only significant variable affecting TE is tip erosion.
The negative coefficient on x-d, implies that tip erosion
at level "1" makes TE go down!  The positive sign on x,d1
means that the level "-1" makes TE increase.

     The proportion of overall variance explained hy the
regression (R )  is 0.87.  This is considered a high value,
indicative of a good fit of the regression.  The standard
d.-viation of replicate FP test runs was 1.305, well within the
range of 2.0 specified in the test procedure.*  The error due to
lack of fit was insignificant, with F = 2.16 (P = 0.36). (Refer
to Appendix G for a glossary of statistical terms.)

Vertical Cylinder Target

     The regression analysis derived for ALC testing of VC
targets is:

       TE = 12.90 - 1.40x2 - 0.78X2


     Like the FP case, tip erosion was the only significant
variable found to affect TE for airless conventional spray
equipment.  In this case the direction of the effect is simi-
larly contingent on selection of level (i.e. dj or dj).

     The proportion of overall variance explained by the re-
gression model is 0.91.  This indicates a good fit of the
regression.

     The standard deviation of replicate VC test runs was
extremely small, at only 0.208.  While this standard deviation
is admirable given the test procedure precision of 2.0, it
raises some question as to why the procedure is so repeatable
for this target configuration.  The answer lies in the very
small overall standard deviation (only 1.4 across the entire
data set) created by intentional introduction of O&M variables.
The insensitivity of this system to intentional attempts to alter
TE demonstrates why the replicate standard deviation is so
small.

     The error due to lack of fit was insignificant at the
0.05 level, with P= 7.76 (P= 0.12).

Graco Target

     The regression analysis derived for ALC testing of Graco
targets is:

          TE = 33.38 - 3.25x2 - 2.98Xj - 0.26x3
*CENTEC Corporation,  "Development of Draft Standard Test Method
 for Spray Painting Transfer Efficiency," for USEPA under Contract
 68-03-1721, Task 2.
                                58

-------
     Tip erosion  is  found significant only at the d- {"+1" level)
for this target configuration.   This is an overwhelmingly large
effect, indicating that the effect on TB is very different at
this tip diameter than at other tip diameters for this system.
Restricted paint  lines are also significant for this case, but
only marginally so.   No interaction between factors is noted for
this system.

     The proportion  of overall  variance explained by the regres-
sion (R ) is high at 0.95.   This indicates a well fitting
model.  The standard deviation  of Graco target replicate test
runs is 0.346.  Like the previous ALC cases, this extremely
low standard deviation is the result of the insensitivity of
this system to the intentional  introduction of O&M factors.

     The F statistics and associated probabilities are given in
Table 24 for each effect included in the regression.


    TABLE 24.  Air  F-STATISnCS (P)  AND ASSOCIATED PROBABILITIES (P)*


              Flat Panel        Vertical  Cylinder
  Effect       P        P           P        P
  X2d1        11.39    .077         19.99     .0466

  X2d2        89.56   0.00         255.33     .0044    383.63   .0000

  x3            -                  -                16.59   .0002

     Tables 25,  26,  and  27 present a comparison of predicted
and observed transfer efficiency values, along with associated
significance limits,  for the flat panel, vertical cylinder, and
Graco targets,  respectively.

ALC CONCLUSIONS

     Three O&M  variables were selected for testing on ALC spray
painting equipment:   tip erosion, booth air rate, and restricted
paint lines. In  every test case, tip erosion is the overwhelming
variable affecting  TB.   The only other variable identified as signi-
ficant in any ALC test was restricted paint lines for the Graco
target.

     The tremendous response to changes in tip diameter is
indicative of a  very strong relationship between selection of
appropriate tip  diameter and TE.  Tip diameter should be careful-
ly selected.  Table 24 shows that the "+1" level displays by far
the most significant effect for all three target types.
*Refer to Appendix  G for glossary of statistical terms.
                                 59

-------
TABLE 25.  ALC-FP COMPARISON OF PREDICTED VERSUS ACTUAL TRANSFER EFFICIENCIES
Observation
number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
IS
16
17
18
19
20
21
22
Observed
value
76.60
76.60
63.10
79.50
79.20
77.60
75.90
68.30
66.00
69.40
64.80
66.60
80.50
77.50
65.50
75.70
74.80
76.40
70.70
67.70
68.30
70.20
Predicted
value
77.92
77.92
66.99
77.92
77.92
77.92
77.92
66.99
66.99
66.99
66.99
66.99
77.92
77.92
66.99
74.40
74.40
74.40
74.40
66.99
66.99
66.99
Residual
-1.32
-1.32
-3.89
1.57
1.27
-0.32
-2.02
1.31
-0.99
2.41
-2.19
-0.39
2.57
-0.42
-1.49
1.30
0.40
2.00
-3.70
0.71
1.31
3.21
Lower 95% CL
for mean
76.39
76.39
65.61
76.39
76.39
76.39
76.39
65.61
65.61
65.61
65.61
65.61
76.39
76.39
65.61
72.23
72.23
72.23
72.23
65.61
65.61
65.61
Upper 95% CL
for mean
79.45
79.45
68.36
79.45
79.45
79.45
79.45
68.36
68.36
68.36
68.36
68.36
79.45
79.45
68.36
76.56
76.56
76.56
76.56
68.36
68.36
68.36

-------
TABLE 26.  ALC-VC COMPARISON OF PREDICTED VERSUS ACTUAL TRANSFER EFFICIENCIES
Observation
number
1
2
3
4
5
6
7
a
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Observed
value
13.00
13.60
10.40
13.70
13.80
12.90
13.20
11.20
10.40
10.60
10.80
10.50
14.80
13.20
10.80
12.70
13.10
13.30
12.50
10.60
11.00
10.90
Predicted
value
13.52
13.52
10.72
13.52
13.52
13.52
13.52
10.72
10.72
10.72
10.72
10.72
13.52
13.52
10.72
12.90
12.90
12.90
12.90
10.72
10.72
10.72
Residual
-0.52
0.07
-0.32
0.17
0.27
-0.62
-0.32
0.48
-0.32
-0.12
0.08
-0.22
1.27
-0.32
0.08
-0.20
0.20
0.40
-0.40
-0.12
0.28
0.18
Lower 95% CL
for mean
13.20
13.20
10.42
13.20
13.20
13.20
13.20
10.42
10.42
10.42
10.42
10.42
13.20
13.20
10.42
12.44
12.44
12.44
12.44
10.42
10.42
10.42
Upper 95% CL
for mean
13.84
13.84
11.01
13.84
13.84
13.64
13.64
11.01
11.01
11.01
11.01
11.01
13.64
13.84
11.01
13.35
13.35
13.35
13.35
11.01
11.01
11.01

-------
         TABLE  27.   ALC-GRACO COMPARISON OF PREDICTED VERSUS ACTUAL TRANSFER EFFICIENCIES
o\
K)
Observation
number
1
2
3
4
5
6
7
3
9
10
11
12
13
14
15
16
17
IB
19
20
21
22
Observed
value
33.70
33.10
28.40
33.00
33.60
33.40
33.00
26.90
25. 9C
27.60
27.50
27.50
35.00
34.10
26.70
32.20
34.50
32.40
34.40
27.30
27.30
27.90
Predicted
value
33.38
33.38
26.88
33.90
33.38
33.90
33.33
27.40
26.88
27.40
26.88
27.40
33.90
33.64
27.14
32.85
33.89
33.37
33.37
27.66
27.66
27.66
Residual
0.31
-0.28
1.51
-0.90
0.21
-0.50
-0.38
0.50
-0.98
0.19
0.61
0.09
1.09
0.45
-0.44
-0.65
0.60
-0.97
1.02
-0.36
-0.36
0.23
lower 95% CL
for mean
32.77
32.77
26.20
33.25
32.77
33.25
32.77
26.89
26.20
26.89
26.20
26.89
33.25
33.08
26.61
31.87
32.91
32.58
32.58
27.02
27.02
27.02
Upper 95% CL
for mean
33.99
33.99
27.56
34.55
33.99
34.55
33.99
27.91
27.56
27.91
27.56
27.91
34.55
34.20
27.67
33.83
34.87
34.16
34.16
28.30
28.30
28.30

-------
     Prom the data generated during this test program, very
little can be said about the effects of other variables on TE
for ALC spray systems.  The response of TE to tip erosion is
so dramatic that it may obscure other potentially important
variables.
                                 63

-------
                            SECTION 7

              AIRLESS ELECTROSTATIC SPRAY EQUIPMENT
EQUIPMENT DESCRIPTION

     A Graco Mo<'sl AL-4000 was selected as the ALE spray equip-
ment for TE testing.  The AL-4000 is operated like conventional
airless spray equipment except the spray is electrically charg-
ed.  The electrical charge is an attractive agent pulling the
paint towards the nearest ground, the target.  Electrical power
is supplied at a controlled voltage on the electrode at the gun
tip.  Fluid flows through the gun at high pressure and is
atomized through a carbide tip.  The atomized paint picks up an
electrical charge as it is sprayed past the charged electrode.
The spray pattern of ALE equipment is determined primarily by
tip orifice size.  Fluid flow cannot be adjusted at the gun
(as it can in conventional and conventional electrostatic
equipment); it is either full on or full off.

     Graco standard black enamel was used as the test paint.  It
was cut to 25.5 seconds on a Shell #3 cup at 25*C.  A 16 L
(4 gal) batch of paint was mixed and stored in a 20 L (5 gal)
Graco paint pressure pot.  This batch was not enough to complete
all ALB testing and was made up on the second and third days of
testing.  The paint was kept in a constant temperatjre booth
along with the paint pump (Model 207-707, K3D 30:1), stirrer,
viscosity measurement equipment, and som* supply lines.  All
paint supply lines were insulated.

     ALE tests were conducted in the Dynaprecipitator water wash
booth described in Section 4.  Booth characteristics were
identical to earlier test runs, with only two air speeds avail-
able.

     Foil weights were determined on Precise laboratory scales
accurate to 0.01 g.  Height-percent-solids samples weie weighed
on 0.0001 g accuracy scales.

     A forced-air, gas-fired oven was used for curing weight-
percent-solids samples and TE samples.  The oven was cleaned
daily to prevent contaminants from adhering to the samples.  All
samples were cured at 171.1"C (340*F) for 20 minutes.  This
is a more severe cure than for previous experiments.  It was
                                54

-------
instituted  to  ensure a complete  cure  for  the heavier  laydown of
paint expected for  this equipment  type.   Trial  and error weight-
percent solids determinations  were made to document the point of
assured complete  curing.

     The mass  flow  meter described in Section 3 was used for
paint ma»s  flow determinations.  The  test method presented
in Appendix A  was strictly  adhered to for ALB testing, as were
the QA/QC requirements of the  test.

      ALB equipment specifications for this test series are
included in Appendix P.

OPERATING AND  MAINTENANCE FACTORS

     Variables had  been previously identified through interviews
and literature search that  were  considered to have an important
effect in achieving optimum TE for ALE equipment.  These 14 var-
iables are presented in Table  28.   Five variables were selected
for ALE testing on  the basis of  the number of times it was iden-
tified for ALE by different sources.-  the  ability to simulate the
variable within the prescribed test methodology, and finally,
the limitation of laboratory time. The five selected test
variables were:

     o  Booth  air rate (linear velocity)

     o  Tip erosion (substituted for  gun  cleanliness)

     o  Restricted  paint  lines

     o  Voltage

     o  Electrode position


     A dummy variable was also included to provide a measure of
th«» inherent error  in the experiment.

EXPERIMENTAL DESIGN

     An experimental design was  developed to accommodate the
limitations of  testing while addressing the effects of each
variable as completely as possible.

     As before, the first restraint on experimental design
was the availability of  laboratory time:  only about 30 test
runs could be  reasonably  expected  during  a week of tescing.
The second limitation was the  number  and  type of simulation
levels for each variable.   Only  two booth air rates were
possible in the test laboratory, while three levels of fan

-------
    TABLE  28.   OPERATING AND MAINTENANCE VARIABLES
               FOR ALE SPRAY EQUIPMENT*
               Booth air rate
               Booth configuration
               Cure schedule (time,  temperature)
               Paint discharge technique
               Equipment design
               Flash off
               Gun cleanliness
               Gun condition
               Gun-to-target distance
               Operator error
               Paint mass flow rate
               Paint characteristics
               Restricted paint supply
               Target configuration
*as mentioned by industry sources contacted
                            66

-------
air were achievable, and  ifive or more levels of some other
variable could be simulated.  Table 29 presents the type of
variable (quantitative/qualitative) and levels to be accom-
modated in the experimental design.

     Table 29 shows the use of a dummy variable.  This variable
represents the effect of  a totally unrelated action on TE.  If
the data analysis shows any significant effect for the dummy
variable it is indicative of some type of problem with the test
method or test performance.

     Table 30 presents the ALE experimental design.  In this
figure, the abbreviations "a," "1," "0," "-I," and "-a" denote
the level of each variable to be tested.  Level "a" denotes the
base level with a good spray pattern.  Level "-a" denotes the
poorest level of a factor to be tested.  The intermediate levels
"1," "0," and "-1" are determined along equal spacing from "a"
to "-a" for the particular variable.  Variable levels for ALE
testing were determined in pretest trials as described in the
following subsection.

     The first 16 test runs in the experimental design are the
fractional factorial portion of the design.  When the results of
several variables are to  be studied, a factorial design is
usually the nost efficient method to use.*  The basic idea o£
factorial design is to alter several aspects of a test at a
time, but in  such a way  that the effects of individual alter-
ations can bos determined. Fractional factorial designs sacri-
fice some ability  to test for interaction between variables but
are able to test for main effects very efficiently.

     Runs 17 through 26 in Table 30, are the "star" portion of
the experimental design.  This portion o: the experiment tests
the eftects of variables  at the extremes of their range (for the
system under test, at "-a" and "a").  The star design broadens
the range of information  gathered in the test.  The star portion
of the design allows extra degrees of freedom in order to assess
lack of fit.

     As in the case of previous designs, the design used hare
entails a central composite design for variables C, D, P, «md G.
However, G contains only  3 levels while C, D, and ? contain 5
levels.  As a result, the design is not a standard central
composite design.
*Youden, W. J.  and  Steiner,  E.  H.,  Statistical Manual of the
 Association of Official  Analytical Chemists, Arlington, Va.,
 1982; and Davies,  O.  I.. ,  Design  and AnalyaTa of Industrial
 Experiments, Great  Britain,  1979.
esig
7T9
                                 67

-------
       TABLE 29.  EXPERIMENTAL VARIABLES SELECTED FOR
                  TESTING ALF. SPRAY EQUIPMENT
 Factor
   ID    Factor description
Quant./
qual.
  No.  of
test levels
   B     Booth air rate (linear
           velocity)                     Quant.
   C     Gun cleanliness (tip erosion)   Quant.
   D     Restricted paint lines          Quant.
   P     Voltage                         Quant.
   G     Electrode position              Qual.
Dummy    Dummy action or variable        Cual.
               2
               5
               5
               5
               3
               2
                                68

-------
        TABLE 30.   ALE EXPERIMENTAL DESIGN
Run  number   B     C

     1        -1    -1
     2         1
     3        -1
     4        -1
     5         1
     6         1
     7        -1
     8         1
     9        -1
    10         1
    11         1
    12        -1
    13        -1
    14         1    -1
    15         1     1
    16        -1     1
    17         1    -a
    18         1     a
    19-0
    20-0
    21               0
    22               0
    23-0
    24-0
    25               0
    26               0
    27               a
    28               a
    29               a
    30               a
    31               a
    32               a
                          Variable
 0
 0
-a
 a
 0
 0
 0
 0
 0
 0
 a
 a
 a
 a
 a
-1
-1
-1
 1
-1
 1
 1
 1
 1
 1
-1
 1
-1
-1
 1
-1
 0
 0
 0
 0
 0
 0
-a
 a
 0
 0
 a
 a
 a
 a
 a
 a
            _G

            -1
 1
 1
 1
 0
 0
 0
 0
 0
 0
 0
 0
-1
-1
-1
 1
-1
 1
-1
 1
 1
 1
-1
 1
-1
 1
-1
 1
-1
 0
 0
 0
 C
~ 1
 1
 0
 0
 0
 0
 1
 1
 1
 1
 1
 1
Where:
   B = Booth  air  rates—test at 2 levels:  1,-1
   C » Tip  Erosion—teat at 5 levels:   a,1,C,-1,-a
   D = Restricted paint lines—test at 5 levels:
       a,1,0f-1,-a
   F =• Voltage—test at 5 levels:  a,1,0,-1,-«;
   G * Electrode  position—test at 3 levels:   1,0,-1
   Dummy »  Dummy  variable not expected to impact TE
                             b9

-------
     The last six runs of the test design are replicates.
Replicates are provided at the base condition of all variables
to provide a measure of the precision of the test.

ALE TEST PERFORMANCE

     ALE testing began February 27, 1984.  Equipment set up,
target assembly and hanging, foil cutting and preweighing, and
other preparatory tasks were completed earlier in the week.  The
paint was adjusted to desired specifications in a 20 L  (5 gal)
paint pot.  Once the paint was adjusted, all equipment  and lines
were checked for proper installation and freedom from obstruc-
tion.  The mass flow meter was calibrated and zeroed.   Mass flow
calibration was double checked against anatomized paint capture
and found to be within 0.4 percent of the meter reading, as
required.

     Base level ("a") for each variable was determined  by
setting the equipment according to Graco experience with the
test paint and spray painting system.  Some adjustments were
necessary to provide a good spray pattern without excessive
laydown.  Final base levels were confirmed by a visual  spray
pattern check.  Base levels thus determined are shown in
Table 31.

     Deteriorated levels were selected by setting all factors
except the subject variable (for each variable in turn) at
the base level.  The subject variable was changed until a sig-
nificantly worse spray pattern could be discerned.  Th-.» spray
pattern was checked by spraying onto a paper target for 5 to 6
seconds, and observing the resulting pattern.  Deteriorated
variable levels ("-a") thus determined are shown in Table 31.

     Intermediate levels were calculated to be evenly spaced
from the base level ("a") *:o the deteriorated level ("-a").
Intermediate levels are also shown in Table 31.

     Electrode position was similarly defined.  Base level was
with the electrode in normal position.  Deteriorated level
("-a") was selected with the electrode clipped off.  An inter-
mediate level was decided as a bent electrode.  All electrode
position levels are shown in Figure 5.  (Tip orientation was
vertical for actual testing.)

     Although gun cleanliness had been selected as an ALE
experimental variable, a partially blocked tip could not be
simulated.  Any blockage affixed to the tip was blown out by
the pressure of the paint during spraying.
                                70

-------
  TABLE  31.   LEVELS OF  OPERATING AMD MAINTENANCE VARIABLES
               TESTED ON  ALE SPRAY PAINTING  EQUIPMENT
          Factor
 Quant/   No. of
 qual.    levels
                       Test levels
B.  Booth air rate*
Quant.
                  1= 0.61m/s  (120 ft/min)
                 -1= 0.36m/s  (70 ft/min)
C.  Tip erosiont
Quant.
                  a= 0.28 inn  (.011 in.) cap
                  1= 0.28 mm  (.011 in.) cap
                  0= 0.33 mn  (.013 in.) cap
                 -1= 0.38 mm  (.015 in.) cap
                     0.38 mm  (.015 in.) cap
D.  Restricted paint lines  Quant.
                  a= 9066.9 kPa  (1300 psig)z
                  1= 8032.4 kPa  (1150 p*,ig)
                  0= 6997.9 kPa  (1000 psig)
                 -1= 5953.4 XPa  (850 psig)
                 -a= 4929.0 kPa  (700 psig)
F.  Voltage*
Quant.
G.  Electrode position
Qual.
                  a= 72 kV
                  1= 63 kV
                  0= 54 kV
                  1= 45 kV
                     36 kV
                  1= normal
                  0= bent ++
                 -1= clipped off tt
   *Actual booth air rates varied frcm 100 to 140 fpm for level "+1" and
    50 to 90 fpm for level B-1."  Average air velocities are used in this
    table.
   1Gun cleaniness was interpreted as "worn tip" for this experiment.
    Progressively wider  .±p hole diameters were used to simulate tip
    wear.
   +The original experimental design called for 5 test levels; in
    practice we were only able to simulate 3 levels.
   zMeasured at gun downstream of all paint filters.
   ttKonitored at power supply.
  •Hfient as shown in Figure 5.
  ttElectrode cut off at plane of cap.
                                    71

-------
      FRONT
                                    SIDE
       Level +a and +1: Normal electrode position
                                    SIDE
               Level 0: Bent electrode
       FRONT
                                     SIDE
        Level -a and -1: Electrode cut off
Figure  5.   Airless electrostatic air cap showing
            test  levels for  electrode position
                            72

-------
     To salvage the variable, it was decided to look at another
identified variable instead.  The only ether tip factor identified
for ALB equipment was tip erosion.  Abrasive paints can erode
the tip orifice after prolonged use.  To simulate tip erosion,
tips at a variety of diameters were obtained and checked for
spray pattern.  Only three tips gave acceptable spray patterns,
at 0.28 mm, 0.33 mm, and 0.38 mm diameters.  With larger tips,
the paint laydown was too heavy to avoid running; smaller tips
were not avialable.

     As in previous experiments, the booth air rate could only
be controlled to two levels.  Voltage and restricted paint lines
were each simulated at  five  levels, as shown in Table 31.

     Variaole level selection was completed on February 27, 1984.
ALE test runs were started on February 28, 1984.  Weight-percent-
sol.xds samples were taken.   The results were in close agreement,
and testing began in randomized order according to the test plan.

     Nine test runs were completed the first day of ALE testing.
One run was thrown out  due to a timer malfunction; one run was
deleted because the booth water wash was not on; and one run was
eliminated because grounding wires had not been attached to the
foils (even though the  flat  panel was grounded).

     Paint had to te added and adjusted for the second day of
testing.  All preparatory steps were taken, but on the first run
the mass flow meter totalizer stuck.  Mass flow measurements
were lost, and the run  was repeated immediately after repair of
the malfunctioning switch.   After eight runs were completed, the
laboratory experienced  a 2-1/2-hour power failure.  When power
was restored, all start-of-test QA/QC measures were repeated
before resuming testing.  Final weight-percent-solids determina-
tions were made after  15 runs were completed.  The morning and
evening weight-percent-solids determinations agreed nicely, but
the power failure sample was several weight-percent higher.
The power failure sample had not been stirred during the power
failure, and probably was not adequately stirred prior to sampl-
ing.  This weight percent solids was omitted from TE calculations
as a suspect sample.

     Paint was added and viscosity adjusted for the final day of
ALE testing.  All prescribed preparatory steps were taken ac-
cording to the test plan  (Appendix A) and QA/QC plan (Appendix B).
The rest of the experiment was completed without incident on
March 1, 1984.

TEST RESULTS

     Tests were run and calculations performed  
-------
               TABLE 32.   ALE TEST RESULTS

                       Percent transfer efficiency
        Run number        FP        VC      Graco

                         91.5      47.5     83.6
            *            87.8      43.2     64.6
            3            93.9      42.2     61.0
            4            89.8      68.8     77.6
            5            83.3      44.1     58.0
            6            93.3      65.1     77.2
            7            96.1      76.2     74.5
            8            90.2      60.2     70.8
            9            90.9      69.4     76.3
           10            93.7      48.2     64.3
           11            94.8      60.0     70.0
           12            87.6      68.0     71.9
           13            90.8      66.1     71.9
           14            84.7      48.0     63.7
           15            90.6      71.9     74.2
           16            88.5      66.4     72.1
           17            91.8      53.3     63.7
           18            89.6      56.7     67.7
           19            88.6      66.2     75.1
           20            91.9      57.0     68.3
           21            90.5      55.3     68.3
           22            84.3      51.0     65.8
           23            89.2      44.1     59.4
           24            93.2      74.9     76.2
           25            88.3      47.3     62.5
           26            88.8      65.9     72.3
           27            91.7      78.6     78.2
           28            92.5      77.8     79.1
           29            89.1      73.3   74.2, 76.
           30            90.4      75.9     79.3
           31            92.6      80.8     78.7
           32            93.4      76.6     76.8
           33            	      	     78.2*
*An extra replicate using only the Graco targets was made for
 Grace's own purposes.  The data is included here for com-
 pleteness.
                                74

-------
STATISTICAL ANALYSIS
     The terminology shown below is used in the regressions
that follow:
            x.=booth air rate  (linear velocity)
            x.=tip erosion
            x3=restricted paint lines
            x.=voltage
            x-=electrode position
            TE=transfer efficiency
     In the case of the airless electrostatic, the following
linear, quadratic, and interaction effects were chosen for
the regression model.
     o  Linar in x.
     o  Linear and quadratic in x-
     o  Linear and quadratic in x.
     o  Linear and quadratic in x.
     o  Dummy variables in x_
     o  Interaction between x. and x.
     o  Interaction between x_ and x.
     o  Interaction between x- and x^
     o  Interaction between x. and x.
    A discussion of how to use the regression equations is
presented in the AAC Statistical Analysis section of this
repo.rt.  The derived regression for each target type follows.
Flat Panel Target
      TE » 90.30 - 1.12x2 + 1.37x4 - 0.68x^3
     Only linear effects are significant for ALE testing of flat
panel targets.  Tip erosion, x2, has a negative effect on TE.
The positive coefficient 
-------
     The proportion of overall variance explained by the re-
gression (R ) is 0.28.  This is the poorest case of all equip-
ment types and target configurations tested.  The raw data was
reviewed to locate any test errors contributing to this un-
usually low R , but no experimental source was found. The low
R  may be attributed to the low overall variation in this test
series.  The variation of TE over all of the experimental com-
binations was only 3.0. This value is barely above the standard
deviation of the test procedure (2.C).  The lack of variance
demonstrates the insensitivity of the system to O&M factors. The
standard deviation of replicated test runs was 2.287, high and
not quite within the specified range of the test procedure.  The
error due to lack of fit is insignificant, with P » 0.38.

Vertical Cylinder Target

      TE = 57.31 - 3.77x1 + 2.85x2 - 1.55x3

         + 7.02x, - 4.66x_d1 + 8.62x,.d_
                4        51        3 I
     More than twice as many variables are significant for VC
testing than were identified for FP testing.  The negative
coefficient on x. indicates that TE decreases with increasing
booth air rates.  The positive coefficient or. tip erosion indi-
cates that as tip diameter decreases, TE increases.  But neither
the booth air rate nor the tip erosion trends are constant
because of the interaction between the two.  The coefficient of
the interaction is negative.  Thus the negative effect of booth
air is enhanced at the high level of tip erosion but is moderated
at the low level of tip erosion.  TE is adversely affected by
increasing restrictions in the paint lines (x3).  However this
negative trend is not constant, due to the interaction with tip
erosion.  Increasing voltage ..ends to increase TE, dramatically.
The magnitude and direction of the effects of different levels of
electrode position changes with the selection of electrode
position.  Figure 5 shows the various test levels for electrode
position.

     Two interactions are significant for this case, tip erosion
with booth air and tip erosion with restricted paint lines.
Each effect acts in a different direction.  Nevertheless, it is
clear that tip erosion is the overwhelming factor for this
case.

     The proportion of overall variance explained by the regres-
sion (R  ) is 0.95, a respectable value.  The standard deviation
due to repeats is 2.55, slightly over the 2.0 value specified
in the test procedure.  The error due to lack of fit is insigni-
ficant, with F= 2.16 (P- 0.20).

-------
Graco Target
       TE - 69.05  -  2.83x1 -

          - G.84xJ + 4
     The Graco target  configuration  identified the most signifi-
cant O&H variables  for ALE testing.  Like  the other rases,  in
creasing booth air  rates  (x. ) causes a drop  in TE.  Restricted
paint lines  (x.) also  cause  a drop in TEr  while  increasing
voltage (x.) raises TE linearly and  causes it to slightly drop
quadratically.  Electrode position is significant at tne d2
("+1") level only.  At this  level TE is  increasing with changes
in electrode position.  Apart from the linear effects, inter-
actions between x.  and x_, x. and x_, and  x. and x, complicate
the system.  The linear trends described above are distinct but
the rates of change of TE with respect to  x. , x-, and x.
are not constant.   As  an  example, the booth  air  rate effect is
negative but is moderated at the high level  of x_, restricted
paint lines.

     The proportion of overall variance  explained by the regres-
sion (R > is a modest  O.B3.  The standard  deviation of rep-
licate test  runs on Graco targets was 1.274, well within the
2.0 specified by the test procedure.  The  error  due to lack
of fit was insignificant, with F» 3.04 (P=» 0,14). Table 33
gives the value of  the P-statistic and associated probability
for each effect of  significance.

     Tables  34, 35, and 36 presents  compariaona  of predicted
and observed values of transfer efficiency for each experimental
condition for the flat ptnel, vertical cylinder, and Graco
targets respectively.

ALE CONCLUSIONS

     ALE test results  sho^-d the most difference in discrimi-
nation among target C'.v.cigurations.  Only  two variables were
identified as significant for the flat panel target (tip
erosion and  voltage),  and these were only  marginally signifi-
cant.  The Graco and vertical cylinder targets identified
four and five significant variables, respectively, not includir.g
several interactions between v& -tables.

     Voltage, booth air rate, restricted paint lines, and
electrode position  were significant  factors  for  Graco and VC
targets types.  These  results are consistent with findings from
AAE and ALC  experiments:  where electrostatic forces are  involv-
ed, voltage, booth  air rate, and electrode position are impor-
tant to establishing optimum TE.  The higher the voltage and the
                                 77

-------
TABLE 33.  ALE F-STAT1STICS (P) AND ASSOCIATED
           PROBABILITIES (P)
Effect
X1
x2
X3
X4
S5d1
X5d2
X1X2
X1X3
X2X3
X3X4
x2
Flat Panel
F P
6.64 .05
-
4.85 .07
-
-
-
6.42 .05
-
-
- -
Vertical
F
54.31
22.94
9.53
191.31
10.09
30.86
8.61
-
7.00
-
-
Cylinder
P
.0007
.0049
.0272
.0000
.0246
.0026
.0325
-
.0457
-
_
Gzaco

F P
117.16 .
-
111.94 .
166.26 .
-
72.09 .
7.67 .
-
30.94 .
32.82 .
10.30 .
0000
-
0000
0000
-
0000
0006
-
0000
0000
OC01
                             78

-------
         TABLE  34.   ALE-FP  COMPARISON OF PREDICTED VERSUS ACTUAL TRANSFER EFFICIENCIES
-j
so
Observation
number
1
2
3
4
5
6
7
8
9
10
It
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Observed
value
91.50
87.80
93.90
89.80
83.30
93.30
96.10
90.20
90.90
93.70
91.80
87.80
90.80
84.70
90.60
63.50
91.80
89.60
88.60
91.90
90.50
84.80
89.20
93.20
88.30
88.80
91.70
92.50
87.10
90.40
92.60
93.40
Pradicted
value
69.39
88.45
90.69
*2.13
87.16
91.19
93.42
93.42
89.89
92. 1 3
90.69
91.19
87.16
89.39
89.89
88.45
91.41
89.17
89.00
91.59
90.29
90.29
87.55
93.03
90.29
90.29
90.62
90.62
90.62
90.62
30.62
90.62
Residual
2.10
-0.65
3.20
-2.33
-3.86
2.10
2.67
-3.22
1.00
1.56
4.10
-3.59
3.63
-4.69
0.70
0.04
0.38
0.42
-0.40
0.30
0.20
-5.99
1.64
0.16
-1.99
-1.49
1.07
1.87
-3.52
-0.22
1.97
2.77
Lower 95% CL
for mean
87.08
06.20
88.78
90.08
84.94
88.98
91.12
91.12
38.52
90.08
88.73
88.98
84.94
87.08
88.52
86.20
89.73
87.63
86.96
89.27
P9.25
8*.?5
85.2
9t. o::
89.25
89.25
88.72
88.72
88.72
86.72
88.72
88.72
Upper 95% CL
for mean
91.70
90.70
92.49
94.16
89.37
93.40
95.73
95.73
91.27
94.18
92.59
93.40
89.37
91.70
91.27
90.70
93.09
90.72
91.01
93.90
91.33
91.33
89.86
95.04
91.33
91.33
92.51
92.51
92.51
92.51
92.51
92.51

-------
         TABLE  35.  ALE-VC  COMPARISON OP  PREDICTED VERSUS  ACTUAL TRANSFER EFFICIENCIES
oo
o
Observation
number
1
2
3
4
5
G
7
8
9
10
11
12
13
14
15
16
1?
18
19
20
21
22
23
24
25
26
27
28
29
30
31
J2
Observed
value
47. M)
43.20
42.20
68.80
44.10
65.10
76.20
60.20
69.40
48.20
60.00
68.00
68.10
48.00
71.90
66.40
53.30
56.70
66.10
57.00
55.30
51.00
44.10
74.90
47.30
65.90
78.60
77.80
73. 3C
75.90
80.80
76.60
Predicted
value
47.86
42.90
44.63
75.18
42.93
70.22
68.95
57.95
68.07
51.73
57.20
68.10
67.32
50.98
70.25
67.34
52.49
54.60
64.17
57.98
53.54
53.54
47.04
75.11
48.88
62.16
77.28
77.28
77.28
77.28
77.28
77.28
Residual
-0.36
0.29
0.56
-6.38
1.16
-5.12
7.24
2.24
1.32
-3.53
2.79
-0.10
0.77
-2.98
1.64
-0.94
0.80
2.09
2.32
-0.98
1.75
-2.54
-2.94
-0.21
-1.58
3.73
1.31
0.51
-3.98
-1.38
3.51
-0.68
Lower 95% CL
for mean
43.52
38.90
37.40
71.01
39.15
66.15
64.65
53.82
63.83
47.58
53.05
64.07
63.20
46.86
68.14
63.18
49.09
51.29
63.58
54.25
50.85
50.85
43.45
71.28
46.38
59.84
74.80
74.80
74.60
74.80
74.8J
74.80
Upper 95% CL
for mean
52.20
46.91
45.87
79.34
46.72
74.30
73.26
62.09
72.30
55.88
61.36
72.12
71.43
55.09
72.37
71.51
55.88
57.90
67.76
61.71
56.23
56.23
50.63
78.84
51.38
64.49
79.77
79.77
79.77
79.77
79.77
79.77

-------
TABLE 36.  ALE-GRACO COMPARISON OF PREDICTED VERSUS ACTUAL TRANSFER EFFICIENCIES
Observation
number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Observed
value
83,60
64.60
61.00
77.60
53.00
77.20
74.50
70.80
76.30
64.30
70.00
71.90
71.90
63.70
74.20
72.10
63.70
63.70
75.10
68.30
68.30
65.80
59.40
76.20
62.50
72.30
78.20
79.10
76.20
79.30
78.70
76.80
Predicted
value
74.27
65.34
62.99
82.74
60.59
73.81
76.75
70.77
73.17
64.77
71.71
73.71
74.11
60.43
74.35
69.36
65.41
67.01
77.25
66.51
66.21
66.21
62.09
74.91
66.21
70.92
77.93
77.93
77.93
77.93
77.93
77.93
Residual
9.32
-0.74
-1.99
-5.14
-2.59
3.38
-2.25
0.02
3.12
-0.47
-1.71
-1.81
-2.21
3.26
-0.15
2.73
-1.71
0.68
-2.15
1.78
2.08
-0.41
-2.69
1.28
-3.71
1.37
0.26
1.16
-1.73
1.36
0.76
-1.13
Lower 95* CL
for mean
70.36
61.42
59.39
78.48
56.65
69.63
72.77
66.84
69.83
61.08
67.78
70.08
69.85
56.32
71.81
65.14
62.81
64.37
73.78
62.65
64.08
64.08
57.16
70.44
64.08
68.35
75.35
75.35
75.35
75.35
7£.35
75.35
Upper 95% a,
for mean
78.18
69.25
66.59
87.01
64.53
77.99
60.73
74.70
76.51
68.47
75.63
77.33
78.37
64.53
76.89
73.58
68.01
69.66
80.72
70.36
68.35
68.35
67.02
79.38
68.35
73.48
80.50
80.50
80.50
80.50
80.50
80.50

-------
lower the booth air rate, the better TE is likey to be.  Thus,
ALB spray painting equipment should be maintained to supply the
maximum allowable voltage to the tip.  Periodic checks of power
supply are recommended to assure tip voltage remains at the
desired level.  Booth air rate should be kept to the lowest
level acceptable for safety, environment, and worker comfort.

     The effect of the position of the electrode in the atomized
paint field is less clear, appearing significant in some cases
and insignificant in other similar cases.  It seems prudent,
however, to maintain the electrode position well into the atom-
ized paint field.  Trimming the electrode is not recommended.

     Restricted paint lines have a significant effect on TE for
Graco and VC target types.  This is a shared phenomenon with
other equipment types.  Pressure of the paint supply to the
spray gun should be monitored to avoid degeneration through
clogging or other restrictions.  If the paint pressure is not
monitored, trie operator may notice a loss of spray quality, but
he is likely to take an inappropriate action to remedy the
problem.  This situation is espei:'.ally true for air-atomized
spray systems where the operator may adjust the fan air or the
atomizing air to counteract the effects of lower paint pressure.
It is equally applicable for ALE spray systems.
                                 82

-------
                           SECTION 8

                     COMPARISON OP TARGETS
BACKGROUND

     The Draft Standard Test Method  (Appendix A) specifier two
sets of targets for spray painting in each test run.  These
targets are described in detail in Appendix A.  The test targets
consisted of a set of foil-covered aluminum vertical cylinders
(VC) mounted in certain positions inside a wooden frame, and a
set of foil strips mounted at certain spacing on a large flat
stainless steel panel (PP).  Both targets were suspended from an
overhead conveyor for the test.  The VC targets were designed to
be somewhat representative of smaller, more open <*nd intricate
substrates.  The PP targets were designed to be representative
of large, relatively flat and closed substrates.  The test
results from a single transfer efficiency determination include
a VC result and a PP result.  These results have quite different
values.

     During the test program at Graco, a third set of targets
(called Graco targets) were painted at the same conditions as
the Draft Standard Test Method targets.  These targets consisted
of a set of ten 15.24 cm (6 in) wide metal panels mounted
15.24 cm (6 in) apart, and hanging 121.92 cm (48 in) long.  The
TE results from the center six cargets were averaged to obtain a
single TE value.  The TE value obtained for this uarget type was
different from the values obtained for VC or PP targets.

     This chapter evaluates the transfer efficierry character-
istics of all three target types for four equipment types to
determine if any of the designs has special advantages over
other targets for future testing.

COMPARISON OP FACTORS IDENTIFIED AS SIGNIFICANT

     Table 37 presents a summary of the variables identified as
significant for each target type and each equipment type.  The
Graco target configuration was the most sensitive, identifying
23 significant O&M variables (cr interactions) over all equip-
ment types.  VC targets carr.e in a close second by identifying
19 significant variables, followed by FP targets at only  13
significant variables.
                                83

-------
        TABLE 37.  COMPARISON OF SIGNIFICANT FACTORS IDENTIFIED
                    BY  THREE TARGET  CONFIGURATIONS

Eauipment
type
ALE







R squared
ALC

R squared
AAE







R squared
AAC






VC
Booth air
Tip Eros.
Paint lines
Voltage *
Elect, pos.
Booth air x tip
Tip x paint lines

0.95
Tip eros. *

0.91
Air lines
Booth air
Gun cleanliness
Paint lines
Fan air
Voltaqe *
Electrode pos.
Booth air x fan air
0.92
Air lines *
(Booth air-close)

Paint lines
Fan air
Target con-
figuration
Graco
Booth air *

Paint lines *
Voltage *
Elect, pos. *
Booth air x tip
Tip x paint lines
Paint lines x volt.
0.83
Tip eros. *
Paint lines
0.95
Air lines
Booth air
Gun cleanliness
Paint lines
Fan air
Voltage *
Electrode pos.

0.94
Air lines *
Booth air
Gun cleanliness
Paint lines
Fan air *


FP

Tip eros. (marg)

Voltage (marg)




0.28
Tip eros. *

0.87
Air lines *


Paint lines

Voltage
Electrode pos.

0.67
Air lines *
Booth air

Paint lines *
Fan air *
	 air lines x booth air 	


	 booth air x
fan 	
R squared
0.79
0.96
0.99
* Strong response, overriding  factor influencing TE
(marg)  Marginally significant  response
                                  84

-------
     The variables  identified by the Graco targets match up
fairly consistently with those identified by VC's; FP targets
presented some anomolies.

     Graco targets had the highest correlation coefficients,
averaging 0.92,  followed closely by VC at 0.89 and FP at 0.70.

WORTH ASSESSMENT OF THREE TARGET CONFIGURATIONS

     A Worth Assessment Model* was constructed to evaluate the
relative merits of  each target type.  Six criteria were selected
for this evaluation:

           1. High  correlation coefficient (ability to fit
              mathematical models)
           2. Target discrimination (ability to identify
              significant effects)
           3. Ease  of fabricating the target
           4. Ease of transporting/storing the target
           5. Ease of use during testing
           6. Target cost

     Each of the targets was ranked from 0 (low) to 1 (high) for
these criteria.  The rank was multiplied by weighting factors
and summed to generate a score.  Several different weighting
factor combinations were used in calculations to compare the
effects en the final score.

     In every case, the Graco target configuration scored high-
est.  The Graco  target scored consistently higher in almost all
categories than VC  or FP targets.  The Graco targets were easier
to handle, provided the best sensitivity to significant factors,
and demonstrated a  very good correlation coefficient.

     The worth assessment scores for evenly weighted criteria
were:

                    Graco         0.79
                    VC            0.50
                    FP            0,50

     This case is the closest competition between target types
Detailed computer printouts of this analysis, is shown in Table
38.
*CENTEC Corporation,  ""Worth Assessment Model," computer
software, Copyright  1979.
                                 85

-------
                   TABLE  38.  WORTH  ASSESSMENT MODEL COMPARING TARGET CONFIGURATIONS
      (1)  VC  target
        Factor  name
                            Ranking   Selection description
                              Weight
         Value
      1  Correlation  coefficient      1.0
      2  Target  TE  discrimination     0.8
      3  Ease  of fabrication          0.3
      4  Ease  of transport/storage    0.5
      5  Ease  of use  during test      0.0
      6  Cost  of target               0.5
                                      Very high R squared
                                      Identifies many factors
                                      Difficult to make
                                      Moderately difficult to T*S
                                      Difficult to use and handle
                                      Moderate cost
                               0.170
                               0.170
                               0.170
                               0.170
                               0.170
                               0.150

                               Score
        0.17000
        0.12750
        0.04250
        0.08500
        0.07500
        0.07500

        0.50000
      (2)  Graco target
        Factor  name
                            Ranking   Selection description
                              Weight
         Value
00
Oi
1  Correlation coefficient     0.8
2  Target TE discrimination    1.0
3  Ease of fabrication         0.8
4  Ease of transport/storage   0.8
5  Ease of use during test     0.8
6  Cost of target              0.8
High R squared
Identifies most factors
Fairly easy to fabricate
Fairly easy to trans. & store
Fairly easy to use and handle
Relatively inexpensive
,170
,170
,170
,170
,170
,150
                                                                          Score
0.12750
0.17000
0.12750
0.12750
0.12750
0.11250

0.79250
      (3)  FP target
        Factor name
                            Ranking   Selection description
                              Weight
         Value
      1  Correlation coefficient     0.8
      2  Target TE discrimination    0.3
      3  Ease of fabrication         1.0
      4  Ease of transport/storage   0.3
      5  Ease of use during  test     0.3
      6  Cost of target              0.5
                                      High R squared                 0.170     0.12750
                                      Identifies a few factors       0.170     0.04250
                                      Easy to fabricate              0.170     0.17000
                                      Difficult to trans. & store    0.170     0.04250
                                      Very inconvenient to use       0.170     0.04250
                                      Moderate cost                  0.150     0.07500
                                                                           Score
                                                                               0.50000

-------
                           APPENDIX A

             DRAFT STANDARD METHOD FOR SPRAY PAINTING
     TRANSFER EFFICIENCY OPERATIONS AND MAINTENANCE TESTING*
1.   SCOPE

    1.1  This method covers testing to determine the effects of
         certain operating and maintenance factors on transfer
         efficiency.  Four types of spray equipment, air
         atomized conventional (AAC), airless conventional
         (ALC), air atomized electrostatic (AAE), and airless
         electrostatic (ALE) are to be tested.

    1.2  The factors selected for testing and the levels of each
         factor to be tested are summarized in the experimental
         design matrix for each type of spray equipment (Subtask
         Re. ort, Tables 5, 6, 7, and 8).

    1.3  This test program is estimated to take 4-5 weeks of
         laboratory work.

    1.4  This method is applicable only to solvent or wafer-
         borne coatings applied in a single pass.  The same
         coating shall be used for all tests in this program.

2.   APPLICABLE DOCUMENTS

    2.1  ASTM Standards:

         •  D-1200-70 Viscosity of Paints, Varnishes, and
            Lacquers by Ford Viscosity Cup

         •  D-2369-81 Standard Test Method for Volatile Content
            of Coatings

         •  D-1005-51 Measurement of Dry Film Thickness of
            Organic Ccatings
*Many conventional  industrial units are used throughout the
 test procedure to  accommodate participating laboratories and to
 minimize conversion errors on site.  Metric conversions are
 made as required as shown in the conversion list at the front
 of the report.
                                 87

-------
         •  ASTM D1212-79 Measurement of Wet Film Tnickness of
            Organic Coatings

         •  ASTM D2353-68 Flow Rating of Organic  Coatings  Using
            the Shell Flow Comparator

         •  ASTM D1475-6U Density of  Paint,  Varnish,  Lacquer,
            and Related Products

    2.2  ANSI/IEEE Metric Practice

3.   SUMMARY OF METHOD

    3.1  A battery of specially designed targets  are  covered
         with preweighed, labeled foil,  then spray painted in  a
         single pass under rigidly controlled conditions as
         specified in the test matrix.   The  foils are removed
         from the targets, cured, and weighed.  The net weight
         gain is divided by the weight of paint sprayed at the
         targets to yield a single transfer  efficiency
         determination.

    3.2  The battery of targets is composed  of  2  sets of 4
         targets each.   The first set of targets  consists  of 4
         foils mounted in prescribed  positions  on a large  steel
         plate.  The mean weight  gain for these 4 foils is used
         to calculate the transfer efficiency.  This  target con-
         figuration is intended to be representative  of large,
         relatively flat industrial workpieces.  The  second set
         of targets consists of 4 foils  mounted on widely  spaced
         vertical cylinders.  The mean weight gain for these 4
         foils is used to calculate the  transfer  efficiency.
         This target configuration is designed  to be  repre-
         sentative of smaller, more intricate and open
         industrial workpieces.

    3.3  A transfer efficiency determination shall be made for
         each set of conditions shown in each test matrix, ex-
         cept that runs will be performed in randomized order
         within each matrix.

    3.4  base conditions ("a") shall  be  established through a
         set of pre-test runs to  determine levels of  each
         factor at good spray conditions.  Deteriorating levels
         {"1,0,-1,-a")  of each factor will be determined from
         the base levels.  The base level and deteriorating
         levels of each factor shall  be  determined prior to
         beginning each test matrix.
                                  88

-------
4.   TEST TARGETS

    4.1  Test targets consisting of a set of 1-1/4-inch diameter
         aluminum cylinders and a large stainless steel flat
         pan.^1, configured as shown in Figure A-l or Figure A-2,
         shall be used for this test.

5.   TEST APPARATUS

    5.1  Spray painting booth, preferably back-drawn with
         100-fpm linear air velocity at the plane of the tar-
         gets or, if not available, any booth meeting regula-
         tions for the type of spray system being tested may be
         used.  The same spray booth shall be used for all tests
         of a series.  The spray booth must be large enough to
         accommodate che prescribed targets.  The spray booth
         must be equipped with a conveyor system capable of
         carrying the test panels past the spray equipment at
         the desired speed, and capable of at least 2 linear air
         rates.

    5.2  Four complete systems (AAE, ALE, AAC, and ALC) for
         spray painting application, including spray gun,
         paint supply pot, power supply (if electrostatic),
         air supply lines, paint supply lines, power cables
         (if electrostatic), regulators, and pressure gages
         shall be used in this test.

    5.3  Scales of suitable size and accuracy shall be used for
         paint mass flow rate determinations.  Laboratory scales
         of suitable size and accuracy shall *e used for
         weighing test foils.  Accuracy of 0.01 percent is
         recommended as a minimum accuracy for scales.

    5.4  Foil, mounted to cover vertical cylinder and flat test
         panels as shown in Figure A-l and Figure A-2 shall be
         used.  Six-inch wide 1.5-mil medium temper alloy foil
         shall be used for covering the test panels.  The shiny
         side of the foil shall always face out.

    5.5  A standard 10-minute stopwatch with 0.1-second accuracy
         shall be used.

    5.6  Tape measure, graduated in 1/16 of an inch, 10 feet
         long, such as a rigid carpenters' rule, may be used.

    5.7  Aluminum foil dish, 58 mm in diameter by 18 mm high
         with a smooth bottom surface shall be used.

    5.8  Syringe, 5 ml, capable of dispensing the coating under
         test at sufficient rate that the specimen can be
         dissolved in solvent shall be used.
                                 89

-------
                                                       Target movement toward gun
                  hi-t film
                  I lucl in-sj
vO
o
                                                       conveyor
                                              Foil Number

                                             4321
   Foil Number

4321
                                         Vertical Cylinder

                                          Target (VC)
































1







6"



C"
*"*'



h"

6"







6"
M
V
C
o
ID
U
•A


d"
  Flat Panel Target (FP)
                                   Figure A-l.   Target Configurations for  Air Atomized
                                                  Conventional and Electrostatic Spray Guns

-------
                                                    Target movement toward qun
                                       conveyor
vo
\
\

1


.0"












1
•
x •









M
U


>

-------
    5.9   Forced  draft  curing  cwen,  sufficient  to hold a  com-
         plete set of  test  foils  and  aluminum  dishesr shall
         be used.

    5.10 Wet film measurement gage.

    5.11 Thermometer,  with  suitable range  for  spray  and  cure
         conditions, accurate to  0.2°F  shall be used.

    5.12 Anemometer, with suitable  range  for booth  linear
         velocity, accurate to 3  percent  of reading  shall  be
         used.

    5.13 Test Notebook,  a bound test notebook  containing the
         test procedure, data sheets, reference methods, and
         QA/QC Plar  shall be  provided to  the  laboratory  by
         CENTGC.

6.   PROCEDURE AND CALCULATIONS

    6.1   Perform calibration of the platform  scale  once  per week
         or each time  that  it is  moved and leveled,  whichever
         occurs more frequently.   Perform calibration of the
         laboratory  scale once every test series.   Calibrate  all
         pressure gages per standard operating procedure prior
         to test.

    6.2   Select test equipment for first  test  series.  Using
         Data Sheet  1, document the test  equipmpnc  specifica-
         tions.   Be  suio to check all information and sign the
         form.  Each data  sheet shall be  double checked by a
         second party, either engineer or technician, and  signed
         off.
                                  92

-------
                           Data Sheet 1

                   Test Equipment Specifications

      Test Date:             Test No.:      Data by/Checked by;
A. Weight Percent Solids Measurement Equipment
   1.  Laboratory Scales
       a. Manufacturer
       b. Model No.                           ~
       c. Serial No.                          ~
       d. Capacity, g                         ~
       e. Rated accuracy, g                   ~
   2.  Foil Dishes
       a. Type
       b. Size
   3.  Syringe
       a. Type
       b. Capacity, mL                        ~
   4.  Solvent Type                           "
B. Conveyor Speed Measurement Equipment
   1.  Rule
       a. Type
       b. Graduations
   2.  Electronic Timer
       a. Type
       b. Manufacturer
       c. Model No.
       d. Serial No.
       e. Rated accuracy, s
C. Mass Flow Measurement Equipment
   1.  Platform Scales
       a. Manufacturer                       	
       b. Model No.                          	
       c. Serial No.                         	
       d. Capacity, kg                       	
       e. Rated accuracy, g                  	
   2.  Stopwatch
       a. Manufacturer                       	
       b. Model No.                          _
       c. Serial No.                         __
       d. Rated accuracy, s                  	
D. Target Foil
   1.  Type                                  	
   2.  Nominal Thickness, mils               	
   3.  Temper                                	
E.  Wet Film Measurement Equipment
       a. Manufacturer                       	
       b. Model No.
                                 93

-------
    6.3  Select coating.  The same coating shall be used for all
         tests in this projram.  Using Data Sheet 2- document
         the paint characteristics.  Paint characteristics shall
         be documented daily, at each addition of paint, and at
         other times as requested by the CENTEC engineer or GRACO
         representative.  Again, check your information and sign
         the form.

                             Dita Sheet 2

                         Paint Specifications
Test Date:
Test No.:
                                          Data by/Checked by:
 1.  Paint  Type

 2.  Resin  Type

 3.  Manufacturer

 4.  Manufacturer's Paint ID No.

 5.  Lot No.

 6.  Color

 7.  Recommended Cure Schedule

 8.  Viscosity (uncut)

 9.  Reducing Solvent

10.  Vol. of  Solvent Put into
    Vol. Paint

11.  Viscosity - Spray (cut)*

12.  Wt./Gallon - Spray

13.  Wt. Solids - Spray

14.  Resistivity or Conductance
                                            min. 9     "F	

                                         sec.*    Ford Cup 8  °F
                                                (vol) solvent in
                                                (vol) paint	
               see.*
                                                  Ford Cup g
                                                   Ibs/qal
                                                         «A
*Use ASTM D-2353-68, ASTM D-1200-70, or ASTM D-3794 part 6.
                                 94

-------
6.4  Set up paint supply eqii'pment and platform scale.
     Using Data Sheet 3, document the paint supply equip-
     ment specifications.  Be sure to check your informa-
     tion and sign the form.
                              95

-------
                         Data Sheet 3

      Paint Spray and Peripheral Equipment Specifications

Test Date:                  Tesc No.:           Data by/Chkd by:
A. Paint Supply Tank

   1. Type
   2. Manufacturer
   3. Model No.
   4. Serial No.
   5. Rated Capacity, gal

B. Paint Spray Equipment

   1. Type
   2. Manufacturer
   3. Model No.
   4. Serial No.
   5. Rated Capacity, cc/min
   6. Air Cap
   7. Fluid Tip
   8. Needle

C. Paxnt Spray Booth
   1. Type
   2. Manufacturer
   3. Model No.
   4. Serial No.
   5. Rated Capacity, cfm

D. Conveyor
   1. Type
   2. Manufacturer
   3. Model No.
   4. Serial No.

E. Forced Draft Oven
   1. Type
   2. Manufacturer
   3. Model No.
   4. Serial No.

F. Paint Heaters
   1. Type
   2. Manufacturer
   3. Model No.
   4. Serial No.
                                  96

-------
6.5  For electrostatic spray equipment only, ground paint
     supply equipment and platform scale per Figure A-3.
     NOTE:  In accordance with Section 9-8 of NFPA 33 for
     fixed electrostatic apparatus, measure resistance of
     equipment to ground (conveyor frame) to insure resist-
     ance is less than 1 x 10& Ohm.

6.6  Using a small glass jar with an airtight lid, take
     paint grab sample from paint pot.  ASTM D-3925-81 pro-
     vides a good standard practice guide for paint sampling.
     Record test series number on label of jar.

6.7  Measure weight solids from paint sample.  Use syringe
     weight difference technique as described in A.STM
     D-2369-81.  Document the cure oven bake sch-adule and
     temperature on Data Sheet 4.  Be sure you use the cure
     schedule recommended by the manufacturer on Data Sheet
     2.  Record raw data and results on Data Sheet 5.
     Paint weight percent solids should be determined before
     each test series, at the start of each test day,
     periodically between tests, and at the end of each test
     day.  The participating laboratory shall store all weight
     percent solids samples until notified by CENTEC that the
     data analysis is complete.
                              97

-------
                                 nc A

I
Paint
Suppl
Tank



K Digital electronic
platform scale

                        Arrangenent B
                  Electrostatic Equipment
Gi




oindir.g Cable
gMflMHgggkggj



Paint
I_
Supply
Tank




.^Electrical Insulation Block*


r
•Slock -ust be capaole  of preventing current flow from supply tank to
 grour.a t.-.rougn t^e platfor— scale.
    Figure A-3.   Set-up  for Paint Supply Equipment and
                    Platform Scales
                                  98

-------
                            Data Sheet 4

                   Equipment Operating Conditions

Test Date:                    Test No.:         Data by/Chkd by:
A. Paint Spray Equipment
   1. Paint Pressure at Paint Pot, psig
   2. Paint Pressure at Spray Gun, psig
   3. Atomizing/Turbine Air Pressure at
       Spray Gun, psig
   4. Operating Voltage, kV
   5. Disk or Bell Speed, rpm
      a. With Paint Applied
      b. Without Paint Applied
   6. Shaping Air for Bell, psig
   7. Paint Temperature at Paint Pot, °F
   8. Gun to target distance, cm
   9. Pump Setting

B. Paint Spray Booth
   1. Ambient Temperature, °F
   2. Relative Humidity, %
   3. Air Flow Velocity, fpm
   4. Air Flow Direction

C. Target Parameters
   1. Average Wet Film Thickness, mils
   2. Average Dry Film Thickness
   3. Vertical Paint Coverage, cm (in)
   4. Target Height, cm (in)
   5. % Vertical Coverage
   6. Resistance to Ground, Ohm

D. Forced Draft Oven*
   1. Cure Time, minutes
      a. Foil Dish  (sample)
      b. Target Foil

   2. Cure Temperature, °F
      a. Foil Dish  (sample)
      b. Target Foil

E. Paint Heaters
   1. Temperature In, °F
   2. Temperature Out, °F

F. Conveyor Speed Setpoint, fpm  (cm/sec)
*Same cure schedule as foils.
                                 99

-------
                          Data Sheet 5

               Weight Solids Test Data & Results
Test Date:
Test No.:
   Data by/Chkd by:
1.  Syringe Weight
    a.  Full, g
    b.  Empty, g
    c.  Net Wet Paint, g

2.  Dish Weight
    a.  After Drying, g
    b.  Empty, g
    c.  Net Dry Solids, g

3.  % Weight Solids (2c/lc)
                                Sample
                                  A
Sample
  B
                      Average
                             A3
NOTES:

    1.  Actual Cure Schedule
         min.
Refer to ASTM 2369-81, Procedure B of "Standard Test Method  for
Volatile Content of Coatings."
                                 100

-------
6.8  Set up the paint spray equipment.  Using Data Sheet 3,
     document specifications for the paint spray equipment
     and spray booth used in this test.  Check your infor-
     mation and sign the data sheet.
         NOTE;  Equipment selection, equipment
         condition, paint selection, target con-
         figuration, and operating conditions have
         a substantial effect on transfer efficiency.
         Care should be taken to use the same booth
         and spray equipment, paint, targets, and
         operating conditions as specified for the
         run in the test matrix (Data Sheet 4, Sec-
         tions A, B, C, and 6a, Sections Id, Ic,
         3, and 4) from test to test for comparable
         results.
6.9  Set up the conveyor speed measuring equipment.  This
     equipment may consist of photoelectric cells or limit
     switches used in conjunction with an automatic digital
     timer.  Alternatively, the conveyor speed may be nea-
     sured using timing marks (chalk marks) on the conveyor
     in conjunction with a hand held stopwatch.  Figure A-4
     shows the permissible methods for conveyor speed mea-
     surement.  Using Data Sheet 6a, record the horizontal
     distance between the photo cell or limit switch on/off
     positions.

6.10 Determine base level of each test factor which will
     produce a reasonably good spray pattern and finish.
     If base level has already been determined for test
     series, proceed to 6.13.  The CENTEC engineer in
     agreement with the laboratory representative shall
     determine "reasonably good spray pattern and finish."
     Base level shall be determined only once for each
     test series.  Base level is denoted by "a" in the
     test matrix.

6.11 Determine deteriorating levels of each test factor to
     be examined in this test series.  Selection shall be
     niad<» by reducing the level of each factor to a point
     where an obvious impact on spray pattern and finish
     is noted.  Again, the CENTEC engineer in agreement
     with the laboratory representative shall determine
     the level where spray pattern and finish is obviously
     poor.  This level  is  the poorest value of each
     factor.  It is denoted by "-a" in the test matrix.
                              101

-------
                             METi:02 A
       Tarqet
                                                            E.'.ictronic
                                                              Timer
e ,
 n
              A • Stationary photoelectric cell or limit switch

              3 - Stationary photoelectric cell or limit switch

              C • Moving plate of KJIOWH width




                             METHOD B
Known Distance
                      Conve"or
              £ • Fixed timing nark

              r « Moving tvsunq nark




Target.



Target



=
Target
IF
                                                            G
                                                            Stop Hatch
     Figure A-4 •   Permissible Methods  for Measuring
                      Conveyor  Speed
                                     102

-------
6.12 Levels "1," "0," and "-1" shall be selected at even
     absolute spacing from the value of each variable "a"
     to "-a."

6.13 The value of "a", "1", "0", "-l"r and "-a" shall remain
     fixed for each variable through a test series.

6.14 Set up targets in accordance with Figure A-l or A-2, as
     appropriate.  Target configuration, material, and
     spacing is critical.  Scavengers are metallic, as is
     the FP target.  Cut 6-inch-wide aluminum foil strips to
     required length for each target.  Label each foil strip
     with the appropriate nomenclature.  (Nomenclature is
     shown on Table A-l.)  Weigh each foil strip and record
     value on foil and on Data Sheet 6b.  Check your infor-
     mation and sign the data sheet.

6.15 Attach foils to the vertical cylinder and/or flat panel
     targets as shown on Figure A-5 or A-6, as appropriate.
     Perform resistance check to verify adequacy of ground-
     ing.  Per NFPA  33 Section 9-8, resistance shall be  less
     than 1 x 10* ohms.

6.16 In accordance with Figure A-l or A-2, attach shim stock
     to scavenger  in order to measure wet  film thickness.

6.17 Adjust all equipment operating parameters,  i.e., gun to
     target distance, paint pot pressure,  turbine air pres-
     sure, etc.,  to  base values.  Set  factor  levels  to values
     required for  this test run in  the matrix.   Record equip-
     ment operating  parameters on Data sheet  4.   Check your
     data and sign the data sheet.  NOTE:   In accordance with
     Section 9-7  of  NFPA  33 for tised electrostatic  apparatus,
     the gun to target distance shall be at  least  twice  the
     sparking distance.
          NOTE;   Equipment selection,  equipment
          condition,  paint selection,  target  con-
          figuration, and operating conditions have
          a substantial effect on transfer efficiency.
          Care should be taken to use  the same booth
          and spray equipment, paint,  targets, and
          operating conditions (Data Sheet 4, Sec-
          tions  A,  B, C, and 6a,  Sections Id, Ic, 3,
          and 4) from test to test for comparable
          results.
 6.18 Check spray equipment and parameters to assure they
      are correct for this run.
                             103

-------
           TABLE A-l.   NOMENCLATURE FOR SPRAY PAINTING
                       TRANSFER EFFICIENCY TESTS   	
Each test foil will be labeled in 5 segments as follows:

1.  Spray Equipment Type

    Air atomized conventional     :    AAC

    Airless conventional          :    ALC

    Air atomized electrostatic    :    AAE

    Airless electrostatic         :    ALE


2.  Target Configuration

    Flat Panel       :  FP
    Vertical Cylinder:  VC


3.  Target Position:  1, 2, 3, or 4.


4.  Test Serirs Identifier (letter or number)




    Example:  AAC-FP2-12

      where AAC = air atomized conventional spray equipment

            FP2 « the second  flat panel target

              12 • test rur. identifier
                                  104

-------
        f: rt:-al Cylinder
                                        Direction cf  -vrar
                                           Hold edgo  of  foil in place against
                                           cylinder »hilc wrapping leading edge
                                           arour.d cylinder.
                                           roil
                                II |  '
                                HI  j
  i.   '.-.'far  vortis.ii cylinder taracts with cylinders nounted-on  target bracket
      iSco  riqurc  ;   inc  J  i .   Wrap  so the leadinq edge  foms a seaw away
      frcn  t:-.o cir:rtion of srrav.
                                         "Cri="
                                             "Crip"
                                                               "Grip"
                                                           "Grip-
As leading  odTC overlaps starting
oilgo,  solidly  "grip" foil into piaco
by tjraspinc foil-covered cylinder.
Secure foil on cylinder by gripping
the length of the cylinder.  Foil  will
have a uniformly wrinkled surface.
        Figure  A-5.   Vertical Cylinder Wrapping  Technique
                                        105

-------
      FLAT PANEL TAR&ZT
Flat Panel Target
   7oil-Peady for
   attachment -
                                       Double-sided  tape
    Figure A-»>.   Flat  Panel Foil  Attachment  Technique
                             106

-------
                           Data Sheet 6a

                     TE Test Data and Results


Test Date:                  Test No.:        Data by/Checked by:
A. Weight Percent Solids  (from Data Sheet 5)         	 A3


B. Total Solids Sprayed


   1. Paint Spray Plow Rate

      a. Beginning Weight, g           	

      b. End Weight, g                 	

      c. Time Between Weighings/ s     	
      d. Flow Rate, g/s                	Bid


   2. Conveyor Speed

      a. Distance Between Marks, cm     	

      b. Time Between Narks, s         	
      c. Speed, cm/a                              B2c
   3. Total Effective Target
      Width, cm*                         15.24    63
   4. Total Solids Sprayed at Each
      Target, g
      (A3 x Bid x B3/B2c)              	B4
   5. Micromotion-metered paint
      mass flow rate, g/s                         Bid"
   * Total effective  target width is six inches per foil on flat
     panel target  (on 6" centers), and six inches per cylinder
     on vertical cylinder target  (also on 6* centers).  Six
     inches •  15.24 cm.
                                  107

-------
    6.19 For electrostatic spray equipment,  measure  the gun  tip
         operating voltage (with lines full  of  paint,  but  gun
         not operating).   Adjust to desired  voltage  and record
         on Data Sheet 4.

    6.20 Check conveyor clock,  stopwatch,  micromotion  meter  and
         platform scale to ensure that all have been zeroed
         (reset) and that  the scales are  in  the tare mode.*

    6.21 Turn on conveyor.  As  the leading edge of the first
         scavenger passes  in front of the  gun,  turn  on paint
         spray equipment and initiate flow;  simultaneously,
         start stopwatch and record scale  reading.

    6.22 As the trailing edge of the last  scavenger  passes in
         front of the gun  area,  stop stopwatch  and paint spray
         flow simultaneously.  Turn off conveyor.  Record
         platform scale, conveyor clock,  micromotion meter
         flow rate, and stopwatch readings on Data Sheet 6a.
         Check the data and sign the data  sheet.

    6.23 Measure wet film  thickness on shim  plate and  record on
         Data Sheet 4. line C-l.

    6.24 Remove foils from targets, making sure no tape has
         stuck to the targets and no paint is lost.  Securely
         attach coated foils to oven racks so all painted
         surfaces are exposed for uniform  drying.  Spring
         clips or tacks may be used to.mount wet targets on
         racks.  Insert racks in oven And  bake  at recommended
         schedule per Data Sheet 2.  Flash time (the time  be-
         tween spraying and getting the targets into the oven)
         should be kept to a minimum.  Set oven timer  per
         recommended schedule.

    6.25 Remove foils from oven and record actual bake schedule
         on Data Sheet 4.   Weigh foils and record weight on  each
         foil and on Data  Sheet 6b.  After weighing, store foils
         in appropriately  labeled plastic  bags, i.e.,  bags that
         have test run number identified.  The  laboratory  shall
         retain all samples until data analyses are  complete.
         Check all data for correctness and  completeness.  Both
         the engineer and  technician must  check and  sign all
         data sheets before proceeding.

         Replicates ot each test run shall be made immediately
         after the original run, if required.
*Du.ing 10 tasts,  ch^ck micromotion meter  vs  manual  deter-
 minations.   If  within precision requirements (see QA/QC  Plan),
 use only micromotion meter thereafter.
                                 108

-------
                                  Data Sheet 6b

                            IE Test Data and Results
Test Date:
Test No.:
Data by/Checked by:
C. Total Solids on Target
   Flat Panel Target

   Foil Weight After Drying, g:

Foil f  1               2
   Foil Weight Before Spraying, g:
                                    Total
D. Vertical Cylinder Target

   Foil Weight After Drying, g:

Foil *  1               2
                                             Net Dry Solidsf g:
                                    Total
    Foil Weight Before Spraying, g:
 E. Transfer Efficiency  (by weight]1

   Flat  Panel Target

   Vertical Cylinder Target
                                             Net Dry Solids,  g:
 1. TL =
                            (Net  Dry Solids, g) x 100%
         ("Total Solids  Sprayed at Each Target, g" fron Data Sheet 6a) x  (4 targets)
                                         109

-------
6.26 Perform TE calculations as indicated on Data  Sheet  5,
     6a, and 6b using the weight solids determined for
     the test series.  Document results on Data  Sheet  6b,
     noting that each transfer  efficiency observation  is
     the mean of the transfer efficiency for 4 foils.

6.27 Repeat above steps (6.2 through 6.26) for each test
     run.

6.28 Be sure all data sheets have been properly  completed,
     checked, and signed.

6.29 Record transfer efficiency in appropriate column  on
     Data Sheet 7.   When roughly 70 percent of the runs
     in a series are complete,  the CENTEC engineer shall
     transmit the TE results to Dr. Ray Myers at 703-
     961-5638.   Dr.  Myers shall perform an outlier
     analysis and respond to the engineer within 24
     hours.  Outlier runs will  be repeated as resources
     allow.

6.30 To proceed with the next run in a series, go  to
     6.10.   To begin a new test series, go to 6.1.

6.31 CENTEC shall retain all original data sheets  and  the
     test notebook.
                             110

-------
Where:
                            Data  Sheet 7

                  Air Atomized Conventional Test

                              FACTOR                 TE Result
        Run Number   A      B     C     D     E     PP        VC
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
-1
-1
-1
-1
1
1
1
1
1
1
1
-1
-1
-1
-1
1
-a
a
0
0
0
0
0
0
a
a
a
a
a
a
-1
-1
-1
1
-1
1
1
1
-1
-1
-1
1
1
1
-1
1
-1
-1
1
1
-1
-1
1
1
1
1
1
1
1
1
-1
-1
1
-1
-1
1
»1
-1
1
1
-1
1
1
-1
1
1
0
0
-a
a
0
0
0
0
a
a
a
a
a
a
-1
1
-1
-1
-1
-1
1
-1
1
-1
1
1
-1
1
1
1
0
0
0
0
-a
a
0
0
a
a
a
a
a
a
1
-1
-1
-1
-1
-1
-1
1
-1
1
1
-1
1
1
1
1
0
0
0
0
0
0
-1
1
1
1
1
1
1
1
     A = Restricted air lines—test at 5 levels: a,l,0,-l,-a
     B = Booth air rates—test at 2 levels:  1,-1
     C B Gun cleanliness—test at 5 levels:  a,l,0,-l,-a
     D = Restricted paint lines—test at 5 levels:  a,l,0,-l,-a
     E = Pan air—test at 3 levels:  1,0 ,-1
                                   Ml

-------
Where:
                          Data Sheet 7

                   Airless Conventional Test

                                 FACTOR              TE Result
        Run Number   BCD     Dummy        FP        VC
            1       -1    -1    -1      -1
            2        1    -1            -1
            3       -11-1      -1
            4       -1-11      -1
            5       -1    -1    -1       1
            6        11-1-1
            7        1-11-1
            8        1-1-1       1
            9       -111-1
           10       -11-1       1
           11       -1-11       1
           12        111-1
           13        11-1       1
           14       -111       1
           15        1-11       1
           16        111       1
           17        1    -a     0       0
           18        1     a     0       0
           19       -1     0    -a       0
           20       -1     0     a       0
           21        100-1
           22        100       1
           23        1     a     a       1
           24        1     a     a       1
           25        1     a     a       1
           26        1     a     a       1
           27        1     a     a       1
           28        1     a     a       1
     B » Booth air rates—test at 2 levels:  1,-1
     C = Gun cleanliness—test at 5 levels:  a,1,0,-1,-a
     D = Restricted paint lines—test at 5 levels:  a,l,0,-l,-a
     Dummy = Dummy variable not expected to affect TE
                                  112

-------
                           Data Sheet 7
Run Number

    1
    2
    3
    4
    5
    6
    7
    8
    9
   10
   11
   12
   13
   14
   15
   16
   17
   18
   19
   20
   21
   22
   23
   24
   25
   26
   27
   28
   29
   30
   31
   32
   33
   34
                Air Atomized Electrostatic Test
          B
                                                       TE  Result
                                                       FP      VC
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
a
a
0
0
0
0
0
0
0
0
0
0
9





-1
1
-1
1
-1
1
-1
1
-1
1
-1
1
-1
1
-1
1
-1
-1
1
1
-1
-1
1
1
-1
-1
1
1
1
1
1
1
1
1
-1
1
-1
1
-1
1
-1
1
1
-1
1
-1
1
-1
1
-1
0
0
a
-a
0
0
0
0
0
0
0
0
a
a
a
a
a
a
-1
-1
1
1
1
1
-1
-1
-1
-1
1
1
1
1
-1
-1
0
0
0
0
-a
a
0
0
0
0
0
0






-1
-1
1
1
1
1
-1
-1
1
1
-1
-1
-1
-1
1
1
0
0
0
0
0
0
-1
1
0
0
0
0
1
1
1
1
1
1
-1
-1
1
1
-1
-1
1
1
-1
-1
1
1
-1
-1
1
1
0
0
0
0
0
0
0
0
-a
a
0
0
a
a
a
a
a
a
-1
1
-1
1
1
-1
1
-1
1
-1
1
-1
-1
1
-1
1
0
0
0
0
0
0
0
0
0
0
-1
1
1
1
1
1
1
1
Where:

     A
     B
     C
     D
     Z
     F
Restricted air lines—test at 5 levels:  a,l,0,-l,-a
Booth air rates—test at 2 levels:   1,-1
Gun cleanliness—test at 5 levels:   a,l,0,-l,-a
Restricted paint lines—test at 5 levels:   a,l,0,-l,-a
Fan air—test at 3 levels:  1,0,-1
Voltage—test at 5 levels:  a,l,0,-l,-a
      G  =  Electrode position—test  at  3  levels:  1,0f-l
                                   113

-------
                            APPENDIX B
              QUALITY ASSURANCE/QUALITY CONTROL PLAN
     SENSITIVITY STUDIES ON THE EFFECTS ON TRANSFER EFFICIENCY
   OF IMPROPERLY MAINTAINED OR OPERATED SPRAY PAINTING EQUIPMENT
                       CENTEC Corporation
                    Reston,  Virginia  22090
                 CONTRACT NO. 68-03-1721, Task 1
          INDUSTRIAL ENVIRONMENTAL RESEARCH LABORATORY
               OFFICE OF RESEARCH AND DEVELOPMENT
              U.S. ENVIRONMENTAL PROTECTION AGENCY
                   CINCINNATI, OHIO  45268
APPROVAL SIGNATURES:
Charles H. Darvin
EPA Project Officer
(Original  signed  by Guy Sims)
Guy F. Slines
EPA Quality Assurance Officer
 Edward H.  Comfort
 Quality Assurance Officer
Kerri C. Kennedy
CENTEC Sr. Project Engineer
                         JANUARY  19AA
                                 114

-------
                       TABLE OP CONTENTS
                                                            Paqe
SECTION 1    INTRODUCTION  	   B-3

SECTION 2    PROJECT DESCRIPTION	   B-4

SECTION 3    PROJECT ORGANIZATION AND RESPONSIBILITY  . .   B-6

SECTION 4    QA OBJECTIVES FOR MEASUREMENT DATA IN
             TERMS OP PRECISION, ACCURACY, COMPLETE-
             NESS, REPRESENTATIVENESS AND
             COMPARABILITY	   B-8

SECTION 5    SAMPLING PROCEDURE 	   B-10

SECTION 6    SAMPLE CUSTODY 	   B-11

SECTION 7    CALIBRATION PROCEDURES, ANALYTICAL
             PROCEDURES AND FREQUENCY	   B-12

SECTION 8    DATA REDUCTION, VALIDATION, AND REPORTING  .   B-13

SECTION 9    INTERNAL QUALITY CONTROL CHECKS	   B-15

SECTION 10   PERFORMANCE OP SYSTEM AUDITS	   B-16

SECTION 11   PREVENTATIVE MAINTENANCE	   B-17

SECTION 12   SPECIFIC ROUTINE PROCEDURES TO ASSESS
             DATA PRECISION, ACCURACY AND
             COMPLETENESS	   B-1 8

SECTION 13   CORRECTIVE ACTION  	   B-21

SECTION 14   QUALITY ASSURANCE REPORTS TO MANAGEMENT  . .   B-22
                                   J15

-------
                                         Sect ion	1	
                                         Revision No.    Original
                                         Date    December 1983
                                         Page    1    of    1


                            SECTION 1

                           INTRODUCTION


This quality assurance/quality control (OA/OC) plan assures

collection of high quality data and insures consistent quality

control measures for data developed under "Sensitivity Studies

on the Effects on Transfer Efficiency of Improperly Maintained

or Operated Spray Painting Equipment," Contract No. 68-03-1721.

Under this contract, CENTEC Corporation will be conducting tests

using a draft standardized method to determine the effect of

operating and maintenance parameters on transfer efficiency  (TE),
                                  116

-------
                                         Sect ion	2	
                                         Revision No.    Original
                                         Date    December 1983
                                         Page    1    of    2


                            SECTION 2

                      PROJECT DESCRIPTION


Sensitivity studies on the effects of TE of improperly maintained

or operated spray painting equipment will be conducted in this

test program.  A draft standardized TE method will be  used  for

all tests in this program.  The draft standard TE test method

consists of passing a prescribed set of preweighed targets  in

front of spray equipment  under rigidly controlled conditions in

an industrial laboratory  spray booth.  The cured painted targets

are weighed, and the original weight subtracted from the final

weight to obtain the net  dry solids deposited on the targets.

The net dry solids  is divided by the total solids sprayed at the

targets, which  is then multiplied by 100 percent to determine  TE.

A complete description of the draft standard TE test method is

contained in Appendix A of the Subtask Report for this contract.


Four types of spray equipment will bp tested during this program:

air atomized conventional (AAC), airless conventional  (ALC), air

atomized electrostatic  (AAE), and airless electrostatic  (ALE).

Each type of equipment  has an  individual experimental  design.

Five operation  and  maintenance  (OtM) factors have been selected

for testing on  conventional  spray equipment.  These factors

 include  booth air  rates,  atomizing air pressure,  fan air, paint

pressure, and cleanliness of  the  spray gun.  Qualitative factors

 (booth  air  rate and fan air)  will  be tested  at  a  minimum of

                                   117

-------
                                         Sect ion	2	
                                         Revision  No.   Original
                                         Date    December  1983
                                         Page    2   of     2

levels each, while quantitative factors (atomizing air, cleanli-
ness, and paint pressure)  will be tested at  five levels.   Levels
will be selected for testing based on an original  set-up  with a
good spray pattern.  For electrostatic guns  these  factors will
be tested as described, except two more factors, tip voltage and
electrode position, will also be tested.  Tip voltage will be
tested at five levels, while electrode position will be tested
at three.  Six replicates are provided for each equipment type.

The four experimental designs are planned to provide enough data
to support development of a response surface and regression
equations to describe the response surface.   A complete descrip-
tion of the experimental design  is included in the Subtask
Report for this contract.
Negotiations are  underway with r.n  industrial laboratory  to
arrange the test  program.  Testing is scheduled to begin in
February H34, and will last approximately one month.
                                    113

-------
                                         Secti on
                                         Revision No.
                                         Date    January
                                         Page""""" 1    or
                            SECTION  3
            PROJECT ORGANIZATION AND RESPONSIBILITY

This project  is administered  through CENTEC Corporation structure,
as shown in Figure B-1.   Day-to-day  test program activities will
be managed on-site by  a  CENTEC  Senior  Project Engineer in direct
contact with  CENTEC QA management  personnel.
At the test site,  the  CENTEC  engineer  is responsible for  imple-
menting QA throughout  the test  program.  The engineer conducts
onsite evaluations to  verify  the degree of  implementation,
assures that  appropriate QA records  are kept, provides QA
direction  to  the  laboratory staff, and reports  regularly  to  the
Project Manager on the status of QA.
The  Project  Manager,  Ed Comfort,  is the Quality Assurance
Officer.   He  continuously monitors the implementation of  QA  and
provides  feedback to the CENTEC engineer  onsite and to CENTEC
management.   QA records Jeep1-  by the engineer (onsite)  and by the
Quality Assurance Officer (offsite) serve as resources  for
preparing reports and documenting adherence to QA procedures
and  specifications.

-------
                                      CHAIRMAN AND
                                        PRESIDENT

                                      PAUL S. MINOR
                                        (Director)
            VICE PRESIDENT
                 AND
       CHIEF TECHNICAL OFFICER
          CHARLES S. MATHENY
       	(Director)	
 VICE PRESIDENT
  FINANCE AND
ADMINISTRATION
ROBERT D. SMITH
 VICE PRESIDENT
PROCESS SYSTEMS

CHARLES M. ROWAN
 VICE PRESIDENT
 CENTEC APPLIED
  TECHNOLOGIES
ROBERT SCHAFFER
   VICE PRESIDENT
MANAGEMENT SYSTEMS

    CURT GRINA
                                                QUALITY ASSURANCE
                                                     OFFICER

                                                   ED COMFORT
                                                 SENIOR PROJECT
                                                    ENGINEERS

                                                   K. KENNEDY
                                                   C. ROBERTS
      Figure B-l. Project Organization as Related to Corporate Structure

-------
                                         Sect ion	4
                                         Revision No.   Original
                                         Date    December 1983
                                         Page    1    of   2
                            SECTION  4

              QA OBJECTIVES FOR MEASUREMENT DATA IN
           TERMS OF  PRECISION, ACCURACY, COMPLETENESS,
               REPRESENTATIVENESS AND COMPARABILITY


For each manor measurement parameter specific QA objectives for

precision, accuracy,  and completeness are  required.  These

objectives are detailed  in Table B-1.


Care must be taken to assure  that all measurements are repre-

sentative of the media and conditions b<»inj measured.  Proven

techniques or methods are therefore  use3 for all measurements.


Data quality objectives  are based on accuracy and precision of

each measurement parameter, as established in Table B-1.  Data

integrity will be validated throuoh  a series of  inspections and

tests described  later in this plan.
                                    121

-------
             Table  B-1.   Spray Painting Transfer Efficiency Precision,
                           Accuracy and Completeness Objective
Itoaaurenent Paranetar
(Matted)
s^t
s Grounding
• Vbltaga
s units
s Distance-length
a Tine (Stopwatch.
tUer)
• tot Pilis Thickness
• Dry nim Thickness
s VlacoBity (Ford cup)
* teslstivity
• Pressure
s Temperature
• Linear Air Velocity
(rotating vem or
haatsd wire anenoej
s Density
S Mt% SollC*
s> Paint Stapling
Reference
Matted

IfflR Std 12-1972
AKSt/IEES Std 142-1972
ANSI C2
IEEE Std 4-1918
ASTM IS 380-76/
1BEB Std 26»-1976

{SM AST* 1200-70)
AS™ D-1212-79
AS™ 0-1005-51(1079)
AS« D 1200-70(1976)
AS™ D 2353-68(1978)



ACEIH Mecoenndad
Practice. Section 9*
ter)
AiHM D 1475-60(1980)
ASM D 2369-81
ASM 0 3925
Exuerlnental Conditions
Laboratory conditions
Laboratory conditions
Laboratory conditions
Laboratory conditions
Laboratory conditions
laboratory oondltiona
Laboratory conditions
Laboratory conditions



TMt conditions
In accordance with Mm



Precision
(Std. Deviation)
lab scale 0.01 g
plat, seals 5 g
™ "
0.05 HV
—
1/32 in
O.ls
0.265 •!!
(2%)
+0.1 mil
1.5s
0.1 MQ
5%
0.1«C
33 (3%)
«C.001g/W.
(1.5%)
—
Accuracy
lao scale *0.01 g
plat, seals «,5 g
^™ *
*0.1 W
^
1/64 In
0.2s
0.85 mil
2%
2s
0. 1 MU
Air atomlMd «0.5 kPa
Airless 0.5 kPs
0.1-C
±3%
0.002 gyW.
4.7%
—
CotRDletaneas
100%
1CO%
190%
100%
100%
100%
110%
100%
100%
100%
100%
100%

100%
100%
100%

s Condition in Container ASTM D 3011-1 — —
•Industrial \tentllatlon - A Manual oC
                                     Practice. American Oonfervnoe at GovemBntal Induatxial Hygenlats, 1972.

-------
                                         Sect ion	5	,
                                         Revision No.   Original
                                         Date    December 1983
                                         Page    1    of   1


                            SECTION 5

                        SAMPLING PROCEDURE


A description of the sampling procedure is provided in the

Subtask Report, Appendix A, Draft Standard Test Method for Spray

Painting Transfer Efficiency.  The draft standard test method

includes:

     •  A description  of the test method, including references
        to standard methods

     •  Figures illustrating specific operations

     •  Description of  sampling and test equipment

     •  Data sheets

     •  Other special  conditions and considerations in
        performing the  test

-------
                                         Section	§	
                                         Revision No.    Original
                                         Date    December 1983
                                         Page    1    of   1


                            SECTION 6

                          SAMPLE CUSTODY


Sample custody procedures are addressed in Draft Standard Test

Method for Spray Painting Transfer Efficiency, Subtask Report

Appendix A.  The CENTEC engineer and laooratory technician will

check and sign all data sheets.  The laboratory will retain all

weighed foils, as described in the draft test method,  until the

data analysis is complete.  CENTEC will retain all original data

sheets.
                                  1241

-------
                                         Sect i on	7	
                                         Revision No.   Original
                                         Date    December 1983
                                         Page	1	of   1


                            SECTION 7

 CALIBRATION PROCEDURES, ANALYTICAL PROCEDURES AND FREQUENCY


Calibration procedures, analytical procedures* end frequency

requirements are included in Draft Standard Test Method for

Spray Painting Transfer Efficiency, Subtask Report,  Appendix A.
                                  125

-------
                                         Sect icn	8	
                                         Revision No.    Original
                                         Date    December 1983
                                         Page    1    of   2
                            SECTION 8

            DATA REDUCTION, VALIDATION, AND REPORTING


8.1  GENERAL

Data will be collected at the test laboratory under the guidance

of a CENTEC engineer.  The data will be collected and documented

according to the requirements of the Draft Standard Test Method

for Spray Painting Transfer Efficiency (Subtask Report, Appendix

A).  Equations for reducing the data are also contained in the

draft standard test Method.  Figure B-2 shows the responsible

parties for each data validation and reduction step.


8.2  DATA REDUCTION, VALIDATION, AND REPORTING

Data reduction will be performed using standard statistical

practices as described in Draft Standard Test Method for Spray

Painting Transfer Efficiency, Subtask Report, Appendix A.  Any

data generated by test runs with known discrepancies in perform-

ance will be labeled as suspect for later evaluation.  Duplicate

data for all suspect runs will be obtained whenever resources

permit.


For each experimental design, the reduced data will be subjected

to a series of  t tests using studentized residuals to evaluate

outliers.  This evaluation will be performed onsite when 75 per-

cent of a test series is complete.  Outliers will be replaced by

duplicate runs as resources permit.  Any remaining outliers will

be eliminated from the data set where possible without rendering

the data set useless.

-------
         CENTEC ENGINEER ON-SITE
                                  CENTEC PROJECT MANAGER
DATA COLLECTION
 TE, SD, COV
CALCULATIONS
 OUTLIER
EVALUATION
 STATISTICAL
ANALYSIS AND
  MODELING
                          Figure B-2. Data Validation Resposibilities

-------
                                        Sect ion	9	
                                        Revision No.   Original
                                        Date    December 1983
                                        Page    1    of    1

                            SECTION  9
                 INTERNAL QUALITY CONTROL  CHECKS

Internal quality control checks are  incorporated into the
experimental design and Draft  Standard Test Method  for Spray
Painting Transfer Efficiency (Subf.sk Report,  Appendix A).
These checks include a battery of replicates  for each type  of
spray equipment to be tested.   Calibration requirements  also are
specified in the Subtask Report, Appendix  A.   All data  is
subjected to two inspections for error  (by the CENTEC engineer,
and by a laboratory representative), with  concurring signatures
required on each data sheet.
                                   128

-------
                                         Sect ion	10	
                                         Revision No.   Original
                                         Date    December 1983
                                         Page    1    of   1


                            SECTION 10

                    PERFORMANCE OF SYSTEM AUDITS


The performance of the TE tests will be monitored constantly as

described in Draft Standard Test Method for Spray Painting

Transfer Efficiency, Subtask Report, Appendix A.
                                    129

-------
                                         Sect ion	11	
                                         Revision No.   Original
                                         Date    December 1983
                                         Page    1    jt   1

                            SECTION 11
                    PREVENTAT1VE MAINTENANCE

Certain preventative maintenance (PM) procedures must be followed
to keep downtiine to a miniimjm.  Kost PM practices are recommended
by the manufacturer to the spray equipment user.  These practices
include keeping the spray equipment and spray area clean, handling
equipment carefully to avoid damage, and using appropriate
equipment for the given job.  These general practices must be
observed to prevent inadvertent deterioration of spray equipment
condition and to minimize downtime.

In addition to these PM practices, extra electrodes  and air caps
should be kept on hand.  Ample supplies for performing TE tests
should be available to avert, shortages.  These  include foil,
paint, tape, solvent, and others outlined  in the draft standard
TE test method.
                                   130

-------
                                         Sect ion	12
                                         Revision No.    Original
                                         Da t e    December 1983	
                                         Page    1    of   3
                            SECTION 12
      SPECIFIC ROUTINE PROCEDURES TO ASSESS DATA PRECISION,
                     ACCURACY AND COMPLETENESS


After the spray painting system is operational,  performance

audits will be conducted to assure continued acceptable  pre-

cision and accuracy during testing.  It is the nature of the

experimental design for this program that TE results cannot be

tested for outliers until three-quarters of a test  series is

complete.  To minimize the likelihood of obtaining  poor  TE

results prior to outlier analyses, performance audits are re-

quired twice daily for each major measurement contributing to TE:


     •  Net solids on target, g

     •  Conveyor speed, cm/*

     •  Paint weight fraction solids

     •  Paint mass flow rate, g/s

     •  Effective target width, cm


These measurements are subject to the precision, accuracy, and

completeness criteria in Table B-l.  They will be audited for

precison and accuracy at the beginning and completion of each

test day.  Periodic audits also may be conducted during  the test

day as deemed appropriate by either the laboratory technician or

CENTEC engineer on site.  Performance audit requirements are

detailed in Table B-2 and in the Draft Standard Test Method for

Spray Painting Transfer Efficiency.
                                   131

-------
                                         Sect ion	12	
                                         Revision No.    Original
                                         Date    December 1983
                                         Page    2    of   3


          TABLE B-2.  PERFORMANCE AUDIT REQUIREMENTS

  Measurement                                                  When
Parameter (units)            Performance Audit Method         Raquired


Net solids on target(g)    o Measure known control weight    A,B,C,D


Conveyor speed (CT/S)      o Blank run using electronic
                             timer                           A,B,C,D
                           o Chalk mark and stopwatch


Paint weight fraction      o Conduct duplicate analyses      A,B,C,D
  solids                     per ASTM 2369 at manufac-
                             turers recommended cure
                             schedule


Paint mass flow  rate  (g/s) o Spraying, using stopwatch       A,B,C,D
                             and scales


Effective target width     o Ruler or tape measure              C
A » Start of each day
B • At change of paint or spray equipment
C • As requested by  lab technician or eng.
D « End of each day
                                  132

-------
                                         Sect ion	12	
                                         Revision No.   Original
                                         Da t e    December 1983
                                         Pane    3    of   3
The precision and/or accuracy of the total measurement system will

be documented at least twice daily.  Problems identified by the

performance audit will be corrected before continuing with the

test program.


Completeness requirements are audited continuously and automati-

cally by the dual check off procedures required on each data

sheet in the draft standard test method.
                                   133

-------
                                         Sect ion	1J	
                                         Revision  No.   Original
                                         Date    December 1983
                                         Page    1    of    1


                            SECTION 13

                        CORRECTIVE ACTION


Performance audits are required twice daily for each major

measurement contributing to TE.  Should any measurement not  meet

the precision or accuracy requirements laid out in Table  B-l,

corrective action must be taken.  Corrective action includes

recalibration, repair, or replacement of the measurement  system

in question.  The CENTEC engineer on site is responsible  for

initiating the appropriate corrective action, with concurrence

from the participating required in writing in the  next QA report

to management.


Corrective action may also be taken to replace data identified

as erroneous by the required data outlier analysis.  The CENTEC

Project Manager is responsible for initiating corrective action

to replace outlier data.


Other corrective action may be taken at the request of onsite

CENTEC or  laboratory personnel whenever suspect or undocumented

conditions occur.  The CENTEC engineer  is responsible for all

such corrective actions.

-------
                                         Sect i on	14	
                                         Revision No.   Original
                                         Date    December 1983
                                         Page    1    of   l


                            SECTION  14

             QUALITY ASSURANCE REPORTS TO MANAGEMENT


The CENTEC engineer on  site will report daily via telephone to

the CENTEC Project Manager regarding the results of all perform-

ance audits, measurement  system accuracy, and measurement system

precision.  Significant OA problems  and recommended solutions

will be discussed.  Brief records  of these reports will be kept

for later inclusion in  the final test report OA section.
                                   135

-------
                            APPENDIX C

     AAC TEST  EQUIPMENT  AND  PAINT  SPECIFICATIONS


                            Data Sheet 1

                    Test Equipment Specifications

      Test Datei              Test No.:      Data  by/Checked by:
         -. 5...,w                 /M,'..,;         "^-f l*( r
A. Height Percent  Solids Measurement Equipment
   1.  Laboratory  Scales
       a. Manufacturer                         »reei««
       b. Model  NO.                             240-21
       C. Serial  NO.                            .-.574227
       d. Capacity,  g                          0-300/0-1000 CRAMS
       e. Rated  accuracy, g                    .PIG t»fff\atinr>/.\e,
   7.  Poil  Dish*s
       a. T>pe                                  Aluminum
       b. Size                                   5Bar round
   3.  Syringe
       a. Type                                   Class
       b. Capacity,  nL                          5
   4.  Solvent  Type                              Xvi. v
B. Conveyor  Speed Measurement Equipment
   1.  Rul»
       •. Type                                   conventional
       b. Graduations                           1/32*
   2.  Electronic Timer
       a. Type                                   Elactro-Heehanieal Digital
       b. Manufacturer
       c. Model No.
       d. Serial No.
       e. Rated accuracy, s
C. Mass Plow Measurement Equipment
   1.  Platform Scales
       a. Manufacturer                        y»«« now n«t«r.  mn-nmntion
       b. Model No.                            C12*T with DIP *t Indicator
       c. Serial No.                          .2425	
       d. Capacity,  kg                        o to IQOQ e/mn. .68 to  13 jtG/Min.
       e. Rated accuracy, g                   »i - .4% B»«I»M
   2.  Stopwatch
       a.  Manufacturer                        ^gnm Praeiaion Produeta. Inc.
       b.  Model No.                           _^	
       e.  Serial No.                          _=	
        d.  Rated accuracy, s                    .ei
 C. Target  Poil
    1.   Type                                   Alum. Alley 1145-
    2.   Nominal Thickness, ailo                  1.5
    3.   Temper                                 *«diua	
 E.  Wet Pi1» Measurement Equipment
        a.  Manufacturer                        Cardce 0-4 Mils
        b.  Model No.                           Precision Direct Readme.

 P.  Dry Film Measurement Equipment
        a.  Manufacturer                      DsFelske Corp.	
        b.  Model  NO.                          Poaiteeter 2000  .1 mils Accuracy


                                     136

-------
    6.3  Select coating.   The  eanc coating shall be used for all
         teats in  this  program.   Using Data Sheet 2,  document
         the paint characteristics.   Paint characteristics shall
         be documented  daily,  at  each addition of paint, and at
         other times as requested by the CENTEC engineer or GRACO
         representative.   Again*  check your information and sign
         the torn.
                              Data Sheet 2
                          Paint Specifications
Test Datei
                          Test  No. i
                          AfV (s
       Data by/Che Aed by:
       O-v    /  K'O^
   1. Paint Type
   2. Resin Typs
   3. Manufacturer
   4. Manufacturer's  Paint ID No.
   5. Lot No.
   «. Color
   7. Recommended Cure Schedule
   8. Viscosity  (uncut)
   9. Reducing Solvent
  10. Vol. of Solvent Put into
      Vol. Paint
  il. Viscosity  - Spray (cut)*
  12. Ht./Gallon -  Spray
  13. Ht. Solids -  Siray
  14. itosistlvity o.-  ' (inductance
 fr  f f*f,i *>'	if—ff /*. •»•> £'**
Black Enanal 
-------
                          Data  Sheet 3

      Paint Spray and  Peripheral  Equipment Specifications
Teat Datei
                             Teat  No.t

                             ft'V Lf
                                                 Data by/Chkd bys
A. Paint Supply  Tank
      Type
      Manufacturer
      Model No.
      Serial No.
      Rated Capacity,  gal

B. Paint Spray Equipment

   1. Type
   2. Manufacturer
    . Model No.
    . Serial No.
    . Rated Capacity,  cc/aUn
    . Air Cap
    . Fluid Tip
    . Needle

C.  aint Spray Booth
      Type
      Manufacturer
      Model No.
      Serial No.
    . Rated Capacity,  cfn

D.  onveyor
 E.
                                                    *o»
                                              BO76O5 (DKO. Ho.)
       Manufacturer
       Model  No.
       Serial Ho.

    orci*d Draft  Oven
       Type
       Manufacturer
       M— A ^ |  Ujk
       nCySV 1  1*0 •
       Serial No.

    aint  Heaters
                                              »«li«nc« Zltctrie Co.
       Manufacturer
       Model No.
       Serial No.
                                              1I»41 Cet»troll«r  1-3M1-A Own
                                                 n/a
                                      138

-------
                            APPENDIX D


     AAE  TEST  EQUIPMENT  AND PAINT SPECIFICATIONS


                            Data  Sheet 1

                    Test  Equipment Specifications

      Test Datei              Test No. t       Data  ^Checked  by:
                                 ~ - 2.3L
A. Height Percent  Solids  Measurement  Equipment
   1.  Laboratory  Scales
       a. Manufacturer                         pr»ci«»
       b. HoJel No.                            340-?1
       e. Serial  No.
       d. Capacity,  g                          o-ioo/o-iooo CBAMS
       e. Rated accuracy, g                    .oir. i^.oiutinn/.ir
       Foil Dishes
                                               Almdnia
       b. Size                                 Sflpa round
   3.  Syringe
       •• Type
       b. Capacity, mL
   4.  Solvent Type
B. Conveyor Speed  Measurement Equipment
   1.  Rule
       b. Graduations                          i/n-
   2.  Electronic Timer
       a. Type                                 EI«ctro-M«eh«nie«l Dioital
       b. Manufacturer                         »»rn«.
       b. Model No.                           CHAT with DIP itt
       e. Serial No.
       d. Capacity,  kg                        a tn 1000 c/mn. .KB «a 13 u/Mia.
       • . Rated accuracy, g                   »i - .4% •».•»<,»
   2.  Stopwatch
       a. Manufacturer                        ammu »y«p«««nn 9*04.,***. jj,c
       b. Model NO.                            -	
       e. Serial No.                           -	
       d. Rated accuracy,  s                    .01
D. Target Foil
   1.  Type                                  MUB. Mlov 1145-0
   2.  Nominal Thickness, mils
   3.  Temper                                mdiia
E.  Wet Film Measurement Equipment
       a. Manufacturer                        earteo o-4 nil*
       b. Model No.                           Fraction Diraet
F.   Dry Film Measurement Equipment
        a.   Manufacturer                      n»r«i»ko Corp.
        b.   Model No.                        ": »n«etor 2000  .1 »xl» Aeeuraev



                                       139

-------
    6.3  Select coating.  The same coating shall be used  for  all
         tcsti in thii program.  Using tat  Sheet  2, document
         the paint characteristics.  Paint characteristics shall
         be documented daily, at each addition of  paint,  ar.d  at
         other times as requested by the CEHTEC vnginser  or CRACO
         representative.  Again, check your information and sign
         the torn.
                             Data Sheet 2
                         Paint Specifications
Test Date t
 J
Test No.i
 MC-11
   1. Paint Type
   2. Resin Type
   3. Manufacturer
   4. Manufacturer's Paint  ID No.
   5. Lot No.
   6. Color
   7. RecoKBsndeJ Cure  Schedule
   S. Viscosity  (uncut)
   9. Reducing Solvent
  10. Vol. of Solvent Put Into
      Vol. Paint
  11. Viscosity - Spray (cutI*
  12. Ht./Gallon -  Sprey
  13. Wt. Solids -  Spray
  .i. Resistivity or Conductance
                                             Data  by/Checked by:
             •lack &«•»! 5PTito»e> CUD t  *P
              \/
                        (vol) solvent in
                        (vol1  paint
                               CUP
                           Ibt/oal
                yt.
                    nn.
 •Use  ASTM D-23SJ-68, ASTN D-1200-70. or ASTM D-3794 part  6.
                                         140

-------
                          Data Sheet  3

      Paint Spray  and  Peripheral Equipment Speeifieationa

                            Teat HO.i           Data by/Chkd byt
Test Datei
a. Paint Supply Tank
      Type
      Hanuf acturer
      Model Ho.
      Serial No.
      Rated Capacity,
gal
jCji.
B. Paint Spray  Equipment

   1. Type
   2. Manufacturer
   3. Model No.
   4. Serial  No.
   5. Rated Ctpacity, ec/min
   (. Air O?
   7. Fluid Tip
   8. Needle

C. Paint Spray  Booth
   1. Type
   2. Manufacturer
   3. Model No.
   4. Serial  No.
   5. Rated Capacity, cln

D. Conveyor
   1. Type
   2. Manufacturer
   3. Model No.
   4. Serial  Ho.

C. Forced  Draft Oven
   1. Type
   2. Manufacturer
   3. Model No.
   4. Serial  No.

r. Paint Heater*
    1. Type
    2. Manufacturer
    3. Model  No.
    4. Serial  No.
                             tiff.  J.lf\t7t3
                            Lit/ S
                            5 f - 7/.C
                                        ****>>
                       B0760S  (Dun. He.)
                       IUli*nc« El«etric Co.
                       m-^.fc y j Drive	
                                            O*«n
                                      14 i

-------
                       APPENDIX E

ALC TEST EQUIPMENT AND  PAINT  SPECIFICATIONS
  Test  Date:
   2 -ii-s-i
                        Data Sheet 1

                Test Equipment Specifications
                          Test No. I
                                                  by^Checked by:
           Model No.

 F.  Dry Film Measurement Equipment
        a.  Manufacturer
        b.  Model No.
                                            240-21
                                            Aluminum
                                            58mm round
                                            conventional
                                            1/32'
A. Height Percent Solids  Measurement  Equipment
   1.  Laboratory Scales
       a. Manufacturer
       b. Model Ko.
       c. Serial No.
       d. Capacity, g
       e. Rated accuracy, g
   2.  roil Dishes
       a. Type
       b. Size
   3.  Syringe
       a. Type
       b. Capacity, «L
   4.  Solvent Type
B. Conveyor Speed Measurement  Equipment
   1.  Rule
       a. Type
       b. Graduations
   2.  Electronic Timer
       a. Type
       b. Manufacturer
       c. Model No.
       d. Serial No.
       e. Rated accuracy, a
C. Mass  Plow  Measurement  Eqjipment
   1.  Platform Scales
       a. Manufacturer                        f
       b. Model Ko.
       c. Serial No.
       d. Capacity,  kg
       e. Rated accuracy, g                   *i - .«%
   2.  Stopwatch
       a. Manufacturer
       b. Model No.
       e. Serial No.
       d. Rated accuracy, s
D. Target roil
   1.  Type
   2.  Nominal Thickness, •  is
   3.  Temper
t.  Wet  Film  Measurement  Equipment
       a. Manufacturer
       b.
                                            El«etro-?larti«nieal Dieital
                                            Prccuion Scientific Co.
                                            69333
                                               flaw jtfT;
                                                          f1irrril>Ct ion
                                                with P10 Bt Indicator
                                           O *n 100O e/Htn. .68 to 13 KC/Mifl.
                                                                  Inc.
                                           -fll.
                                           »\um. fclloy 1K5-0
                                            1.5
                                           Madtuo	
                                           Carclco 0-4
                                    142
                                           Praciaien Direct R»«dlna
                                           OtFeliko Corp.
                                           Pooiteetor 2000  .1 miIt Accuracy

-------
    6.3  Select coating.  The  same  coating  shall  be  used  for all
         tests in this  program.   Uaing  Data Sheet 2,  document
         the paint characteristics.   Paint  characteristics ahall
         be documented  daily,  At  each addition of paint,  and at
         other tines as requested by  the CCNTEC engineer  or CRACO
         representative.  Again,  check  your information and sign
         the form.
                             Data Sheet 2
                          Paint Specifications
Test Datei
                          Test  No.i
 1.  Paint Type
 2.  Resin Type
 3.  Manufacturer
 4.  Manufacturer's Paint  ID  No.
 5.  Lot No.
 6.  Color
 7.  Recommended Cure Schedule
 8.  Viscosity (uncut)
 9.  Reducing Solvent
10.  Vol. of Solvent Put  into
    Vol. Paint
11.  Viscosity - Sprsy  (cut)*
12.  Wt./Gallon -  Spray
13.  Mt. Solids -  Spray
14.  Rtaistivity or Conductance
                                          Dsta by/Checked  by:
                                      •lack Ena»»l (Craco •077-001)
                                                 (.'A//763  *-
                                      M»li«nc«
                                      210-3159
                                       I: H 12
                                                     4. -*»-
                                      Black
                                             min. 9  3»9 »f
                                      MilUOVO 66 S«C.  I2ZAHN » 77'P
                                          sec.i    Ford Cup •
                                                 (vol) solvent in
                                                 (vol) paint
                                                    Ibs/aal
 •Use  ASTM D-23S3-68, AST* D-l200-70, or ASTM 0-3794 part 6.
                                       143

-------
                    Data  Sheet 3

Paint Spray and  Peripheral Equipment Specifications
Teat Datet                   Teat No.i
                                            Data  by/Child byi
A. Paint  Supply Tank

   1  Type
      Manufacturer
      Model  No.
      Serial No.
      Rated  Capacity, gal

B. Paint  Spray equipment

   1. Type
   2. Manufacturer
   I. Model  No.
   4. serial No.
   S. Rated  Capacity, cc/aun
   6. Air Cap
   7. Fluid  Tip
   8. Needle

C. Paint  Spray Booth
   1. Type
   2. Manufacturer
   1. Model  No.
   4. Serial No.
   5. Rated  Capacity, cfn

0. Conveyor
   1. Type
   2. Manufacturer
   3. Model  No.
   4. serial NO.

B. Forced Draft Ov«n
   1. Type
   2. Manufacturer
   3. Model  No.
   4. Serial No.

F. Paint  Heatera
   1. Type
   2. Manufacturer
   3. Model  No.
   4. serial No.
                                        Dvt.jtsrnpicll.or IUt»r Wash
                                        B07605 CPwo. No.)
                                        OV»rh«»d
                                        »>li«ne« Electric Co.
                                        Mltmjtk V.S Driva
                                              Air C«« Ovtn
                                              SvBt»m« Connanv


                                           .   .-ntroller 1-2341-* Or n
                                           n/a
                                           n/a
                                           n/a
                               144

-------
                           APPENDIX F

      ALE TEST EQUIPMENT AMD  PAINT  SPECIFICATIONS
                            Data Sheet 1

                    Te«t Equipment Specifications

      Test Dstei
          *•                                      x
A. Height Percent  Solids Measurement Equipment
   1.  Laboratory  Scales
       a. Manufacturer                         PT«.
       b. Model  No.                             J4p. 31
       c. Serial  No.
       d. Capacity,  g
       e. Rated  accuracy,  g                    .me p..Biu^M/.ie »..niut<,»
   2.  Foil  Dishis
       a. Type                                  fcluainuBi
       b. Site                                  iBm round
   3.  Syringe
       a. Type
       b. Capacity,  mL                        _S
   4.  Solvent  Type                            _J
B. Conveyor Speed  Meaaurement Equipment
   1.  Rule
          Type                                  e
          Graduations
   2.
C. Haaa
   1.
        lectronic  Timer
          Type                                  Il»ctro-H»ehanic«l Digital
          Manufacturer                          frtemcn Scientific Co.
          Model No.                             1*330
          Serial No.                             -
          Rated accuracy,  a                    .1 »te.
        flow Measurement  Equlpnent
        letters Scales
          Manufacture!                        iu«« MB»  M»»-«T.  mrl.»-^n~.
          Model No.                           P?*f yritfc C1° *"
          Serial No.                          Mas	
          Capacity,  kg                        fl_t
          Rated accuracy,  g                  «i	.
   2.   topvateh
          Haimia. ACtUiTtteT                        ^^gQ^*B^^2*B£n^^£f^^Q£££^£tl) t XllCe
          Model No.                             -
          Serial No.                            -	
          Rated accuracy,  a                   .01
D. Target r>! 1
   1.  Type                                   Mi». Alloy 1145-C
   2.  NO»I.-» Thickness,  mils                 i.i	
   3.  Temper                                Mtdiua
t.  Wet rilm Meaaurement Equipment
       a.  Manufacturer                       cardce 0-4 MI•
       b.  Model NO.                          »r«ei«ioq Pjrjrt P"J;'"

 r.   Dry rilin Measurement Equipment
        a.  Manufacturer                     Dtr«i«ke Corp.
        b.  Model No.                         >o»it«ctor 2000  .1 «il» Accuracy


                                         145

-------
    6.3  Select coating.  The same coating  shall  be used for all
         tests in this progran.  Using  Data Sheet 3,  docunent
         the paint characteristics.   Paint  characteristics shall
         be documented daily, at each addition  of paint, and at
         other times as requested by  the  CENTEC engineer or GRACO
         representative.  Again, check  your Information and sign
         the torn.
                             Data Sheet 2
                         Paint Specification!
Test Datei
                         Test No.i
                         &iP 7-3
       Beta by/Cheeked byt
   1. Paint Type
   2. Resin Type
   3. Manufacturer
   4. Manufacturer's Paint 10 No.
   5. Lot No.
   6. Color
   7. Recommended Cure Schedule
   I. Viscosity (uncut)
   9. Reducing Solvent
  10. Vol. of Solvent Put into
      Vol. Paint
  11. Viecoeity - Spray  (cut)*
  12. Mt./Gallon - Spray
  13. *t. Solids - Spray
  14. Resistivity or Conductance
                                      »l«cfc Enaaal (Grace 1077-0011
Alkyd •«••
mtliance
21C-31S3
•lack
       min. a
       0* MC. I2ZM1N ff 71
    sec.t    Pord Cup •
           (voll eel vent In
                             . Itt't -t"
           (vol) paint
»ee.>
             Ford Cup f  *r
     .•',-o
•Use ASTH D-23S3-68, ASTN  0-1200-70,  or ASTN P-3TV4 pert t.
                                         146

-------
                         Data Sheet 3

      Paint Sorav and Peripheral Equipment Specifications
Test Date:
Teat .No. t
J£L_
                                               Data by/Child Oyi
A. Paint Supply Tank
      Type
      Manufacturer
      Model No.
      Serial No.
      Rated Capacity, gal
                                                 •aflfe
                                                  «3gP
B. Paint Spray Equipment

   1. Type
   2. Manufacturer
   1. Model No.
   4. Serial No.
   5. Hated Capacity, ce/min
   7. Fluid Tip
   B. Needle

C. Paint Fpray Booth
   ». Type
   2. Manufacturer
   1. Model No.
   4. Serial No.
   5. Rated Capacity, eta

D. Conveyor
   1. Type
   2. Manufacturer
   ). Model He.
   4. Serial No.

E. Forced Draft Oven
   1. Type
   2. Manufacturer
   >. Model No.
   4. Serial No.

F. Paint >i.ater»
   1. Type
   2. Manufacturer
   3. Model No.
   4. serial No.
                                         XP"~
                                                          jC^^U^o
                                                ^^B-^arAwT
                                           •O7605 (Bug. Bo.)
                                            O»«rh«»d
                                                    laetric Co.
                                                 V.S t>riv«  	
                                            rare.
                                              •ft
                                     147

-------
                                APPENDIX G
                        Glossary of Statistical  Terms
Regression
  The procedure of fitting a model to a set of data using the method of
  least squares.  The product is a prediction equation for predicting a
  f""nendent response as a function of independent "input" variables.
Residuals
  The error in fit of a regression equation.  The residual is the difference
  between the observed response and a predicted response from the regression
  model.
t-tests

  t-tests are used in the present context to test the hypothesis that a
  regression coefficient is zero.  The t-statistic Is a ratio

                             regression coefficient	
                      C " standard error of coefficient
  Small values of t are evidence of a  coefficient that does not differ
  significantly from zero.
 F-tests

   F-tests are used  in a manner very similar to the  t-tests.  For a specific
   regression coefficient, and thus for a particular variable, the F-
   statistlc represents the  ratio of the variance explained by the variable
   being  tested  to  the variance attributed  to experimental error.
                                        148

-------
Significance level (p value)

  The significance level  is used in Che context of significance testing.
  If a regression coefficient (or a model variable) is significant at the
  0.02 level, the value 0.02 is the probability of obtaining a t-statistic
  (or F) as large as that observed, when in fact the model variable, plays
  absolutely no role in the system.  In other words, a p valuo is the prob-
  ability of obtaining such information due to chance alone.  Clearly, a
  small p value is evidence of a strong model term or variable.
Standard Deviation
  A standard deviation is a measure of spread in a statistical distribution
  or a set of data.  Given x., x_, ..., x , observations in a set of data
                            ii        n
  and x, the mean, the sample standard deviation is given by
R?- (Coefficient of Determination)

  The coefficient of determination R2 is a measure of quality of fit of
  fitted model.  The statistic R2 is defined as

                R2 o variation tn response explained by model
                          variation in response observed
                                        149

-------
Confidence Limits on Mean Response

  In a regression context, Che 95% confidence limits on the mean response
  represent "bounds" around the fitted regression that are defined such
  that
       "we are 95% confident that the mean response, at rhe data
       locations in question, is covered by the bounds."
Lack of Fi'. Test

  The lack of fit test is an F-test for ascertaining whether or not a
  fitted model is adequate.  The test essentially tests for the signifi-
  cance of higher order terms in the  egression.  If the F-statistic is
  nonsignificant, the conclusion is that there is no evidence that a more
  complicated model would improve the regression.
Dummy Variables

  The use of dummy variables is a standard way of accommodating "categories"
  in a regression situation that also contains the ordinary continuous
  variables.

-------