United States EPA-600/2-88-Q26a
Environmental Protection
A8encv April 1988
<&EPA Research and
Development
DEVELOPMENT OF PROPOSED
STANDARD TEST METHOD FOR
SPRAY PAINTING TRANSFER EFFICIENCY
Volume I. Laboratory Development
Prepared for
Office of, Air Quality Planning and Standards'
Prepared by
Air and Energy Engineering Research
Laboratory
Research Triangle Park NC 27711
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution. This work
provides the new or improved technology required for the control and treatment
of pollution sources to meet environmental quality standards.
EPA REVIEW NOTICE
This report has been reviewed by the U.S. Environmental Protection Agency, and
approved for publication. Approval does not signify that the contents necessarily
reflect the views and policy of the Agency, nor does mention of trade names or
commercial products constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-88-026a
April 1988
DEVELOPMENT OF
PROPOSED STANDARD TEST METHOD
FOR
SPRAY PAINTING TRANSFER EFFICIENCY
VOLUME I. LABORATORY DEVELOPMENT
BY
K. C. KENNEDY
CENTEC CORPORATION
Reston, Virginia 22090
EPA Contract Number 68-03-1721, Task 2
EPA Project Officer
Charles H. Darvin
Air and Energy Engineering Research Laboratory
Research Triangle Park, NC 27711
Prepared for:
U. S. ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF RESEARCH AND DEVELOPMENT
WASHINGTON, D. C. 20460
-------
ABSTRACT
This research proqram was initiated with the overall ob-
jective of developing a standardized spray painting transfer
efficiency determination methodology. A Steering Committee made
up of industry and EPA representatives was assembled to provide
consultation and guidance for the TE program.
Research into existing TE methods and an exhaustive litera-
ture search were conducted. TE methodologies identified by this
research were compared and evaluated. The best characteristics
of these methods were used in developing the standardized
laboratory TE method.
In initial tests the standardized laboratory method
comprised three major equipment types, two paint types, and
specially designed, spray targets. Paint was applied to the
targets under rigidly specified conditions. The amount of
solids deposited on the target was divided by the net solids
sprayed at the target to arrive at transfer efficiency-
According to ASTM 691-79, the first requirement for the
"existence of a valid, well-written test method [is that the
test method] has been developed in one or more competent
laboratories and has been subjected to a screening procedure or to
ruggedness testing." To fulfill this requirement, the test
method was tested at PPG Industries and at Ransburg Electro-
static Equipment. In each laboratory the TE results were
tightly grouped, exhibiting high precision. The standard
deviations were 2.5 and 1.8 for PPG and Ransburg, respectively.
The differences between TE's for the two tests indicated
that some important factors may not have been adequately
controlled. A third test, performed at Nordson Corporation, was
designed to intentionally vary those factors suspected of not
having been adequately controlled in the two previous tests; the
objective of this design was to see if those factors were
responsible for the variation in TE's. The results of the third
test were evaluated for their relative impact on TE. Six
factors were determined to have a significant effect on TE for
at least one spray gun type. The effect of these factors was
enough to have been responsible for the differences between the
first two tests. Recommendations are made in Section 8 to
revise the test plan to properly specify all significant
factors.
11
-------
CONTENTS
Abstract in
Figures v
Tables .Y1.
Abbreviations and Unit Conversions v]1
Acknowledgments ix
1. Introduction 1
2. Literature Search 3
3. Steering Committee 4
4. Transfer Efficiency Test Methods . 6
Industrial Test Methods 6
Standardized EPA TE Test Method 6
Improvements to Test Procedure 12
Draft Transfer Efficiency Test Procedure . 14
5. Phase I Laboratory Test 35
Facilities 35
Description of Paints 35.
Mass Flow Comparison Test 40
Test Parameters 42
Phase I Test Results 49
Transfer Efficiency Test Results ..... 53
Weight Percent Solids Test Results .... 61
Mass Flow Comparison Test Results 61
6. Phase II Laboratory Test 67
Facilities 67
Description of Paints 67
Test Parameters 70
Test Sequence 70
Foil Handling Procedures 73
QA/QC Procedures 73
Test Results 73
7. Test Comparison 91
Effect of Foil Wrap Technique - Vertical
Cylinder 91
Comparison of Interlaboratory Variances . . 93
Analysis of Variance for Paint,
Laboratory 95
Test Reproducibility 98
m
-------
CONTENTS
(Continued)
8. Third Laboratory Test „.....« 102
Facilities ....... 1°2
Description of Paint 102
QA/QC Procedures ....... • 102
Test Design 104
Test Parameters ..... 106
Test Sequence ..........«•••• 106
Solvent-Only Run .........«••• H7
Data Analysis .......... H7
9. Conclusions and Recommendations ........ 124
Bibliography ....... • 126
Appendix A - Worth Assessment Model of TE Test Methods . 130
Appendix B - Screening Procedure/Multiple Linear
Regression .................. 138
-------
FIGURES
Number Pag*
1 Set-up for paint supply equipment and platform
scales 20
2 Permissible methods for measuring conveyor speed . . 24
3 Target configurations for air atomized conventional
and electrostatic spray guns 28
4 Target configuration for high speed bell 29
5 Ransburg vertical cylinder wrapping technique ... 30
6 Flat panel foil attachment technique 31
7 Foil attachment techniques for vertical cylinder and
flat panel targets, PPG test, September 1982 ... 45
8 Comparison of vertical cylinder and semitubular
target configurations 57
9 Foil attachment techniques for vertical cylinder
(VC) and flat panel (FP) targets proposed for
Ransburg test 74
-------
TABLES
Number ?*21
1 Summary of Measurement Methodologies for Paint
Spray Transfer Efficiency .... 7
2 Summary of Paint Spray Transfer Efficiency Test
Methods Used by Industry ..... ... 9
3 Summary of Paint Spray and Peripheral Equipment
Specifications . - . . « 36
4 Test Equipment Specifications . 37
5 Measured Parameters and Rated Measurement Accuracy . 38
6 Summary of Recorded Paint Specifications ...... 39
7 Nomenclature for Spray Painting Transfer
Efficiency Tests 46
8 Transfer Efficiency Test Sequence at PPG
September 1982 ..... ....... 48
9 Summary of Transfer Efficiency Test Results .... 50
10 Results of Bartlett's Test for Evaluation of Pool-
ability of Variances at 95% Level of Confidence . 52
11 Summary of Equipment Operating Conditions for Air
Atomized Electrostatic and Conventional
Spray Guns .- 54
12 Summary of Equipment Operating Conditions for
High Speed Bell . . . . 55
13 Summary of TE Test Results for Air Atomized
Electrostatic Spray Equipment ..... 56
14 Summary of TE Test Results for Air Atomized
Conventional Spray Equipment ........... 59
15 Summary of TE Test Results for High Speed
Bell Coating Equipment .............. 60
16 Summary of Weight Solids Test Results ....... 62
17 Test Equipment Specifications for Mass Flow
Comparison Tests ........... 63
18 Paint Specification for Mass Flow Comparison Tests . 64
19 Equipment Specifications and Operating Conditions
for Mass Flow Comparison Test .......... 65
20 Mass Flow Comparison Test Results 66
21 Summary of Reported Paint Spray and Peripheral
Equipment Specifications for Ransburg Test .... 68
va
-------
TABLES (continued)
Number Pag6
22 Summary of Paint Properties 69
23 estimation of Required TE Test Sample Size for
Future Laboratory Tests 71
24 Test Matrix for Phase II Laboratory Tests of
TE Standard Test Method 72
25 Summary of TE Test Results - Ransburg Test 77
26 Results of Bartlett's Test for Evaluation of Pool-
ability of Variances at 95% Level of Confidence . 78
27 Transfer Efficiency Data, Ransburg Laboratory Test . 79
28 AAE Paint Spray and Peripheral Equipment
Specifications 81
29 AAC Paint Spray and Peripheral Equipment
Specifications ..... 84
30 HSB Paint Spray and Peripheral Equipment
Specifications 86
31 Transfer Efficiency Data 89
32 Tests to Determine Effect of Foil Attachment Method
AAC-67-VC: Tape Wrap'Method 92
33 Test to Determine Effect of Foil Attachment Method
AAC-67-VC: Crimp Method 92
34 Comparison of Variances for Tests at each Laboratory 94
35 Analysis of Variance - Transfer Efficiency 95
36 ANOVA Results for Paint and Laboratory 97
37 Transfer Efficiency Comparison 98
38 Paint Specifications • 103
39 AOAC Screening Test Design 105
40 Air Atomized Electrostatic Test Matrix and Results
(Nordson) 107
41 High Speed Bell Test Matrix and Results (Nordson) . 108
42 Air Atomized Conventional Test Matrix and Results
(Nordson) 109
43 Nordson Test Equipment Specifications 110
44 AAE Paint Spray and Peripheral Equipment
Specifications (Nordson) Ill
45 AAE Equipment Operating Conditions (Nordson) .... 112
46 HSB Equipment Operating Conditions (Nordson) .... 113
47 HSB Paint Spray and Peripheral Equipment
Specifications (Nordson) „ 114
48 AAC Equipment Operating Conditions (Nordson) .... 115
49 AAC Paint Spray and Peripheral Equipment
Specifications (Nordson) 116
50 Screening Experiment Analysis Tabulated Values of
Variance and Contrasts 118
51 ANOVA Results for Screening Test 119
52 Regression Models Derived from the Screening Tests . 121
53 Predicted Transfer Efficiency Results
(PPG and Ransburg) 123
vii
-------
LIST OF ABBREVIATIONS AND UNIT CONVERSIONS
ABBREVIATIONS
AOAC
ASTM
AAC
AAE
EPA
FP
HSB
QA/QC
TE
VC
VOC
Association of Official Analytical Chemists
American Society for Testing and Materials
air atomized conventional paint spray equipment
air atomized electrostatic paint spray equipment
United States Environmental Protection Agency
flat panel (target configuration)
high speed bell paint spray equipment
quality assurance/quality control
transfer efficiency
vertical cylinder (target configuration)
volatile organic compounds
UNIT CONVERSIONS
To go from
cm
g
kg
kg/L
kPa
L
m
m
m/s
mVs
rps
s
To
op
in
Ib
Ib
Ib/gal
gal
ft
mils
ft/min (fpm)
ftvmin
rpm
min
Multiply by
1,
2,
0,
2,
8,
0,
0,
3,
3,
196,
2118,
0,
60
8°C + 32
54
0022
204
328
145
264
281
937 x 10
86
8
017
kPa -14.7
-------
ACKNOWLEDGMENTS
The contributions of PPG Industries, Ransburg Electrostatic
Equipment, and Nordson Corporation are gratefully acknowledged.
These companies donated laboratory facilities, test equipment
and supplies, and provided technical support for the tests
described herein.
The Spray Painting Transfer Efficiency Steering Committee
has provided considerable help and constructive suggestions for
this project. This committee's contributions have been
invaluable.
IX
-------
SECTION 1
INTRODUCTION
Spray Painting Transfer Efficiency (TE). is a measurement of
how much paint actually coats a surface compared with the total
paint available to coat that surface. Historically, the spray
painting industry has developed its own methods for determining
TE. These methods vary from company to company and in their
objectives. The measurements are used by the coatings industry
to optimize on-line spraying; they are used by manufacturers to
develop more efficient spray equipment; and they are used to
minimize losses throughout the industry. More recently the need
to determine TE has a new aspect: TE can be used to quantify
emissions from these sources.
The U.S. Environmental Protection Agency (EPA) has been
charged by Congress via the Clean Air Act to quantify and
control emissions of volatile organic compounds (VOC). Acting
in response to this congressional edict, EPA has designed a
program to standardize TE test methods. Phase I of this effort
examined the TE methods currently used. Based on these methods,
a standardized procedure was proposed and tested at PPG Industries
in September 1982. Three types of spray painting equipment and
two types of paint were tested. The results are documented in
Section 5. The -PPG test established the test method as a viable laboratory
procedure, with standard deviation of 2.5.
A second laboratory test was performed in March 1983, at
Ransburg Corporation to further develop the proposed test method
(Phase II). The same paints and equipment configurations were
tested. The results of these tests are included in Section 6.
The standard deviation of the second test was 1.8.
The TE's varied considerably between laboratories. Analy-
sis of the test data from Phase I and Phase II showed some
factors were not adequately controlled from test to test. (The
data comparison and explanation is made in Section 7 of this
report.)
-------
A third laboratory test was conducted at Nordson Corporation
as a screening procedure to establish the relative effects of
previously uncontrolled factors. In the screening test each of
these factors was intentionally varied in such a way as to
facilitate evaluation of the effect of each factor. The evalua-
tion found six factors to have significant effects on TE for at
least one of the three spray gun types tested;
• Booth air rate (at target plane), fpm
• Shaping air, psig
• Atomizing air, psig
• Voltage at tip, kv
• Paint discharge technique
• Paint mass flow, g/s
Recommendations are made in Section 8 to revise the draft
test plan to properly specify all significant factors. The
draft TE test method thus developed is included in Section 4.
-------
SECTION 2
LITERATURE SEARCH
An extensive literature search was performed to document
the state of the art in TE determination. Considerable infor-
mation was collected on current TE methodologies, available
equipment, and necessary measurements. This information was
collected from vendors, manufacturers, journals, DIALOG (on-line
search in Compendex), EPA Publications Bibliography, and others.
A list of the most useful references is in the Bibliography.
Considerable research into current TE methods was required
in developing the standardized laboratory method presented in
Section 4. As part of developing the test procedure, appro-
priate paint formulations and equipment configurations were
evaluated. Personal visits were made to General Motors,
Ransburg, Nordson, and PPG.
Some of the people contacted during the literature search
were selected to the Steering Committee for this project. The
members of the Steering Committee represent divergent experience
and interest in the.TE project. Their names and affiliations
appear in Section 3.
-------
SECTION 3
STEERING COMMITTEE
CF.NTEC assembled a Steering Committee made up of industry
and EPA representatives to provide consultation and guidance for
the TE program. T"he industrial members of this committee were
unpaid and contributed their time to the project at their own expense
The Steering Committee was responsible for reviewing the
technical aspects and results from this project. The participa-
tion of a diverse Steering Committee assures the interest of
each group was recognized within the technical objectives of this
project.
The following people are members of the Spray Painting
Transfer Efficiency Steering Committee:
Charles H. Darvin
Physical Scientist
Industrial Processes Branch
U.S. EPA, Air and Energy Engineering
Research Laboratory
Research Triangle Park, NC 27711
David I. Salman
U. S. EPA, OAQPS, MD-13
Research Triangle Park, NC 27711
Charles M. Rowan, Jr.
CENTEC Corporation
11260 Roger Bacon Drive
Reston, VA 22090
Gregory A. Conner
Davidson Rubber Division
Ex-Cell-O Corporation
Industrial Park
Dover, NH 03820
John Lipscomb
Milwaukee Solvents &
Chemicals Corp.
P- O.Pox 444
Butler, WI 53007
EPA Project Officer
Representing EPA
Representing
Project Contractor
Representing Industry,
Coatings Chemist
Representing the
Chemical Coaters
Association
-------
Albert Hungerford
Butler Manufacturing Co.
1020 South Henderson St.
Galesburg, IL 61401
Mike Tersillo
Project Leader
Applications Group
PPG Industries, Inc.
R&D Center, P- O. Box 9
Allison Park, PA 15101
Steven J. Gunsel
Nordson Corpation
555 Jackson Street
P. O. Box 151
Amherst, OH 44001
Robert Lukes
General Electric Co.
Bldg. 35, Room 1117
Appliance Park
Louisville, KY 40225
Eugene A. Praschan
Engineering Group Manager
Fisher Body
Division of General Motors Corp.
30001 Van Dyke
Warren, MI 48909
Al Chasan
Coatings, Paints & Preservation
Branch, Code 2841
Naval Ship Research
and Development Center
Annapolis, MD 21402
E. W. Drum
Director. Environmental Affairs
Ransburg Electrostatic Equipment
3939 West 56th Street
Indianapolis, IN 46254
Representing Industry,
Paint Applicator
Representing Industry,
Paint Manufacturer
Representing Industry,
Spray Equipment
Manufacturer
Representing Industry,
Paint Applicator
Representing Industry/
Paint Applicator
Representing the Navy,
High Performance
Representing Industry,
Spray Equipment
Manufacturer
-------
SECTION 4
TRANSFER EFFICIENCY TEST METHODS
INDUSTRIAL TEST METHODS
A number of TE test methods are currently used by industry.
In all of these test methods, three basic measurements are
necessary:
• The amount of solids in the paint
• The amount of paint sprayed
• The amount of solids deposited on the target
These amounts may be determined by weight or by volume.
Typical methods are shown in Table 1. The paint may be
deposited on a target, shim, or work piece. Some of the test
methods are strictly laboratory procedures, while others are for
production line TE determination.
CENTEC considered over a dozen test methods in developing
the standardized EPA test method. A summary of paint spray TE
test methods considered and comments on these methods are
presented in Table 2.
STANDARDIZED EPA TE TEST METHOD
The EPA test procedure was developed to have the best
qualities of previous methods. Since it was desirable to have a
controlled standardized procedure, production line tests were
ruled out. Production line tests are usually made at production
conveyor speeds using foil-coated dummy targets shaped like the
product to be coated. These characteristics are all site
specif ic.
The EPA test procedure was developed for ease of per-
formance in standard industrial spray paint laboratories. For
verification tests, three major types of equipment (air atomized
conventional, air-atomized electrostatic, and high speed bell)
and two types of paint (65 and 52 weight percent solids) were
selected.
-------
TABLE 1. SUMMARY OF MEASUREMENT METHODOLOGIES FOR PAINT SPRAY TRANSFER EFFICIENCY
Paint Solids Content (I)
By Weight
I. Measure VOC In nonwaterborne
paints per AS7M D-2369-81.
2. Measure VOC and water In
waterborne paints per ASTM
D-2369-81 and ASTM D-3792-79
(or D-40I7-8I).
3. Measure density per ASTM
D-1475-80.
4. Measure weight solids per
ASTM 0-1644-81. (2)
Paint Spray Flow Rate
By Weight
1. Gun nozzle capture (bucket and
stopwatch).
2. Mass flow meter (Micro Motion
type) and time clock.
3. Load cell or scales for paint
mix tank.
4. Dipstick on or other level mea-
suring device on paint mix tank
(convert by density to mass flow).
Paint Solids on Target
By Weight
A. Where target Is product
I. Target weight change (load cell
or scales)
2. Foil on target weight change
(scales).
3. Shlm(s) on target weight change
(scales and weight conversion
via area ratio)
4. Target solvent stripping, dis-
tillation and weighing.
B. Where target Is simulated shape
(flat plate, vertical cylinders,
etc.).
I. Target weight change (load cell
or scales).
2. Foil on target weight change
(scales).
3. Shim on target weight change
(scales and weight conversion
via area ratio).
-------
TABLE 1 (continued)
Paint Soilds Content (!)
By Volume
Paint Spray Flow Rate
By Volume
Paint Solids on Target
By Volume
I. Determine volume fraction of
paint solids by calculation
using manufacturer's formu-
lation.
2, Measure volume solids per
ASTM 0-2697-79.
I. Dipstick on paint mix tank.
2. Paint Inventory change.
A. Where target Is product
1. Target coating thickness and
area.
2. Foil on target thickness area.
3. Shlm(s) on target thickness and
area (convert via area ratio).
CD
Where target Is simulated shape
1, Target coating thickness and
area.
2. Foil on target thickness and
area.
3. Shlm(s) on target thickness
and area (convert via area
ratio).
(1) Reference Method 24 — Revision to Appendix A of 40 CFR Part 60, Final Rule.
(2) Method generally used by Industry.
-------
TABLE 2. SUMMARY OF PAINT SPRAY TRANSFER EFFICIENCY
TEST METHODS USED BY INDUSTRY
Summary of Methods
Foil on Simulated Shaoe; Weight Change (Lab) -Measure
weight solids per ASTM 0-1644-81 (or similar method);
measure paint spray flow rate by capturing atomizer output
in anatomized state; weigh foil before and after drying
(use same curing schedule as for paint samples).
Shim on Product; Thickness and Area (Plant) - Use theo-
retical mileage data (ftvuncut gallon of paint) from
paint supplier; measure paint spray flow rate by cap-
turing atomizer output in anatomized state; measure shim
dry film thickness and overall product (part) area
using GE magnetic thickness gage and flexible tape.
3.
Product Coating and Area (Plant) - Use theoretical
mileage data (ft^/uncut gallon of paint) from
paint supplier; measure paint spray flow rate by
capturing atomizer output In anatomized state;
measure actual product dry film thickness and
overalI product (part) area.
Paint Inventory/Product Coating Thickness and Area
(Plant) - Use production data (parts coated, surface
area, gallons of paint used); use theoretical mileage
data (ft^/uncut gallon of paint) from paint
supplier; use QC data for average film thickness.
Comments
Ability to accurately measure
paint spray flow is question-
able (industry experience with
calibrated mass flow meter In-
dicates that this method has
inherent inaccuracy); overall,
method looks good, e.g., dummy
targets used before and after
main target to simulate pro-
duction situation. Foil and
simulated shape thermal inei—
tla — equivalent bake
schedule problem.
2. Potential variability in coat-
ing thickness means deter-
mination of overalI "average"
film thickness may have In-
herent repeatability problem
— multiple shims helps
minimize this problem;
question of reliability on
manufacturer's data for
theoretical paint mileage;
paint spray flow measurement
in question (see comment I);
questionable ability to mea-
sure film thickness accurate-
ly — repeated calibration
may help minimize this prob-
lem; ability to measure area
accurately Is a concern.
3. Same as Comment 2; additional
disadvantage is limitation
to ferromagnetic or aluminum
substrate (magnetic or eddy
current applications).
4. Question of reliability on
manufacturer's data for theo-
retical paint mileage; ability
to measure film thickness
accurately Is In question;
ability to measure area
accurately is a concern.
-------
TABLE 2 (continued
Summary of Methods
Foil on Simulated Shape; Weight Change (Lab) -
Measure weight solids per AS.TM D-1644-81;
measure paint spray flow rate by mass flow meter
(Micro Motion); weigh foil attached to flat panel
before and after drying (using same curing schedule
as for paint samples).
6. Foil on Simulated Shape; Weight Change (Lab) -
Measure weight solids per ASTM 0-1644-81;
measure paint spray flow rate by capturing
atomizer output In unatomlzed state; weigh
foil attached to flat panel before and after
drying — remove foil from rack prior to drying
to better simulate same bake schedule as paint
samples.
7. Foil on Simulated Shape; Weight Change (Lab) -
Same as Method 6 except that simulated shape
is 2" x 4" lumber.
8. Foil on Product; Weight Change (Lab) -Measure
weight solids per ASTM 0-1644-81; measure
paint use by weighing paint supply apparatus
before and after spraying; weigh foil before
and after drying — initial foil weight de-
termined by subtraction process for amount
of foil remaining on roll.
9. Foil on Product; Volume (Lab) -Measure paint
density per ASTM D 1475-60 (1980); measure
paint use by weighing apparatus before and
after spraying; measure film thickness on
foil with micrometer; tape measure for area;
measure volume fraction solids per ASTM D 2697-73.
10. Product Coating; Weight (Plant) -Measure
paint solids on target by solvent stripping,
distillation (evaporation, and weighing of
dried sol ids).
Comments
5. Only part of foil panel is coated
and thus lab test does not accurate-
ly simulate plant production situ-
ation — this shortcoming, however.
Is true of essentially a I I lab
tests; other problem is that of
thermal Inertia (foil & rack) and
equivalence of bake schedule.
6. Problem with accurate measurement
of paint spray flow (see Comment I);
good approach on paint drying
procedure (target and sample).
7. Same advantages and disadvantages
as Comment 6.
8. Problem with bake schedule since
entire car body dried with foil
attached (equivalalence problem
with sample drying method);
hysteresis (repeatability) con-
cern with use of load cell.
9. Average film thickness determi-
nation requires multiple
(measurements to obtain good
statistical average); complexity
of product shape makes area
determination a tedious process;
load cell hysteresis problem.
10. Load cell hysteresis (repeatability)
problem; accuracy was supposedly
good on load cell (1,000 I b _+_ 0.02
Ib), however, as confirmed by good
correlation with lab TE weight
method (foil on body — refer to
Method 8).
10
-------
TABLE 2 (continued)
11.
Summary of Methods
Product Coating Thickness i Area (Plant) -
Measure paint volume solids per ASTM D 2697-73;
measure average film thickness with eddy
current Instrument; measure area with flexible
tape; measure paint use be dipstick in mix tank.
12.
Product Coating; Weight Change (Plant) -
Measure paint density per ASTM 0 1475-60
(1980); measure paint use by dipstick In
mix tank; measure weight gain of car body
by load cell; measure paint weight solids
per ASTM D 1644-81.
Comments
11. Average film thickness determined
by numerous measurements; car
body divided Into two general area
categories to help account for
for variances in film thickness
for low appearance (large variation
in film thickness) and high appear-
ance (small variation In film thick-
ness) areas.
12. Incomplete method, however, concept
may have some merit for small parts
applicatlons.
13. Same as Comment 4.
11
-------
Two target configurations were selected for the standard-
ized test. A set of four flat panel targets was selected to
typify TE's for large, relatively flat industrial targets. A
set of four vertical cylinder targets was selected to typify
TE's for coating smaller, more intricate targets. These
targets, mounted in a prescribed configuration, constitute the
test panels for the EPA draft TE laboratory method.
The following test procedure is similar to the one used for
the two laboratory tests described in Sections 5 and 6, but
reflects some improvements made to the procedure as a result of
subsequent test experience.
IMPROVEMENTS TO TEST PROCEDURE
As a result of actual test experience, several recommenda-
tions can be made to clarify and standardize the test procedure.
First, it is recommended that the booth air flow rate be stand-
ardized. Since most booth exhaust fans are not rate-adjustable,
booths with O.bl-m/s (100-ft/min) linear flow rates should be
selected for testing. This is a standard booth size that should
be available to testers. The. exhaust fans should be used at all
times to keep air-borne solvent and paint levels at a minimum.
Second, the test procedure must be modified to provide
identical paint delivery to the targets. The paint flow should
be initiated from the same position (relative to the targets)
from test to test. A consistent starting point and delivery is
expected to lead to more consistent TE's between laboratories,
especially for electrostatic equipment. It is recommended to
use the timing marks at the first and last scavengers as paint
start/stop marks (refer to the following section, "Draft
Transfer Efficiency Test Procedure"). The rationale for this
recommendation is in Section 7.
Third, the electrostatic equipment voltages must be con-
sistent between laboratories. Fixed power supplies may vary by
30 percent in output voltage, and cable, gun, and paint charac-
teristics vary considerably between laboratories. Variable
power supplies will be required to keep equivalent voltages on
subsequent tests.
Fourth, calibration of all pressure gages involved before
the test is advisable. Gages on the paint supply tank, paint at
the spray gun, atomizing (or turbine) air pressure, and shaping
air pressure should be calibrated before the test. The
requirement assures equivalent operating pressures between
laboratories.
Fifth, spray guns should be equipped with new fluid tips at
the beginning of the test.
12
-------
The test procedure should provide more guidance in accept-
able techniques for determining paint characteristics. The
recommended technique must be readily available to labora-
tories. Consistent documentation will avert confusion during
tne test and uncertainty when comparing results.
Finally, there has been some speculation about the effect
of having other grounded equipment near the target assembly.
Although this effect has not been directly observed, it seems
prudent to keep all grounded objects away from the targets. A
minimum 3 meters (10 feet) is recommended as the distance to
walls or other grounds.
13
-------
DRAFT TRANSFER EFFICIENCY TEST PROCEDURE1
Equipment Calibration
Perform calibration of the platform scale once per week or each
time that it is moved and leveled, whichever occurs more fre-
quently. Perform calibration of the laboratory scale once every
3 months. Calibrate all pressure gages.
Test Procedure
A. Select test equipment. Using Data Sheet 1, document
the test equipment specifications. Be sure to check the
information and sign the form.
1Many 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.
14
-------
Data Sheet 1
Test Equipment Specifications^
Test Date: Test No.: Data by/Checked by:
A. Vfeiyht 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. ~
15
-------
B. Select coating type to be used. Using Data Sheet 2,
document the paint characteristics. Again, check your
information and sign the form.
Test Date:
Data Sheet 2
Paint Specifications
Test No. s
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. @
sec. #
Ford Cup @
(vol) solvent in
(vol) paint
sec
Ford Cup @
Ibs/gal
*Use ASTM D-1200-70, "Viscosity of Paints, Varnishes, and Lacquers
by Ford Viscosity Cup." Viscosity may also be determined by
ASTM D-3794, Part 6 (Zahn Cup method) in_ addition _to the Ford Cup
measurements.
16
-------
C. Set up paint supply equipment 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.
17
-------
Data Sheet 3
Paint Spray and Peripheral Equipment Specifications
Test Date: Test 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. Paint Spray Booth
1. Type
20 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.
18
-------
D. For electrostatic spray equipment only, ground paint
supply equipment and platform scale per Figure 1.
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 106 Ohm.
E. Using a small glass jar with an airtight lid, take
paint grab sample from paint pot. Record test series
number on label of jar.
F. Measure weight solids from grab sample using syringe
weight difference technique as described in ASTM
D-2369-81. Document the cure oven bake schedule 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, and
periodically between tests.
19
-------
Arrangement A
Nonelectrostatzc Eauiomen
Paint
Suppl-_.
Tank
Arrangement B
Electrostatic Equipment
-Digital electronic
platform scale
Grounding Cable
•Electrical Insulation Block*
Digital electronic
Platform Scale
*31oc.k must be capable of preventing current flow from supply tank to
ground through the platfora scale.
Figure 1. Set-up for Paint Supply Equipment and
Platform Scales
20
-------
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.
21
-------
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."
22
-------
G. Set up the paint spray equipment. Using Data Sheet 3,
document specifications for the paint spray equipment
and spray booth used in this test.
H. 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 mea-
sured using timing marks (chalk marks) on the conveyor
in conjunction with a hand held stopwatch. Figure 2
shows tne 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.
23
-------
METHOD A
Target
Target
Target
^
B
r~
c
Electronic
"imer
A = Stationary photoelectric cell or limit switch
B = Stationary photoelectric cell or limit switch
C = Moving plate of known width
METHOD B
Known Distance
.onvevor
Stop Watch
E = Fixed timing mark
F = Moving timing mark
Figure 2. Permissible Methods for Measuring
Conveyor Speed
24
-------
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
8. Total Solids Sprayed
1. Paint Spray Flow 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 Marks, s
c. Speed, cm/s B2c
3. Total Effective Target
Width, cm* 15.24 B3
4. Total Solids Sprayed, g
(A3 x Bid x B3/B2c) B4
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.
25
-------
Data Sheet 6b
TE 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 #1 2 3 4 Total
Foil Weight Before Spraying, g:
Net Dry Solids, g:
Foil vfeight Before Spraying, g:
Net Dry Solids, g:
E. Transfer Efficiency (by weight)^
Flat Panel Target
Vertical Cylinder Target
1. Net Dry Solids x 100 = TE
Total Solids Sprayed
26
D. Vertical Cylinder Target
Foil Weight After Drying, y:
Foil * 1 2 3 4 Total
-------
I. Set up targets in accordance with Figure 3 or 4, 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. (See Section 5 for
nomenclature.) Weigh each foil strip and record value
on foil and on Data Sheet 6b.
J. Attach foils to the vertical cylinder and/or flat panel
targets as shown on Figure 5 or 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 106 ohms.
K. In accordance with Figure 3 or 4, attach shim stock
to scavenger in order to measure wet film thickness.
L. Adjust all equipment operating parameters, i.e., gun to
target distance, atomizing air press, paint pot pres-
sure, shaping air pressure, turbine air pressure, etc.,
to desired values. Record equipment operating parame-
ters on Data Sheet 4. NOTE: In accordance with Sec-
tion 9-7 of NFPA 33 for fixed electrostatic apparatus,
the gun to target distance shall be at least twice the
sparking distance.
M. Check spray gun condition. Install new fluid tip, air
cap, and needle.
N. 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.
O. Check conveyor clock, stopwatch, and platform scale to
ensure that all have been zeroed (reset) and that the
scales are in the tare mode.
P. 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.
Q. 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, and stopwatch readings
on Data Sheet 6a.
R. Measure wet film thickness on shim plate and record on
Data Sheet 4, line C-l.
27
-------
Target movement toward gun
Wet Film
Thickness
Plate
NJ
oo
Note:
lin.=
2.54cm.
Conveyor
12"
Foil No.
4321
Foil No.
3 2 1
f • • • 1
kl
K
0
1
o
V
c
o
>
o
if)
yl%MDia
aluminum
pipes
Stainless
St«el
^^
1 1
1 1
^
L
1
Vertical Cylinder (VC)
Target
6" 6
2- 2" 2"
Flat Panel (FP) Target
Figure 3 Target configuration for air atomized conventional
and electrostatic spray guns.
36"
-------
Target movement toward gun
conveyor
Foil Number
Foil Number
T
60"
\ / \ 4321 / \ 4 3 2 1 / /
I /
M
O
U'
c
O
£»
ID
U
W
t!"
r.
h
01
tr
C
OJ
^
ra
u
to
'1
^
ID
CP
c
a)
^
(d
u
CO
1
12"
6"
_
6"
6"
6"
6"
IV dia
aluminum
pipes
Stainless
\ 1
Steel
.
^^^
^^x^.
1
12"
1
6"
6-
i'
-'il
T
f."
r-«-
J'
6""
G"
6"
^ /
^
c>
Cn
C
O
flj
u
CO
8"
Vortical Cylinder Taryot
(VC)
wet film
thickness
plate
Flat Panel Target (FP)
Figure 4 . Target Configuration for High Speed Bell
-------
Leading edge
of foil
Vertical Cylinder
-.-.-ran
Hold edge: of foil in place against
cylinder while wrapping leading edge
arour.d cylinder.
Foil
V.'rap vertical cylinder targets with cylinders mounted on target bracket
(See Figure 3 and 4 j. wrap so the leading edge forras a seam away
fron the direction of sorav.
•Gric"
'"rap
'•••'rap
"Grio"
"Grip"
•Grin"
"Grip"
-_-. "Grip"
As leading edge overlaps starting
edge, solidly "grip" foil into place
by grasping foil-covered cylinder.
Secure foil on cylinder by gripping
the length of the cylinder. Foil will
have a uniformly wrinkled surface.
Figure 5. Ransburg Vertical Cylinder Wrapping Technique
30
-------
FLAT PANEL TARGET
Flat Panel Target
Foil-Ready for
attachment
Double-sided tape
Figure 6. Flat Panel Foil Attachment Technique
31
-------
S. Remove foils from targets and securely attach to oven
racks so all painted surfaces are exposed for uniform
drying. Spring clips or tacks may be used to mount wet
targets. 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. Five minutes is con-
sidered excessive flash time- Set oven timer per
recommended schedule.
T. 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.
U. Perform TE calculations using the average weight solids
for all the samples taken for the test run. Document
results on Data Sheet 6b.
V0 After completing a test series, perform statistical
analysis as indicated on Data Sheet 7. Document
results of analysis on Data Sheet 7.
W. Repeat above steps (A through S) for each test series.
X. After completing all 12 test series, transfer results
of each TE test to Data Sheet 8.
32
-------
Data Sheet 7
vfork Sheet for Calculating Standard Deviation
Test Series:
Calculation by:
Checked by:
Test Run No.
1
2
3
4
5
6
7
8
Totals:1 n =
TE|yi = B/n
s = [ C / (n-1)] °
TE, %
B
.5
(TE-TE,), %
•
=
=
=
(TE-TEtf)2, %2
C =
(1)
(2)
1. TEjj = mean value for transfer efficiency, %:
vdiere B = TEi + TE2 + ... TEn
and n = Total number of test runs for one
test series
s = Standard deviation, %
where C = [(TE^TE^)2 + (TE2-TEN)2 + ... +
33
-------
Data Sheet 8
Statistical Analysis Sunmary for TE Test Results
Prepared by:
Checked by;
Number of Standard Coefficient of
Test Series Test Runs TE,^, % (1) Deviation, % (2) Variation, % (3)
1. TEM = mean value for transfer efficiency
(Refer to Data Sheet 7)
2. Refer to Data Sheet 7
3. (Standard Deviation / TE^) x 100
34
-------
SECTION 5
PHASE I LABORATORY TEST
FACILITIES
The original laboratory test was performed September 24
through 30, 1982, at PPG Industries' Allison Park, Pennsylvania,
laboratory. PPG donated the facilities, paints, and technicians
for the test.
The Phase I laboratory tests were run in two booths. The
air atomized conventional and electrostatic equipment was run in
a closed water-wash booth rated at 1.9 m3/s (4000 ft3/min).
The high speed bell equipment was run in a closed dry-filter
booth rated at 4.2 m3/s (9200 ft3/min). Exhaust was normal
to the targets in both booths. Paint spray and peripheral equip-
ment specifications used for this test are shown in Table 3.
The equipment sensitivities are documented in Table 4 and 5. In
addition to this equipment, the following items were used for
measuring electrical resistance and electrostatic voltage
potential:
• Volt-ohm meter; Nordson 790-108 Hand kV Meter.
• High voltage DC meter; 0-100 kV; PCF Group, Inc.;
5,000 megaohm probe; Model HV-100A.
DESCRIPTION OF PAINTS
A 65-percent solids coating and 52-percent solids coating
were supplied for testing by PPG. The paint properties are
summarized in Table 6. The higher solids coating is typical of
a compliance solvent-borne coating for the metal finishing
industries. The lower solids coating is an example of an
acrylic enamel used in the automotive industry. Paint weight
percent solids content was determined by ASTM D-2369-81, except
for cure times.
The ASTM method uses syringes to dispense the wet paint
into the aluminum weight dishes. Initially, five grab samples
were processed for each coating using two weight dishes for each
grab sample. At the end of the testing (Thursday, September 9,
1932), an additional grab sample was taken from the 5-gallon can
35
-------
TABLE 3. SUMMARY OF PAINT SPRAY AND PERIPHERAL EQUIPMENT SPECIFICATIONS (1)
A. Paint Supply Tank
1. Type
2. Manufacturer
3. Model No.
4. Serial No.
5. Rated Capacity, gal
B. Paint Spray Equipment
'. Type
2. Manufacturer
3. Model No.
4. Serial No.
5. Rated Capacity, cc/mln
6. Air Cap
7. Fluid Tip
8. Needle
9. Power supply
C. Paint Spray Booth
1. Type
2. Manufacturer
3. Model No.
4. Serial No.
5. Rated Capacity, m-Vs
(ft3/mln) estimated
Oe 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.
Air Atomized
Electrostatic
Pressure Pot
Blnks
83-5501
Not aval lable
2
Air atom elec
Nor d son
AN-8
Not aval lable
Not available
987
Not available
Not ava 1 lable
EPU-8; 76 kV rated;
Water wash
Binks
Dy naprec 1 p 1 tor*
Not aval lable
1.9 (4,000)
Overhead
Not available
Not aval lable
Not available
Convect ion
Despatch
V-29 HD; V-15 HD (3)
Not aval lable
Air Atomized
Conventional
Pressure Pot
Blnks
83-5501
Not ava 1 lable
2
Air atom conv
Blnks
610
Not aval lable
Not aval lable
63 PB
63 C
Not aval lable
Not appl 1 cable
Water wash
Blnks
Dynapreclpltor®
Not aval lable
1.9 (4,000)
Overhead
Not available
Not aval lable
Not available
Convect Ion
Despatch
V-29 HD; V-15 HD (3)
Not aval lable
High Speed
Bell
Positive displacement pump
Ransburg
9966-9
062702
1 (2)
High speed bel 1
Ransburg
Turbobel 1
Not aval lable
Not aval lable
Not appl icabl e
Not appl (cable
Not appl Icable
Not available
Dry filter
Blnks
Not available
Not aval lable
4.2 (9,200)
Overhead
Not available
Not aval lable
Not available
Convection
Despatch
V-29 HD; V-15 HD (3)
Not aval lable
(I) No paint heaters were required for these tests.
(2) A l-gailon can was used to provide paint to the pump suction.
(3) Two different ovens used to dry foils.
-------
TABLE 4. TEST EQUIPMENT SPECIFICATIONS
A.
B.
C.
Percent Weight Solids
1. Laboratory Scales
a. Manufacturer
b. Model No.
c. Serial No.
d. Capacity, g
e. Rated accuracy
2. Foil Dishes
a. Type
b. Size
3. Syringe
a. Type
b. Capacity, raL
4. Solvent Type
Measurement Equipment
Mettler
P12UUN
558312
1,200
, Q + 0.01
Aluminum
Standard
Luer
5
High Solids - Solvesso® 100
Low Solids - Acetone
Conveyor Speed Measurement Equipment
1. Rule
a. Type Flexible Steel Tape
b. Graduations
2. Electronic Timer
a. Type
b. Manufacturer
c. Model No.
d. Serial No.
e. Rated Accuracy
Mass Flow Measurement
1/16 inch
Digital
Lab - Line Instruments, Inc.
1405
Not available
, s Not available (1)
Equipment
Platform Scales
a. Manufacturer NCI, Inc.
b. Model No. 5780
c. Serial No. C790482
d. Capacity, kg 90 (2)
e. Rated Accuracy, g £ 5
2. Stopwatch ~"
a. Manufacturer Markson Science,Inc.
b. Model No. Digital
c. Serial No. 26119
d. Rated Accuracy, s + 0.001% (3 )
D. Target Foil ~"
1. Type Aluminum
2. Nominal Thickness, mils 1.5
3. Temper Medium
E. Wet Film Measurement Equipment
1. Distributor Paul N. Gardner Co.
2. Model No. Nordson Type
(1) Per telephone conversation with manufacturer on 10/4/82.
(2) Fitted with 45 kg load cells.
(3) +0.001% of reading.
37
-------
TABLE 5. MEASURED PARAMETERS AND RATED MEASUREMENT ACCURACY,
Parameter
Test
Device
Weight
- Aluminum dish
- Syringe
- Target foil
- Paint supply tank
Time
- Paint capture
- Conveyor speed
Distance
- Target width
- Conveyor spe.ed
lab scale
lab scale
lab scale
platform scale
stop watch
digital timer
steel rule
steel rule
Rated
Accuracy
+_ 0.01 g
+ 0,01 g
_+ 0.01 g
+ 5 g
_+ 0.001 % (1)
Unknown (2)
4- 0.001 m
(1/32 inch) (3
i 0.001 m
(1/32 inch)
(1) Of reading.
(2) Not available from manufacturer; resolution is +_ 0.05 sec
(1/2 scale reading of 0.1 sec). ~~
(3) Equivalent to 1/2 of smallest reading.
38
-------
TABLE 6. SUMMARY OF RECORDED PAINT SPECIFICATIONS
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
Paint Type
Resin Type
Manufacturer
Manufacturer's Paint ID No.
Lot No.
Color
Reconmended Cure Schedule 10 min (
Viscosity (uncut paint)
Reducing Solvent
% Reduction to Spray
Viscosity - Spray
wt/gal - Spray
wt Solids - Spray
Resistance or Conductance
Paint 1
Enamel
Polyester
—
W4533D
Not available
Beige
§ 177°C (350°F)
28.0 sec (1)
Solvesso® 100
2700 cc/15. 1L
14.5 sec (1)
64.8% (4)
64.8% (4)
2.3 Mf>
Paint 2
Enamel
Acrylic
—
DXH82445
Not available
Gold
30 min @ 121°C (250°F)
27.2 sec (2)
DxD reduce (3)
1400 CC/11.4L
18.0 sec (2)
52.0% (5)
52.0% (5)
2.2 M!2
(1) #3 Zahn cup @ 23°C (74°F).
(2) *4 Ford cup @ 23°C (74°F).
(3) DxD reducer composition: 5U% acetone, 30% Cellosolve® acetate,
12% toluene, 8% Solvesso^ 150.
(4) Mean value for 19 weight dish samples.
(5) Mean value for 10 wt dish samples.
39
-------
containing the cut 65-percent coating. Another nine dishes were
processed from this grab sample and the following bake schedule
was used:
* 3 dishes for 10 minutes at 177°C (350°F)
• 3 dishes for 20 minutes at 177°C (350°F)
• 3 dishes for 30 minutes at 177°C (350°F)
Nineteen (5x2-1-3x3) sample dishes were thus obtained
for the 65-percent solids coating.
MASS FLOW COMPARISON TEST
On Monday, September 27, 1982, the scheduled laboratory
tests began with a comparison of two paint mass flow measurement
techniques:
• Method A - Pressurized paint pot weight change
(atomizing air on)
• Method B - Capture and weighing of unatomized paint flow
Method A would determine actual atomized paint flow during TE
test runs; Method B would predetermine a paint mass flow rate
for unatomized paint to be used in TE calculations. Method B,
commonly used by industry, was suspected to be unrepresentative
of paint flows during spray operations. The purpose of this
test was to determine if there was a statistically significant
difference between the two techniques. If a significant
difference is detected, Method A will be selected as most
representative for TE testing.
The following procedure was used to obtain the desired test
data:
1. Select test equipment. Using Data Sheet D-l, document
the test equipment specifications.
2. Select coating type to be used. Using Data Sheet D-2,
document the manufacturer supplied paint character-
istics .
3. Select spray equipment type and paint supply equipment
to be used. Using Data Sheet D-3, document the equip-
ment specifications.
4. Set up all equipment and adjust paint pot pressure and
atomizing air pressure to desired settings. Record
these settings on Data Sheet D-3.
40
-------
5. Check stopwatch and platform scale to ensure that both
have been zeroed (reset) and that the scale is in the
tare mode.
6. Turn on paint spray equipment and initiate flow.
Simultaneously, start stopwatch. After approximately
500 g have been sprayed, stop stopwatch and paint
spray flow simultaneously. Record platform scale and
stopwatch readings on Data Sheet D-4.
7. Repeat steps 5 and 6 until a total of eight samples
have been recorded. Proceed to Step 8.
8. Turn off atomizing air pressure to spray gun.
9. Tare a paper cup on laboratory scale.
10. Check stopwatch to ensure that it has been zeroed.
11. Turn on paint spray equipment and initiate unatomized
paint flow into paper cup. Simultaneously, start
stopwatch. After cup is approximately 3/4 full, stop
stopwatch and paint flow simultaneously.
12. Weigh filled (and previously tared) paper cup on
laboratory scale. Record weight gain and stopwatch
readings on Data Sheet D-4.
13. Repeat steps 9 through 12 until eight samples have
been recorded-.
14. Perform statistical analyses of test data using the
following two techniques:
• Wilcoxon test for two independent samples
• "Prob-Value" using t-statistic
In performing the above tests the platform scale was
initially calibrated and then checked for accuracy using a paper
cup that had been preweighed at 5 g on the laboratory scale.
The results indicated that the platform scale was capable of
measuring 1,000 _+ 5 g and 500 _+ 5 g and was thus within the
desired precision band.
After checking out the platform scale, the first test run
was performed and aborted because the stopwatch was accidentally
stopped at about 300 g without shutting off the spray gun. A
total of five test runs had been performed when the entire test
series was aborted. This termination of testing occurred be-
cause it was discovered that there was insufficient paint
available of the same type already in the pot to complete the
41
-------
tests. A switch was made to another spray booth and another
paint and the Method A test series was begun again. At this
time it was suggested that a calculation be made to determine
the impact of displacement of the paint in the pot by pres-
surized air at 322 kPa (32 psig). The calculation indicated
that 500 g of paint was displaced by about 1.4 g of air. In
effect, therefore, about 501.4 g of paint was actually being
sprayed. Since the accuracy and resolution of the platform
scale was _+ 5 g, the 1.4-g value was below this threshold limit
and for all practical purposes was undetectable. Therefore, no
adjustment was made in the spray flow rate measurements for
Method A (paint pot weight change).
After completing eight measurements for Method A, a ninth
measurement was made in the event any of the data was later re-
jected using an outlier analysis. For example, it was noted in
the Log Book that data point A-8 may have been suspect due to
the paint line possibly having an air bubble in it.
After completing the Method A test series, the testing
proceeded to Method B for capturing the unatomized output. This
test series went smoothly until data point B-8 was ready to be
taken. At this time it was noted that although the spray gun
had been shut off, a small stream of paint was leaking from the
nozzle. The gun was taken apart, cleaned and reassembled for
data point B-8. An additional test (B-9) was run in the event
8-8 or any other data point in the Method B test series was
later rejected on an outlier basis.
The analysis of the data, which led to the selection of
Method A for all tests, is discussed on p. 61.
TEST PARAMETERS
Determination of Sample Size
The selection of the recommended sample size for each test
series was based on the use of small sampling theory and the t-
statistic. Small sampling theory uses the following terms and
relationships:
U = X + t x s/(n)°-5
where U = True population mean value of transfer
efficiency, %
X = Measured mean value of transfer efficiency for n
samples, %
t = t-statistic representing area under the
t-distribution probability curve with n-1 degrees
of freedom for some specified confidence interval
42
-------
s = Standard deviation for n observations
n = Number of TE sample values
The last term in the above relationship represents the
error in estimating the true mean value of transfer efficiency
(U) using the measured mean (X). To estimate the sample size it
was assumed that the error term should be less than or equal to
the standard deviation at the 95-percent level of confidence.
The sample size can then be evaluated using the following
relationship:
t/(n)0-5 <_ l
or n >_ t2
where t = t-statistic at 95 percent confidence level for n-1
degrees of freedom
The results of the sample size evaluation are as follows:
H n-1 t @ 95% t2
3 2
4 3
5 4
6 5
7 6
8 7
This evaluation shows that a minimum of seven observations
should be selected to ensure that the measured mean TE values
are within one standard deviation of the true population mean at
the 95-percent level of confidence. For these laboratory tests
a sample size of eight was selected for each test series.
Foil Handling Procedures
Of all the steps in the transfer efficiency test procedure,
foil handling proved to be the most time consuming. Foil
handling involves the following activities:
• Cutting foil to desired target length
• Labeling foil with appropriate test run nomenclature
• Preweighing foil and labeling foil with the beginning
weight
• Mounting foil on flat panel or vertical cylinder targets
43
4.303
3.182
2.776
2.571
2.447
2.365
18.5
10.1
7.7
6.6
6.0
5.6
-------
o Removing foil from flat panel or vertical cylinder tar-
gets
o Attaching foil to oven drying racks and inserting racks
in oven
o Removing foils from oven and racks, weighing and
labeling each foil with the ending weight
Of all these activities the foil mounting and removal from
the targets proved to be the most difficult, particularly for
the 3.2 cm (1-1/4 inch) diameter vertical cylinder targets.
Figure 7 shows the attachment techniques for the two target
configurations. For the flat panel targets the foil attachment
and removal were relatively easy. Double-sided tape was
attached to the flat panel targets to allow four sample foils to
be placed on the targets.
In the case of the vertical cylinder targets, however, foil
attachment and removal were difficult. First, a strip of mask-
ing tape was placed the entire length of one of the seams. Next
a strip of double-sided tape was placed over the masking tape.
The foil was then lined up as vertically plumb as possible and
the front edge (refer to Figure 7) attached to the vertical
cylinder using a small piece of masking tape at the top, middle,
and bottom of the foil. Then the foil was tightly wrapped until
the rear seam was barely touching the vertical cylinder. At
this point a vertical mark was made to allow a piece of masking
tape to be lined up correctly and tjien attached to the vertical
cylinder. The rear seam with double-sided tape was then se-
curely attached to the masking tape. The purpose of the masking
tape was to minimize the potential for tearing the wet foil
while removing it prior to oven rack attachment. Experimenta-
tion with the foil attachment had shown that the double-sided
tape was too sticky to allow its direct placement on the foil
surface.
QA/QC Procedures
As part of the overall Quality Assurance Program Plan,
QA/QC procedures were developed and implemented for the
laboratory TE tests. For example, to provide ease in data
handling, the nomenclature shown in Table 7 was adopted and used
extensively throughout the tests. To facilitate recording of
data, three notebooks were kept during the tests:
• Log Book
• Master Data Book
• Weight Solids Data Book
44
-------
••.asking
Tace
Oouble-sided tape
-Double Sided Tape
Foil
-Target
Expanded Top View
Foil "wrapping VC Target
Foil-Ready for attachment
to VC target
VC Target with
Foil Attached
FLAT PANEL TARGET
0.15 m (6 in.)
Flat Panel Target
Foil-Ready for
attachment
Double-sided tape
Figure 7. Foil Attachment Techniques for Vertical
Cylinder and Fl.-it Panel Targets, PPG Test,
September 1982
45
-------
TABLE 7. NOMENCLATURE FOR SPRAY PAINTING
TRANSFER EFFICIENCY TESTS
Parameter Nomenclature
I. Spray Equipment Type
A. Air atomized conventional AAC
B. Air atomized electrostatic AAE
C. High speed bell HSB
II. Coating Type (1)
A. Automotive enamel 52
B. General purpose high solids 65
III. Target Configuration
A. Flat panel FP
B. Vertical cylinders VC
Test No.
Example: AAC - 52 - FP - 1
rthere AAC - 52 - FP = Test series number
and 1 = Observation number
(1) For clarity, the numerical weight percent solids is used
to identify the coating.
46
-------
The Log Book was used to keep track of miscellaneous data
and to record any test anomalies. The Master Data Book
contained the following information:
• All data sheets for all tests
• Up-to-date copy of the Test and Evaluation Plan,
including Appendices A through E
• Summary of test matrix components
• Test sequence
• Test flow charts
The Master Data Book served as the primary reference docu-
ment for all the TE tests. The height Percent Solids Data Book
contained all the necessary data sheets and a copy of ASTM
D-2369-81 (Standard Test Method for Volatile Content of Coat-
ings). All of the weight percent solids analyses were recorded
in this book.
To help ensure correctness, thoroughness, and completeness,
all data sheets were independently reviewed, checked, and signed
off.
As part of the overall QA/QC Plan, equipment with high
rated accuracy was used wherever possible. Table 5 summarizes
the major measured parameters and the associated equipment rated
accuracy-
Test Sequence
Table 8 depicts the actual test sequence used for the
transfer efficiency laboratory tests. As this table shows, the
platform scale was initially set up on September 24, 1982. A
quick check of the scale indicated that it was a 90-kg (200-lb)
model as opposed to the desired 45-kg (100-lb) model. The scale
rental supplier advised that the desired 45-kg load cells had
been installed in the scale with the desired _+ 5-g accuracy and
readability (resolution). A paper cup was cut to size and
weighed on the Mettler laboratory scale (accuracy of _+ 0.01 g)
to obtain a "calibration" weight. The scale was then loaded
with 1,000- and 500-g weights and the paper cup was added to
determine if the scale would read 1,005 and 505 g. This
preliminary check-out was successful.
After checking out the scale, the vertical cylinder targets
were assembled from the prefabricated pieces. The target
assembly went smoothly and all preliminary test arrangements
were completed in preparation for the following week's tests
(9/27 - 9/30/82), described below.
47
-------
TABLE 8. TRANSFER EFFICIENCY TEST SEQUENCE
AT PPG SEPTEMBER 1982
DATE
Friday (9-24-82)
Monday (9-27-82)
ACTIVITY
o Set up platform scale and
made preliminary adjustments
o Set up vertical and flat panel
targets and performed "dry run"
check out
o Performed mass flow comparison
test
TEST RUNS
o Began TE test runs
Tuesday (9-28-82) o Continued TE test runs
o Performed additional TE
test run
Wednesday (9-29-82) o Changed to HSB booth
o Completed TE test runs
o Booth cleanup
Thursday (9-30-82) o Rechecked target foils
o Performed additional weight
solids tests
o
0
o
o
o
o
o
0
0
o
o
o
o
o
o
o
o
o
0
o
o
o
o
o
o
o
AAE -
AAE -
AAE -
AAE -
AAE -
AAE -
AAE -
AAE -
AAC -
AAC -
AAC -
AAC'-
AAC -
AAC -
AAC -
AAC -
AAE -
AAE -
HSB -
HSB -
HSB -
HSB -
HSB -
HSB -
HSB -
HSB -
65
65
65
65
52
52
52
52
52
52
52
52
65
65
65
65
65
65
65
65
65
65
52
52
65
65
- VC -
- FP -
- VC -
- FP -
- VC -
- FP -
- VC -
- FP -
- VC -
- FP -
- VC -
- FP -•
- VC -
- FP -
- VC -
- FP -
- VC -
- ST -
- FP -
- FP -
- VC -
- VC -
- FP -
- FP -
- VC -
- VC -
1
1
5
5
1
1
5
5
1
1
5
5
1
1
5
5
to
to
to
to
to
to
to
to
to
to
to
to
to
to
to
to
4
4
8
8
4
4
8
8
4
4
8
8
4
4
8
8
1A to 4A
1
1
5
1
5
1
5
1
5
to
to
to
to
to
to
to
to
to
3 (1)
4
8
4
8
4
8
4
8
1) ST = semi-tubular target configuration
48
-------
PHASE I TEST RESULTS
General
Transfer efficiency measurements were made for a total of
12 test series. These series encompassed a test matrix of three
types of spray equipment, two coatings, and two target config-
urations. For each test series, eight foils were coated to
determine transfer efficiency. The eight observations per test
series were obtained by coating two separate groups of four
foils per target assembly (refer to Figures 3 and 4). For
comparative purposes, however, Table 9 summarizes the results of
the tests based on evaluating the data under the following two
scenarios:
• Each test series is composed of eight observations from
two groups of four foils each
• Each test series is composed of two observations; each
observation is the average of a group of four foils
These two different scenarios for evaluating the data are
shown because of the question of whether or not a group of four
foils should be treated as four separate observations or as one
observation averaged from four foils. Treating the test data
for each test series as two observations composed of the
averages of four foils each is a conservative approach.
Realistically, the data obtained in the tests are equivalent to
a sample size somewhere between two and eight. The more conser-
vative approach (data treated as two observations of four foils
each) is the preferred technique for estimating the overall
precision of the test method.
In both scenarios, the measured mean TE values are equal;
however, as expected, the standard deviations and hence
coefficients of variation are different. For example, the
measured mean TE values ranged from a low of 10.2 percent for
the air atomized conventional gun spraying 65-percent solids
paint at a vertical cylinder target, to a high of 93.9 percent
for a high speed rotating bell delivering 65-percent solids
paint to a vertical cylinder target.
For the standard deviations, however, the two different
data handling scenarios produced noticeably different results.
In the first case (data treated as eight individual observa-
tions), the standard deviations ranged from 0.6 to 5.2 percent.
The standard deviations in the second case (data treated as two
observations of four foils each), however, ranged from 0.04 to
6.1 percent. Similarly, the coefficients of variation, which
are a measure of the relative precision among test series,
ranged from 1.2 to 5.3 percent for the first case and 0.1 to 6.1
percent for the second case.
49
-------
TABLE 9. SUMMARY OF TRANSFER EFFICIENCY TEST RESULTS
CASE 1: CASE 2:
Data treated as 8 Data treated as 2
individual observations observations of 4 foils each
Test Series
AAE-52-VC
52-FP
65-VC
65-FP
AAC-52-VC
52-FP
65-VC
65-FP
HSB-52-VC
52-FP
65-VC
65-FP
30.8
81.2
25.7
81.5
10.8
61.7
10.2
60.8
91.3
91.3
93.1
92.8
S
1.7
1.3
1.6
1.9
0.7
1.2
0.6
1.6
2.8
1.9
5.2
1.1
COV, %
5.5
1.6
6.2
2.3
6.5
1.9
5.9
2.6
3.1
2.1
5.3
1.2
S
1.7
0.2
0.4
1.7
0.1
0.5
0.04
0.04
2.4
0.5
6.1
0.6
COV, %
5.5
0.4
1.6
2.1
0.9
0.8
0.4
0.1
2.6
0.5
6.1
0.7
NOTES:
TEjYj = measured mean transfer efficiency, %
S = sample standard deviation
COV = coefficient of variation (S/TEvj) x 100, %
50
-------
Having calculated the standard deviations for each test
series, the analyses proceeded to an evaluation of the
poolability of the variances to determine if a statement could
be made concerning the overall precision of the test method.
Bartlett's Test was used to examine the homogeneity of the
variances of the test series (i.e., are the variances of the
test series equal?).
From the Bartlett's Test it was found that the variances of
all 12 test series could not be considered equal for either of
the cases in Table 9. The analyses then proceeded to an
examination of the following various groups of data:
• Compare poolability of variances for TE's obtained from
the same spray equipment type (4 TE's = 2 targets x 2
coatings); 3 gun groups to be evaluated.
» Compare poolability of variances for TE's obtained for
the same target configuration (6 TE's = 3 spray equip-
ment x 2 coatings); 2 target groups to be evaluated.
• Compare poolability of variances for TE's obtained for
the same coating (6 TE's = 3 spray equipment x 2 tar-
gets); 2 coatings groups to be evaluated.
• Several combinations of the first group.
These analyses were performed for-only the second case,
i.e., treating the data for each test series as two observations
of four foils each. The results of these analyses are shown in
Table 10. The pooled standard deviations ranged from 0.3 to
3.3; however, the pooled standard deviation for the maximum
sample size of 16 observations was 2.5. This value is the
preferred value for use in estimating the required sample size
for future laboratory tests.
An examination of the results in Table 10 shows that the
data were not poolable, primarily due to the very large standard
deviation (6.1) for one test series (high speed bell, 65-
percent-solids, vertical cylinder) coupled with the very low
variance exhibited by the AAC configuration. As discussed
later, in this particular test series (HSB-65-VC), one sample
resulted in a mean TE of 102.4 percent and the other a mean of
93.9 percent, with a resultant average value of 98.1 percent.
Since transfer efficiencies greater than 100 percent are
impossible, the mean TE of 102.4 percent is clearly in error.
However, .the data exceeding 1QO percent could not be rejected on an outlier
basis. Sufficient data points were hgt available to perform an outlier
evaluation to permit removal of the data from the statistical analysis.
51
-------
TABLE 10. RESULTS OF BARTLETT'S TEST FOR EVALUATION OF POOLABILITY
OF VARIANCES AT 95% LEVEL OF CONFIDENCE
Test Series
Evaluated
I. All test series
Pooled
No. of No. of Standard
Test Series Observations (1) Deviation
12
24
Not poolable
II. Groups of test series
A. Sane spray equipment type
1. Air atomized electrostatic
2. Air atomized conventional
3. High speed bell
B. Sane target configuration
1. Vertical cylinders
2. Flat panels
C. Same coating
1. 52% wt solids
2. 65% wt solids
4
4
4
6
6
6
6
8
8
8
12
12
12
12
1.2
0.3
3,3
Not poolable
0.8
1.2
Not poolable
III. Combinations of groups
A. Group Al & A2
B. Group Al & A3
C. Group A2 & A3
8
8
16
16
16
Not poolable
2.5
Not poolable
(1) Data treated as two observations of four foils each.
52
-------
Coating and Equipment Specifications and Equipment
Operating Conditions
Tables 3, 11, and 12 show the specifications of the equip-
ment and the equipment operating conditions used in these tests.
Table 6 shows the coating specifications.
Table 3 shows the specifications of the three types of
paint spray equipment used in this test program. In addition,
the table provides information on the peripheral equipment used
in the tests.
TRANSFER EFFICIENCY TEST RESULTS
Air Atomized Electrostatic Spray Equipment Test Results
Table 13 summarizes the results of all the tests for the
air atomized electrostatic paint spray equipment. Transfer
efficiencies ranged from 23.5 to 33.7 percent for the foils on
the vertical cylinders to 78.5 to 84.4 percent for the foils on
the flat panel targets. The resulting standard deviations were
low and ranged from 0.2 to 1.7, indicative of very good
precision.
In addition to these results, one supplementary test was
performed in which four additional vertical cylinder foils and
three semi-tubular target foils were coated using the high sol-
ids coating. The supplementary test was performed because the
first sample, AAE-HS-VC-1 to 4, in particular, had not used the
foil attachment technique shown in Figure 7- As a result, wet
paint was deposited on masking tape used to hold the rear seam
down. This tape was then removed, subsequently affecting the
dry solids weight for the cured foil. Figure 8 shows the
difference between the vertical cylinder and semitubular target
configurations.
If this additional observation is included in the AAE-65-VC
test series, the new calculated results are as follows:
• TEM = 27.9%
• S = 3.9
• COV = 14.0%
These results are based on the four foils in the third
observation having individual TE's of 33.9, 31.5, 32.1, and 32.1
percent. The new standard deviation and COV are significantly
higher than the results of the other test series for the air
atomized electrostatic spray gun. The bias introduced by the
tape removal problem (discussed above) clearly contributed to
53
-------
TABLE 11. SUMMARY OF EQUIPMENT OPERATING CONDITIONS FOR AIR ATOMIZED
ELECTROSTATIC AND CONVENTIONAL SPRAY GUNS
Air Atomized Electrostatic
Air Atomized Conventional
A. Paint Spray Equipment (1)
1. Paint Pressure at Paint Pot, kPa (psig)
2. Paint Pressure at Spray Gun, kPa (psiy)
3. Atomizing Air Pressure at Spray Gun, kPa (psig)
4, Operating Voltage, kV
5. Paint Temperature at Paint Pot, "C (°F)
6. Gun to Target Distance, in (an)
B. Paint Spray Hooth
1. Ambient Temperature, °C ("F)
2, Relative Humidity, %
3. Air Flow Velocity, m/s (fpm)
4. Air Flow Direction
C. Target Parameters
1. Average Wet Film Thickness, 10~*» meters (mils)
2. Vertical Paint Coverage, cm (in)
3. Target Height, cm (in)
4. % Vertical Coverage
5. Resistance to Ground, Ohm
D. Forced Draft Oven
1. Cure Time, sec (min)
a. Foil Dish (sample)
b. Target Foil
2. Cure.Temperature, °C (°F)
a. Foil Dish (sample)
b. Target Foil
E. Conveyor Speed Setpoint, m/s (fpm)
Low Solids
294 (28)
Not measured
446 (50)
45 (2)
23 (74)
12 (30); 13 (33)
(3)
High Solids
252 (22)
Not measured
446 (50)
45 (2)
23 (74)
12 (30); 13 (33)
Low Solids
294 (28)
Not measured
446-453 (50-51)
Not applicable
23 (74)
12 (30); 13 (33)
High Solids
225 (18)
Not measured
453 (51)
Not applicable
23 (74)
12 (30); 13 (33)
23 (74) 24 (75)
52 49
0.762-1.02 (150-200) 0.762-1.143 (150-225)
Normal to target Normal to target
23 (73) '
54
0.762-1.02 (150-200)
Normal to target
23 (73)
54
0.762-1.02 (150-200)
Normal to target
i.27-2.54 (1/2 to 1)
Not measured
86-89 (34-35)
75 (est)
< 1 million
1600 (30)
1800 (30)
121 (250)
121 (250)
0.15 (30)
3.81 (1 1/2)
Not measured
86-89 (34-35)
75 (est)
< 1 million
600, 1200, 1800 (4)
(10, 20, 30)
600 (10)
177 (350)
177 (350)
0.15 (30)
J.81 (1 1/2) 5.08-6.35 (2 to 2 1/2
Not measured Not measured
86-89 (34-35) 86-89 (34-35)
75 (est) 75 (est)
< 1 million < 1 million
1800 (30)
1800 (30)
12l'(250)
121 ,(250)
0.15 (30)
600, 1200, 1800 (4)
(10, 20, 30)
10
177 (350)
177 (350)
0.15 (30)
(1) Paint heaters not used in these tests.
(2) Measured at gun tip with paint in lines.
(3) 12 inches for vertical cylinder target; 13 inches for flat panel target.
(4) Some sample dishes were baked for 20 and 30 minutes (refer to discussion in this report).
-------
TABLE 12. SUMMARY OF EQUIPMENT OPERATING CONDITIONS FOR HIGH SPEED BELL
ui
01
Vertical Cylinders
Low Solids
High Solids
Flat
Low Solids
Panel
High Solids
A. Paint Spray hijuipment (1)
1.
2.
3.
4.
5.
6.
Turbine Air Pressure, KPa (psig)
Operating Voltage, M
Disk or Bell Speed, rps (rpm)
a. With faint Applied
b. Without Paint Applied
Shaping Air for Bell, kPa (psig)
Paint Temperature at Paint Pot, "C (°F)
Gun to Target Distance, cm (in)
308 (30)
80 (2)
Not measured
333 (20,000)
308 (30)
21 (70)
30 (12)
308 (30)
80
Not measured
333 (20,000)
308 (30)
21 (70)
30 (12)
308 (30)
80
Not measured
333 (20,000)
308 (30)
21 (70)
JO (12)
308 (30)
80
Not measured
333 (20,000)
308 (3U)
21 (70)
30 (12)
B. Paint spray booth
1. Ambient Temperature, °C (°F)
2. Relative Humidity, %
3. Air Flow \telocity, m/s (fpm)
4. Air Flow Direction
C. Target Parameters
1. Average Wit Film Thickness, 10~6 meter (mils)
2. Vertical Paint Coverage, cm (in)
3. Target Height, on (in)
4. * Vertical Coverage
5. Resistance to Ground, Ohm
24 (75) 21 (70) 21 (70) 21 (70)
50 56 56 b6
0.635-0.889 (125-175) 0.508-0.762 (100-150) 0.508-0.762 (100-150) 0.508-0.762 (100-150)
Normal to target Normal to target Normal to target Normal to target
3.81 (1 1/2) 3.81-5.08 (1 1/2 - 2) 3.81-5.08 (1 1/2 - 2) 3.81 (1 1/2)
Not measured Not measured Not measured Not measured
142 (56) 142 (56) 142 (56) 142 (56)
90 (est) 90 (est) 90 (est) 90 (est)
.< 1 million < 1 million < 1 million < 1 million
D.
Forced Draft oven
1. Cure Time, s (mint
E.
a. Foil Dish (sample)
b. Target Foil
2. Cure Temperature, °C (°F)
a. Foil Dish (sample)
b. Target Foil
Conveyor Speed Setpoint, m/s (fpm)
1800 (30)
1800 (30)
121(250)
121(250)
9.1(30)
600, 1200, 1800 (3)
(10, 20, 30)
600 (10)
177(350)
177(350)
9.1(30)
1800 (30)
1800 (30)
121(250)
121(250)
. 9.1(30)
600, 1200, 1800 (3)
(10, 20, 30)
600 (10)
177(350)
177(350)
9.1(30)
(1) Paint heaters not used in these tests.
(2) Measured at bell high voltage take-off connection with paint in lines.
(3) Sooc sample dishes were baked for 20 and 30 minutes (refer to discussion in this report).
-------
U1
TABLE 13. SUMMARY OF TE TEST RESULTS FOR AIR ATOMIZED ELECTROSTATIC
SPRAY EQUIPMENT
Transfer Efficiencies, % (1)
Coeffi-
Foil No. Foil No. Std Devi- cient o£
Coating Target ation, Variation,
Ht Solids Configuration 1111 TElU) ^ j> II TE^O) TCfi(4) (5) t (6)
52 Vertical Cylinder 31.0 28.8 29.2 29,6 2«.7 33.7 32.2 30.3 31.8 32.0 30.8 1.66 5.5
52 Hat Panel 82.5 82.5 81.4 79.2 81.4 80.5 80.5 80.5 82.8 81.1 81.2 0.23 0.2
65 Vertical Cylinder 26.9 2-1.2 25.8 26.9 26.0 27.4 24.0 23.5 26.8 25.4 25.7 0.37 1.6
65 Flat Panel 81.2 80.6 80.6 78.5 80.2 80.4 84.4 82.7 83.2 82.7 81.5 1.73 2.1
(1) Values are shown rounded to nearest 0.1%; statistical analyses were performed with values rounded to nearest 0.01%.
(2) TEj is Observation No. 1 and represents the arithmetic average of foil Nos. 1, 2, 3 and 4.
(3) TE2 is Observation No. 2 and represents the arithmetic average of foil Nos. 5, 6, 7 and 8.
(4) Ttfj is the neasurerl mean TE for each test series and represents the arithmetic average of Observation No. 1 (TEi) and Observation
No. 2 (TE2>.
(5) For these tests which assume two observations, see test series? the sample standard deviation is as follows? S = ((TEi -
(6) The coefficient of variation (COV) is determined as follows! OOV = SA^M? standard deviations were rounded to nearest 0.1 before can-
put ing the Cov.
-------
3.2cm
(I3* in.)
~H
(l in.)
Vertical.
Cylinder (VC).
Target
Semitubolar (STi
Target
Figure 8. Comparison of Vertical Cylinder and
Semitubular Target Configurations
57
-------
the calculation for the mean transfer efficiency and the
relatively large standard deviation for this test series.
As expected, due to the increased exposed surface area, the
semitubular target had a higher transfer efficiency than the
vertical cylinder target. The three foils for the semitubular
target had individual TE's of 34.5, 35.7, and 40.5, for an
average of 36.9 percent.
The coefficients of variation for the test series in
Table 13 ranged from 0.2 to 5.5 percent, indicative of good
relative precision. The COV for the AAE-65-VC test series using
three observations instead of two was a poor 14 percent.
Air Atomized Conventional Spray Equipment Test Results
Table 14 summarizes the results of all the test series for
the air atomized conventional paint spray equipment. Transfer
efficiencies ranged from 9.7 to 11.9 percent for the foils on
the vertical cylinders to 58.6 to 63.4 percent for the foils on
the flat panel targets. The resulting standard deviations were
very low and ranged from 0.04 to 0.5, indicative of excellent
precision for these test series. Similarly, the relative preci-
sion expressed by the COV's was excellent and fell in the range
of 0.1 to 0.9 percent.
Hj.gh Speed Bell Spray Equipment Test Results
Table 15 summarizes the results of all the test series for
the high speed bell coating equipment. Transfer efficiencies
ranged from 86.7 to 106.3 percent for the foils on the vertical
cylinders to 89.3 to 94.4 percent for the foils on the flat
panel targets. The resulting standard deviations were rela-
tively low and ranged from 0.5 to 6.1, indicative of good
precision for these test series. Similarly, the COV's were
relatively low and ranged from 0.5 to 6.2 percent.
Note that the transfer efficiency in one of the vertical
cylinder/65-percent solids test series was over 100 percent.
Transfer efficiencies over 100 percent are clearly erroneous and
most likely are the result of bias in one or more of the
following measurements:^
• Incorrect end weight for platform scale (mass flow rate
calculation)
• Incorrect stopwatch reading (mass flow rate calculation)
• Incorrect conveyor time reading (conveyor speed measure-
ment calculation)
1Refer to Section 6 for further explanation.
58
-------
TABLE 14. SUMMARY OF TE TEST RESULTS FOR AIR ATOMIZED CONVENTIONAL
SPRAY EQUIPMENT
ui
Coating
Wt Solids
Foil No.
Target
Configuration 123
Transfer Efficiencies
4 TCi(2) 5
, % (1)
Foil No.
678 TE2(3) TEj)
Std Devi-
ation,
(4) (5)
Coeffi-
cient of
Variation,
» (6)
b2 Vertical Cylinder 11.9 10.9 10.4 10.2 10.9 10.5 11.7 10.5 10.0 10.7 10.B 0.12 0.9
52 Flat Panel 61.9 62.4 60.4 60.9 61.4 63.4 63.2 60.9 60.7 62.1 61.7 0.46 0.8
65 Vertical Cylinder 9.8 9.8 10.8 10.3 10.2 11.3 10.2 9.7 9.7 10.2 10.2 0.04 0.1
65 Flat Panel 61.3 61.9 58.6 61.3 60.8 58.6 61.3 60.8 62.4 60.8 60.8 0.04 0.1
(1) Values are shown rounded to nearest 0.1»; statistical analyses were performed with values rounded to nearest O.OH.
(2) TE^ is Observation No. 1 and represents the arithmetic average of foil Nos. 1, 2, 3 and 4.
(J) TE2 is Observation No. 2 and represents the arithmetic average of foil Nos. 5, 6, 7 and B.
(4) Tfc^ is the measured nean TE for each test series and represents the arithmetic average of Observation No. 1 (TEj) and observation No. 2
(TE;;).
(5) tor these tests which assuire two observations, see test series; the sample standard deviation is as follows: S » l(TEi - T
(6) The coefficient of variation (COV) is determined as follows: COV = S/Tk^; standard deviations were rounded to nearest 0.1 before com-
puting the CUV.
-------
TABLE 15, SUMMARY OF TE TEST RESULTS FOR HIGH SPEED BELL COATING
EQUIPMENT
Transfer Efficiencies, % (1)
Coetti-
Foil No. Foil No. Std Uevi- cient of
Coatiny Target ation. Variation,
hit Solids Configuration 1111 TEj<2) 11 11 TE^P) TE^H) (5) % (bj
S2 vertical Cylinder 93.9 91.4 92.8 93.9 93.0 90.2 86.7 87.8 93.5 89.6 91.3 2.44 2.6
52 Flat Panel 94.0 89,3 89.3 91.2 91.0 94.2 90.6 91.7 90.3 91.7 91.3 0.53 0.5
65 Vertical Cylinder 100.0 100.9 102.5 106.3 102.4 92.8 92.8 91.9 97.9 93.9 98.1 6.07 6.2
65 Flat Panel 93^ 92.2 93.1 94.4 93.2 93.6 92.4 90.5 92.7 92.3 92.8 0.64 0^6
OT>
° (1) Values are shown rounded to nearest 0.1%; statistical analyses were performed with values rounded to nearest 0.01%.
(2) TEj is Observation No. 1 and represents the arithmetic average of foil Nos. 1, 2, 3 and 4.
(3) TE;> is Observation No. 2 and represents the arithmetic average of foil Nos. 5, 6, 7 and 8.
(4) Tt+j is the iteasured maan TE for each test series and represents the arithmetic average of observation No. 1 (TEj) and Observation
No. 2 (TE2).
(b) For these tests which assume two observations, see test series; the sample standard deviation is as follows! S » ((TEj - Tt^j)2 +
(TE2 - TEM»2)0.5
(6) The coefficient of variation (COV) is determined as follows! COV = S/TEMi standard deviations were rounded to nearest 0.1 before com-
puting the COV.
-------
• Incorrect foil weight before spraying (net solids
deposited calculation)
WEIGHT PERCENT SOLIDS TEST RESULTS
Table 16 summarizes the results of the weight percent
solids tests performed as part of the transfer efficiency
measurements. The high solids coating used in these tests had
an average weight solids content of 64.8 percent and a standard
deviation of 2.4 percent for 19 samples. The low solids coating
had an average weight solids content of 52.0 percent and a
standard deviation of 1.9 for 10 samples. These average values
were used in the calculations for transfer efficiency.
MASS FLOW COMPARISON TEST RESULTS
Tables 17, 18, and 19 summarize the specifications for the
test equipment, paint, and paint spray and peripheral equipment
used in the mass flow comparison test. Table 20 summarizes the
results of the nine test runs performed for Method A (platform
scale weight change) and Method B (capture unatomized output).
Using this information, statistical analyses were performed to
determine if both mass flow measurement techniques yielded
results that could be considered essentially equal. Two
independent statistical techniques showed that the two mass flow
measurement techniques yielded significantly different
results.
The first method applied to evaluate the test data was the
Wilcoxon test for two independent samples. In applying this
method, it was decided not to adjust the flow rate in Method A
to account for the air weight change in the pressure pot. This
adjustment is not recommended since the air weight change is
approximately 1.4 g per 500 g paint sprayed and represents only
28 percent of the 5-g accuracy and resolution of the platform
scales.
The second method applied to evaluate the test data was the
t-statistic. The same data were used as for the Wilcoxon eval-
uation. The results show that the measured mean flows, 2.902 x
10~3 and 3.021 x 10~3 kg/s, are significantly different with
a significance level of 0.0005, i.e., 0.05 percent.
As a result of this comparison, Method A was selected as
the most representative mass flow determination method. Method
A is used in the Draft Transfer Efficiency Test Procedure.
61
-------
TABLE 16. SUMMARY OF WEIGHT SOLIDS TEST RESULTS
High Solids Coating (1)
Sample
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
(1)
(2)
Cure
Time, s (min)
1200 (20)
1200 (20)
600 (10)
600 (10)
600 (10)
600 (10)
600 (10)
600 (10)
600 (10)
600 (10)
600 (10)
600 (10)
600 (10)
1200 (20)
1200 (20)
1200 (20)
1800 (30)
1800 (30)
1800 (30)
Ave.
S
Cure temperature =
Cure temperature =
Low Solids Coating (2)
Weight Cure vfeight
Solids, % Time, s (min) Solids, %
64.9
64.7
67.3
69.2
65.1
62.1
65.6
66<, 7
61.4
62.3
66.7
66.7
64.0
66.7
66.7
64.4
64.2
59.3
64.1
64.8%
2.4
177°C (350°F);
121°C (250°F);
1800
1800
1800
1800
1800
1800
1800
1800
1800
1800
(30) 51.9
(30) 54.2
(30) 48o2
(30) 50.0
(30) 53.7
(30) 52.0
(30) 53.7
(30) 53.7
(30) 51.1
(30) 51.9
Ave. = 52.0%
S = 1.9
all samples.
all samples.
62
-------
TABLE 17. TEST EQUIPMENT SPECIFICATIONS FOR MASS FLOW COMPARISON TESTS
A. Laboratory Scale
1. Manufacturer
2. Model No.
3. Serial No.
4. Capacity/ g
5. Rated Accuracy, g
B. Platform Scale
1. Manufacturer
2. Model No.
3. Serial No.
4. Capacity, g
5. Rated Accuracy, g
C. Capture Container
1. Type
2. Approximate dry weight, g
D. Stop vtotch
1. Manufacturer
2. Model No.
3. Serial No.
4. Capacity, g
5. Rated Accuracy, g
Mettler
P1200N
558312
1,200
0.01
NCI, Inc.
5780
C790482
90 (1)
+ 5.
Paper Cup
Markson Science, Inc.
Digital LCD
26119
Not available
+ 0.001% of reading
(1) Fitted with 45-kg load cells.
63
-------
TABLE 18. PAINT SPECIFICATION FOR MASS FLOW COMPARISON TESTS
1. Paint Type
2, Resin Type
3. Manufacturer
4. Manufacturer"r Paint ID No.
5. Lot No.
6. Color
1. Recommended Cure Schedule
8. Viscosity
9,, Reducing Solvent
10. % Reduction to Spray
Enamel
Polyester
PPG
AG452D1331-A
Not Available
White
Not Required
Not Required
Xylene
Unknown
11. Viscosity - Spray
12. Wt/Gallon - Spray
13. Wt Solids - Spray
14. Resistance or Conductance
19 sec wi/»3 Zahn cup @ 23°C (74°F)
1.33 kg/L (11.1 Ib/gal)
Unknown
Unknown
64
-------
TABLE 19. EQUIPMENT SPECIFICATIONS AND OPERATING
CONDITIONS FOR MASS FLOW COMPARISON TEST
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. Equipment Operating Parameters
1. Atomizing Air Pressure at
Spray Gun, psig
2. Paint Pressure at Paint
Pot
3. Initial Pot Loading
Pressurized Pot
Binks
83-5501
Not Available
AAE
Nordson
AN-8
Not Available
Not Available
987
228
Not Available
9.38 kPa (50 psig)
6.77 kPa (32 psiq)
3.8L (1 gal)
65
-------
TABLE 20. MASS FLOW COMPARISON TEST RESULTS
Method A:
Test Run
No.
A-l
A-2
A-3
A-4
A-5
A- 6
A-7
A- 8
A-9
Method Bs
Test Run
No.
B-l
B-2
B-3
B-4
B-5
B-6
B-7
B-8
B-9
(Platform scale;
Beginning
Wt, kg
0
0
0
0
0
0
0
0
0
(Laboratory scale
Beginning
Wt, kg
0
0
0
0
0
0
0
0
0
atomized.. output )
Ending
Wt, kg
0.505
0.505
0.505
0.500
0.505
0.505
0,505
0.500
0.500
I unatomized output)
Ending
Wt, kg
134.17
135.44
133087
134;71
135.49
131.30
131.03
140.82
137.85
Elapsed
Time, s
174.23
173.81
172.01
172.06
174.59
173.23
174.80
169.40
171.99
Elapsed
Time, s
44.33
44.99
44.97
45.08
45.14
43.81
44.09
45. 11
44.51
Flow Rate,
10-3 kg/s (1)
(g/min )
2.898 (173.91)
2.905 (174.33)
2.936 (176.15)
2.906 (174.36)
2.892 (173.55)
2.915 (174.91)
2.889 (173.34)
2.952 (177.10)
2.907 (174.43)
Mean Flow = 2.902 (174.68) (2)
Flow Rate,
10-3 kg/s (1)
(g/min)
3.027 (181.60)
3.010 (180.63)
2.977 (178.61)
2.988 (179.29)
3.002 (180.09)
2.997 (179.82)
2.972 (178.31)
3.122 (187.30)
3.097 (185.82)
Mean Flow = 3.021 (181.27) (2)
(1) Method A results not corrected for pressurized paint pot air weight change.
(2) Arithmetic mean values for nine observations.
-------
SECTION 6
PHASE II LABORATORY TEST
FACILITIES
The Phase II Laboratory test runs were done using the
original test plan. They were run in an open booth approxi-
mately 4.6 x 7.6 meters (15 feet by 25 feet) with the exhaust
fan along the back wall normal to the targets. Paint spray and
peripheral equipment specifications used for this test are shown
in Table 21. Every effort was made to duplicate the equipment
specifications used in the original test. The equipment rated
accuracy matched or exceeded the requirements of the Test Plan
(Table 5).
For the voltage measurements, Ransburg supplied a Singer
Kilovolt Meter, which does not draw current while sensing the
voltage.
DESCRIPTION OF PAINTS
PPG supplied two paints from the same batch as the first
test. These paints were recut to test specifications at PPG
before shipment. Table 22 documents the paint properties
measured at Ransburg prior to the test. In the Phase I test,
the paint viscosities were 18 seconds (Ford C4 cup) and 14.5
seconds (Zahn C3 cup) for the 51-percent and 67-percent solids
paints, respectively. PPG reports identical results for the
paint they supplied for Phase II. Values of 22 and 20.5
seconds were measured for the corresponding paints at Ransburg.
Since agreement with specifications was documented by PPG before
shipment, the difference is attributed to small variations from
cup to cup and in measurement technique.
A Ransburg Model 234 meter was used to measure the resis-
tance of the paint. The resistance was 1.5 megohms for the 51
percent solids paint, and 2.5 megohms for the 67 percent solids
paint. These resistances also differ from the original test
observations.
Paint weight percent solids determinations were made
according to the Test Plan and ASTM D-2369-81 (see Section 4).
The paint manufacturers' recommended cure schedules were used
for the weight percent solids determination and for curing
67
-------
TABLE 21. SUMMARY OF REPORTED PAINT SPRAY AND PERIPHERAL
EQUIPMENT SPECIFICATIONS FOR RANSBURG TEST (1)
CD
A 1 r Atom 1 zed
E 1 ectrostat 1 c
A. Pa
1.
2.
3.
4.
5.
B. Pa
1.
2.
3.
4.
5.
6.
7.
a.
9.
C. Pa
1.
2.
3.
4.
5.
Int Supply Tank
Type
Manu f actur er
Model No.
Serial No.
Rated Capacity, gallons
Int Spray Equipment
Type
Ma nu f actur er
Model No.
Serial No.
Rated Capacity, cc/mln
A 1 r Cap
Fluid Tip
Need le
Power supply
Int Spray Booth
Type
Ma nuf act urer
Model No.
Serial No.
Rated Capacity, cfm
Press ur e
Bl nks
83-5508
Not ava 1
2
Air
a torn
Pot
fa b 1 e
el
ec
A 1 r Atom 1 zed
Convent 1 ona 1
Pressure
Bl nks
83-5508
Not ava 1 1
2
Air
Nordson
Not
Not
Not
EPU-8;
150 u
Not
Not
Not
Not
AN-8
ava 1
ava 1
987
228
ava 1
1 ab
lab
le
le
Not
a torn
High Speed
Bel 1
Pot Positive dlsplacemen
Ransbur g
9966-9
able 062702
2
conv
Bl nks
21 {
361 1
ava 1
2
3
1
)
able
63 PB
1 ab
1 e
76 kV rated ;
amp
Open
ava 1
ava 1
ava 1
ava 1
Not
Not
63 C
ava 1
appl
1
1
abl e
cable
High speed be 1
Ransbur g
Turbobe 1 1
Not ava liable
Not aval (able
Not appl 1 ca b 1
Not appl 1 cab 1
Not appl 1 ca b 1
Not ava liable
t pump
1
e
e
e
rated
1 ab
lab
lab
lab
le
le
le
le
Not
Not
Not
Not
Open
ava I
ava 1
ava 1
ava 1
1
1
1
1
able
able
able
able
Open
Not ava 1 1 ab 1 e
Not available
Not a va 1 1 ab 1 e
Not aval (able
Conveyor
t. Type
2. Manufacturer
3: Model No.
4. Serial No.
Forced Draft Oven
1. Type
2. Ma nu factur er
3. Model No.
4. SerI a I No.
Overhead
Rlchards-WIIcox
Not available
Not ava I table
Electric
Young Brothers
Not ava I IabIe
Not aval table
Overhead
Rlchards-WIIcox
Not avaI(able
Not avaliable
Electr Ic
Young Brothers
Not ava I (able
Not aval IabIe
Overhead
R I chards-WI I cox
Not ava liable
Not ava I lable
Electrl c
Young Brothers
Not a va I (able
Not aval I a bIe
(1) No paint heaters were required for these tests
(2) Model 21 equivalent to Model 610 except made of different materials. All Interior
dimensions identical.
-------
TABLE 22. SUMMARY OF PAINT PROPERTIES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
Paint Type
Resin Type
Manufacturer
Manufacturer's Paint ID No.
Lot No.
Color
Recommended Cure Schedule
Viscosity (uncut)
Reducing Solvent
% Reduction to Spray
Viscosity - Spray
wt/gal - Spray
wt Solids - Spray
Resistance or Conductance
Paint 1
Enamel
Polyester
PPG
W4533D
Not available
Beige
600 s @ 177°C
(10 min e 350°F)
Not available
Solvesso® 100
Not available
20.5 s (2)
Not available
66.8% (4)
2.5 Mf2
Paint 2
Enamel
Acrylic
PPG
DXH82445
Not available
Gold
1800 s @ 121°C
(30 min @ 250°F)
Not available
DxD reduce (1)
Not available
24 s (3)
Not available
51.4% (5)
1.5 Mfi
(1) DxD reducer composition: 50% acetone, 30% Cellosolve® acetate,
12% toluene, 8% Solvesso® 150.
(2) #3 Zahn cup @ 24°C (76°F).
(3) #2 Zahn cup @ 25°C (77°F), converts to 22 sec on Ford #4.
(4) Mean value for 32 weight dish samples.
(5) Mean value for 15 wt dish samples.
69
-------
the targets. Weight percent solids determinations were made
between each test series, and at the start of each day.
During this test there was an insufficient supply of small
aluminum dishes for the paint weight percent solids determination,
The ASTM method was compared with Ransburg's own method using
the remaining dishes. The two methods differ only in the
vehicle the paint is put into for drying and in that Ransburg
uses no solvent. The Ransburg method fabricates a small dish of
aluminum foil and places the paint in it as a thin film for
drying. The two techniques yielded almost identical results,
and since the subsequent TE's fell well within our method
standard deviation, the paint samples were not rerun.
TEST PARAMETERS
Based on the results of Phase I testing (refer to Section
5), the standard deviation for this test method was determined
to be 2.5. Based on an allowable difference of 2 (absolute)
between the measured mean TE and population mean TE, this
standard deviation indicates a sample size of six observations
for each test series. Table 23 provides the basis for this
estimation.
A test matrix was designed to provide six runs for each of
12 configurations. The 12 configurations are described in
Table 24. Each run has both the vertical cylinder and flat
panel targets, for a total of eight foils per run (one observa-
tion is the average of four foils).
TEST SEQUENCE (1)
The actual test sequence was:
March 18 - Set up equipment, check out scales
March 21 - Began TE test runs, completed HSB-67-FPVC 1-3
March 22 - Ran AAE-67-FPVC 1-6
(AAE-67-FPVC-6 used the original VC
wrapping technique)
March 23 - Ran AAE-51-FPVC 1-8
March 24 - Ran HSB-51-FPVC 1-6
Ran HSB-67-FPVC 3-6
(1) See Table 7 for nomenclature.
70
-------
TABLE 23. ESTIMATION OF REQUIRED TE TEST SAMPLE SIZE
FOR FUTURE LABORATORY TESTS
Estimated Required
Allowable
Error, +/-(!)
0.5
1.0
1.5
2.0
2.5
3.0
0.5
4
1
1
1
1
1
for
1.0
16
4
2
1
1
1
Sample
Standard Deviation
1.5
36
9
4
3
2
1
2.0
62
16
7
4
3
2
Size
of
2.5
96
36
11
6
4
3
3.0
138
62
16
9
6
4
Basis:
o 95% level of confidence
o Sample standard deviation = population standard deviation
o Estimated Sample Size (n) = (1.96 x s/e)2
where s = standard deviation
and e = allowable error
(1) The allowable error is the size of the difference between
the means (measured mean TE and population mean TE) that
is considered acceptable.
71
-------
TABLE 24. TEST MATRIX FOR PHASE II LABORATORY TESTS
OF TE STANDARD TEST METHOD
Spray Coating
Equipment Wt. Solids
51
Air Atomized
Electrostatic
67
51
Air Atomized
Conventional
67
51
High Speed
Bell
67
Target
Configuration
Vert. Cyl.
Flat Panel
Vert. Cyl.
Flat Panel
Vert. Cyl.
Flat Panel
Vert. Cyl.
Flat Panel
Vert. Cyl.
Flat Panel
Vert. Cyl.
Flat Panel
Test Series No. of
Nomenclature Observations (1)
AAE-51-VC
AAE-51-FP
AAE-67-VC
AAE-67-FP
AAC-51-VC
AAC-51-FP
AAC-67-VC
AAC-67-FP
HSB-51-VC
HSB-51-FP
HSB-67-VC
HSB-67-FP
6
6
6
6
6
6
6
6
6
6
6
6
(1) One observation is the average of four target foils.
72
-------
March 25 - Ran AAC-67-FPVC 1-6
Ran AAC-67-FPVC 1A-4A using original
VC wrapping technique
March 28 - Ran AAC-51-FPVC 1-6
Clean up
FOIL HANDLING PROCEDURES
The original vertical cylinder (VC) wrapping technique was
cumbersome during the PPG test. For the second test, a candy-
cane wrapping technique for VC's was proposed. The foil was to
be secured at each end by 0-rings (Figure 9).
The recommended candy-cane wrapping technique was tried for
several preliminary runs. This method also was cumbersome and
it allowed underspray through the seams onto the VC post. It
was abandoned in favor of Ransburg's own technique.
Ransburg cuts foil to the length of the target, then hand
wraps the foil vertically around the length of each cylinder
(Figure 5). The foil is crimped into place by hand. No tape or
0-rings are required.
This VC wrapping technique was used throughout the Ransburg
tests. Four runs (16 cylinders) were made using the original
wrapping technique for comparison.
QA/QC PROCEDURES
All quality assurance/quality control procedures in the
approved Test Plan were followed (See Section 4). In addition to
the recommended practices, it is recommended that all associated
pressure gages be calibrated before testing. Calibration
devices are usually standard equipment for testing laboratories
but may be obtained at a reasonable cost from suppliers.
Other recommendations are detailed in Section 9.
TEST RESULTS
General
Transfer efficiency measurements were made for a total of
12 test series. These series encompassed a test matrix of three
types of spray equipment, two coatings, and two target configura-
tions. Six runs of eight foils (four on vertical cylinder, four
on flat panel) were coated to determine transfer efficiency,
constituting each test series. The rationale for selecting a
73
-------
VERTICAL CYLINDER TARGET
oaring to secure foil
3.2cm (1-1/4 inch)
diameter aluminum pipe,
wrapped candy-cane style
with 6 in.s 1»5 mil, foil
Top View of Cylinder
secure foil with double-
sided tape until 0 ring
is installed
double sided tape to secure
foil before 0 ring is placed
bottom O ring
FLAT PANEL TARGET
15.24cm(6 in.)-*!
foil
\
Flat Panel Target
X
^
s"
h-
s
s
^ douh
\s"
)le -sided tape
Figure 9. Foil Attachment Techniques for Vertical
Cylinder (VC) and Flat Panel (FP) Targets
Proposed for Ransburg Test
74
-------
sample size of six observations for each test series was dis-
cussed earlier. Forty-eight measurements per test series were
obtained by coating eight foils per target assembly for each of
the six runs. Table 25 summarizes the results of the tests
based on evaluating the data under the following two scenarios:
• Each test series is composed of 24 observations from
six groups of four foils each.
• Each test series is composed of six observations; each
observation is the average of a group of four foils.
These two different scenarios for evaluating the data are
shown to reconfirm the original assessment of whether or not a
group of four foils should be treated as four separate observa-
tions or as one observation that is the average of four foils.
As outlined in Section 5, treating the test data for each test
series as two observations composed of the averages of four
foils each is a conservative approach. Realistically, the data
obtained in the tests are equivalent to a sample size somewhere
between 6 and 24. The more conservative approach (data treated
as six observations of four foils each) is the preferred
technique for estimating the overall precision of the test
method.
In both scenarios, the measured mean TE values are equal.
However, the standard deviations and coefficients of variation
are different. For example, the measured mean TE values ranged
from a low of 14.7 percent for the air atomized conventional gun
spraying 67-percent solids paint at a vertical cylinder target,
to a high of over 100 percent (this is explained later) for a
high speed rotating bell delivering 67-percent solids paint to a
vertical cylinder target.
For the standard deviations, however, the two different
data handling scenarios produced noticeably different results as
expected. In the first case (data treated as 24 individual
observations), the standard deviations ranged from 0.8 to 5.2.
The standard deviations in the second case (data treated as six
observations of four foils each), ranged from 0.2 to 2.4.
Similarly, the coefficients of variation, which are a measure of
the relative precision among test series, ranged from 1.8 to 9.5
percent for the first case and 1.0 to 3.4 percent for the second
case.
Having calculated the standard deviations for each test
series, the analyses proceeded to an evaluation of the pool-
ability of the variances to determine if a statement could be
made concerning the overall precision of the test method.
Bartlett's Test was used to examine the homogeneity of the
variances of the test series.
75
-------
From the Bartlett's Test it was found that the variances of
all 12 test series could not be considered homogeneous for
either of the cases in Table 25. The analyses then proceeded to
an examination of the following various groups of data:
• Compare poolability of variances for TE's obtained from
the same spray equipment type (4 TE's = 2 targets x 2
coatings); 3 gun groups to be evaluated.
• Compare poolability of variances for TE's obtained for
the same target configuration (6 TE's = 3 spray equip-
ment x 2 coatings); 2 target groups to be evaluated.
• Compare poolability of variances for TE's obtained for
the same coating (6 TE's = 3 spray equipment x 2 tar-
gets); 2 coatings groups to be evaluated.
• Several combinations of the first group.
These analyses were performed for only the second case,
i.e., treating the data for each test series as six samples of
four foils each. The results of these analyses are shown in
Table 26. The pooled standard deviations ranged from 1.5 to
2.0; however, the pooled standard deviation for the maximum
sample size of 48 was 1.8. This value is lower than the 2,5
estimated from the Phase I laboratory test.
The results on Table 26 do not include HSB-67-FPVC-2,
which was eliminated through an outlier analysis (ASTM E180).
Confirmation runs using the original VC wrapping technique are
also not included.
Air Atomized Electrostatic Test Results
Table 27 presents the results of six test runs using air
atomized electrostatic paint spray equipment. Transfer effi-
ciencies for flat panel targets ranged from 87.4 (AAE-51-FP-6)
to 93.4 (AAE-51-FP-2) percent. The standard deviation was 1.7
to 2.4. The relative precision of these runs as expressed by
coefficient of variation was 1.9 to 2.6 percent.
Vertical cylinder TE's ranged from 50.4 (AAE-67-VC-6) to
63.5 (AAE-51-VC-5) percent. The standard deviation was 1.8 to
2.1. The coefficient of variation was 3.4 to 3.5 percent.
A summary of equipment operating conditions for the air
atomized electrostatic test series is shown in Table 28.
76
-------
TABLE 25. SUMMARY OF TE TEST RESULTS
RANSBURG TEST
Test Series
AAE-51-VC
51-FP
67-VC
67-FP
AAO51-VC
51-FP
67-VC
67-FP
HSB-51-VC
51-FP
67-VC
67-FP
RANGE
TEM, %
61.3
90.7
51.5
90.3
15.8
76.5
14.7
75.5
103.0
.101.6
101.0
102.1
CASE 1:
24 indiv
S
2.3
3.7
1.9
2.6
0.8
3.3
1.4
3.2
5.2
1.8
4.7
2.0
0.8-5.2
Eata Treated as
. observations
COV,%
3.8
4.1
3.7
2.9
5.1
4.3
9.5
4.2
5.0
1.8
4.6
2.0
1.8-9.5
CASE 2: C&ta
observations
S
2.1
2.4
1.8
1.7
0.2
0.8
0.5
2.1
1.5
1.4
1.5
1.6
0.2-2.4
treated as 6
of 4 foils ea.
COV,%
3.4
2.6
3.5
1.9
1.3
1.0
3.4
2.8
1.5
1.4
1.5
1.6
1.0-3.4
NOTES:
TEM = Mean Transfer Efficiency
S = Standard deviation = [TEi-TEM)2 + (TE2-TEM)2 + ...]°*5
COV = Coefficient of Variation (S/TEM) * 100%
77
-------
TABLE 26. RESULTS OF BARTLETTS'S TEST FOR EVALUATION OF POOLABILITY
OF VARIANCES AT 95% LEVEL OF CONFIDENCE
PHASE II (RANSBURG)
Test Series
Evaluated
Pooled
No. of No. of Standard
Test Series Observations(1) Deviation,%
I. All Test series
II. Groups of test series
A. Same spray equipment type
12
72
Not poolable
1. Air atomized electrostatic
2. Air atomized conventional
3. High speed bell
B. Same target configuration
1. Vertical cylinders
2. Flat panels
C. Sane coating
1. Paint 2 (51% wt)
2. Paint 1 (67% wt)
III. Combinations of groups
A. Group Al & A2
B. Group Al & A3
C. Group A2 & A3
4
4
4
6
6
6
6
8
8
8
24
24
24
36
36
36
36
48
48
48
2.0
Not poolable
1.5
Not poolable
1.7
Not poolable
1.6
Not poolable
1.8
Not poolable
(1) Data treated as six observations of four foils each per test series.
78
-------
TABLE 27. TRANSFER EFFICIENCY DATA, RANSBURG LABORATORY TEST,
MARCH 21-29, 1983 (1)
Run 1
Faint
Ut »
Solids
AAE
51
51
67
67
AAC
51
51
-J 67
VO
67
HSB
51
51
67
67
foil No.
Target
vc
FP
vc
FP
vc
FP
vc
FP
VC
FP
VC
FP
1
63.3
88.6
50.9
87.6
16.5
75. B
15.1
70.4
106.6
102.2
103.2
102.5
2
60.6
92.6
52.0
92.0
14.7
76.7
13.7
7J.4
103.1
101.4
96.8
99.3
3
62.3
94. U
51.6
91.6
15.2
72.2
15.1
72.2
99.3
101.7
98.6
98.9
4
63.2
92.1
54.0
90.1
16.4
75.0
13.9
76.0
110.9
103.8
103.5
100.5
TE1 (2) 1
62.4
92.0
52.1
90. J
15.7
74.9
14.5
73.0
105.0
102.3
100.5
100.3
63.2
85.9
52.7
90.3
17.4
68.5
16.9
71.5
104.1
100.4
Run 2
Foil No.
2
61.2
92.4
54.3
94.2
14.9
78.0
14.1
78.7
97.2
98.0
3
61.0
98.5
54.1
95.9
15.7
16.1
14.1
78.9
96.1
98.6
OUTLIER
4
64. U
96.8
56.0
93.0
15.7
82.8
15.5
82.3
108.2
101.8
TE2 (2)
62.6
93.4
54.3
93.3
15.9
76.9
15.2
77.9
101.4
99.7
1
61.2
94.3
52.1
87.2
17.2
79.5
17.7
75.0
105.7
100.6
101.1
103.3
Run 3
Foil No.
2
60.9
94. J
51.4
91.5
15.7
80.0
15.2
77.9
93.4
98.9
102.5
101.0
3
61.2
92.0
50.9
92.3
15.1
74.9
13.4
72.1
96.3
98.5
95.3
99.4
4
63.7
88.8
53.1
90.1
16.1
74.6
15.4
78.0
106.4
101.5
105.1
103.5
TE3 (2)
61.8
92.4
51.9
90.3
16.0
77.3
15.4
75.8
101.7
99.9
101.0
101.8
1
60.4
86.8
48.7
85.6
16.3
73.8
16.3
73.8
106.6
102.9
104.3
102.9
Run 4
Foil No.
2
57.5
90.5
4U.2
89.9
15.8
77.7
14.4
76.1
99.5
102.1
95.1
102.0
3
59.4
93.5
48.8
91.1
14.5
76.1
13.6
73.9
97.3
100.7
94.4
101.7
4
61.2
91.6
50.4
88.6
16.1
78.1
15.8
77.5
109.1
103.5
106.4
104.2
TC4 (2)
59.6
90.6
49.0
88.8
15.7
76.4
15.0
75.3
103.1
102.3
100.0
102.7
-------
TABLE 27 (continued)
CO
O
Ha i nt
Wt %
solids
AAE
51
51
67
67
AAC
51
51
67
67
MSB
51
51
67
67
Run 5
Foil No.
Target
VC
FP
VC
FP
VC
FP
VC
FP
VC
FP
VC
FP
1
63.5
84.9
49.5
84.7
16.0
70.8
12.7
78.0
108.2
103.1
103.0
102.0
2
61.4
89.0
51.7
90.2
15.7
78.7
17.3
80.9
101.1
101.4
95.6
99.9
3
63.5
91.0
51.3
90.6
15.4
77.7
13.4
77.2
111.3
101.2
94.1
100.4
4
65.7
88.6
51.6
88.3
14.8
81.0
13.4
74.6
97.6
103.2
105.8
101.3
TC5 (2)
63.5
8U.4
51.0
88.5
15.5
77.0
14.2
77.7
104.6
102.2
99.6
100.9
1
57.5
83.6
50.0
88.2
17.3
72.9
15.5
75.1
103.8
104.6
106.6
105.3
Run
Foil
2
57.6
87.5
50.4
90.2
14.9
77.8
13.1
75.1
97.5
102.1
100.3
102.7
6
No.
3
58.6
90.3
50.3
93.2
15.4
74.9
13.6
70.6
97.0
104.0
99.5
104.2
4
58.8
88.3
51.0
89.9
15.9
79.9
14.7
71.9
110.2
101.3
109.2
107.0
TE6 (2) 1
58.1
87.4
50.4(6)
90.4
15.9
76.4
14.2
73.2
102.1
103.0
103.9 103.4
104.8 103.6
Run 7
Foil No.
2 3 4 TE7 (2) TEm (3) S (4) COV (5)
61.3 2.1 3.4
90.7 2.4 2.6
51.5 1.8 3.5
90.3 1.7 1.9
15.8 0.2 1.3
76.5 0.8 1.0
14.7 0.5 3.4
75.5 2.1 2.8
103.0 1.5 1.5
101.6 1.4 1.4
94.9 97.0 108.1 101.0 101.0 1.5 1.5
99.5 101.2 103.3 101.9 102.1 1.6 1.6
(1) TE calc'd to 0.1%, statistics calc'd to 0.1%. (4) S= standard deviation = [(TEl-TUn)2 + (TE2^rE,)2 + ...j
(2) TEn= arithjietic avg o£ foils (l-4)n. (5) COV= coefficient of variation = S/TEm X 100.
(3) Tthi= arithmetic avy of all TEn. (6) This run using Tape Wrap Tech.
-------
TABLE 28. AAE PAINT SPRAY AND PERIPHERAL
EQUIPMENT SPECIFICATIONS
A. Paint Supply Tank
1. Type
2. Manufacturer
3. Model No.
4. Serial No.
5. Rated Capacity, gallons
Pressure Pot
Binks
83-5508
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
Electrostatic Hand Gun
Nordson
AN-8
987
228
C. Paint Spray Booth
1. Type
2. Manufacturer
3. Model No.
4. Serial No.
5. Rated Capacity, cfm
Open
D. Conveyor
1. Type
2. Manufacturer
3. Model No.
4. Serial No.
Over-Head
Richards-Wilcox Mfg,
Forced Draft Oven
1. Type
2. Manufacturer
3. Model No.
4. Serial No.
Paint Heaters
1. Type
2. Manufacturer
3. Model No.
4. Serial No.
Electric
Young Brothers
81
-------
TABLE 28 (continued)
AAE EQUIPMENT OPERATING CONDITIONS
A. Paint Spray Equipment
1. Paint Pressure at Paint Pot, kPa (psig) 5.32 (22)
2. Paint Pressre at Spray Gun, kPa (psig) N/A
3. Atomizing Air Pressure at Spray Gun,
kPa (psig)
4. Operating Voltage, kV 63
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 (in) 30.5 (12)
B. Paint Spray Booth
1. Ambient Temperature, °C (°F) 24 (76)
2. Relative Humidity, % 19
3. Air Flow Velocity, ra/s (fpm) 0.5 (100)
4. Air Flow Direction Normal to target
C. Target Parameters
1. Average Wet Film Thickness, cm (mils) 5xlO~3 (2.0)
2. Vertical Paint Coverage
3. Target Height, cm (in) .
4. % Vertical Coverage 45 Jest)
5. Resistance to Ground, Ohm >10b
D. Forced Draft Oven (1)
1. Cure Time, minutes
a. Foil Dish (sample) 67% 600 s @ 177°C (10 min at 350°)
b. Target Foil 51% 1800 s @ 121°C (30 min at 250°)
2. Cure Temperature
a. Foil Dish (sample) See above
b. Target Foil ^
Paint Heaters
1. Temperature In
2. Temperature Out
F. Conveyor Speed Setpoint, cm/s (fps) 15.8 (0.52)
(1) The same cure schedules were used for foils and dishes.
82
-------
Air Atomized Conventional Test Results
Table 27 summarizes the results of all six test series for
air atomized conventional paint spray equipment. Flat panel
transfer efficiencies ranged from 73.0 (AAC-67-FP-1) to 77.9
(AAC-67-FP-2). Vertical cylinder transfer efficiencies ranged
from 14.2 (AAC-67-VC-5 and 6) to 16.0 (AAC-51-VC-3). The
standard deviations were 0.2 to 2.1, indicating good precision.
In addition to the six scheduled runs, four additional runs
were made using the original vertical cylinder wrapping tech-
nique (Figure 7) with conventional flat panel targets accompany-
ing as controls. These runs are compared with the new (Ransburg)
wrapping method in Section 7.
A summary of operating conditions for the air atomized
conventional equipment is presented in Table 29.
High Speed Bell Test Results
Table 27 summarizes the results from seven test runs on
the high speed bell paint spray equipment. Test conditions are
summarized on Table 30. One set of runs (HSB-67-FP-VC-2)
was eliminated by an outlier analysis, leaving six complete
runs. Transfer efficiencies for flat panel targets range from
99.7 (HSB-51-FP-2) to 104.8 (HSB-67-FP-6), with a standard
deviation of 1.4 to 1.6. The coefficient of variation is low,
from 1.4 to 1.6 percent.
Transfer efficiencies for vertical cylinder targets range
from 99.6 (HSB-51-VC-1) to 105.0 (HSB-51-VC-1), with a standard
deviation of 1.5. The coefficient of variation is a very low
1.5 percent.
It is obvious that transfer efficiency cannot exceed 100
percent. The HSB data points to a consistent system or method
error in determining TE. When TE's began appearing at 100+
percent, every possible physical testing error was examined.
Several more weight percent solids determinations were made,
scales were recalibrated, and new runs were made. No equipment
error was found, but an explanation for the high TE's was
discovered .
The original test method called for 400 g of paint to be
sprayed before passing the targets in front of the gun. In the
Ransburg test about 100 g was sprayed before the target was
passed through. (Paint was in short supply.) When the 100 g
spray was initiated, the targets were about 5 feet down the
conveyor in the open booth, moving towards the gun. Some of
this paint had the capability of remaining suspended long enough
to eventually coat the targets considering the electrostatic
83
-------
TABLE 29. AAC PAINT SPRAY AND PERIPHERAL
EQUIPMENT SPECIFICATIONS
A. Paint Supply Tank
1. Type
2. Manufacturer
3. Model No.
4. Serial No.
5. Rated Capacity, gallons
Pressure Pot
Binks
83-5508
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
Conventional Air Spray
Binks
21
36113
63 PB
63C
C. Paint Spray Booth
1. Type
2. Manufacturer
3, Model No.
4. Serial No.
5. Rated Capacity, cfm
Open
D. Conveyor
1. Type
2. Manufacturer
3. Model No.
4. Serial No.
Over-Head
Richards-Wilcox Mfg.
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.
Electric
Young Brothers
84
-------
TABLE 29 (continued)
AAC EQUIPMENT OPERATING CONDITIONS
3.73 (11)
453 (51)
0
30.5 (12)
A. Paint Spray Equipment
1. Paint Pressure at Paint Pot, kPa (psig)
2. Paint Pressre at Spray Gun, kPa (psig)
3. Atomizing Air Pressure at Spray Gun,
kPa (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 (in)
B. Paint Spray Booth
1. Ambient Temperature, °C (°F)
2. Relative Humidity, %
3. Air Flow Velocity, m/s (fpm)
4. Air Flow Direction
C. Target Parameters
1. Average Wet Film Thickness, cm (mils)
2. Vertical Paint Coverage, cm (in)
3. Target Height, cm (in)
4. % Vertical Coverage
5. Resistance to Ground, Ohm
D. Forced Draft Oven (1)
23 (73)
25
0.5 (100)
Normal
7.6xlO-3 (3.0)
(16)
varied
45
1. Cure Time, minutes
a. Foil Dish (sample)
b. Target Foil
67%-60Q s @ 177°C (10 min at 350°F)
51%-1800 s § 121°C (30 min at 250°F)
2. Cure Temperature
a. Foil Dish (sample)
b. Target Foil
E. Paint Heaters
1. Temperature In
2. Temperature Out
F. Conveyor Speed Setpoint, cm/s (fpm)
See above
16.0 (0.52)
(1) The same cure schedule was used for foils and dishes.
85
-------
TABLE 30. HSB PAINT SPRAY AND PERIPHERAL
EQUIPMENT SPECIFICATIONS
A. Paint Supply Tank
1. Type DC Pump
2. Manufacturer Ransburg
3. Model No. 9966-01
4. Serial No. ___________
5. Rated Capacity, gallons _____________
B. Paint Spray Equipment
1. Type High Speed Bell
2. Manufacturer Ransburg
3. Model No. 20 865-04
4. Serial No.
5. Rated Capacity, cc/min
6. Air Cap _______________
7. Fluid Tip _____________
8. Needle
Paint Spray Booth
1. Type Open
2. Manufacturer ________
3. Model No. _,______,
4. Serial No. ________
5. Rated Capacity, cfm (ft3/min) _______
D. Conveyor
1. Type Over-Head
2. Manufacturer Richards-Wilcox Mfg,
3. Model No. _______________________
4. Serial No.
Forced Draft Oven
1. Type Electric
2. Manufacturer Young Brothers
3. Model No. .
4. Serial No.
Paint Heaters
1. Type
2. Manufacturer
3. Model No.
4. Serial No,
86
-------
TABLE 30 (continued)
HSB EQUIPMENT OPERATING CONDITIONS
A.
B.
C.
D.
Paint Spray Equipment
1. Paint Pressure at Paint Pot, kPa (psig
2. Paint Pressre at Spray Gun, kPa (psig)
3. Atomizing Air Pressure at Spray Gun,
kPa (psig)
4. Operating Voltage, kV
5. Disk or Bell Speed, rps (rpm)
a. With Paint Applied
b. Without Paint Applied
6. Shaping Air for Bell, kPa (psig)
7. Paint Temperature at Paint Pot, °C ( 8F
8. Gun to Target Distance, cm (in)
9. Pump Setting
Paint Spray Booth
1. Ambient Temperature, °F
2. Relative Humidity, %
3. Air Flow Velocity, m/s (rpm)
4. Air Flow Direction
Target Parameters
1. Average Wet Film Thickness, cm (mils)
2. Average Dry Film Thickness
2. Vertical Paint Coverage, cm (in)
3. Target Height, cm (in)
4. % Vertical Coverage
5. Resistance to Ground, Ohm
Forced Draft Oven (1)
1. Cure Time, minutes
a. Foil Dish (sample) 67%-600 s @ 177
b. Target .Foil 51%-1800 s § 121
2. Cure Temperature, °F
a. Foil Dish (sample)
b. Target Foil
)
-
205 (15)
80
167 (10000)
333 (20000)
308 (30)
)
28 (11)
1.45
75°
18
0.5 (100)
Normal to target
S.lxlO-3 (2.0)
1.2
81 (32)
152 (60)
53
<1 million
°C (10 min at 350°F)
°C (30 min at 250°F)
See above
E. Paint Heaters
1. Temperature In, °F
2. Temperature Out, °F
F. Conveyor Speed Setpoint, cm/s (fpm)
16 (0.52)
(1) Same cure schedule for foils and dishes.
87
-------
nature of the process. In the original test, the targets were
outside the closed booth while 400 g were sprayed. Thus, the
targets in the Ransburg test were probably exposed to more paint
than was measured.
Table 31 represents an approximation of the TE's when cor-
rected for the paint sprayed 2 feet before the first timing
mark. This approximation is based on observations during the
tests. As can be seen, the correction brings HSB TE's into a
range more comparable to Phase I.
Based on the correction, HSB flat panel TE's were 88,7 to
93.2 percent with standard deviation of 1.2 'to 1.5. The co-
efficient of variation was 1.3 to 1.7 percent. Vertical
cylinder TE's were 88.9 to 93.3 percent with standard deviation
of 1.4. The VC coefficient of variation was 1.5 to 1.6 percent.
88
-------
TABLE 31. TRANSFER EFFICIENCY DATA
CO
vo
Faint
wt %
Sol ids
51
51
67
67
Target
VC
FP
VC
FP
1
96.2
91.7
91.6
90.7
Run 5
Foil No.
234
89.8 98.9 87.8
90.1 89.9 91.7
85.0 83.7 94.0
88.8 89.2 90.0
Ransburg Laboratory Test March 21-29, 1983
Corrected HSU Transfer Efficiency I&ta (1)
Run 6 Run 7
Foil No.
TE5 (2)
93.2
90.9
88.6
89.7
1
92.2
93.0
94.8
93.6
2
86.6
90.8
89.2
91.3
3
86.2
92.5
88.5
92.7
4 TE6 (2) 1
97.9 90.7
90.0 91.6
97.1 92.4 92.0
95.1 93.2 92.1
Foil No.
2 3 4 TE7 (2) TEm (3)
- - - 91.6
- 90.3
84.3 86.2 96.1 89.7 89.8
88.5 90.0 91.8 90.6 90.8
S (4) COV (5)
1.4 1.5
1.2 1.3
1.4 1.6
1.5 1.7
(1) These calculations assume that the paint sprayed 2 ft prior to target arrival has the capability of staying dispersed long enough to attach
to the targets. Constant flow rate and conveyor speed are assumed tor the additional 2 ft. (See Section 7 for further discussion.)
TE calc'd to 0.1», statistics calc'd to 0.1%.
(2) TBn= arithmetic avg of foils (l-4)n.
(3) TEm= arithmetic avg of all TEn.
(4) S= standard deviation « l(TEl-TEm)2 + (TE2-TEm)2 + ...)0.5
(5) COV= coefficient of variation = S/TEm X 100.
-------
TABLE 31 (continued)
Paint
hit $
Solids Target 1
51 VC 96.2
51 t'P 91.7
67 VC 91.6
67 fP 90.7
Ransbury Laboratory Test March 21-29, 1983
Corrected HSH Transfer Efficiency L&ta (1)
Run 5 Run 6 Run 7
Foil No. Foil No. Foil No.
234 TE5 (2) 1 2 3 4 TE6 {2) 1 2 3 4 TE7 (2> TCm (3) a (4) CUV (5)
89.8 98.9 87.8 93.2 92.2 86.6 86.2 97.9 90.7 ----- 91.6 1.4 1.5
90.1 89.9 91.7 90.9 93.0 90.8 92.5 90.0 91.6 ----- 90.3 j.2 1.3
85.0 83. 7 94.0 88.6 94.8 89.2 88.5 97.1 92.4 92.0 84.3 86.2 96.1 89.7 89.8 1.4 1.6
a«.8 89.2 90.0 89.7 93.6 91.3 92.7 95.1 93.2 92.1 88.5 90.0 91.8 90.6 90.8 1.5 1.7
(1) These calculations assume that the paint sprayed 2 ft prior to target arrival has the capability of staying dispersed long enough to attach
to the targets. Constant flow rate and conveyor speed are assumed for the additional 2 ft. (See Section 7 for further discussion.)
TB calc'd to 0.1», statistics calc'd to 0.1».
(2) TEn= arithjiEtic avy of foils (l-4)n. (4) S= standard deviation = ((TEl-TEm)2 * (TE2-TEro)2 + ...)0-5
(3) Ttln= arithmetic avg of all TEn. (5) COV= coefficient of variation » S/fEm X 100.
-------
SECTION 7
TEST COMPARISON
EFFECT OF FOIL WRAP TECHNIQUE - VERTICAL CYLINDER
The method used to attach the foil to the vertical cylinder
targets during Phase I testing proved highly unsatisfactory. As
discussed in Section 5, the foil could be removed from the
double-sided tape without losing paint only with great diffi-
culty. Therefore, it was decided that the "tape wrap" method
would not be used in subsequent tests. Instead the "crimp"
method described in Section 6 was used in Phase II.
It was decided to conduct a series of tests to compare
transfer efficiency results obtained with each method of foil
attachment. The objective of these tests was to provide data
necessary for the determination of the extent to which the foil
attachment method influenced the resulting transfer efficiency,
if at all. The t-test of the difference between two means was
then applied to the resulting data as follows.
The air atomized conventional (AAC) spray gun was selected
for this purpose in order to minimize the number of observations
required to detect a small difference in transfer efficiency.
The standard deviation of the transfer efficiency results ob-
tained during Phase I testing using the AAC spray gun (taking
each cylinder as an individual observation) had been found to be
the least of any spray apparatus.
Thus, using the AAC configuration, four runs of four
cylinders each were performed as a special test using the tape
wrap method of foil attachment. Each of the four runs was made
under the same conditions using the air atomized conventional
gun with the high solids paint. Resulting transfer efficiencies
are given in Table 32.
91
-------
TABLE 32. TESTS TO DETERMINE EFFECT OF FOIL ATTACHMENT METHOD
AAC-67-VC: TAPE WRAP METHOD
Transfer Efficiency
Foil No.
Run No.
TE
1A
2A
3A
4A
14.0
15.1
16.8
15.1
12.5
13.9
14.3
13.7
13.1
13.8
11.8
12.2
13.6
13.8
15.3
13.0
13.4
14.2
14.6
13.5
These results may be compared with the following values,
reproduced here from Table 27, showing transfer efficiencies
resulting from tests carried out under identical conditions
using the crimp method of foil attachment.
TABLE 33. TESTS TO DETERMINE EFFECT OF FOIL ATTACHMENT METHOD
AAC-67-VC; CRIMP METHOD(l)
Transfer Efficiency
Foil No.
Run No.
TE
1
2
3
4
5
6
15.1
16.9
17.7
16.3
12.7
15o5
13.7
14.1
15.2
14.4
17.3
13.1
15.1
14.1
13.4
13.6
13.4
13 06
13.9
15.5
15e4
15.8
13.4
14 c 7
-n
14.5
15.2
15.4
15oO
14.2
14,2
(1) From Table 27.
The standard deviation of the 16 observations using the
tape wrap method (Table 32) is 1.25. The standard deviation of
the 24 observations using the crimp method is 1.37. The ratio
of variances (F-statistic) then is 1.19. Since the critical
value of F at the 0.1 level, for 23 and 15 degrees of freedom,
is 1.9, these variances are homogenous and therefore poolable.
The pooled standard deviation may be shown to be 1.32.
92
-------
Given a standard deviation of 1.32, the number of observa-
tions required for the t-test to detect a difference of 2 while
controlling both the a risk and the /5 risk to the 0.05 level
is 13. Eighteen observations are required to reduce the
risk to 0.01. With 16 observations of the tape wrap method and
24 of the crimp method, the /3 risk may be said to be controlled
at a level below 0.05 when the t-test is performed at the 0.05
level. The oi risk is the risk of deciding that two samples
were drawn from two different populations when in reality they
came from the same population. The ft risk, perhaps more
crucial for the present consideration, is the risk of deciding
that the samples are drawn from the same population when in
reality there is a real difference of at least the magnitude
specified.
The value of the t-statistic is given by:
TE (crimp) - TE (wrap)
t = — = 1.99
\l ±- + -L-
\f 24 + 16
with 38 degrees of freedom.
The critical value of t at the 0.05 level with 38 degrees
of freedom is 2.03. Thus, the value of t for this test is not
significant at the specified level. It is therefore concluded
that the transfer efficiency results from the two methods of
foil attachment do not differ significantly, with the risk of
their actually differing by 2 or more equal to less than 0.05.
This result is further verified by the single run (Run 6)
performed using the tape wrap technique in the AAE-HS-VC series
(see Table 27, footnote 6). That run resulted in a transfer
efficiency of 50.4, compared to the mean transfer efficiency for
all 6 runs taken together of 51.5.
COMPARISON OF INTERLABORATORY VARIANCES
The results from each of the two participating laboratories
may be compared to determine if a significant difference in test
precision exists between laboratories. Table 34 shows the
standard deviations at each laboratory for each test series,
treating each target as a separate observation.
Davies, 0. L., The Design and Analysis of Industrial Experi-
ments, 2nd Edition, Table E-l, pp 609-611. Imperial Chemical
Industries Ltd., London.
93
-------
The F-statistic is equal to the ratio of the variances, the
squares of the standard deviations. Degrees of freedom associ-
ated with numerator and denominator are 23 and 7, or 7 and 23,
depending on whether Phase I results or Phase II results ex-
hibited the greater variance. The test for significance is a
two-tailed test, because the question being tested is whether
the precision at either laboratory is significantly different
(greater or less) than the precision at the other.
TABLE 34. COMPARISON OF VARIANCES FOR TESTS AT EACH LABORATORY (1!
AAE-51-VC
* 51-FP
67-VC
67-FP
AAC-51-VC
* 51-FP
* 67-VC
67-FP
HSB-51-VC
51-FP
67-VC
67-FP
ST(8)
1.7
1.3
1.6
1.9
0.7
1.2
0 = 6
1.6
2,8
1.9
5.2
1 = 1
SIl(24)
2.3
3.7
1.9
2.6
0.8
3o3
1.4
3,2
5.2
1.8
4.7
2.0
F
1.83
8.10
1.41
1.87
1.31
7.56
5.44
4.00
3 = 45
1.11
1.22
3.31
F.05
4.4
4.4
4.4
4.4
4.4
4.4
4.4
4 = 4
4.4
2.9
2,9
4.4
(1) Each Target Considered a Separate Observation.
* Significant at 0.05 level
Table 34 shows that variances for 3 test series are found
to differ significantly between laboratories. In each of these
three cases, the precision exhibited in the Phase II testing at
Ransburg is poorer. No significant difference at the 0.05 level
is found for the other nine test series.
The precision of the test at each laboratory was found to
be 2.5 standard deviation for Phase I and 1.8 standard deviation
for Phase II based on the largest number of poolable observa-
tions in each case. (For both laboratories it was found that
all air atomized electrostatic and all high speed bell results
could be pooled.) This pooling was performed using standard
deviations based on treating the mean of four targets as a sin-
gle observation. In contrast^ Table 34 is based on treating
each target as a separate observation, because the single degree
of freedom associated with Phase I results when treating four
targets as a single observation is not sufficient to provide for
a meaningful test.
94
-------
The pooled estimates of precision for each laboratory may
be tested to determine if-they differ significantly. The ratio
of variances is (2.5/1.8) = 1.93, with 15 and 47 degrees of
freedom. The critical value of F for a double-tailed test at
the 0.05 level is 2.12. Thus, the overall test precision
(repeatability) at each of the two laboratories is not signifi-
cantly different. Pooling the two estimates of standard devia-
tion gives an overall test precision of 2.0.
ANALYSIS OF VARIANCE FOR PAINT, LABORATORY
This section presents a two-way analysis of variance (ANOVA)
for the transfer efficiency results obtained using each combina-
tion of spray gun and target. The objective of the analysis is
the determination of the relative influence of laboratory and
paint type on transfer efficiency, and the estimation of intra-
laboratory precision (repeatability) and the interlaboratory
precision (reproducibility).
Table 35 presents values of transfer efficiency obtained
under each test condition. Considering a single observation to
be the mean of four targets, two observations were obtained for
each treatment during Phase I, and six observations for Phase
II, which constitutes an unbalanced experimental design.
TABLE 35. ANALYSIS OF VARIANCE - TRANSFER EFFICIENCY
Laboratory
II
Laboratory
I II
51
Paint
67
51
Paint
67
51
Paint
67
29.7
32.0
26.0
25.4
a.
10.9
10.7
10.2
10.2
c.
93.0
89.6
102.4
93.9
62.4
62.6
61.8
52.1
54.3
51.9
AAE-VC
15.7
15.9
16.0
14.5
15.2
15.4
AAC-VC
105.0
101.4
101.7
100.5
101.0
100.0
59.6
63.5
58.1
49.0
51.0
50.4
15.7
15.5
15.9
15.0
14.2
14.2
103.1
104.6
102.1
99.6
103.9
101.0
81.4
81.1
80.2
82.7
b.
61.4
62.1
60.8
60.8
d.
91.0
91.7
93.2
92.3
92.0
93.4
92.4
90.3
93.3
90.3
AAE-FP
74.9
76.9
77.3
73.0
77.9
75.8
AAC-FP
102.3
99.7
99.9
100.3
101.8
102.7
90.6
88.4
87.4
88.8
88.5
90.4
76.4
77.0
76.4
75.3
77.7
73.2
102.3
102.2
103.0
100.9
104.8
101.9
e. HSB-VC
f. HSB-FP
95
-------
Table 36 presents the ANOVA results in standard format.
Those effects significant at the 0.01 level are noted by an
asterisk. The column effect (laboratory) is significant for
every configuration. The row effect (paint) is significant^for
two of the six configurations, but considerably less significant
than the laboratory effect even in those cases. Interaction is
essentially absent, although it is just significant at the .01
level for the single instance of the high speed bell - vertical
cylinder case. Although there is no effect of paint for this
case, the fact that interaction is present means that there is a
real effect, but the effect is in opposite direction at each
laboratory. Having noted this, it is interesting that the paint
exhibits a real effect on transfer efficiency in all three
vertical cylinder target configurations and in none of the flat
panel target configurations.
Although the paints used for each phase of testing were not
split from identical 67 percent solids and 51 percent solids
batches, as would have been ideal, they were formulated to be
similar as delivered to each laboratory. Nevertheless, because
the physical properties of the two higher solids and the two
lower solids paints differ to some extent between laboratories,
there is concern that such a difference may be the major cause
of the difference observed in transfer efficiency results at
each laboratory. The analysis of variance presented here shows
this to be possible but unlikely. The effect of paint on
transfer efficiency, even when the paint weight percent solids
is purposely varied, has been shown to be either absent or to be
much less significant than the observed effect of the laboratory.
Thus the reasons for the difference in results between labora-
tories must be sought elsewhere, as is done in the next section.
The error mean square (sum of squares divided by degrees of
freedom) is an estimate of the variance associated with repeated
measurements under identical conditions. Although these vari-
ance estimates are not all poolable by Bartlett's test at the
Oe05 level, the variances associated with five of the six test
configurations (all except the air atomized conventional spray
with the vertical cylinder target) can be pooled. This pooled
variance estimate is 3.33. The precision estimate associated
with this variance is 1.82 (standard deviation), the test
repeatability.
Reproducibility, the precision of interlaboratory testing,
for these results is clearly not acceptable. The inter-
laboratory variance estimates (columns) shown in Table 36 are
large and are poolable, with a pooled variance of 630.6, corre-
sponding to a reproducibility estimate of 25 (standard devia-
tion) . The next section analyzes the causes of this poor test
reproducibility.
96
-------
TABLE 36. ANOVA RESULTS FOR PAINT AND LABORATORY
Rows
Cols
Int
Error
Total
Row
Col
Int
Error
Total
Row
Col
Int
Error
Total
Row
Col
Int
Error
Total
Row
Col
Int
Error
Total
Row
Col
Int
Error
Total
*Significant at
SS
302.
2371.
16.
40.
2731.
0.
250.
0.
45.
296.
3.
649.
0.
26.
679.
3.
68.
0.
1.
73.
0.
158.
58.
65.
282.
2.
286.
0.
22.
311.
.01 level
8
6
8
0
2
30
25
31
72
58
90
01
00"
14
05
42
16
15
54
27
20
40
54
15
29
10
16
61
90
77
•
AAE-VC
1
1
1
12
15
AAE-FP
1
1
1
12
15
AAC-FP
1
1
1
12
15
AAC-VC
1
1
1
12
15
HSB-VC
1
1
1
12
15
HSB-FP
1
1
1
12
15
MS
302.
2371.
16.
3.
0.
250.
0.
3.
3.
649.
0.
2.
3.
68.
0.
0.
0.
158.
58.
5.
2.
286.
0.
1.
8
6
8
3
30
25
31
81
90
01
00
18
42
16
15
13
20
40
54
43
10
16
61
91
F
91.
718.
5.
0.
65.
0.
1.
247.
0.
26.
524.
1.
0.
29.
10.
1.
149.
0.
75*
66
09
08
*
68*
08
79
90
00
30
30
15
04
17
78
10
*
*
*
*
*
82*
32
97
-------
TEST REPRODUCIBILITY
As presented in Sections 5 and 6, high precision was ob-
served for the results from each laboratory test. When inter-
laboratory results were compared, however, the reproducibility
was very poor. The combination of high individual precision with
low interlaboratory reproducibility pointed to systematic
differences between the two laboratory tests. The test proce-
dure, materials, and equipment for the two sets of tests were
methodically compared to define any systematic differences.
Three major differences between tests were documented. These
differences qualitatively account for the observed discrepancies
between laboratory results.
The first major difference between Phase I and Phase II
testing was the paint spray booth air flow rate. During Phase I
testing, the linear air velocity was measured up to 1.0 m/s
(200 fpm) in the plane of the target. The air flow in the plane
of the target in Phase II was undetectable. Directly in front
of the fan it was only 0.5 m/s (100 fpm). This is a large dif-
ference between Phase I and Phase II conditions.
At higher booth air velocity the paint may flow more
quickly past the targets towards the exhaust* Thus the atomized
paint has less opportunity to coat the targets at higher booth
air flow rates. This effect is observed in the significantly
higher TE's for all Phase II test configurations compared to
Phase I results (see Table 37).
TABLE 37- TRANSFER EFFICIENCY COMPARISON
vc
AAE-51-VC
AAE-67-VC
AAC-51-VC
AAC-67-VC
HSB-51-VC
HSB-67-VC
Average
FP
AAE-51-FP
AAE-67-FP
AAC-51-FP
AAC-67-FP
HSB-51-FP
HSB-67-FP
Average
Phase I
TEM
30.8
25,7
10.8
10.2
91.3
93.9
43.8
81.2
81.5
61.7
60e8
91.3
92.8
78.2
Phase II
TEM
61.3
51.5
15.8
14.7
103,0
101.0
57.9
90.7
90.3
76.5
75.5
101.6
102.1
89»5
Net Difference
%
99*0
94e2
46.3
44.1
12.8
7.6
50.7
11.7
10.8
24.0
24 . 2
11.3
10.0
15.3
98
-------
As confirmation of the nature of this phenomenon, the VC
target results would be expected to be more affected than FP
target results. The VC targets have open air space between them,
which allow paint to be drawn past each foil; the FP foils have
no interstice to allow easy air (and atomized paint) flow past
the foils. The interlaboratory data support this expectation,
with the VC results being 50.7 percent different and the FP
results only 15.3 percent different between laboratories (Table
37).
Another logical outcome of low booth air rates in Phase II
would be more of an effect for runs with electrostatic equip-
ment. The electrostatic attraction would help attract slow
moving atomized paint to the targets. This effect is seen for
AAC and AAE runs, but is obscured by other differences in the
HSB runs.
Interlaboratory test results for air atomized electrostatic
equipment varied more than any other test series (refer to
Table 37). Vertical cylinder TE's were dramatically more
affected than the flat panel TE's, pointing to a difference in
attractive forces (operating voltage). Although the same model
power supply was used in both laboratories, the operating volt-
age was about 47 kV in Phase I and 62 kV in Phase II. Higher
TE's (especially for VC targets) are the reasonable consequence
of higher voltages in the Phase II tests.
Finally, the test procedure was slightly altered for Phase
II testing. The test procedure calls for 0.4 kg of paint to be
sprayed before initiating the conveyor to pass the targets
through the gun. (This requirement was made to minimize the
error of having a +_ 5-g paint scale accuracy for mass flow
determination.) There was not enough paint left from Phase I to
complete Phase II and meet the 0.4-kg requirement, so the de-
cision was made to use 0.1 kg. It was expected that any in-
accuracy introduced here would be offset by more test repeti-
tions, as was shown.
However, the 0.4 kg in Phase I was discharged with the
targets out of the booth. In Phase II, the 0.1 kg was dis-
charged as the targets moved towards the gun, inside the
booth. There was no way to determine mass flow with the targets
on the conveyor (but outside the booth) in Phase II because of
the booth configuration.
Although this procedural change should not have affected
the AAC runs, the electrostatic equipment was affected. The
paint sprayed as the target moved towards the gun had the
capability of staying airborne long enough to be attracted to
the targets as they passed through. The effect on AAE runs is
probably small because the paint is sprayed directly towards the
targets and exhaust system. The horizontal mass velocity of the
99
-------
paint tends to carry it quickly towards the exhaust. The effect
on the HSB runs is larger because the HSB produces a fog of
paint with very low horizontal mass velocity. Thus the paint
sprayed by the HSB is capable of staying suspended in the
vicinity of the targets long enough to coat the targets. The
adherence of extra paint to the HSB targets threw the TE ' s over
100 percent, a clear physical impossibility.
In an effort to salvage the HSB data, engineering judgment
was used to correct the data for the additional paint the tar-
gets were exposed to. Based on observations during the Ransburg
test, the paint was sprayed 5 feet in front of the target with
the target in the booth. At PPG, the targets entered the booth
3 feet from the spray gun. The difference of 2 feet was con-
verted to mass and subtracted from the HSB data. Corrected HSB
data were presented in Section 6, Table 31. The correction,
though based solely on engineering judgment, aligns Phase I re-
sults more closely than Phase II. This alignment is considered
corroborative of the described effect.
Four AAE runs (AAE-67-1-4) were made during Phase II timing
the flow of paint using marks at the beginning of the first
scavenger and the end of the last scavenger. These runs were
made to determine the effect of reduced accuracy in the paint
weighing operation on the transfer efficiency determination.,
The standard deviation of the four runs AAE-67-FP-1 through 4
(treating the mean of 4 targets as a single observation) was
2.18. In the case of the four runs AAE-67-VC-1 through 4, the
standard deviation was 1.89. A test was performed to see if
either of these represents a different level of precision than
that of the overall transfer efficiency test.
The pooled standard deviation, S", for Phase II testing was
1.8 based on 48 observations. While Bartlett's test is appropri-
ate for comparison of more than 2 variances, the F-statistic is
used to determine if a pair of sample variances differ signifi-
cantly.
The following are the calculations.
2 —c^ /•?'2
~S /S
3,47
AAE-67-FP(l-4) 2.18 4.73 3.24 1.46
AAE-67-VC(l-4) 1.89 3.56 3.24 1.10
With 3 and 47 degrees of freedom, the critical value of F for a
two-sided test at the 0.1 significance level is 2.84. Therefore
the null hypothesis cannot be rejected, leading to the conclu-
sion that the variance associated with the altered test method
is poolable with that of the standard method.
100
-------
It is therefore recommended that the test procedure be revised
to set the timing and mass flow rate marks at the scavengers.
This recommendation does not have a significant effect on the
test precision, while it avoids the problem of premature paint
flow initiation biasing the transfer efficiency results.
101
-------
SECTION 8
THIRD LABORATORY TEST
FACILITIES
The third laboratory test was performed at Nordson using
the original test plan with modifications to accommodate varying
several parameters for each successive run. The test was run in
a 3 m (10 ft) open-faced booth contained in a 4.5 x 6.1 m (150 x
20 ft) closed room. Ventilation was provided by a multispeed
fan drawing outside air once through the booth. Test equipment
specifications matched or exceeded the requirements of the test
plan.
DESCRIPTION OF PAINT
A 70-percent solids enamel paint with polyester resin was
obtained from PPG for this test (see Table 38). The paint
differs slightly from the 67-percent solids paints used in
earlier tests. The weight percent solids was higher, and a
small additive present in the original test paints had been
deleted by the manufacturer.
A SANES Model 877 was used to measure paint resistance.
ASTM D-1200-70 was used for viscosity determination. Paint
weight percent solids was determined by ASTM D-2369-81 using the
manufacturer's recommended cure schedule.
QA/QC PROCEDURES
All quality assurance/quality control procedures in the
approved Test Plan were followed. As in Phase I and Phase II,
a Master Data book was prepared prior to testing. All test
documentation and data were recorded (and checked) in the Master
Data book. A log book was also maintained for special notations
and observations.
TEST DESIGN
According to ASTM 691-79, "The first requirement is the ex-
istence of a valid, well written test method that has been
developed in one or more competent laboratories and has been
102
-------
TABLE 38. PAINT SPECIFICATIONS
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) (1)
9. Reducing Solvent
10. Vol. of Solvent Put into
Vol. Paint
11. Viscosity - Spray (cut) (1)
12. wt/gal - Spray
13. wt. Solids - Spray
14. Resistance
Enamel
Polyester
PPG
W45458
Tan
10 min @ 350 °F
— Sec.tt — Ford Cup @ °F
Solvesso
_ 700 cc (vol) solvent in
1 1/2 (vol) paint
15 1/2 s Zahn 3 _
28.5 s #4 Ford Cup @ 75°F
_ ™ _ Ibs/gal _
1.7X108
(1) Use ASTM D-1200-70, "Viscosity of Paints, Varnishes, and
Lacquers by Ford Viscosity Cup." Viscosity may also be
determined by ASTM D-3794, Part 6 (Zahn Cup method)
in addition to the Ford Cup measurements.
103
-------
subjected to a screening procedure. . .". The third laboratory
test was developed to provide the required screening procedure
to determine the extent to which previously uncontrolled factors
influence the resulting transfer efficiency for the three equip-
ment types previously tested. Variables were selected primarily
through an analysis of the differences between Phase I and Phase
II test conditions. Unknown or uncontrolled factors were
selected for further testing. In addition to these variables,
the Steering Committee suggested other parameters for screening
testing. The following parameters were selected for screening:
AAE AAC
Booth air rate, fpm • Booth air rate, fpm
Paint mass flow, g/s • Paint mass flow, g/s
Lag discharge distance, ft • Lag discharge distance, ft
Shaping air • Shaping air
Atomizing air, psig • Atomizing air, psig
Tip voltage, kV
HSB
Booth air rate, fpm
Paint mass flow, g/s
Log discharge distance, ft
Shaping air
Bell rpms
Tip voltage, kV
When the results of two or more 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 procedure at a time, but in such a way that the effects of
individual alterations may be determined. Fractional factorial
designs sacrifice the ability to test for some or all inter-
actions but are able to test for a number of main effects very
efficiently.
For TE screening testing, the effect of five or six parameters
needed to be evaluated. The Association of Official Analytical
Chemists (AOAC) design (the 1/16 fraction of a 2 factorial)
for examining the effects of seven variables in eight trials was
selected for the third TE test, and is presented in Table 39.
One or two "dummy" variables were added to complete the AOAC
matrix for TE testing. (Dummy variables introduce a meaningless
action, such as checking your watch prior to a test run, into
the matrix. If this action is later demonstrated to be signifi-
cant, it points to other difficulties with the procedure that
must be addressed.)
104
-------
TABLE 39. AOAC(1^ SCREENING TEST DESIGN
Eight combinations of seven factors used to test the ruggedness
of an analytical method
Combination or Detn No.
Factor
A
B
C
D
E
F
G
or
or
or
or
or
or
or
Value
a
b
c
d
e
f
g
Observed result .
i
A
B
C
D
E
F
G
s
2
A
B
c
D
e
f
g
t
3
A
b
C
d
E
f
g
u
4
A
b
c
d
e
F
G
V
5
a
B
C
d
e
F
g
w
6
a
B
c
d
E
f
G
X
7
a
b
C
D
e
f
G
y
8
a
b
c
D
E
F
g
z
A, B, C, D, E, F, G = nominal values for 7 different factors
that might influence the result if their
nominal values are slightly changed.
a, b, c, d, e, f, g = the alternative value of A, B, C, D, E,
F, G
Youden, W. J. and E. H. Steiner, "Statistical Manual of the
Association of Official Analytical Chemists," AAOC, Arlington,
VA, 1975.
Note: This test design does not address the interactions of the main effects
to each other. For example, paint flow, although not significant at
the 90and 95 percent confidence level is confounded with the inter-
actions of voltage and atomizing air or lag distance is confounded
with the interactions of voltage and booth air and others.
105
-------
From Table 39 it is evident that each factor can be evalu-
ated by some combination of the eight determinations. More is
presented on the data analysis later in this section.
Following the AOAC design, a test matrix was developed for
each equipment type. These matrices are presented in Tables 40,
41, and 42. The performance order for each set of eight runs
(for each equipment type) was randomized. This randomized set
was then assigned sequential run numbers to identify the pre-
scribed order of testing. Tables 40 through 42 show the actual
values of each factor attained during the testing, as well as
the resulting values of transfer efficiency for each equipment
configuration.
TEST PARAMETERS
Each equipment type, HSB, AAC, and AAE, was tested accord-
ing to the AOAC test design. The equipment specifications (with
the exception of intentionally varied factors) are presented in
Tables 43 through 49.
TEST SEQUENCE(1)
The actual test sequence was?
Monday (6/27/83)
o Obtained platform scale
o Set up HSB
o Assembled targets
o Cut and weighed foils for HSB runs
o Set conveyor speed
o Fabricated scavengers
Tuesday (6/28/83)
o Completed HSB set up
o Cut paint and documented
o Conducted trial mass flow and spray pattern runs
o Conducted dummy run
o Conducted HSB 5-8
Wednesday (6/29/83)
o Performed HSB 1-4
o Performed AAE 1-8
Thursday (6/30/83)
o Performed AAC 1-8
o Performed blank/solvent run
o Completed all weighing
o Clean up
See Table 7 for nomenclature.
106
-------
TABLE 40. AIR ATOMIZED ELECTROSTATIC TEST MATRIX
AND RESULTS (NORDSON)
FLAT PANEL
Combination No.
Factor Value
A
B
C
D
E
F
G
= Voltage at tip, kV
= Atomizing air, psig
= Booth air at target, fpm
= Paint mass flow, g/s
= Lag discharge distance, ft
= Shaping air, high=l, Lcw=0
= Dummy one, Yes=l, No=0
TE
1
60
50
65
3.8
3
1
1
83.9
2
60
50
45
3.7
0
0
0
91.8
3
60
30
65
3.0
3
0
0
89.0
4
60
30
45
2.9
0
1
1
90.4
5
45
50
65
2.9
0
1
0
80.5
6
45
50
45
2.4
3
0
1
87.5
7
45
30
65
3.7
0
0
1
89.2
8
45
30
45
3.8
3
1
0
88.3
VERTICAL CYLINDER
Combination No
Factor Value
A
B
C
D
E
F
G
= Voltage at tip, kV
= Atomizing air, psig
= Booth air at target, fpm
= Paint mass flow, g/s
= Lag discharge distance, ft
= Shaping air, high=l, Low=0
= Dummy one, Yes=l, No=0
1
60
50
65
3.8
3
1
1
2
60
50
45
3.7
0
0
0
3
60
30
65
3.0
3
0
0
4
60
30
45
2.9
0
1
1
5
45
50
65
2.9
0
1
0
*
6
45
50
45
2.4
3
0
1
7
45
30
65
3.7
0
0
1
8
45
30
45
3.8
3
1
0
TE
21.4 23.9 28.3 36.7 21.1 17.8 26.8 30.6
107
-------
TABLE 41. HIGH SPEED BELL TEST MATRIX AND RESULTS (NORDSON)
FIAT PANEL
Combination No.
Factor Value
A =
B =
C =
D -
E =
F =
G =
Shaping air, psig at gun
Voltage at tip, kV
Paint mass flow, g/s
Lag discharge distance, ft
Booth air at target, fpm
Thousand rpms
Dummy one, Yes=l, No=0
TE 96
1
40
90
6.5
3
105
11
1
.4
2
40
90
3.8
3
80
9.3
0
104.4
3
40
72
6.5
0
80
9.2
0
91.1
4
40
72
3.6
0
50
11
1
97.6
5
30
90
6.4
0
65
11
0
101.1
6
30
90
3.4
0
105
9.3
1
101.3
7
30
72
7.0
3
65
9.3
1
107
8
30
72
3.6
3
67
11
0
.1 100.5
VERTICAL CYLINDER
Combination No.
Factor Value
A =
B =
C =
D =
E =
F =
G =
Shaping air, psig at gun
Voltage at tip, kV
Paint mass flow, g/s
Lag discharge distance, ft
Booth air at target, fpm
Thousand rpms
Dummy one, Yes=l, No=0
1
40
90
6.5
3
105
11
1
2
40
90
3.8
3
80
9.3
0
3
40
72
6.5
0
80
9.2
0
4
40
72
3.6
0
50
11
1
5
30
90
6.4
0
65
11
0
6
30
90
3.4
0
105
9.3
1
7
30
72
7.0
3
65
9.3
1
8
30
72
3.6
3
67
11
0
TE 97.8 103.1 92.7 98.6 109.7 103.4 107 101.2
108
-------
TABLE 42. AIR ATOMIZED CONVENTIONAL TEST MATRIX
AND RESULTS (NORDSON)
FLAT PANEL
Combination No.
Factor Value
A
B
C
D
E
F
G
= Atomizing Air, psig
= Paint mass flow, g/s
= Booth air at target, fpm
= Lag discharge distance, ft
= Shaping air, High=l, Lcw=0
= Dummy one, Yes=l, No=0
= Dummy two, Yes=l, No=0
TE
1
50
6.2
75
3
1
1
1
76.6
2
50
6.2
40
3
0
0
0
90.3
3
50
1.8
75
0
1
0
0
70.1
4
50
2.4
40
0
0
1
1
85.0
5
30
6.3
75
0
0
1
0
86.9
6
30
6.4
40
0
1
0
1
86.0
7
30
3.0
75
3
0
0
1
85.4
8
30
2.6
40
3
1
1
0
79.9
VERTICAL CYLINDER
Combination No.
Factor Value
A
B
C
D
E
F
G
= Atomizing Air, psig
= Paint mass flow, g/s
= Booth air at target, fpm
= Lag discharge distance, ft
= Shaping air, High=l, Lcw=0
= Dummy one, Yes=l, No=0
= Dummy two, Yes=l, No=0
1
50
6.2
75
3
1
1
1
2
50
6.2
40
3
0
0
0
3
50
1.8
75
0
1
0
0
4
50
2.4
40
0
0
1
1
5
30
6.3
75
0
0
1
0
6
30
6.4
40
0
1
0
1
7
30
3.0
75
3
0
0
1
8
30
2.6
40
3
1
1
0
TE 14.7 17.4 13.3 15.3 18.3 17.2 15.9 14.8
109
-------
TABLE 43. NORDSON TEST EQUIPMENT SPECIFICATIONS
A. We
1.
2.
3.
4.
ight Percent Solids Measurement
Laboratory Scales
a. Manufacturer
b. Model No.
c. Serial No.
d. Capacity, g
e. Rated accuracy, g
Foil Dishes
a« Type
b. Size
Syringe
a. Type
b. Capacity, ml
Solvent Type
Equipment
Mettler
H78AR
623434
160
2" dia. aluminum
2" dia.
Plastic
10
Solvesso
( paint )
B. Conveyor Speed Measurement Equipment
1. Rule
a. Type
b. Graduations
2. Electronic Timer
NA
NA
Autotron Electronic
Autotron
a. Type
b. Manufacturer
c. Model No. Countomatic (Digital Display) A579SA
d. Serial No.
e. Rated accuracy, s
Mass Flow Measurement Equipment
364
0.01
c.
d.
Target
1.
2.
3.
Wet
Platform Scales
a. Manufacturer Metrodyne
b. Model No. SS-100
c. Serial No. 100#
d. Capacity, kg
e. Rated Accuracy, g
Stopwatch
a. Manufacturer
b. Model No.
Serial No.
Rated Accuracy, s
Foil
Type
Nominal Thickness, mils
Temper
(Platform) Filina Scales-Con-
Film Measurement Equipment
a. Manufacturer
b. Model No.
Voltage Measuring Equipment Electrostatic
a. Manufacturer
b. Model No.
Linear Air Velocity Measurement
a. Anemometer Mfg.
b. Model No.
solidated Controls
UMC2000 AAAA
5521
45
Marcel & Cie
1/5
1.5
Med
System Analyzer
Nordson
P160102B
Alnor
8500 +1
f pm
110
-------
TABLE 44. AAE PAINT SPRAY AND PERIPHERAL
EQUIPMENT SPECIFICATIONS (NORDSON
A. Paint Supply Tank
1. Type
2. Manufacturer
3. Model No.
4. Serial No.
5. Rated Capacity, gallons
Pressure Pot
Binks
2 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
Conventional
Nordson Electrostatic
Nordson
AN8A 2465028
A2K01472
245-987
228(59/1000
C. Paint Spray Booth
1. Type
2. Manufacturer
3. Model No.
4. Serial No.
5. Rated Capacity, cfm
Open 10X17 ft
D. Conveyor
1. Type
2. Manufacturer
3. Model No.
4. Serial No.
Overhead
Unibuilt
Forced Draft Oven
1. Type
2. Manufacturer
3. Model No.
4. Serial No.
Paint Heaters
1. Type
2. Manufacturer
3. Model No.
4. Serial No.
2 door 8'XIO'
Grieve Co.
SC550
49037
NA
NA
NA
NA
111
-------
TABLE 45. AAE EQUIPMENT OPERATING CONDITIONS (NORDSON
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. Horn 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,
at plane of target
4. Air Flow, fpm, measured 4 ft
in front of fan
5. 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)
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/s)
12
50
60
74
12"
74
61%
30-60 fpm
Normal
100 microns
15"
30"
50%
24
^
10
350
350
NA
NA
NA
(1) Same cure schedule as foils.
112
-------
TABLE 46. HSB EQUIPMENT OPERATING CONDITIONS (NQRDSON
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
B
9. Pump Setting
Paint Spray Booth
1. Ambient Temperature, °F
2. Relative Humidity, %
3. Air Flow Velocity, fpm,
at plane of target
4. Air Flow, fpm, measured 4 ft
in front of fan
5. Air Flow Direction
C. Target Parameters
1. Average Wet Film Thickness, mils
2. Average Dry Film Thickness
Vertical Paint Coverage, cm (in)
Target Height, cm (in)
% Vertical Coverage
3,
4.
5,
6. Resistance to Ground, Ohm
D. Forced Draft Oven (1)
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/s)
20.5
90
9300
12700
40
74
12"
.062
74
61%
Normal
20 microns
30
60
50%
P.452
10
10
350
350
NA
NA
NA
(1) Same cure schedule as foils.
113
-------
TABLE 47. HSB PAINT SPRAY AND PERIPHERAL
EQUIPMENT SPECIFICATIONS (NORDSON
A. Paint Supply Tank
1. Type
2. Manufacturer
3. Model No.
4. Serial No.
5. Rated Capacity, gal
2 gal STD
Sinks
2 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
Turbobell
Ransburg
3655 01
20074- 12 head
C. Paint 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.
Open Room 10X17 ft
50-80 fpm at plane of tgt.
Overhead-square run
Unibuilt
Forced Draft Oven
1. Type
2. Manufacturer
3. Model No.
4. Serial No.
Paint Heaters
1. Type
2. Manufacturer
3. Model No.
4. Serial No.
2 door 8BX10'
Grieve Co.
SC55Q
49037
NA
NA
NA
NA
114
-------
TABLE 48. AAC EQUIPMENT OPERATING CONDITIONS (NORDSON;
A. Paint Spray Equipment
1. Paint Pressure at Paint Pot, psig 9
2. Paint Pressure at Spray Gun, psig —
3. Atomizing/Turbine Air Pressure at
Spray Gun, psig 50
4. Operating Voltage, kV —
5. Disk or Bell Speed, rpm —
a. With Paint Applied —
b. Without Paint Applied —
Horn Air for Bell, psig
7. Paint Temperature at Paint Pot, °F 85
8. Gun to Target Distance, cm 12"
9. Pump Setting —
B. Paint Spray Booth
1. Ambient Temperature, °F 85
2. Relative Humidity, % 65%
Air Flow Velocity, fpm, 30-50
at plane of target
4. Air Flow, fpm, measured 4 ft —
in front of fan
5. Air Flow Direction Normal
C. Target Parameters
1. Average Wet Film Thickness, mils 350 microns
2. Average Dry Film Thickness —
3. Vertical Paint Coverage, cm (in) 8
4. Target Height, cm (in) 30
5. % Vertical Coverage —
6. Resistance to Ground, Ohm
D. Forced Draft Oven (1)
1. Cure Time, minutes
a. Foil Dish (sample) 10
b. Target Foil 10
2. Cure Temperature, °F
a. Foil Dish (sample) 350
b. Target Foil 350
E. Paint Heaters
1. Temperature In, °F NA
2. Temperature Out, °F NA
F. Conveyor Speed Setpoint, fpm (cm/s) NA
(1) Same cure schedule as foils.
115
-------
TABLE 49. AAC PAINT SPRAY AND PERIPHERAL
EQUIPMENT SPECIFICATIONS (NORDSON)
A. Paint Supply Tank
1. Type
2, Manufacturer
3. Model No.
4. Serial No.
5. Rated Capacity, gal
Pressurized Std
Sinks
2 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 (Had to change fluid tip)
8,, Needle
Air Atom.Conventional
Sinks
610
56789
63P3 or 63PB
C. Paint Spray Booth
1. Type
20 Manufacturer
3. Model No.
4. Serial No.
5c Rated Capacity, cfm
Open 10X17 ft
D. Conveyor
1. Type
2. Manufacturer
3. Model No.
4. Serial No.
Overhead
Unibuilt
Forced Draft Oven
1. Type
2. Manufacturer
3. Model No.
4. Serial No.
Paint Heaters
1. Type
2o Manufacturer
3. Model No.
4. Serial No.
8'XIO1 Conventional
Grieve Co.
SC550
49037
NA
NA
NA
NA
116
-------
Friday (7/1/83)
o Completed clean up
o Prepared test materials for shipment
SOLVENT-ONLY RUN
A test run was conducted using Solvesso (no paint) as the
spray in order to determine if the solvent or tape were con-
tributing any weight gain to the foils. The foils averaged a
0.001-g weight gain. This weight gain is undetectable in
reported TE's calculated to tenths.
DATA ANALYSIS
During the data analysis an error was found in the record-
ed cured foil weights for AAE runs E and F, which happened to
have been run consecutively in the randomized order. It was
discovered that the cured foil weights for runs E and F were
inadvertently recorded on the incorrect sheets. The following
data analysis is based on the corrected weights.
Table 50 presents the analysis of the experiment. The
computer printout gives the contrasts and variances associated
with each factor. The contrast is the difference between the
mean transfer efficiency values obtained when each factor is at
its high and low values.
The variance associated with the dummy variable or vari-
ables is small in each case, as is to be expected. At least
one other factor exhibits a variance similar to that exhibited
by the dummy variable(s) for each experiment. Such variances
were pooled with the dummy variance(s) to get an estimate of
error for each experimental configuration.
Table 51 presents the results of an analysis of variance
based on the variances from Table 50. The degrees of freedom
associated with the error term is noted in each case. Those
factors found to be significant at the 0.1 level are marked
with an asterisk.
A regression equation may be written to show the manner
in which transfer efficiency varies with each factor over the
range tested. Regressions for each experimental configuration
with the factors found to be significant as independent vari-
ables are presented in Table 52. The coefficients are simply
the contrasts (1) from Table 50.
(1) Myers, Raymond H., Response Surface Methodology pp 47-48,
1976.
117
-------
TABLE 50= SCREENING EXPERIMENT ANALYSIS
TABULATED VALUES OF VARIANCE AND
CONTRASTS
AAC -VC
Factor Variance
c
B
F
A
E
D
G
C
B
F
A
E
D
G
C
B
F
A
E
D
G
C
B
F
A
E
D
g
0.78125
8.61125
0.06125
3.78125
5.95125
0.21125
0.06125
Contrast
0.625
-2.075
-0.175
1.375
-1.725.
-0.325
0.175
AAC -FP
Variance
61.605
47.045
1.445
32.805
153.125
2.205
4.205
Contrast
5.55
-4.85
-0.85
4.05
-8.75
1.05
-1.45
MSB -VC
Variance
0.10125
26.28125
0.15125
105.85125
67.86125
2.76125
0.00125
Contrast
-0.225
-3.625
0.275
7.275
-5.825
1.175
-0.025
HSB -FP
Variance
8.20125
5.95125
8.61125
52.53125
54.60125
37.41125
3.51125
Contrast
2.025
-1.725
-2.075
5.125
-5.225
4.325
-1.325
AAE -VC
Variance
16.245
182.405
21.125
24.5
13.52
0.13
0.18
Contrast
2.85
9.55
3.25
-3.5
-2.6
-0.3
0.3
AAE -FP
Variance
29.645
21.78
25.92
11.52
1.28
4.205
0.245
Contrast
3.85
3.3
-3.6
-2.4
-0.9
1.45
-0.35
118
-------
TABLE 51,
ANOVA RESULTS FOR SCREENING TEST
Source
Regression
on mean
on A
on B
on C
on E
error
Total
Regression
on mean
on A
on B
on C
on E
error
Total
Regression
on mean
on A
on B
on D
on E
error
Total
Regression
on mean
on A
on D
on E
error
Total
AAC-VC
SS
2012.9512
3.78125
8.61125
.78125
5.95125
.33375
2032.41
AAC-FP
54483.005
32.805
47.045
61.605
153.125
7.855
54785.44
HSB-VC
82722.781
105.85125
26.28125
2.76125
67.86125
.25375
82925.789
HSB-FP
79900.031
52.53125
37.41125
54.60125
26.275
80070.849
DF
1
1
1
1
1
3
8
1
1
1
1
1
3
8
1
1
1
1
1
3
8
1
1
1
1
4
8
MS
3.78125
8.61125
.78125
5.75125
.11125
32.805
47.045
61.605
153.125
2.618
105.85125
26.28125
2.76125
67.86125
.085
52.53125
37.41125
54.60725
6.56875
F
33.99*
77.40*
7.02*
53.49*
12.53*
17.97*
23.53*
58.48*
3724.8*
924.8*
97.2*
2388.0*
8.00*
5.69*
8.31*
119
-------
TABLE 51. (Continued)
Source
AAE-VC
SS
DF
MS
F
Regress ion
on
on
on
on
on
on
error
mean
A
B
C
E
F
Total
5335
24
182
16
13
21
5593
.445
.5
.405
.245
.52
.125
.36
.6
AAE-FP
1
1
1
1
1
1
2
8
24
182
16
13
21
.5
.405
.245
.52
.125
.18
136
1013
90
75
117
.1*
.4*
.3*
. 1*
.4*
Regression
on
on
on
on
on
on
error
mean
A
B
C
D
F
Total
61355
11
21
29
4
25
1
61449
.045
.52
.78
.645
.205
.92
.525
.64
1
1
1
1
1
1
2
8
11
21
29
4
25
.52
.78
.645
.205
.92
.7625
15
28
38
5
33
ol*
.6*
.9*
.5
.9*
*Significant at .10 level
120
-------
TABLE 52. REGRESSION MODELS DERIVED FROM THE SCREENING TESTS
' AAC-VC '
TE = 15.862 - A + B - - C -
TE = 82.525 - A
A = Atomizing air +1 = 50 psig -1 = 30 psig
B = Paint mass flow +1 = 6.275 -1 = 2.45
C = Booth air at target +1 = 75 fpm -1 = 40 fpm
E = Shaping air +1 = high -1 = low
HSB-VC
TE = 101.6875 - 14™ A + *l™ B + ii££ D - 1
TE = 99.9375 -
HSB-FP
4.325 5.225
TE = 87.575 -I-
A = Shaping air, psig at gun +1 =40 -1 = 30
B =. Voltage at tip, kV +1 = 90 -1 = 72
D = Lag disch. dist., ft +1=3 -1=0
E = Air rate at target, fpm +1 = 89.25 -1 = 65
AAE-VC
TE = 25.825 +!5A-iB_2c_26E +
A = Voltage at tip, kV +1 =60 -1 = 45
B = Atomizing air, psig +1 =50 -1 = 30
C = Booth air at target, fpm +1 =65 -1 = 45
E = Lag disch. dist., ft +1 = 3 -1=0
F = Shaping air +1 = high -1 = low
121
-------
The differences observed between transfer efficiency
results at PPG and at Ransburg for the same test configurations
can now be seen to be explained by the uncontrolled variables.
Table 53 presents, for each configuration, two sets of test
conditions which, according to the appropriate regression, would
result in transfer efficiency values similar to those actually
observed at each laboratory. The values of each variable
utilized in construction of Table 53 are the actual experimental
values for those variables for which data was recorded.
The predicted transfer efficiency values shown in Table 53
are similar to those actually obtained during the Phase I and
Phase II testing. In general, they are within one standard
deviation of the observed values, where the precision of the
test method has been specified as a standard deviation of 2.0.
Appendix B presents the results of a multiple linear re-
gression computer analysis of the screening experiment, utiliz-
ing as independent variables for each regression the factors
found to be significant by ANOVA (see Table 51). Actual experi-
mental values of each factor for each experiment were used in
Appendix B rather than the mean of each high and low setting as
was done here.
122
-------
TABLE 53. PREDICTED TRANSFER EFFICIENCY RESULTS
(PPG AND RANSBURG)
Air Atomized Conventional
PPG Ransburg
Independent Variable
A - atomizing air, psig 50 51
B - paint mass flow, g/s 3.04 2.76
C - booth air at target, fpm 100 50
E - shaping air 2 0
Predicted TE
Vertical cylinder 11.97 14.37
Flat panel 63.33 79.45
High Speed Bell
Independent Variable
A - shaping air, psig 35 45
-B - voltage at tip, kV 80 80
D - lag distance, ft 0 5
E - booth air at target, fpm 100 50
Predicted TE
Vertical cylinder 95.4 102.1
Flat panel 92.8 105.7
Air Atomized Electrostatic
Independent Variable
A - voltage at tip, kV 47 63
B - atomizing air, psig 30 10
C - booth air at target, fpm 100 50
E - lag distance, ft 0 5
F - shaping air 0 4
Predicted TE
Vertical cylinder 24.2 46.8
Flat panel 79.7 88.0
123
-------
SECTION 9
CONCLUSIONS AND RECOMMENDATIONS
The results of Phase I and Phase II testing demonstrated
the viability of the draft spray painting TE method. While
these tests exhibited good precision, tney produced clearly
different TE results. Subsequently, a third test was performed
to define the factors contributing to the differences between
Phase I and II results.
The following factors were determined to have a significant
impact on TE for at least one type of spray equipment tested:
o Booth air rate at target, fpm
o Shaping air
o Atomizing air, psig
o Voltage at tip, kV (for electrostatic equipment)
o Paint discharge technique (lag discharge distance)
o Paint mass flow, g/s
A complete TE method must specifically address control of
these factors (in addition to those already controlled) to avoid
introduction of significant errors. The draft test method pre-
sented in Section 4 incorporates these findings as summarized
below:
o The booth air rate at the targets should be controlled.,
While a certain air rate may be specified in the TE
method, most laboratories do not have the capability to
adjust air rates. A conservative alternative would be to
set a minimum acceptable air rate for all tests.
o Shaping air pressure must be controlled. Generally
shaping air is adjusted to provide a properly shaped
spray. The adjustment is qualitative.
o Atomizing air pressure must be controlled.
o Tip voltage must be controlled; an adjustable power
supply is strongly recommended.
124
-------
o The paint discharge technique must be standardized to
provide identical opportunity for paint to adhere for
each run.
o The paint mass flow rate must be controlled for air
atomized equipment.
The following changes are also recommended in the draft TE
procedure to provide a more precise, simpler test performance:
o The test panels and cylinders can be mounted in a less
complex manner, provided dimensions and materials are
consistent with the original targets.
o A single source of paint should be used for each series
of tests.
o The same oven and timer should be used for all curing.
o Plastic syringes are acceptable for use in weight
percent solids determination.
o All pressure gages should be calibrated before
performing tests.
It is essential to control all previously listed factors to
the same level for verification testing of the draft TE method.
The draft test method arrived at through development and
screening tests as described in this report is now ready to be
subjected to a systematic program of interlaboratory testing.
Such a program will enable the determination of the reproduc-
ibility component of precision. It is recommended that an
interlaboratory test program including from 6 to 10 laboratories
be undertaken, using as a basis the draft test method developed
here.
125
-------
BIBLIOGRAPHY
Air Pollution Control Board of the State of Indiana, VOC
Regulations-General Provisions, Section 325 IAC 8-1.1.
Aldorfer, D.M. A Report on Transfer Efficiency of Waterborne
Enamel, Preliminary Results. Warren, Michigan: General
Motors Technical Center, Fisher Body Division, April 24,
1979.
Antonelli, J. A. "Liquid High Solids-Organic Coatings," Society
of Manufacturing Engineers, 1974 Technical Paper FC74-654.
Brewer, G.E.F. "Calculations of Painting Wasteloads Associated
with Metal Finishing," for Industrial Environmental
Research Laboratory, Cincinnati, EPA-600/2-80-144, June 1980,
Brewer, G.E.F. "Mathematics of Emissions and Transfer Effici-
ency," U. S. Environmental Protection Agency Seminar on
Volatile Organic Compound Control in Surface Coating
Industries. Brighton, Michigan: George E.F. Brewer
Coating Consultants, 1979.
Bublick, Timothy "Robots for Spray Finishing," Plating and
Surface Finishing, November 1980, V. 67, No. 11, PSFMDH 67.
Chapman, G.R. "High Solids Coatings," 1982 Products Finishing
Directory, September 1981, V. 45, No. 12A.
DeVilbiss Company, "The ABC's of Spray Equipment," Third
Edition, Toledo, Ohio., 1954.
Drum, E.W. Letter to Dave Salman, U. S. Environmental Protec-
tion Agency, Research Triangle Park, N.C.. Indianapolis:
Ransburg Electrostatic Equipment, A Division of Ransburg
Corporation, October 31, 1978.
126
-------
Farrell, Ron "Powder Coatings," 1982 Products Finishing
Directory, September 1981, V. 45, No. 12A.
General Motors Corporation. GMAD Oklahoma City, Oklahoma Trans-
fer Efficiency Experiments. Warren, Michigan: General
Motors Technical Center, Fisher Body Division, December 13,
1979.
Baseline Transfer Efficiency of Waterborne Enamel Top-
coat, Warren, Michigan: General Motors Technical Center,
Fisher Body Division, October 1, 1979.
Godovoy. A. "Deposition Efficiency in Electrostatic Spraying
of Powder Coating." Journal of Paint Technology, Vol. 15
No. 580. May 1973.
Gunsel, S.J. Letter (SG-51) to Central Docket Section (A-130),
U. S. Environmental Protection Agency, Washington, DC.
Amherst, Ohio: Nordson Corporation, February 19, 1981.
Harding, S.K. Trip Report—April 19-21, 1982. Ft. Lauderdale,
Florida: CENTEC Corporation.
Health and Human Services, U. S. Department of. An Evaluation
of Engineering Control Technology for Spray Painting, NIOSH
Technical Report, -Publication No. 81-121. "Washington, DC:
U. S. Government Printing Office, June 1981.
Hungerford, A.O. Transfer Efficiencies of Painting Method, A
Case History. Galesburg, Illinois: Butler Manufacturing
Company, March 1, 1982.
Industrial Finishing, October 1978, Part Two.
Industrial Finishing, August 1982, page 43.
Industrial Finishing, February 1982, pages 9, 10, 16.
McCrodden, B.J. Final Trip Report — Ransburg Corporation.
Research Triangle Park, N.C.: Research Triangle Institute,
June 19, 1980.
Metal Finishing, January 1980, page 112.
Metal Finishing Annual Review Issue, February 1981, February
1982, and April 1982.
Miller, E.P- "Powder Coating Compared with Other Techniques
Available for Meeting OSHA and EPA Regulations," Society of
Manufacturing Engineers, 1973 Technical Paper, FC73-553.
127
-------
Myers, Raymond H. Response Surface Methodology, 1976.
Nickerson, R.S. "Applying High Solids Coatings," Products
Finishing, November 1981, pages 81-88.
Plating and Surface Finishing, "EPA Reveals New Policy on
Hydrocarbon Emissions," No date.
Powell, G. and Bussjaeger, S. "Water-bourne Coatings," 1982
Products Finishing Directory, September 1981, V. 45, No.
T2A^
Product Source Listings, 1982 Product Finishing Directory,
September 1981, V. 45, No. 12A.'
Roobol, N.R. "Focus on Automotive Coatings," Metal Finishing,
September 1980, pages 31-35.
Scarbrough, D. "Electrostatic Spray Coating . . . Focus on
Profitability," Finishers' Management, March 1981.
Sloan, E.M. Trip Report -- GMAD - Fremont, California Transfer
Efficiency Experiments. Warren, Michigan: General Motors
Technical Center, Fisher Body Division, April, 1980.
Vehicle Paint Transfer Efficiency Test -- A New Tool to
Compare Solvent Emissions (GMMD - 79-065), SME Paper No.
FC80-610. Dearborn, Michigan: Association for Finishing
Processes, Society of Manufacturing Engineers, 1980.
manable, Y. Feasibility Study for Base Coat/Clear Coat Painting
System Using New High Solid Paints. Marysville, Ohio:
Honda of America Manufacturing, Inc., n.d.
U. S. Department of Health and Human Services, "An Evaluation of
Engineering Control Technology for Spray Painting," June
1981 .
U. S. Environmental Protection Agency, "Industrial Surface
Coating Appliances-Background Information for Proposed
Standards," Office of Air Quality Planning and Standards,
November 1930, EPA-450/3-80-037a (NTIS PB 82-152174).
U. S. Environmental Protection Agency, "Standards of Performance
for New Stationary Sources: Industrial Surface Coating:
Appliances," Proposed Rule and Notice of Public Hearing,
Federal Register, V. 45, No. 249, December 24, 1980.
128
-------
U. S. Environmental Protection Agency, "Standards of Performance
for New Stationary Sources: Industrial Surface Coating:
Appliances," Correction, Federal Register, V. 46, No. 18,
January 28, 1981.
U. S. Environmental Protection Agency, "Standards of Performance
for New Stationary Sources: Industrial Surface Coating:
Appliances, Correction," Federal Register, V. 46, No. 107,
June 4, 1981.
U. S. Environmental Protection Agency, "Measurement of Volatile
Organic Compounds," Office of Air Quality Planning and
Standards 1979, EPA-450/2-78-041 (NTIS PB80-221674).
Walberg, A.C. "Painting Efficiency of Automatic Electrostatic
Systems," A paper presented at Chemical Coaters Association
Meeting, June 12-14, 1979.
Welch, R.A. "Cyrogenic Paint Stripping," Industrial Finishing,
May 1984.
Youden, W.J. and E.H. Steiner- "Statistical Manual of the
Association of Official Analytical Chemists," AAOC,
Arlington, VA, 1975.
Ziegeweid, J.E. "Applying Organic Coatings, Tools of the
Trade," Metal Finishing, October 1980, pages 57-61.
Ziegeweid, J.E. "Applying Organic Coatings — Airless Comes
of Age," Metal Finishing, June 1981, pages 37-41.
129
-------
APPENDIX A
WORTH ASSESSMENT MODEL OF TE TEST METHODS
Nine TE test methods were evaluated according to the
following criteria:
1. How representative is the test method expected to be of
actual line TE's?
1.0 Identical
0.8 Very close
0*5 Defines range of TE's
0.3 Poor
0.0 Very poor
2. How complex will the procedure be to perform?
1.0 Easy
008 Fairly easy
0.5 Moderate
0.3 Fairly difficult
0.0 Difficult
3. How precise is the method expected to be?
1.0 Excellent
0,8 Very good
0,5 Fair
0.3 Poor
0.0 Very poor
4. How expensive will the method be to perform?
1.0 Little expense involved
0.8 Operating expense only
0.5 Little operating and capital expense
0.3 Moderate capital and operating expense
0.0 Substantial capital and operating expense
130
-------
5. Does the method bias against certain equipment
manufacturers?
1.0 None
0.8 Little
0.5 Moderate
0.3 Substantial
0.0 Severe
6. Does the method bias against certain target geometries?
1.0 None
0.8 Little
0.5 Moderate
0.3 Substantial
0.0 Severe
7. How well developed is the method? (How much more work will
be required for a final method?)
1.0 Ready for use
0.8 Ready for verification testing
0.5 Needs some development
0.3 Needs much development
0.0 Develop from scratch
8. How much process modification is required to use the method?
1.0 None
0.8 Little
0.5 Moderate
0.3 Substantial
0.0 Severe
9. Are there special limitations or restrictions to the method
which must be considered?
1.0 None
0.8 Little
0.5 Moderate
0.3 Substantial
0.0 Severe
131
-------
Each criteria was weighed according to its relative
importance as shown:
Criteria Weighting
1 Representation 0.20
2 Complexity 0.09
3 Precision 0.15
4 Expense 0.08
5 Mfg. Bias 0.12
6 Tgt. Bias 0.11
7 Level of Dev. 0.07
8 Process Mod. 0.08
9 Restrictions 0.10
Each TE test method is described by a three section
descriptor:
AAA/BBB/CCC
where
AAA = plant method or laboratory method
BBB = target type; actual workpiece (opr) target,
standardized targets, or shim
CCC = test parameters; actual operating (line) conditions,
standardized spray conditions
The relative rankings of the methods are shown in Table A-l.
The individual components of the assessment follow in
computer print-out form.
132
-------
Table A-l
TEST METHOD ALTERNATIVES
UJ
U)
METHOD TYPE
PLANT
LAB
LAB
LAB
LAB
PLANT
LAB
LAB
PLANT
TARGET TYPE
WORKPIECE
WORKPIECE
STANDARD
STANDARD
WORKPIECE
STANDARD
SHIM
SHIM
SHIM
TEST PARAMETERS
ACTUAL OPERATING
ACTUAL OPERATING
ACTUAL OPERATING
STANDARD
STANDARD
ACTUAL OPERATING
ACTUAL OPERATING
STANDARD
ACTUAL OPERATING
SCORE
72.5
71.5
67.3
61.8
60.8
60.3
58.0
53.0
18.3
-------
8 PLANT/OPR TGT/OPR COND .725
3 LAB/OPR TGT/OPR COND .715
2 LAB/STD TGT/OPR COND .6725
1 LAB/STD TGT/STD COND .6175
5 LAB/OPR TGT/STD COND .6075
7 PLAN/STD TGTS/OPR COND .6025
4 LAB/SHIM/OPR COND .58
6 LAB/SHIM/STD COND .58
9 PLANT/SHIM/OPR COND .4825
134
-------
(1) LAS/STD TCT/STD
* FACTOR NAME
1 REPRESENTATION
2 COMF1. EX I TV
3 PRECISION
4 EXPENSE
: MFC BIAS
6 TGT EIA3
7 LEVEL OF DEV
8 PROCESS MOD
9 RESTRICTIONS
TOTAL
COND
VALUE SELECTION DESCRIPTION
•3 3 POOR
0 8 FAIRLY EASY
1 0 EXCELLENT
0 . 3 CPR ONLY
0 3 SUBSTANTIAL
0.3 SUBSTANTIAL
0 3 AEADY FOR VEftIF
1.0 NONE
1.0 NONE
WE I CHT
Y / N
ft VALUE
0
0
Q
0
0 ,
0
0
0
0
200
090
100
030
. 1 20
1 10
070
.080
.100
YES
YES
V E S
YES
YES
YES
YES
YES
YES
0
0
0
0
0
0
0
0
0
03000
06750
15000
06000
03000
02750
.05250
. 08000
.10000
0 61750
(2) LAB/STD TCT/OPR COND
* FACTOR NAME VALUE
SELECTION DESCRIPTION
WEIGHT Y/N ff VALUE
1 REPRESENTATION
2 COMPLEXITY
3 PRECISION
4 EXPENSE
5 MFC BIAS
6 TCT BIAS
7 LEVEL OF DEV
8 PROCESS MOD
9 RESTRICTIONS
0 3 DEFINES RANGE
0 3 FAIRLY EASY
1 0 EXCELLENT
0 8 OPR ONLY
C.5 MODERATE
0.3 SUBSTANTIAL
0.9 READY FOR VERIF
i.0 NOME
0.8 LITTLE
0
0 ,
0 .
0 ,
0
0
0 .
0
0
200
090
150
, 080
120
1 10
, 070
.080
1 00
YES
YES
YES
YES
YES
YES
YES-
YES
YES
0
0
0
o
0
0
0
0
0
10000
06750
15000
, 06000
06000
02750
05250
08000
07500
rUTAL
0 67250
13) LA3/OPR TGT/OPR COIID
« FACTOR NAME VALUE
SELECTION DESCRIPTION
WEIGH:
Y/N * VALUE
1 REPRESENTATION
2 COMPLEXITY
3 PRECISION
4 EXPENSE
5 MFC BIAS
& TGT BIAS
7 LEVEL OF DEV
3 PROCESS MOD
9 RESTRICTIONS
rOTAL
0.3 VERY CLOSE
0.8 FAIRLY EASY
0 5 FAIR
0.8 OPR ONLY
0.8 LITTLE
0.8 LITTLE
0.5 NEEDS DEV
1.0 NONE
0.8 LITTLE
0 .
0
0 ,
0
0
0
0
0
0
200
. 090
.150
080
. 120
.110
. 070
080
100
YES
YES
YES
YES
YES
YES
YES
YES
YES
0 .
0
0 .
0
C
0
0
0
0
15000
06750
07500
06000
C9000
08250
33500
.03000
07500
0 71500
135
-------
(4) LAE/SHIM/CrR
# FACTOR NAME
VALUI SELECTION DESCRIPTION
WEIGHT Y/N - VALUE
1 REPRESENTATION
2 COMPLEXITY
3 PRECISION
4 EXPENSE
3 MFC BIAS
6 TGT SI AS
7 LEVEL OF LEV
8 PROCESS MOD
9 RESTRICTIONS
TOTAL
(5) LAS /OPR TGT/STD
4i FACTOR NAME
1 REPRESENTATION
2 CO?!?LEXITY
3 PRECISION
4 EXPENSE
3 MFC BIAS
6 TGT BIAS
7 LEVEL OF DEV
8 PROCESS MOD
9 RESTRICTIONS
TOTAL
(6) LAB/SHIM/STD CO
# FACTOR NAME
1 REPRESENTATION
2 COMPLEXITY
3 PRECISION
4 EXPENSE
5 MFC BIAS
6 TGT BIAS
7 LEVEL OF DEV
6 PROCESS MOD
9 RESTRICTIONS
0 3
0 . 8
o :
0 8
a . s
0 . 3
0 S
1 0
0 8
COND
VALUE
0 5
o a
0 S
0 8
0 3
0 . 3
0 3
: . o
0 8
NO
VALUE
0 3
0 . 8
0 S
0 8
0 3
0 3
0 5
1 0
0 3
DEFINES RANGE
FAIRLY EASY
FAIR
OPR ONLY
MODERATE
SUBSTANTIAL
NEEDS DEV
NONE
LITTLE
SELECTION DESCRIPTION
DEFINES RANGE
FAIRLY EASY
FAIR
OPR ONLY
MODERATE
MODERATE
NEEDS DEV
NONE
LITTLE
SELECTION DESCRIPTION
POOR
FAIRLY EASY
FAIR
OPR ONLY
MODERATE
SUBSTANTIAL
NEEDS DEV
NONE
LITTLE
0
c
0
0
0
a
0
0
0 ,
200
090
i SQ
080
1 20
1 10
070
080
ICO
YES
YES
YES
YES
YES
YES
YES
YES
YES
0
0
0
0
0
0
0
0
0
10000
06730
07500
04 000
06000
02750
03500
.08000
075 00
0 S3000
WEIGHT Y/N S VALUE
0
Cr
0
0
0
0
0
0
0
WE
0
0
0
0
0
0
0
0
0
200
090
.150
. 030
120
1 1 0
070
030
.100
IGHT
.200
090
1 SO
.080
. 120
.110
. 070
030
100
YES
YE 3
YES
YES
YES
YES
YES-
YES
YES
Y/N
YES
YES
YES
YES
YES
YES
YES
YES
YES
0
0
0
0
0
0
0-
0 .
0
0 .
ft
0
0
0
0
0
0
0
0
0
10000
06750
07500
06000
06000
OS500
OSS 00
08000
07^500
60750
VALUE
05000
06750
07500
06000
06000
02750
Q3500
08000
07500
TOTAL
0 53000
136
-------
<7) PLANT/3TD TG'
# FACTOR NAME
1 REPRESENTATION
2 COMPLEXITY
3 PRECISION
4 EXPENSE
3 MFC BIAS
6 TGT BIAS
7 LEVEL OF DEV
8 PROCESS MOD
9 RESTRICTIONS
TOTAL
,/OFR COND
VALUE SELECTION DESCRIPTION
0 5 DEFINES RANGE
0 3 FAIRLY EASY
1 Q EXCELLENT
0.5 LITTLE CAP&OPR
0.5 MODERATE
0 5 MODERATE
0.5 NEEDS DEV
0.3 SUBSTANTIAL
0.S LITTLE
WEIGHT Y/N » VALUE
0
0
0 .
0
0
0
0
0
0
200
090
1 50
080
120
1 1 0
070
. G30
100
YES
YES
YES
YES
YES
YES
YES
YES
YES
0
0
0
0
0
0
0
0
0
10000
04730
15000
04000
06000
05500
03500
02000
07500
0 60250
<8> PLANT/OPR TGT/OFR COND
S FACTOR NAME VALUE
.ELECTION DESCRIPTION
WEIGHT Y/N * VALUE
1 REPRESENTATION
2 COMPLEXITY
3 PRECISION
4 EXPENSE
5 MFC BIAS
6 TCT BIAS
7 LEVEL OF DEV
8 PROCESS MOD
9 RESTRICTIONS
1 0 IDENTICAL
0 . 8 FAIRLY EASY
0.5 FAIR
0 5 LITTLE CAP&OPR
1.0 NONE
1 0 HONE
0 3 NEEDS MUCH DEV
0.3 SUBSTANTIAL
0.8 LITTLE
0
0
0
0
0
0
0
0
0
200
0 90
1 50
080
1 20
.110
.070
. 080
. 100
YES
YES
YES
YES
YES
YES
YES"
YES
YES
0
0
0
0
0
0
0
0
0
20000
06750
07500
04000
12000
11000
01750
02000
07500
TOTAL
0 72500
(9) PLANT/SHIM/OPR COND
# FACTOR NAME VALUE
1 REPRESENTATION 0 5
2 COMPLEXITY 0 8
3 PRECISION 0.5
4 EXPENSE 0 5
5 MFC BIAS 0 S
6 TGT BIAS 0.3
7 LEVEL OF DEV 0.3
8 PROCESS MOD 0.3
9 RESTRICTIONS 0 8
SELECTION DESCRIPTION
DEFINES RANGE
FAIRLY EASY
FAIR
LITTLE CAF&OFR
MODERATE
SUBSTANTIAL
NEEDS MUCH DEV
SUBSTANTIAL
LITTLE
WEIGHT Y/N * VALUE
0
0
0
0
0 .
0
0
0
0
200
090
150
080
120
1 10
070
080
100
YES
YES
YES
YES
YES
YES
YES
YES
YES
0 10000
0.06750
0 07500
0 04000
0.06000
0 02750
0.01750
0 02000
0 07500
TOTAL
0 48250
137
-------
APPENDIX B
SCREENING PROCEDURE
MULTIPLE LINEAR REGRESSION FOR TRANSFER EFFICIENCY
The regressions presented in Section 8 of this report were
derived from the contrasts associated with the independent
variables. Only those independent variables found to be signi-
ficant based on an analysis of variance were included. The
resulting regression coefficients were based on normalized
values of the experimental factor levels, 1 = the high value and
-1 = the low value.
This appendix presents nearly equivalent regression expres-
sions derived from a conventional multiple linear regression
program. Again, only independent variables found to be signifi-
cant by ANOVA are included. Whereas the Section 8 analysis
considered each factor to have taken on only two values (the
mean of the four actual high values and the mean of the four
actual low values), the analysis presented here used the actual
experimental data in every case. Differences are slight
because, for the most part, the planned high and low values of
each factor were actually achieved.
Results are presented in Table B-l.
138
-------
Table B-l
SCREENING PROCEDURE--MULTIPLE LINEAR REGRESSION
Air Atomized Electrostatic
Means
Standard Deviations
Correlation Coefficients
Vertical Cylinder
52.5 xA
40.0 xB
55. xC
1.5 xE
0.5 xF
6.072831771 SO
.5345224838 S5
1.603567451 S4
10.69044968 S3
10.69044968 S2
8.017837257 SI
.3080652921 RO1
-.8405781542 R02
-.2508531664 RO3
-.2288485027 RO4
.2860606284 RO5
Flat Plate
52
40
55
0
3
10
10,
8,
-0
676080988
5345224889
69044968
69044968
017837257
3489732586
4798382306
559811269
-.5234598879
xA
xB
•xC
xF
SO
34
S3
S2
SL
ROL
RO2
RO3
R04
Determinant of the Corre-
lation Matrix Inverse
Beta Coefficients
Regression Coefficients
Intercept
t Statistic
Standard Error
0
1
2
3
4
5
1.
.3080652921
-.8405781542
-.2508531664
-.2288485027
.2860606284
.2333333333
-0.4775
-0.1425
-.8666666667
3.25
40.1875
11.66666666
-31.83333331
-9.499999992
-8.666666659
10.33333332
27.02921615
Bl
B2
B3
B4
B5
Bl
B2
B3
B4
B5
BO
Tl
T2
T3
T4
T5
TO
1.
.3489732586
-.4798382306
-0. 559811269
-.5234598879
0
-0
-0
-3
.9986054889 R2
.4242640691 SE
Transfer Efficiency 0
A Voltage 1
B Atomizing Air 2
C Booth Air 3
E Lag Distance 4
F Shaping Air
139
16
165
1925
6
98.1625
2.455893648
-3.376853767
-3.939662728
-3.683840473
20.37178647
-9394259739
Bl
B2
B3
34
Bl
B2
B3
B4
BO
Tl
T2
T3
T4
TO
R3
1.382027496 SE
Transfer Efficiency
A Voltage
B Atomizing Air
C Booth Air
F Shaping Air
-------
Table B-l (continued)
Air Atomized Conventional
Vertical Cylinder
Means
Standard Deviations
Correlation
Determinant of the Corre-
lation Matrix Inverse
Beta Coefficients
Regression Coefficients
Intercept
t Statistic
40.
4.3625
57.5
0.5
1
18
2
10
Standard Error
667279649
5345224838
70828693
071533249
69044968
4408189206
7291604789
2003722366
5530273731
.9842287046
-.3689709194
.6551665119
-.1876931776
-0.514990196
-.0575446051
.5273127008
-.0167271871
-1.606354642
17.62887312
-3.292647716
5.835660816
-1.684830484
-4.615799271
16.72299965
0.962783067
.4913230413
xA
xB
xC
xE
SO
34
S3
S2
SI
RO1
RO2
RO3
R04
Bl
B2
B3
B4
Bl
B2
B3
B4
BO
Tl
T2
T3
T4
TO
R3
SE
Flat Plate
40.
4.3625
57.5
0
6
18
2
10
-0
573051042
5345224838
70828693
071533249
69044968
3293472157
4894084014
451327666
7115526266
.9842287046
xA
xB
xC
xE
SO
S4
S3
S2
SI
RO1
RO2
R03
R04
-.2844171654
.4097074919
-.4433988336
-.6877661293
-.1748746406
1.300016912
-0.155785678
-8.457496195
97.03508643
-3.833560108
5.511954203
-6.011676679
-9.310694592
35.26575728
.9836862467
1.282426085
Bl
B2
B3
B4
81
B2
B3
B4
BO
Tl
T2
T3
T4
TO
R3
SE
0 = Transfer Efficiency
1 = A Atomizing Air
2 = B Paint Mass Flow
3 = C Booth Air
4 = E Shaping Air
140
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO
EPA-600/2-88-026a
2.
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
Development of Proposed Standard Test Method for
Spray Painting Transfer Efficiency; Volume I.
Laboratory Development
5. REPORT DATE
April 1988
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO.
K. C. Kennedy
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Centec Corporation
11260 Roger Bacon Drive
Reston, Virginia 22090
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-03-1721, Task 2
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Air and Energy Engineering Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final: 1/82 - 1/87
14. SPONSORING AGENCY CODE
EPA/600/13
16. SUPPLEMENTARY NOTES AEERL project officer is Charles H. Darvin, Mail Drop 62b,
919/541-7633. Volume II describes the verification program.
16. ABSTRACT
The two-volume report gives results of a program to develop and verify
a standardized spray-painting transfer-efficiency test method. Both review of the
literature and laboratory research were conducted. Transfer efficiency measure-
ment methods presently used by industry were evaluated and compared. The best
characteristics of these methods were incorporated into the final proposed standard
method. The resulting method was determined to be viable for laboratory evalua-
tions. It still awaits adaptation and verification for production line applications.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Pollution
Spray Painting
Tests
Pollution Control
Stationary Sources
Transfer Efficiency
13B
13H
14B
8. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
150
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
141
------- |