153
A Pneumatic Conveying Powder Delivery System For Continuously Heterogeneous Material Deposition In Solid Freeform Fabrication by Shawn Fitzgerald Thesis submitted to the Faculty of the Virginia Polytechnic Institute and State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE IN MECHANICAL ENGINEERING APPROVED: _________________________________ Dr. Jan Helge Bøhn, Chairman ______________________________ ______________________________ Dr. Arvid Myklebust Dr. Ronald Kander July 1996 Blacksburg, Virginia Keywords: Rapid Prototyping, Powder Deposition, Selective Laser Sintering, 3D Printing, Freeform Powder Molding

Transporte Neumatico Thesis

Embed Size (px)

Citation preview

Page 1: Transporte Neumatico Thesis

A Pneumatic Conveying Powder Delivery System For

Continuously Heterogeneous Material Deposition

In Solid Freeform Fabrication

by

Shawn Fitzgerald

Thesis submitted to the Faculty of the

Virginia Polytechnic Institute and State University

in partial fulfillment of the requirements for the degree of

MASTER OF SCIENCEIN

MECHANICAL ENGINEERING

APPROVED:

_________________________________Dr. Jan Helge Bøhn, Chairman

______________________________ ______________________________Dr. Arvid Myklebust Dr. Ronald Kander

July 1996Blacksburg, Virginia

Keywords: Rapid Prototyping, Powder Deposition, Selective Laser Sintering,

3D Printing, Freeform Powder Molding

Page 2: Transporte Neumatico Thesis

A Pneumatic Conveying Powder Delivery System For

Continuously Heterogeneous Material Deposition

In Solid Freeform Fabrication

by

Shawn Fitzgerald

Jan Helge Bøhn, Chairman

Department of Mechanical Engineering

Great improvements are continuously being made in the solid freeform fabrication (SFF)

industry in terms of processes and materials. Fully functional parts are being created

directly with little, if any, finishing. Parts are being directly fabricated with engineering

materials such as ceramics and metals. This thesis aims to facilitate a substantial advance

in rapid prototyping capabilities, namely that of fabricating parts with continuously

heterogeneous material compositions. Because SFF is an additive building process,

building parts layer-by-layer or even point-by-point, adjusting material composition

throughout the entire part, in all three dimensions, is feasible. The use of fine powders as

its build material provides the potential for the Selective Laser Sintering (SLS), Three-

Dimensional Printing (3DP), and Freeform Powder Molding (FPM) processes to be

altered to create continuously heterogeneous material composition. The current roller

distribution system needs to be replaced with a new means of delivering the powder that

facilitates selective heterogeneous material compositions. This thesis explores a dense-

phase pneumatic conveying system that has the potential to deliver the powder in a

controlled manner and allow for adjustment of material composition throughout the layer.

Page 3: Transporte Neumatico Thesis

iii

Acknowledgments

I would like to thank:

• My family for all of their love, understanding, and monetary support.

• My friends for always being there when I wasn’t doing work.

• Dr. Bøhn for introducing me to rapid prototyping and all of his help.

• My committee members, Dr. Myklebust and Dr. Kander, for taking the and effort to

provide assistance during the academic year and summer sessions.

• Scott Houser, for providing help whenever needed and being there to hear my

complaints.

• The Virginia Tech CAD laboratory for providing the hardware and software needed in

the design of the experimental apparatus.

• To everyone in the ME machine shops for helping me build the apparatus.

This research was supported in part by the Naval Surface Warfare Center, Dahlgren

Division under contract N60921-89-D-A239, Order 0045. Any opinions, findings,

conclusions, or recommendations expressed in this thesis are those of the author and do

not necessarily reflect the view of the Naval Surface Warfare Center, Dahlgren Division.

Page 4: Transporte Neumatico Thesis

iv

CONTENTS

ABSTRACT ........................................................................................................................i

ACKNOWLEDGMENTS .................................................................................................iii

CONTENTS ......................................................................................................................iv

LIST OF FIGURES...........................................................................................................vii

LIST OF TABLES.............................................................................................................xii

CHAPTER 1 INTRODUCTION.......................................................................................1

1.1 PROBLEM STATEMENT..............................................................................6

1.2 SOLUTION OVERVIEW.................................................................................6

1.3 THESIS OVERVIEW........................................................................................8

CHAPTER 2 LITERATURE REVIEW...........................................................................10

2.1 SECONDARY SUPPORT MATERIALS.....................................................10

2.2 SECONDARY MICROSTRUCTURE MATERIALS..................................14

2.3 MATERIAL DEPOSITION..........................................................................17

2.4 PNEUMATIC CONVEYING........................................................................21

2.5 OBSERVATIONS .........................................................................................28

CHAPTER 3 METHODS.................................................................................................29

3.1 EXPERIMENTAL APPARATUS.................................................................29

3.1.1 POWDER PREPARATION & LOADING.....................................33

3.1.2 AIR HANDLING SYSTEM............................................................35

Page 5: Transporte Neumatico Thesis

v

3.1.3 MECHANIZED SLIDER COLLECTION SYSTEM......................41

3.2 EXPERIMENTAL SETUP.............................................................................45

3.2.1 ESTIMATING THE VOLUMETRIC FLOW RATE.....................46

3.2.2 EVALUATING THE POWDER PATH CHARACTERISTICS....48

3.2.3 EXPLORING MULTIPLE POWDER DEPOSITION...................50

CHAPTER 4 RESULTS...................................................................................................52

4.1 LINEAR VELOCITY EXPERIMENTS........................................................52

4.2 POWDER VOLUMETRIC FLOW RATE....................................................57

4.3 POWDER PATH CHARACTERISTICS......................................................63

4.4 MULTIPLE POWDER MIXING..................................................................71

4.5 MULTIPLE MATERIAL TRANSITIONS..................................................75

4.6 OBSERVATIONS .........................................................................................76

CHAPTER 5 MULTIPLE POWDER DEPOSITION SYSTEM...................................78

5.1: INTRODUCTION........................................................................................78

5.2: NOZZLE ARRAY AND POWDER DEPOSITION HEAD.......................80

5.3: MOTION SYSTEM......................................................................................83

5.4: POWDER PLUG CREATION & DELIVERY.............................................86

5.5: SOFTWARE & CONTROL..........................................................................89

5.6: POWDER PROPERTY EFFECTS...............................................................90

5.7: PERFORMANCE ANALYSIS.....................................................................92

Page 6: Transporte Neumatico Thesis

vi

5.8: DESIGN OVERVIEW....................................................................................93

CHAPTER 6 CONCLUSION ........................................................................................96

6.1: CONCLUDING REMARKS........................................................................96

6.2: CONTRIBUTIONS.......................................................................................97

6.3: RECOMMENDATIONS FOR FUTURE WORK......................................98

REFERENCES ...............................................................................................................101

APPENDIX A BILL OF MATERIALS FOR EXPERIMENTAL

APPARATUS .............................................................................106

APPENDIX B PARTS LIST FOR EXPERIMENTAL

APPARATUS .............................................................................107

APPENDIX C SUPPLEMENTAL DRAWINGS OF EXPERIMENTAL

APPARATUS .............................................................................108

APPENDIX D SUPPLEMENTAL DRAWINGS OF PROPOSED

SYSTEM ....................................................................................135

VITA ...............................................................................................................................140

Page 7: Transporte Neumatico Thesis

vii

LIST OF FIGURES

1.1: Typical Selective Laser Sintering system.................................................................3

1.2: Typical Three Dimensional Printing system...........................................................3

1.3: Freeform Powder Molding process..........................................................................4

1.4: Powder leveling system............................................................................................5

1.4: Schematic of experimental apparatus.......................................................................8

2.1: The Solid Ground Curing (SGC) process...............................................................11

2.2: FDM head dual-material nozzles...........................................................................12

2.3: Model Maker 3D Plotter.......................................................................................13

2.4: Possible methods for powder deposition...............................................................18

2.5: Internal secondary air pipe system........................................................................23

2.6: Plug formation using timer operated air knife........................................................24

2.7: Plug formation using alternating air valves.............................................................24

2.8: Flow pattern of plug phase conveying...................................................................26

3.1: Complete experimental apparatus..........................................................................30

3.2: Air handling and powder deposition components.................................................31

3.3: Mechanized slider collection components.............................................................31

3.4: Multiple plugs passing through parallel tubing system.........................................32

3.5: Loading of powder into polyurethane tubing........................................................35

Page 8: Transporte Neumatico Thesis

viii

3.6: Powder plug moving through tubing......................................................................35

3.7: Main air line division by three port manifold........................................................36

3.8: Screw clamp operating as flow rate controller.......................................................37

3.9: Nozzle with supply tubing and connections.........................................................38

3.10: Nozzle and support stand......................................................................................39

3.11: Severe stair stepping in adjacent layers..................................................................40

3.12: Motor control circuit diagram................................................................................42

3.13: Motor and thread winding system.........................................................................43

3.14: Exploded slider collection system..........................................................................44

3.15: Potentiometer and calibration template..................................................................45

3.16: Parameters tested for deposition characteristics....................................................49

3.17: Deposited powder path characteristics evaluated..................................................49

4.1: Output voltage vs. angular displacement of potentiometer...................................54

4.2: Velocity of slider system vs. motor input voltage.................................................56

4.3: Volumetric flow rate vs. powder plug lengths for a pressure of 34.5 kPa.............59

4.4: Volumetric flow rate vs. powder plug lengths for a pressure of 25.9 kPa.............60

4.5: Volumetric flow rate vs. powder plug lengths for a pressure of 17.2 kPa.............61

4.6: Actual powder path showing width and spread....................................................64

4.7: Powder deposition width vs. nozzle lengths.........................................................65

4.8: Powder deposition width vs. nozzle diameter.......................................................66

Page 9: Transporte Neumatico Thesis

ix

4.9: Powder deposition width vs. nozzle height...........................................................67

4.10: Powder deposition width vs. slider collection system speed................................68

4.11: Powder deposition width vs. air pressure..............................................................69

4.12: Default parameter setup, original image (a), annotated image (b)..........................72

4.13: Nozzle height of 7.0 mm, original image (a), annotated image (b)..........................73

4.14: Nozzle height of 9.0 mm, original image (a), annotated image (b)..........................73

4.15: Nozzle length of 10.0 mm, original image (a), annotated image (b)........................74

4.16: Nozzle diameter of 3.0 mm, original image (a), annotated image (b)...................74

4.17: Transition from red to green powder, original image (a), annotated image (b).....76

5.1: Multiple Powder Deposition system....................................................................79

5.2: Proposed deposition nozzle...................................................................................81

5.3: Nozzle array positioning........................................................................................82

5.4: Powder deposition head.........................................................................................83

5.5: Proposed motion system for the multiple powder deposition system.................85

5.6: Portion of motion pattern for powder deposition.................................................86

5.7: Powder plugs passing through tubing....................................................................87

5.8: Plug loading system, 1 of 54 required....................................................................87

5.9: Operation of proportioning valve for powder composition adjustment................89

A.1: Base board............................................................................................................109

A.2: Track groove........................................................................................................110

Page 10: Transporte Neumatico Thesis

x

A.3: Main support stand.............................................................................................111

A.4: Main air line 1......................................................................................................112

A.5: Main air line 2......................................................................................................113

A.6: Luer fitting............................................................................................................114

A.7: Luer lock...............................................................................................................115

A.8: Middle tubing.......................................................................................................116

A.9: Screw clamp.........................................................................................................117

A.10: Reducer fitting......................................................................................................118

A.11: Teflon tubing........................................................................................................119

A.12: Straight micro-fitting............................................................................................120

A.13: Nozzle .................................................................................................................121

A.14: Nozzle support....................................................................................................122

A.15: Motor spool.........................................................................................................123

A.16: Circuit support.....................................................................................................124

A.17: Slider top ............................................................................................................125

A.18: Slider bottom........................................................................................................126

A.19: Adhesive surface..................................................................................................127

A.20: Eye hook..............................................................................................................128

A.21: Funnel ..................................................................................................................129

A.22: Layout of fixtures attached to base board...........................................................130

Page 11: Transporte Neumatico Thesis

xi

A.23: Layout of fixtures attached to main support.......................................................131

A.24: Exploded manifold assembly................................................................................132

A.25: Exploded nozzle assembly...................................................................................133

A.26: Exploded motor assembly....................................................................................134

A.27: Powder print head................................................................................................136

A.28: Proposed nozzle...................................................................................................137

A.29: Motion system.....................................................................................................138

A.30: Powder loading system........................................................................................139

Page 12: Transporte Neumatico Thesis

xii

LIST OF TABLES

3.1: Laserlite™ LPC3000 polycarbonate properties....................................................34

4.1: Output voltage for potentiometer angular displacement increments.....................54

4.2: Time measurements for calibrated motor input voltage.........................................55

4.3: Linear velocity of slider collection for calibrated motor input voltages.................56

4.4: Deposited powder length for air pressure of 34.5 kPa and a linear velocity

of 192 mm/s............................................................................................................59

4.5: Volumetric flow rate for an air pressure of 34.5 kPa.............................................59

4.6: Deposited powder length for air pressure of 25.9 kPa and a linear velocity

of 192 mm/s............................................................................................................60

4.7: Volumetric flow rate for an air pressure of 25.9 kPa.............................................60

4.8: Deposited powder length for air pressure of 17.2 kPa and a linear velocity

of 192 mm/s............................................................................................................61

4.9: Volumetric flow rate for an air pressure of 17.2 kPa.............................................61

.10: Default parameter values........................................................................................64

4.11: Powder path width and spread for various nozzle lengths....................................65

4.12: Powder path width and spread for various nozzle diameters................................66

4.13: Powder path width and spread for various nozzle heights....................................67

4.14: Powder path width and spread for various slider collection system speeds.........68

Page 13: Transporte Neumatico Thesis

xiii

4.15: Powder path width and spread for various air pressures......................................69

5.1: Laserlite™ LN4010 nylon compound properties.................................................91

5.2: Laserlite™ LNF5000 nylon compound properties...............................................91

5.3: Laserlite™ LWX2010 wax compound properties.................................................92

Page 14: Transporte Neumatico Thesis

1

CHAPTER 1

INTRODUCTION

Solid freeform fabrication (SFF) is a set of modern technologies in which three-

dimensional solid objects are built directly from computer-aided design models. They are

additive processes which build objects by successively adding raw material in particles or

layers to create a solid volume of the desired shape [Burns93]. The major applications

for these objects include quick production prototypes for testing and design verification

and as patterns for casting and tooling. SFF can dramatically reduce the time and costs

required to bring a product to market.

Several technologies have been developed that accomplish these prototyping

objectives. They include photopolymer resin curing by ultraviolet light

(Stereolithograthy and Solid Ground Curing), thermal fusion of powders by laser scanning

(Selective Laser Sintering), joining of powders by ink-jet spreading of a binder (Three-

Dimensional Printing), compacting and sintering of powders (Freeform Powder Molding),

extruding heated thermoplastics (Fused Deposition Modeling), and laser cutting of

laminated sheets of paper (Laminated Object Manufacturing). SFF technologies are

progressing tremendously in terms of processes and materials used for the creation of

conceptual and functional parts, yet versatility in the selection of materials for each

system is still relatively limited and varies substantially for each specific system.

Page 15: Transporte Neumatico Thesis

2

Current SFF technologies offer multiple material choices. However, they only

permit a single material for each region of a part. Changes among material composition are

discrete. These additive build processes have the potential to fabricate parts with

selective heterogeneous material compositions. Realizing this potential, and, in particular,

developing the capability to fabricate parts with continuously variable material

composition will revolutionize existing SFF and manufacturing capabilities.

The ability to fabricate parts with continuously changing material composition

throughout the part would give an endless control over final part characteristics.

Engineers may then begin to design at the microscopic level, optimizing material

composition to provide the final part properties desired. Not all SFF systems can readily

adapt to a heterogeneous material blend. Some are best suited for building with a

premixed uniform material type, at least within each layer. Selective Laser Sintering

(SLS), Three-Dimensional printing (3DP), and Freeform Powder Molding (FPM) on the

other hand, fabricate parts out of powder and hence have the greatest potential for

adapting to heterogeneous material compositions.

SLS and 3DP use similar processes to prototype parts. First, a computer solid

model of the prototype is generated and sliced appropriately to generate the layerwise

data. Next, a powder-leveling mechanism lays a thin (75 µm to 250 µm) uniform layer of

powder. The material is then bonded, by laser thermal fusion (SLS) or by liquid adhesion

(3DP), in the image of the corresponding two-dimensional slice. The unbound powder

Page 16: Transporte Neumatico Thesis

3

Figure 1.1: Typical Selective Laser Sintering system [SFF Group95].

Figure 1.2: Typical Three-Dimensional Printing system [Michaels92].

remains for support. Finally, another layer of loose powder is deposited and the process

is repeated until the three-dimensional solid object is complete. Figures 1.1 and 1.2

illustrate the two fabrication systems.

Page 17: Transporte Neumatico Thesis

4

The FPM operates with a slightly different process to prototype parts. Rock and

Gilman [Rock95] describe the fabrication process. First, the layerwise cross-sectional

data is generated from the CAD model. Next, a thin layer of powder is deposited. One

powder type is deposited in the locations corresponding to the model’s cross-sectional

data, and a second powder is deposited elsewhere for support. No details of the

deposition method for the FPM are provided. After each layer is deposited, it is leveled

and compacted by a vertical press. After the deposition of all of the layers, the entire

powder volume is sintered. The two powders have different thermal responses, thus

only the powder deposited in the model’s location will be sintered. The secondary

powder is then removed. Figure 1.3 illustrates this process.

Figure 1.3: Freeform Powder Molding process [Rock95].

Machines in both the SLS and 3DP processes currently use a powder leveling

drum to distribute uniform layers of powder. First a piston raises a small increment

exposing an amount of powder to the roller mechanism. The roller advances spreading the

Page 18: Transporte Neumatico Thesis

5

powder over the buildspace. The roller returns in the opposite direction compacting the

powder, ensuring a near uniform layer. Figure 1.4 illustrates this process. This

application method allows for the use of multiple materials, but they must be mixed prior

to their distribution and therefore may be varied layer-by-layer at best. Development of

an alternate method of powder distribution is necessary in order to reach the potential of

building parts of heterogeneous material compositions, continuously varying in three

dimensions.

Figure 1.4: Powder leveling system [Lee93].

Page 19: Transporte Neumatico Thesis

6

1.1 PROBLEM STATEMENT

There is a need to develop continuously heterogeneous materials for high-performance

fabrication and physical rapid prototyping. Because SFF is an additive building process,

built layer-by-layer, adjusting material composition throughout the entire part, or even

layer, is theoretically feasible. The use of fine powders as their build material makes SLS,

3DP, and FPM potentially adaptable to a heterogeneous material composition. However,

a new method of powder deposition is required that will allow for the selective placement

of multiple combinations of powder composition to make possible the fabrication of

continuously heterogeneous parts. The objective of this thesis is to explore a new

powder deposition approach and to design, fabricate, and test a low cost experimental

apparatus to illustrate its fundamental operating principles.

1.2 SOLUTION OVERVIEW

A viable replacement to the roller distribution system currently used in SLS is a

pneumatic conveying system. Pneumatic conveying is the movement of dry material

through an enclosed pipeline by the motion of air. It has been used successfully for many

years in the chemical and processing industries for the transportation of materials such as

flour, granular chemicals, lime, soda ash, plastic chips, and coal [Sadler49]. Several

advantages can be gained by pneumatic conveying. They include dust and contamination

free transportation with a great flexibility in routing. It allows for build materials to be

Page 20: Transporte Neumatico Thesis

7

stored outside a prototyping device. It is easily automated and controlled, thus the

proportional mixing of several powders can be performed without manual interference.

A delivery system of this kind provides the potential for depositing continuously

heterogeneous powder layers for use in SLS, 3DP, and FPM. The feasibility of

employing such a system is shown by the design and fabrication of a low cost

experimental apparatus that explores the operating principles and performance

capabilities of a larger scale system. It demonstrates the ability to control the proportion

of three materials deposited through a single nozzle and the results are extrapolated to

estimate the performance of a large scale deposition system of this manner.

Figure 1.5 illustrates the fundamental components of the experimental apparatus

used for testing. A standard connection to a compressed air line supplies the pressurized

air. A filter and pressure regulator cleanse the compressed air of any impurities or surges.

A manifold distributes the pressurized air into three parallel lines. The air then travels

through a flow rate control which adjusts the air flow rate to a desired condition. Each of

three materials is transported through the three parallel lines, respectively. By adjusting

the airflow used in each individual pipe, different rates of mass flow are achieved for each

material. Thus, the proportion of each material deposited by the nozzle can be

controlled. The pipes merge to allow mixing before exiting through the same orifice in a

nozzle. The material is deposited ont a collection system for visual inspection and

Page 21: Transporte Neumatico Thesis

8

Figure 1.5: Schematic of experimental apparatus.

analysis. The linear motion is provided by a motor and slider system capable of

adjustable and constant linear velocity.

1.3 THESIS OVERVIEW

Chapter Two reviews existing SFF material systems, methods of powder deposition, and,

in particular, pneumatic conveying.

Chapter Three provides a detailed description of the design and fabrication of the

experimental apparatus. Also discussed are methods of data collection and how they are

related to the performance of the experimental apparatus and a proposed larger scale

system.

Chapter Four includes the results and data analysis of the deposition achieved by

the experimental apparatus.

Page 22: Transporte Neumatico Thesis

9

Chapter Five develops a conceptual design of a scaled up version and estimates

the performance that can be expected based on the performance results of the

experimental system.

Chapter Six remarks on the performance capabilities of the experimental system

and that expected for a larger scale model. It also analyzes the deficiencies expected in

delivering powder in a pneumatic conveying system and provides recommendations for

future work.

Finally, the Appendix includes a Bill of Materials, a parts list, and dimensioned

component drawings along with assembled and exploded drawings of the experimental

apparatus.

Page 23: Transporte Neumatico Thesis

10

CHAPTER 2

LITERATURE REVIEW

Solid freeform fabrication (SFF) is a relatively new technology, yet tremendous progress

has been made in terms of the systems and materials. This chapter examines several

methods in which SFF systems are incorporating multiple materials and several methods

of material deposition, in particular pneumatic conveying. Section 2.1 reviews a few

processes which have added a secondary material for support generation. Section 2.2

examines methods of controlling the microstructure of fabricated parts. These methods

include dual material blends used for reducing the difficulty associated with fabricating

with ceramics and metals, selective material deposition within each layer, and depositing

material of controlled composition within each layer. Section 2.3 reviews the current

roller distribution system, alternative methods of depositing an entire layer of powder,

and ink-jet-style deposition. Section 2.4 examines the deposition method chosen for this

research, pneumatic conveying. Section 2.5 provides some general observations based

upon the literature reviewed.

2.1 SECONDARY SUPPORT MATERIALS

Layered manufacturing requires supports during fabrication to provide stability for

overhang structures. Some SFF systems have solved this problem by building support

Page 24: Transporte Neumatico Thesis

11

structures using a temporary secondary support material. Examples are Solid Ground

Curing (SGC), Fused Deposition Modeling (FDM), Freeform Powder Molding (FPM),

and the Model Maker 3D Plotting System.

In the SGC process (Fig. 2.1), the part is fabricated using ultraviolet sensitive

photopolymer resin. During the processing of each layer, the uncured resin is removed

and wax is added to fill in the voids. Once solidified, the wax serves as a solid filler

providing complete support throughout fabrication [Burns93]. This petroleum soluble

wax is washed away during post-process finishing.

Figure 2.1: The Solid Ground Curing (SGC) process [Burns93].

Page 25: Transporte Neumatico Thesis

12

The FDM system extrudes two materials using a dual nozzle system (Fig. 2.2).

One nozzle is used for dispensing the model material, and the second nozzle is used for

dispensing material for use as a base and support for the model material. The support

material is designed not to adhere well with the model material and is very brittle, thus it

can easily be removed during post-processing. Tests have also been performed using an

ammonium soluble ABS plastic as a support material [FDM96].

FDM head

Modeler tip Support tip

Model materialsource

Support materialsource

Figure 2.2: FDM head dual-material nozzles.

The FPM system described by Rock and Gilman [Rock95] uses a secondary

powder to contain the model material and provide three dimensional support (Fig. 1.3).

For each cross-section of the fabricated part, a primary powder type is deposited in the

locations corresponding to the part’s cross-sectional data and a secondary powder is

deposited elsewhere for total support. The two powders have different thermal

Page 26: Transporte Neumatico Thesis

13

responses, thus during sintering only the primary powder will be sintered. The unbound

secondary powder is poured out during post-processing.

The Model Maker 3D Plotter (Fig. 2.3) distributes liquid thermoplastics and

waxes through a dual inkjet system [Sanders96]. The dispensed thermoplastic solidifies

to form the model and the dispensed wax solidifies to form the support structure. Similar

to SGC, the petroleum soluble wax is washed away during post-processing.

Figure 2.3: Model Maker 3D Plotter [Sanders96].

The dual material systems previously described have distinct areas of model and

support material. The support material serves no purpose in the final application of the

part and is used only as a means of building a better quality part. It is removed during

finishing and cleaning. However, these systems illustrate the potential of fabricating with

selectively placed materials throughout each layer.

Page 27: Transporte Neumatico Thesis

14

2.2 SECONDARY MICROSTRUCTURE MATERIALS

Secondary materials have also been used in SFF to facilitate the fabrication of parts with

control over their microstructure. This section discusses several methods of gaining this

microstructure control. These include dual material blends used for reducing the difficulty

associated with fabricating with ceramics and metals, selective material deposition within

each layer, and depositing material of controlled composition within each layer.

Using a dual material blend for fabrication in SFF is not a recent advancement. In

this method, a secondary material is used to contain the primary material, usually ceramic

or metal, during fabrication. 3DP utilizes this approach to create ceramic [Yoo93] and

metal [Michaels92] preforms. The binder is selectively spread onto the powder bed,

either ceramic or metal, through ink-jet deposition and captures the powder in the three-

dimensional shape desired. After the preform is fabricated, a firing process removes the

binder and densifies the part.

SLS uses a similar dual material approach. In its indirect processing method, a low

melting temperature binder is uniformly premixed with the material of choice prior to the

sintering process [Bourell92] [Badrinarayan92]. After sintering, the binder is burnt off,

leaving behind the desired material. The sintering and debinding processes result in a

lower density “green” part. Post-processing is necessary to increase the strength and

utility of the part. Several methods have been studied, including infiltration [Deckard93]

Page 28: Transporte Neumatico Thesis

15

[Vail92] [Tobin93] [Sindel94], hot isostatic pressing [Carter93], and heat treatments

[Prabhu93].

FDM and similar systems have also used a similar method. Agarwala et al.

[Agarwala95] have worked with fabricating with a ceramic powder blended within

polymer/wax based binder systems for use with FDM. Greul et al. [Greul95] have

developed a process similar to FDM called Multiphase Jet Solidification (MJS). They

deposit a blend of 50% binder (polymers and waxes) and 50% metal powders. After a

deposition process similar to FDM, the binder is removed in solvent and the porous

metal part is sintered to reach the final part density.

Other research has focused on adjustment of material content layer-by-layer.

Pegna [Pegna95] has been performing research using multiple reactant bulk materials to

form parts. In his experiments, alternate layers of sand and cement were applied. After

the preparation of each layer, they were placed into a pressurized steam chamber of

3 atmospheres and 300ºC. Successful bonding was achieved between the two reactant

materials. Applications of this sort were directed into the construction industry. Pegna

successfully combined multiple materials but they were done layer-by-layer and much

automation must be added to make this system feasible for rapid prototyping for each

layer of material was hand deposited and leveled.

Beck et al. [Beck92] report on a method of altering material content within each

layer. They developed the recursive mask and deposit, or MD*, process to fulfill this

Page 29: Transporte Neumatico Thesis

16

need. The MD* process is a thermal spray shape deposition system. Thermal spray

methods are used to deposit thin, heated planar layers of material, which then solidify.

The part is built by spraying a succession of cross-sectional layers.

The MD* system has the potential for selective material deposition within each

layer, enabling multi-material parts to be produced during a single build process. The

process could be adapted to manufacture complete, integrated electromechanical

assemblies, including mechatronics. Mechatronics are parts that perform both mechanical

and electrical functions.

Selective control of the composition of particle clusters within each layer can also

be achieved by 3DP in its current configuration. The original powder composition

remains uniform, but variance in the particle clustering can be obtained by manipulating

the application of the binder. Multiple jets of different binder composition or

concentration could be used to build components with composition and density variations

on a fine scale [Cima92].

Cima and Sachs have been using a process similar to the original 3D Printing

device to build objects by controlling their microstructure and composition [Ashley95].

Rather than ejecting a binder, the nozzles eject a ceramic slurry. Alteration of the slurry

composition allows for changes in the material composition throughout the building cycle.

One unique application for this is products such as pills that release measured drug doses

at specified times during the day.

Page 30: Transporte Neumatico Thesis

17

SFF devices often build parts on a point-by-point basis, and some devices allow

for the possibility of varying the composition and structure of particle clusters from

position to position with complete freedom. Several potential applications of this

building process include components with anisotropic thermal, electrical, or mechanical

properties or microengineered porosity [Cima92].

2.3 MATERIAL DEPOSITION

SLS and 3DP can benefit from an alternate powder deposition system. Not only could it

potentially allow for heterogeneous material composition, but it could also provide better

packed powder layers. There have been several alternative approaches of depositing

powder layers. This section reviews the current roller distribution system and several

alternative powder delivery methods including simultaneous deposition of entire layers

and deposition jet-style scanning.

The current powder deposition method, shown in Fig. 1.3, uses a powder leveling

drum to distribute a layer of powder. Yoo et al. [Yoo93] describe this deposition

process. First, the build area is lowered and a thin layer of loosely packed agglomerates is

created by spreading the powder over the build area with a counter-rotating spreader rod.

Next, the build area is raised slightly to expose the upper portion of the loosely packed

layer. The spreader rod is then rotated back across the build area to compact the powder

Page 31: Transporte Neumatico Thesis

18

into a uniform layer rather than to sweep away the excess powder. This produces a

dense uniform layer upon which the sintering or binding is performed.

Van der Schueren and Kruth [Van der Schueren95] explored three methods of

powder deposition for a fabrication system based on the principles of SLS. The first

approach used a scraper blade to sweep powder over a build container (Fig. 2.4.a).

Several problems were reported. To ensure that the build plane was completely covered

with each new layer, a surplus of powder was needed at the beginning of each sweep.

The excess powder increased the weight of the material to be pushed, thus, there was an

increase in friction between the transported powder and the underlying layer. Another

problem arose from in the fixed intersection edge between scraper blade and the powder

surface. Any irregularity in the powder would be swept over the build surface and cause

unacceptable grooves in the surface. The final problem was the low powder bed density.

This would result in structurally weak parts with high porosity.

Figure 2.4: Possible methods for powder deposition [Van der Schueren95].

Page 32: Transporte Neumatico Thesis

19

The second approach explored was the use of a counter rolling cylinder to sweep

the powder into place (Fig. 2.4.b). The rotary movement of the cylinder caused any

irregularity in the powder appearing in the intersecting edge between the cylinder and the

powder surface to be dissipated because the irregularity will only remain in the path of

the roller momentarily. This method also allows for compacting the powder while its

distributed by simultaneously applying a vertical vibration during deposition. This

vibratory counter rotating method is similar to the approach taken by Yoo et al. [Yoo93].

The third and final approach explored was the one chosen for implementation into

Van der Schueren and Kruth’s fabrication system. It delivers powder in a slot feed

mechanism (Fig. 2.4.c). The powder flows vertically out of the feeder when there is a gap

between the slot and the substrate. This method significantly reduces the friction created

in the other two methods. However, like the scraper blade method, it minimally

compacts the powder bed. Van der Schueren and Kruth therefore employ a vibratory

counter rolling cylinder in combination with the slot feeder in order to deposit the powder

with minimal friction while compacting the powder. An alternative compaction method

would be to apply a vertical press, as is done in the FPM process [Rock95]. This

method removes the contact edge and sliding friction problems. Any irregularity in the

powder bed will not be propagated throughout the rest of the build plane.

Melvin and Beaman [Melvin91] at the University of Texas at Austin developed

an electrostatic powder application system. It delivers the powder from a storage

Page 33: Transporte Neumatico Thesis

20

container using compressed air, passes the powder particles through an electric field, and

sprays them onto a grounded plate. The experimental apparatus consisted of a powder

canister which holds and delivers powder similar to an aerosol can, flow control devices

that regulate compressed air flow into the canister, and air flow that carries the air from

the powder canister to the feed head. The powder leaves the canister and enters the feed

head which forces it through a spray nozzle and electromagnetically charges it. Two 13.5

kV, 0.31 mA power sources charge the powder in the nozzle. One source delivers the

positive charge, while the other source delivers the negative charge. The layer thickness

ranged from 0.012” (0.30 mm) to 0.016” (0.41 mm) with a variance of 0.002” (0.05 mm)

to 0.003” (0.08 mm). The new application method showed a reduction in the number of

air pockets as compared to the roller deposition method. Clogging was observed each

time the powder flow was cut off. These clogs were easily cleared using a pressurized air

purge.

Rather than depositing the powder by an entire layer, the method of powder

deposition developed in the work of this thesis experiment explores the use of ink-jet

style deposition using an array of nozzles for deposition. Multiple nozzles help to

reduce the time required to deposit a layer of powder over a given area. Their use is not

uncommon in either SFF technologies or in the printing industry. 3D Printing has used an

eight-nozzle print head [Ashley95] and a 32 nozzle print head [Michaels92] to distribute

the binder. Recent printheads have grown in size to 1280 jets [Ashley95]. These nozzle

Page 34: Transporte Neumatico Thesis

21

clusters are moved at linear speeds reported between 1.65 [Sachs92] and 2.5 m/s

[Cima92].

In the printing industry, high resolutions must be maintained without any sacrifice

in speed. Arrays of 128 bubble jet nozzles have been produced on one chip with a nozzle

density of 12 nozzles/mm [Sachs92]. An example of a high speed application is the

“Djit” printer from Diconix, Inc. It uses a line printing bar containing 1500 jets to print at

linear speeds up to 5 m/s [Heinzel85].

The deposition requirements proposed for this thesis are unique. The deposition

systems previously discussed are only capable of depositing single materials, at least

within each layer. Adapting to an ink-jet style of powder deposition would provide the

potential for selective powder placement. Current ink-jet systems cannot accommodate

the deposition of powder due to the larger particle sizes and higher viscosity of the

powder. Alternative powder transportation methods need to be explored.

2.4 PNEUMATIC CONVEYING

The flow of powders represents a greater challenge than that of the low viscosity binder

and printing ink. Geldart [Geldart90] states that powders are not solids, although they

can withstand some deformation. Powders are not liquid, although they can be made to

flow. Powders are not gases, although they can be compressed. The space between

Page 35: Transporte Neumatico Thesis

22

particles is filled with gases and therefore the solid/gas interaction, the interparticle

contact area, and adhesion between particles must be considered during flow analysis.

In the past, solids were commonly transported in suspension form via lean-phase

conveying. Lean-phase conveying is defined by the low volumetric concentration of

solids, typically less than ten percent. Conveying in this form may cause a few

problems. There is a high rate of pipe wear and particle attrition due to the high

velocities. Depositing powder in this manner for use in SLS, 3DP, or FPM would also

have several other undesirous effects. For instance, it is likely to deposit the powder

with too large a force and disturb the underlying powder bed, and the control over the

accuracy of desired powder compositions may be lost due to the local sandstorm effect.

An alternate transportation approach is dense-phase conveying. It is defined as

the conveying of particles by air along a pipe that is filled with particles at one or more

cross-sections [Konrad86]. It offers improvements in slower air speeds and lower

volumes of gas required for transporting the same amount of material, which is of

particular importance if an inert gas is needed to reduce the risk of explosions [Konrad86].

The work of Albright et al. [Albright51] was one of the first to study dense-phase

conveying. The purpose was to minimize the amount of gas required to feed solids into a

coal gasification reactor. Since then, its use has grown substantially.

Dense-phase conveying appears more appropriate for powder deposition in SLS,

3DP, and FPM. The slower air speeds are less likely to disturb the previously deposited

Page 36: Transporte Neumatico Thesis

23

powder bed and greater control over powder placement should be gained. The lower

volumes of air should also benefit the SLS process which uses an inert gas, usually

nitrogen, in the process chamber to prevent oxidation or explosions [Behrendt95].

Dense-phase conveying is not without problems. Since a section of the pipeline is

completely filled with material, clogging can be a common occurrence and the flow

somewhat unstable. One approach to reduce this clogging has been to artificially create

distinct plugs of material. Conveying in this form is known as plug-phase conveying.

Konrad [Konrad86] reports of three commercial systems developed to deal

exclusively with plug-phase conveying. In the first, a bypass system was developed by

Lippert [Lippert66] to stabilize pressure fluctuations in plug-phase conveying (Fig. 2.5).

The conveying pipeline has an additional bypass pipeline in which there are holes at set

intervals. No additional air is blown into this bypass pipe. It merely serves as an

alternative route for the conveying air to break up lengthy plugs.

Inner TubeFluted Nozzles

Figure 2.5: Internal secondary air pipe system [Mainwaring93].

The other two systems artificially induce plugs of solid material separated by the

addition of a secondary air source. There are distinct plugs of material separated by plugs

Page 37: Transporte Neumatico Thesis

24

of air. One such system uses an air knife to provide regular pulses of air that will chop

up the moving solids fed into the pipeline (Fig. 2.6). In the other system, the plugs are

created by using alternating air valves (Fig. 2.7).

Figure 2.6: Plug formation using timer operated air knife [Geldart90].

Figure 2.7: Plug formation using alternating air valves [Geldart90].

Page 38: Transporte Neumatico Thesis

25

SLS, 3DP, and FPM could use a powder delivery system similar to the

commercial plug-phase conveying systems utilizing an additional air source with some

modifications. These systems do not provide a continuos flow of powder, but this can be

overcome by running two parallel pipelines operating at opposite phases. They would

also require the addition of a dispensing nozzle to mix and to accurately deposit the

powder onto the underlying powder bed. Various powders must be delivered to the

dispensing nozzle through different tubing systems for continuously heterogeneous

material deposition. By controlling the velocity of plugs throughout each of the material

supply systems, different proportions of each material could be combined and deposited.

Unfortunately, experimental evidence available in literature regarding plug-phase

conveying is contradictory [Konrad81]. Lippert [Lippert66] measured the pressure drop

across individual moving plugs in a horizontal pipeline. The pressure drop was found to

be proportional to the plug length squared divided by the pipe diameter. Flain [Flain72]

used a fine cohesive material to construct plugs of a given length in a horizontal pipeline.

He then measured the pressure drop required to move these plugs and concluded that this

pressure drop was proportional to the square of the plug length. While they agree that

the pressure drop is proportional to the square of the plug length for fine cohesive

material, Dickson et al. [Dickson78], however, suggest a linear relationship for both

coarse and fine materials. Fine powders are defined as those with particle sizes between 1

Page 39: Transporte Neumatico Thesis

26

and 100 µm [Brown70] which are also commonly used in SLS [Lakshminarayan92] and

3DP [Sachs92].

Konrad [Konrad86] has done extensive work on plug-phase pneumatic conveying.

He found that solid materials flow in discrete plugs that fill the tube cross-section at

approximately maximum packing density. Between the plugs, the upper part of the pipe

contains moving air with some dispersed particles. The lower half is filled with a

stationary bed of particles (Fig. 2.8). Each plug picks up the particle bed in front of it,

while leaving behind a stationary layer of nearly equal thickness behind it.

Air

Stationary particles

Moving particles

Figure 2.8: Flow pattern of plug-phase conveying.

Konrad et al. [Konrad80] developed a theoretical model to predict the pipeline

pressure drop in horizontal dense-phase plug conveying. The theory is based on the

following premises. The material is conveyed only in the plugs and in the regions just in

front of and behind them. There is a layer of stationary material between the plugs. The

Page 40: Transporte Neumatico Thesis

27

flow pattern resembles that of a gas-liquid system. The pressure drop required to move a

single horizontal plug is therefore given by the expression:

( ) ( )∆P

H

F

D

c

D DB w

w w w w w wk k c= + ++ ⋅ +

+24 4 1 4cos cos

(2.1)

where,

H is the plug length. ρB is the bulk density.

g is the acceleration due to gravity. µw is tan φw.

φw is the angle of wall friction. φ is the angle of internal friction.

kw is the Jansen coefficient at the wall. D is the pipe diameter.

F is the stress on the front end of the plug. C is the interparticle cohesion.

ω = sin-1 (sin φw / sin φ). Cw is the particle wall cohesion.

This equation can only be used to predict general flow trends in the experimental

apparatus since it uses multiple tubing systems with several of the above parameters

being unknown. There is therefore a need to collect experimental data for each of these

tubing systems in order to determine these parameters and thereby enable the use of this

equation.

Page 41: Transporte Neumatico Thesis

28

2.5 OBSERVATIONS

The literature has shown that incorporating multiple material systems into SFF

technologies is beneficial in fabricating for a variety of utilities. The use of multiple

materials is evolving from a temporary fabrication aid to a tool for producing parts with

“spatially tailored material properties” [Rock95] for a variety of applications. SLS, 3DP,

and FPM could benefit from an alternative deposition mechanism that will provide

potential for fabrication of three-dimensional heterogeneous parts. For this research, the

deposition mechanism combines principles from ink-jet printing and plug-phase

pneumatic conveying for potentially delivering powder with continuously heterogeneous

material composition.

Page 42: Transporte Neumatico Thesis

29

CHAPTER 3

METHODS

This thesis explores the feasibility of implementing a plug-phase pneumatic conveying

powder delivery system for three-dimensional continuously heterogeneous material

deposition for potential use in SLS, 3DP, and FPM. This chapter describes the

experimental apparatus developed to explore deposition of this kind. Section 3.1

describes the design, fabrication and operation of the experimental apparatus. This

includes the powder preparation, air handling system, and mechanized slider collection

system. Section 3.2 describes the experimental setup and testing procedure for each

evaluation of the performance of the experimental apparatus.

3.1 EXPERIMENTAL APPARATUS

An experimental apparatus (Fig. 3.1) was designed and built to demonstrate the

deposition characteristics achievable by plug-phase pneumatic conveying powder

delivery. The apparatus is capable of depositing a narrow path of powder in a thin layer

with selectively variable powder composition.

Page 43: Transporte Neumatico Thesis

30

Figure 3.1: Complete experimental apparatus.

The experimental apparatus consists of two major functional components. One is

the pressurized air handling system, which is responsible for transporting the multiple

powders through the tubing system, delivering them to the nozzle in appropriate

quantities, and discharging them through the nozzle orifice in the composition desired

(Fig. 3.2). The other major functional component is the mechanized slider collection

system, which is responsible for providing linear motion of constant, yet adjustable,

velocity for the collection of an even distribution of deposited powder (Fig. 3.3).

Page 44: Transporte Neumatico Thesis

31

Figure 3.2: Air handling and powder deposition components.

Figure 3.3: Mechanized slider collection components.

Page 45: Transporte Neumatico Thesis

32

This experimental apparatus demonstrates the feasibility of depositing powder

with selectively continuous heterogeneous composition. It simulates the operation of

commercial plug-phase conveying systems, such as those discussed in Section 2.4, by

manually creating and transporting a single plug. In order to deposit multiple powders

simultaneously, three plug-conveying systems are run in parallel, with each transporting a

different powder. They join at a nozzle prior to being deposited (Fig. 3.4). Pressurized

air is used to propel the plugs of powder through the tubing. Control of the air flow rate

driving each plug may be used to control the rate at which each of the parallel plugs reach

the nozzle, thus controlling the contribution each plug makes to the final deposited

powder composition.

Parallel powder plugs

Intersection of parallel tubing systems at nozzle

Discharge point (nozzle)

Mixed powder

Air

Air

Air

Figure 3.4: Multiple plugs passing through parallel tubing system.

Page 46: Transporte Neumatico Thesis

33

The following sections discuss the material preparation and the major components

and functions of the experimental apparatus used for continuously heterogeneous powder

deposition by plug-phase pneumatic conveying.

3.1.1 POWDER PREPARATION & LOADING

The material used for deposition in the experiments was Laserlite LPC3000

polycarbonate which was supplied by the DTM Corporation. It is one of the materials

available for use in DTM’s commercial Sinterstation 2000 systems (SLS processes). It

was chosen for this experimental work for several reasons. Many deposition methods

require the powder transported to be able to flow with a minimal amount of shear applied

to it [Sachs92]. The polycarbonate’s spherical particle shape allows for easy flow.

Guidelines suggest that the inside diameter of the pipe should be at least three times that

of the largest particle size to avoid potential pipe blockage [Marcus90]. Even with the

miniature tubing inner diameters used, 1.5 mm, the particle sizes, 30 - 175 µm, are well

within this limit. Polycarbonate is not easily affected by environmental conditions, and its

material properties should therefore remain consistent throughout testing. The general

properties of Laserlite LPC3000 polycarbonate are shown in Tbl. 3.1.

Page 47: Transporte Neumatico Thesis

34

Table 3.1: Laserlite™ LPC3000 polycarbonate properties [DTM94c].

General Properties Value Test MethodSpecific Gravity, 20ºC 1.20 g/cm3 ASTM D792Moisture Absorption, 20ºC, 0.35% ASTM D570 65% relative humidityPowder Tap Density 0.62 g/cm3 ASTM D4164Volume Average Particle Size 90 microns laser diffractionParticle Size Range, 90% 30-175 microns laser diffraction

The polycarbonate was also chosen because it is amorphous and hence easily

dyed. This was an important consideration because the deposited powder composition

was to be visually inspected. Quantities of powder were therefore dyed red, green, and

blue. The dying procedure was performed by diluting standard food coloring with water

and mixing it with a quantity of polycarbonate powder. This resulted in a thick slurry

with much conglomeration (i.e., groups of particles adhered together). The mixture was

dried at 70°C for several hours with periodic mixing. During mixing, all conglomerates

were broken to prevent an increase in particle size. The drying and mixing processes

returned the powder to characteristics similar to the original with exception of its color.

The powder was sealed in a bag with desiccant to eliminate the potential effects of

atmospheric humidity during long-term storage.

To simulate the operation principles of the plug-phase pneumatic conveying

systems commercially available, the plugs were manually created. Prior to pressurizing

the tubing system, a single plug of loosely compacted powder was formed in each of the

three polyurethane tubing lines. They were created by bending a section of tubing and

filling it with powder via a funnel (Fig. 3.5). Once pressurized, the plugs were pushed

Page 48: Transporte Neumatico Thesis

35

into the smaller diameter Teflon tubing and their density increased significantly. The

powder would then be ready to flow through the tubing in plug form until discharging out

the nozzle (Fig. 3.6).

Funnel

Polyurethanetubing

Figure 3.5: Loading of powder into polyurethane tubing.

Tubing

Powder plug

Figure 3.6: Powder plug moving through tubing.

3.1.2 AIR HANDLING SYSTEM

Pressurized air is used to propel the transported material within a pneumatic conveying

system. For this experiment, the air was provided through connections to high-pressure

air lines. These connections include a pressure regulator to eliminate pressure surges from

Page 49: Transporte Neumatico Thesis

36

the supplied air and a controller and gage to adjust the pressure of the dispensed air. Two

connections are used to provide two independently controlled pressure sources. An end

of each main air line of the experimental apparatus slips over barbed fitting to connect to

the pressurized air source. These two main air lines are constructed of polyurethane

tubing that are 1.8 m and 3.4 m long, respectively, and each with an inner diameter of 3.2

mm and an outer diameter of 6.4 mm. The other end of each tube is connected to a three

port manifold. The manifold receives pressurized air from the two independently

controlled pressure sources and distributes it among three parallel tubing lines in

proportions determined by desired powder compositions (Fig. 3.7).

Main air lines

Three port manifold

Figure 3.7: Main air line division by three port manifold.

Each of these three parallel tubing lines is loaded with powder prior to

pressurizing the lines (see Section 3.1.1). Screw clamps are used to constrict the diameter

of each of the parallel tubing lines and thereby allowing for variance of the air flow rates in

each tube (Fig. 3.8). They are located prior to the powder plugs. They therefore affect

Page 50: Transporte Neumatico Thesis

37

the air flow rate that reaches the powder plugs rather than adjusting the diameter of tubing

that the plugs pass through. Each tube has a length of 210 mm, with an inner diameter of

3.2 mm and an outer diameter of 6.4 mm, and is constructed of a flexible material

(polyurethane) to allow for easy routing and compression by the screw clamp. Each of

these parallel tubes is connected to smaller tubing line for delivery to the nozzle by a

reducing compression fitting. Each of the three smaller tubes have a 1.6 mm inner

diameter, a 3.2 mm outer diameter, and a length of 120 mm. Each of these smaller tubes is

coated in Teflon since its extremely low coefficient of friction facilitates the easy passage

of the powder plugs.

Screw clamp

Flexibletubing

Figure 3.8: Screw clamp operating as flow rate controller.

Each Teflon tube is connected to the deposition nozzle by straight compression

fittings (Fig. 3.9). The nozzle was fabricated out of ABS plastic with a FDM1600 rapid

prototyping system. The nozzle has three equally spaced tubing extensions, each

Page 51: Transporte Neumatico Thesis

38

extending 30 degrees from the vertical with an inner diameter of 1.5 mm, an outer diameter

of 3.5 mm, and length of 15.0 mm. These tubing extensions attach to the material supply

lines. These extensions join and merge together through a 1.5 mm diameter passage of

length 6.0 mm before reaching the nozzle’s exit orifice. Alternative nozzles were

fabricated for testing of different lengths and diameters. The nozzle is held stationary in a

support bracket (Fig. 3.10). This support bracket provides three positions of various

vertical heights for nozzle placement and can be positioned along the track groove in a

variety of positions using Velcro adhesion.

Nozzle

Straight compressionfittings

Powder supplytubing

Figure 3.9: Nozzle with supply tubing and connections.

Page 52: Transporte Neumatico Thesis

39

Figure 3.10: Nozzle and support stand.

Proper design and fabrication of the nozzle is fundamental to the success of the

powder deposition. The nozzle has three equally-spaced identical extensions to accept

powder from three sources, respectively. They are aligned 30 degrees from the vertical to

provide sufficient space for the fittings connecting it to the Teflon tubing. Too large an

angle could attribute to too severe a deviation in the flow pattern and too small an angle

would require tremendous lengthening of the extensions to provide sufficient space for the

fittings. The extensions join and merge together to allow for mixing of the powders

reaching the nozzle before discharging them through the nozzle orifice.

The fabrication process characteristics inherent to a FDM1600 system were a

significant factor in the design of the nozzle. It is important not to create areas where

support material is required and would be embedded inside the nozzle, and where it

would be difficult, if not impossible, to remove. Also, there is a great concern for the

surface quality of the interior passages of the nozzle. Severe stair-stepping patterns

between adjacent horizontal layers (Fig. 3.11) should be avoided or they will disrupt

Page 53: Transporte Neumatico Thesis

40

Stair-step pattern observed betweenadjoining layers

Figure 3.11: Severe stair stepping between adjacent layers.

powder flow. Stairsteps become more pronounced as the part surface becomes more

horizontal.

After nozzle fabrication, finishing was required to minimize any stray ABS plastic

strands or fabrication inconsistencies from the interior of the nozzles. This was

performed by washing the nozzle, alternatingly with acetone and water. Acetone etched

away the ABS plastic and water cleansed the plastic of the solvent to eliminate any

residual effects. These fluids were flushed through the nozzle interior passages with eye

droppers. Next, the interior passages of the nozzle were bored out to ensure that each

passageway was clear of obtrusions. A final acetone and water wash was performed on

the nozzle interior. Finally, a light sanding was performed to enhance exterior surface

quality.

Page 54: Transporte Neumatico Thesis

41

3.1.3 MECHANIZED SLIDER COLLECTION SYSTEM

A mechanized slider collection system was developed to provide a relative linear motion

of the powder delivery head. The system provides constant, adjustable head velocity and

collects the discharged powder such that it may be saved for subsequent evaluation.

The slider collection system is propelled by a Delco gear-reduced ball-bearing

motor which operates at 200 rpm with no load and a power supply of 12 VDC and

1 amp. A DC power supply plugged into a 115 VAC outlet supplies the requisite power

to run the motor. The velocity of the slider collection system can be controlled so that

the powder deposition may be analyzed at several linear speeds. This is provided

through an adjustable speed control circuit which controls the motor shaft’s angular

velocity and thus the velocity of the slider collection system. The motor control circuit

uses pulse-width modulation through an integrated circuit (IC) chip to provide a control

range of 5% to 98% by simple adjustment of the potentiometer knob. The motor

controller circuit diagram is shown in Fig. 3.12.

Page 55: Transporte Neumatico Thesis

42

Figure 3.12: Motor control circuit diagram.

To translate the angular displacement of the motor shaft into linear motion of the

slider system, a reel system was developed. Attached to the motor shaft is a spool used

for the winding of a thin thread and thereby pulling the slider system (Fig. 3.13). The

spool has an outer diameter of 30 mm and was fabricated out of ABS plastic with a

FDM1600 system. Its diameter was made sufficiently large so that changes in the

diameter caused by the wrapping of the nylon thread would produce only negligible error.

The thread passes through an eye-hook to properly guide it before attaching to the slider

system. Extra thread can be let out to allow the motor to operate for a few seconds

before pulling the slider system. This allows the motor to ramp up and reach its steady

state angular velocity.

Page 56: Transporte Neumatico Thesis

43

MotorSpool

Thread

Eye hook

Figure 3.13: Motor and thread winding system.

The slider system consists of several components. A track is used to guide the

slider platform in a straight line travel path. It has a travel surface area 500 mm wide and

700 mm long, with 10 mm high flanges on the sides acting as retaining walls. The slider

platform slides freely within the track flanges; loose enough to not cause too severe

friction, but securely enough to not deviate from straight line motion. It is attached to the

thread and is pulled with a constant velocity by the motor. The slider platform also is

used to transport the powder collection system.

The powder collection system consists of two parts (Fig. 3.14). One part is the

collection tray. It fits securely on top of the slider platform. The other part is a thin

adhesive surface. This fits underneath the tray and the adhesive surface is exposed to the

discharged powder through an opening in the collection tray. During the linear translation

of the slider system, powder is collected on the adhesive surface for subsequent analysis.

Page 57: Transporte Neumatico Thesis

44

Slider platform

Collection tray

Adhesive surface

Thread

Figure 3.14: Exploded slider collection system.

Calibration of the motor and slider collection system is necessary for powder

deposition data analysis. This was performed by measuring the voltage supplied to the

motor and the corresponding average velocity of the slider system for various increments

of the potentiometer. A template is used to mark positions of the potentiometer knob in

18 degree increments (Fig. 3.15). Voltages are read for each of these positions along with

corresponding average linear velocities of the slider collection system for several of the

positions.

Page 58: Transporte Neumatico Thesis

45

Figure 3.15: Potentiometer and calibration template.

3.2 EXPERIMENTAL SETUP

Several characteristics of the experimental apparatus dictate the performance capabilities

of deposition systems based on similar principles. One such is the volumetric flow rate

of the discharged powder for a variety of pressure and plug length combinations. The rate

at which the powder can be discharged through the nozzle is essential in estimating the

time required for the deposition of an entire layer. Also of importance is the

characteristics of the powder being deposited. The path shape width, height, and spread

of the powder is essential to providing consistent layers of powder. Finally, the ability

to deposit multiple powders needs to be observed. Tests were designed to observe the

mixing capabilities of the system and the ability of changing material composition on the

fly. The following sections describe the purpose of each experiment and the experimental

methods and details. These experiments include estimating the volumetric flow rate,

evaluating the powder path characteristics, and exploring multiple powder deposition.

Page 59: Transporte Neumatico Thesis

46

3.2.1 ESTIMATING THE VOLUMETRIC FLOW RATE

One of the beneficial characteristics of plug-phase pneumatic conveying is its ability to

deliver large volumes of powder of consistent density with a minimal air pressure.

Minimizing the pressure required is necessary to reduce the potential for disturbing the

underlying powder bed. However, minimizing the air pressure also decreases the

volumetric flow rate and increases the time required for depositing a single layer of

powder. Tests were therefore designed to estimate the average volumetric flow rate for

several plug length and pressure combinations.

A single plug of powder is created in one of the supply tubes. Air flow at a set

pressure is provided to this one supply tube only. The reason for testing with a single

tube is that air is delivered to the manifold at a set pressure. The manifold distributes the

air to each of the three supply tubes. The air will take a preferential route in which the

pressure drop is least [Mainwaring93]. This may cause the pressure among the parallel

tubes to fluctuate and be disproportional.

After creating a plug in one of the tubing systems and an input voltage is selected

to be sent to the motor, the volumetric flow rate can be determined for various plug length

and pressure combinations. By measuring the length of the deposited powder, DL, with a

constant linear velocity for the slider and collection system, the average volumetric flow

rate ,Q, can be determined by performing the following calculation:

Page 60: Transporte Neumatico Thesis

47

Q = PL · Ø2· π V (3.1) 4 DL

where PL is the powder plug length, Ø is the tubing inner diameter, and V is the average

linear velocity of the slider collection system.

One of the major performance criteria of implementing a new powder delivery

system is the rate at which each layer of powder is deposited. SFF technologies are

continuously trying to minimize fabrication times, so creating a large delay in the powder

deposition stage would be undesirable. For comparison, Sachs et al. [Sachs92] report of

times on the order of 0.1 - 1.0 second per layer. In the case of the experimental

apparatus, the deposition time, t, for a single layer given a buildspace of width, w, length,

l, and height, h, can be estimated by:

t = w · l · h (3.2) Q

Note that these times are not exact because of the acceleration and deceleration required at

the start and finish of each pass.

This time can be further reduced to:

t = w · l · h (3.3) Q ⋅ n

Page 61: Transporte Neumatico Thesis

48

by using an array of n nozzles. Ideally, the number of nozzles in the powder distribution

head would equal the workspace width divided by the deposition width of the powder

from each nozzle. This would provide for a single-pass, single-direction deposition.

Development of a high speed deposition system is not the focus of this research.

However, selective placement of a controlled powder composition is of greater

importance. Further tests were performed to evaluate the deposition characteristics while

changing several parameters of the experimental apparatus setup.

3.2.2 EVALUATING THE POWDER PATH CHARACTERISTICS

Another area of importance in evaluating the performance of the experimental apparatus

is the characteristics of the path of powder deposited. This is necessary for evaluating

the tolerance to which multiple materials may be placed and in developing the spatial

positioning required for an array of nozzles. Control of the deposited powder is

essential. It should not be randomly sprayed, but be placed in the locations desired.

A test was designed to evaluate the effects that several parameters had on the path

of the deposited powder. For each test run, a single plug of powder is created in one of

the supply tubes. Then the pressure and voltage are set accordingly, and a single pass

deposition test is run to evaluate the effect of changing a single design parameter on the

deposition characteristics of the powder path.

Page 62: Transporte Neumatico Thesis

49

The parameters to be examined include the nozzle length (L), nozzle diameter (Ø),

the deposition height (H), the velocity of the slider collection system (V), and the

volumetric flow rate (Q). Figure 3.16 shows these parameters. The characteristics of the

powder path of interest are the height (h), width (w) of the main proportion of powder,

and the total spread (s) of powder. Figure 3.17 illustrates the observed characteristics.

H

L

Ø

Q

V

Nozzle

Slider/Collection System

Figure 3.16: Parameters tested for deposition characteristics.

h

sw

Figure 3.17: Deposited powder path characteristics evaluated.

Page 63: Transporte Neumatico Thesis

50

3.2.3 EXPLORING MULTIPLE POWDER DEPOSITION

The experiments previously described show the capabilities of depositing a single

material. In order to make a significant improvement over deposition methods previously

done, the simultaneous deposition of multiple materials needs to be achieved. A series of

experiments were designed to evaluate the mixing performance of the experimental

apparatus using a dual powder delivery: Two powder plugs of different color are loaded

in two supply tubes, respectively. The air from the two pressure sources are set equal

and directed through the manifold to each polyurethane tube. Single pass deposition runs

are performed to evaluate the effects that the nozzle height and nozzle diameter each have

on the mixing of the deposited powder. These experiments, however, are not data

intensive because of the limited control the experimental apparatus has over the

pressurized air. Instead they qualitatively show the feasibility of multiple material

deposition in such a manner.

The path of powder is inspected visually for the major mixing characteristics

evident for each of the parameters. Characteristics to be observed are segregation among

the different colored powder and the spatial pattern and thoroughness of the mixing that

occurs.

Another series of trials is attempted to demonstrate transitions among powder

composition. These are performed similar to the mixing experiments with one exception.

Instead of setting the plugs in motion simultaneously, one plug is sent slightly ahead of

Page 64: Transporte Neumatico Thesis

51

the other. The air supplied to the leading plug is stopped at approximately the time air is

supplied to the trailing plug. A transition between the powder composition should be

observed. However, exact measurements are not possible due to the simplicity of the

controls of the experimental apparatus.

Page 65: Transporte Neumatico Thesis

52

CHAPTER 4

RESULTS

An experimental apparatus has been designed and built to explore the feasibility of

employing plug-phase pneumatic conveying to deposit powder in SLS, 3DP, and FPM in

a manner that extends their fabrication capabilities to include continuous heterogeneous

material composition. Several tests were performed to evaluate the experimental

apparatus’s performance both quantitatively and qualitatively. These included

determining the linear velocity of the slider system, the volumetric flow rate of

corresponding air pressure and plug lengths, the powder path characteristics for several

design parameters and its ability for depositing heterogeneous compositions. The

following sections describe these experiments in more detail.

4.1 LINEAR VELOCITY EXPERIMENTS

The mechanized slider collection system was designed to travel at a constant, yet

adjustable, linear velocity. A series of experiments were performed to calibrate the motor

controller with the various linear velocities within the slider collection system’s operating

range. First, the motor input voltages were found in relation to the corresponding angular

increments of the potentiometer knob. Next, the average linear velocity of the slider

collection system was tested for various motor input voltages.

Page 66: Transporte Neumatico Thesis

53

A template was created to divide the angular displacement of the potentiometer

knob into 18 degree increments (Fig. 3.15). For each of these positions, the output

voltage from the control circuit was measured. Table 4.1 shows the average output

voltage for corresponding potentiometer angular displacements from its origin. Figure 4.1

is a plot of the average output voltage with the addition of a maximum and minimum

voltage measured for each data set.

A series of tests were performed to determine the average linear velocity of the

slider collection system for several of the calibrated voltages. A pair of marks were made

within the track groove at an increment of 300 mm. The front of the slider collection tray

was positioned 35 mm behind the starting point to allow the slider collection system to

begin motion before measurements were taken, and thus, diminish the initial acceleration

from the linear velocity measurement. Also, a significant extra amount length of thread

was unwound from the spool to allow the motor to reach steady-state angular velocity

before pulling the slider tray. For each incremented potentiometer position, ten tests

were performed to measure the time required for the slider collection system to travel

300 mm. Note that lower voltages are not included in the data for they were not

sufficient to turn the motor shaft and pull the slider system. Also, higher voltages are not

included for they produced too extreme a velocity and measurement became severely

uncertain. The results are shown in Tbl. 4.2.

Page 67: Transporte Neumatico Thesis

54

Table 4.1: Output voltage for potentiometer angular displacement increments.

Potentiometer Angular

Displacement (degrees)

Average Output

Voltage (Volts)

0 0.03418 0.03836 0.05254 0.5772 1.3790 2.3108 3.2126 4.2144 5.3162 6.4180 7.2198 8.1216 9.2234 9.7252 10.5270 11.2288 11.3306 11.4

Output Voltage vs. Angular Displacement of Potentiometer

0

2

4

6

8

10

12

0 18 36 54 72 90 108 126 144 162 180 198 216 234 252 270 288 306

Angular Displacement of Potentiometer (degrees)

Figure 4.1: Output voltage vs. angular displacement of potentiometer.

Page 68: Transporte Neumatico Thesis

55

Table 4.2: Time measurements for calibrated motor input voltage.

VOLTAGE 4.3 V 5.3 V 6.3 V 7.5 V 8.5 V

time # 1 (s) 4.03 1.57 1.03 0.85 0.78

time # 2 (s) 3.97 1.50 1.00 0.84 0.82

time # 3 (s) 3.97 1.53 1.09 0.88 0.75

time # 4 (s) 3.87 1.50 1.06 0.85 0.72

time # 5 (s) 3.90 1.59 1.03 0.84 0.80

time # 6 (s) 3.94 1.57 1.07 0.88 0.84

time # 7 (s) 3.91 1.59 1.10 0.87 0.72

time # 8 (s) 3.94 1.62 0.99 0.89 0.79

time # 9 (s) 3.91 1.57 1.07 0.88 0.81

time # 10 (s) 3.95 1.60 1.05 0.90 0.81

mean (s) 3.94 1.56 1.05 0.87 0.78

standard deviation 0.0451 0.0412 0.0363 0.0215 0.0414

The average linear velocity of the slider collection system for each calibrated

voltage was then calculated by dividing the distance traveled (300 mm) by the time

measured for the linear translation to occur. The corresponding average linear velocities

are shown in Tbl. 4.3. These values along with maximum and minimum velocities

calculated for each data set are plotted in Fig. 4.2.

Page 69: Transporte Neumatico Thesis

56

Table 4.3: Linear velocity of slider collection for calibrated motor input voltages.

Average Input Voltage (volts)Average Linear Velocity (mm/s)4.2 765.3 1926.4 2867.2 3458.1 385

Velocity of Slider System vs. Average Motor Input Voltage

0

50

100

150

200

250

300

350

400

450

4 5 6 7 8 9

Input Voltage (V)

Figure 4.2: Velocity of slider system vs. motor input voltage.

As seen by the low standard deviations in the time measurements, the linear

velocity calculated for the slider collection system is fairly accurate. This calibration is

not intended for use solely as an evaluation of the performance of the mechanized slider

collection system; rather it is to be used as a means for calculating other data.

Page 70: Transporte Neumatico Thesis

57

4.2 POWDER VOLUMETRIC FLOW RATES

The volumetric flow rate achievable by the experimental apparatus will influence

deposition by plug-phase pneumatic conveying acceptance into the SFF industry. SFF

technologies are steadily trying to improve upon their build times in all stages of the

fabrication process. The time required for the deposition of an entire layer is directly

related to the volumetric flow rate of the powder. A series of measurements were taken

to calculate the volumetric flow rate of the discharged powder for a variety of pressure

and plug length combinations. Air pressure was kept at a minimum to reduce the force

that the powder would have on the underlying powder bed and to ensure that a transition

to lean-phase conveying did not occur.

A variety of plug length and pressure combinations were tested to determine the

volumetric flow rate of the polycarbonate powder. The plug lengths ranged from 40 mm

to 80 mm and the air pressures ranged from 17.2 kPa (2.5 psi) to 34.5 kPa (5.0 psi). The

longer plug lengths were not used in combination with the lower air pressures since there

was not enough pressure to transport the powder.

First, a single plug of powder is created in one of the supply tubes. Air at a set

pressure is provided to this tube only. After creating a plug in one of the tubing systems

and an input voltage is selected to be sent to the motor, the motor is activated and the

tubing pressurized. By measuring the length of the deposited powder path, the average

Page 71: Transporte Neumatico Thesis

58

volumetric flow rate can be determined. This deposited length and the other parameter

inputs are substituted into equation 3.1 to determine the average volumetric flow rate.

This testing method relies on the results of the slider collection system velocity

calibration to determine the value for V at the set potentiometer position. For each of

these experiments, a linear velocity of 192 mm/s was used. The volume of powder

deposited was calculated by multiplying the cross-sectional area of the interior of the

polyurethane tubing by the powder plug length. The cross-sectional area of the interior

of the polyurethane tubing is equal to 8.04 mm2.

The nozzle used was set at a height of 5.0 mm, and it had a length of 6.0 mm and

diameter of 1.5 mm. The deposited powder lengths for each trial for each air pressure and

plug length are shown in Tbl. 4.4, 4.6, and 4.8. The average volumetric flow rate was

calculated for each of these data points using equation 3.1. Tables 4.5, 4.7, and 4.9 show

the average volumetric flow rate for each data set and maximum and minimum calculated

for each data set. The volumetric flow rate values are plotted in Fig. 4.3, 4.4, and 4.5.

Page 72: Transporte Neumatico Thesis

59

Table 4.4: DEPOSITED POWDER LENGTH for air pressure of 34.5 kPa and linear velocity of 192 mm/s.

PLUG LENGTH 40 mm 50 mm 60 mm 70 mm 80 mm

trial # 1 (mm) 35 46 45 66 71

trial # 2 (mm) 38 45 47 64 78

trial # 3 (mm) 32 38 51 62 80

trial # 4 (mm) 42 42 47 63 85

trial # 5 (mm) 40 44 52 66 81

trial # 6 (mm) 38 43 54 68 87

trial # 7 (mm) 36 46 46 72 79

trial # 8 (mm) 41 42 48 70 88

trial # 9 (mm) 38 45 53 74 83

trial # 10 (mm) 40 44 47 67 79

mean (mm) 38 44 49 67 81

std. dev. 3 2 3 4 5

Table 4.5: VOLUMETRIC FLOW RATE for air pressure of 34.5 kPa.

PLUG LENGTH 40 mm 50 mm 60 mm 70 mm 80 mm

mean (mm3/s) 1628 1778 1874 1611 1526

maximum (mm3/s) 1933 2035 2062 1463 1743

minimum (mm3/s) 1473 1681 1718 1746 1406

Volumetric Flow Rate vs. Plug Lengths for a Pressure of 34.5 kPa

1000

1200

1400

1600

1800

2000

2200

40 50 60 70 80

Powder Plug Length (mm)

3/s)

Figure 4.3: Volumetric flow rate vs. powder plug length for an air pressure of 34.5 kPa.

Page 73: Transporte Neumatico Thesis

60

Table 4.6: DEPOSITED POWDER LENGTH for air pressure of 25.9 kPa and linear velocity of 192 mm/s.

PLUG LENGTH 40 mm 50 mm 60 mm

trial # 1 (mm) 43 48 50

trial # 2 (mm) 39 45 55

trial # 3 (mm) 37 46 50

trial # 4 (mm) 42 47 61

trial # 5 (mm) 36 49 53

trial # 6 (mm) 40 47 58

trial # 7 (mm) 41 51 52

trial # 8 (mm) 38 48 62

trial # 9 (mm) 42 44 64

trial # 10 (mm) 37 49 65

mean (mm) 39.5 47.4 57

std. dev. 2.46 2.07 5.75

Table 4.7: VOLUMETRIC FLOW RATE for air pressure of 25.9 kPa.

PLUG LENGTH 40 mm 50 mm 60 mm

mean (mm3/s) 1566 1631 1628

maximum (mm3/s) 1718 1757 1856

minimum (mm3/s) 1439 1516 1428

Volumetric Flow Rate vs. Plug Lengths for a Pressure of 25.9 kPa

1000

1200

1400

1600

1800

2000

2200

40 50 60

Powder Plug Length (mm)

Volumetric Flow Rate (mm

3/s)

Figure 4.4: Volumetric flow rate vs. powder plug length for an air pressure of 25.9 kPa.

Page 74: Transporte Neumatico Thesis

61

Table 4.8: DEPOSITED POWDER LENGTH for air pressure of 17.2 kPa and linear velocity of 192 mm/s.

PLUG LENGTH 40 mm 50 mm

trial # 1 (mm) 48 53

trial # 2 (mm) 45 58

trial # 3 (mm) 44 54

trial # 4 (mm) 43 52

trial # 5 (mm) 41 60

trial # 6 (mm) 46 58

trial # 7 (mm) 47 56

trial # 8 (mm) 49 61

trial # 9 (mm) 45 59

trial # 10 (mm) 42 55

mean (mm) 45 56.6

std. dev. 2.36 2.79

Table 4.9: VOLUMETRIC FLOW RATE for air pressure of 17.2 kPa.

PLUG LENGTH 40 mm 50 mm

mean (mm3/s) 1375 1366

maximum (mm3/s) 1509 1487

minimum (mm3/s) 1263 1268

Volumetric Flow Rate vs. Plug Lengths for a

Pressure of 17.2 kPa

1000

1200

1400

1600

1800

2000

2200

40 50

Powder Plug Length (mm)

3/s)

Figure 4.5: Volumetric flow rate vs. powder plug length for an air pressure of 17.2 kPa.

Page 75: Transporte Neumatico Thesis

62

SFF technologies are continuously trying to improve upon build time without

sacrificing build quality. The additional complexity of depositing heterogeneous powder

layers incur a time penalty but it should not be too significant. For a deposition system

using plug-phase pneumatic conveying, the deposition time hinges upon the volumetric

flow rate reliably achieved within the operating parameters of the system. For each of the

tests performed, the average volumetric flow rate was on the order of 1.5 × 103 mm3/s. In

order to achieve continuous flow, conveying lines will have to be run in parallel 180

degrees out of phase to supply a steady flow of powder plugs. Assuming that the nozzle

has the mechanics to traverse the entire layer area in a raster fashion, the deposition of a

single layer can be estimated by equation 3.2.

Assuming that a 10” by 10” (254 mm by 254 mm) build area with a layer

thickness of 0.008” (0.2 mm) is required, the deposition time for a single nozzle system

with continuous flow would be on the order of nine seconds. Deposition in this manner

may also cause additional time penalties. In order to deliver the powder in the density

necessary for sintering or liquid binder adhesion, a thicker layer may need to be deposited

and then further compacted by a mechanical force to increase its density. Also,

acceleration and deceleration of the powder head during each pass of the raster motion

will cause an additional time loss.

The deposition time, however, can be reduced significantly. For instance, using an

array of n nozzles would reduce the deposition time an order of n. Including a rough

Page 76: Transporte Neumatico Thesis

63

estimate of the time penalties discussed previously, an array of ten nozzles all capable of

simultaneous continuous deposition would result in a deposition time on the order of two

seconds per layer. This is slightly higher than the 0.1 to 1.0 second times reported by

Sachs et al. [Sachs92], but not too high for the greater fabrication capabilities gained.

4.3 POWDER PATH CHARACTERISTICS

The rapid deposition of large quantities of powder can be achieved through several

alternative methods, yet none match a plug-phase pneumatic conveying systems control

of the powder throughout its transportation. This is because when it reaches the nozzle

orifice, it remains intact as a densely packed plug with a known flow rate and can be

deposited with control. Experiments were performed with the experimental apparatus to

observe the effects of the design parameters on the deposited powder path characteristics.

These parameters include the nozzle length, diameter, and height; slider collection system

speed; and air pressure (Fig. 3.16). All parameters were kept at their default values while

one parameter was modified at a time to observe its effect on the powder path

characteristics. The default values used in these experiments are shown in Tbl. 4.10. The

powder characteristics measured were the main width, total spread, and the height

(Fig. 3.17).

Page 77: Transporte Neumatico Thesis

64

Table 4.10: Default parameter values.

PARAMETER DEFAULT VALUEnozzle length 6.0 mm

nozzle diameter 1.5 mmnozzle height 5.0 mm

slider collection velocity 192 mm/sair pressure 17.2 kPa

powder plug length 50 mm

The height of the powder path was measured at its peak (Fig. 3.17). The heights

were on the order of 0.1 mm to 0.3 mm for each test. Heights remained consistent for

each sample taken and their values varied indirectly with the main powder width. Exact

values are not reported because it was not possible to make accurate measurements. For

each test, also measured were the main width and the total spread of the powder path

(Fig. 4.6). Measurements were taken along areas of consistent width and spread. These

reported values varied roughly 10% among each data sample. These results are shown in

Tbl. 4.11-15. The average width and spread along with maximum and minimum values

measured for each data set are plotted in Fig. 4.7-11.

widthspread

mm

Figure 4.6: Actual powder path showing width and spread.

Page 78: Transporte Neumatico Thesis

65

Table 4.11: Powder path width and spread for various nozzle lengths.

NOZZLE LENGTH (mm)2.0 6.0 10.0

width (mm) spread (mm) width (mm) spread (mm) width (mm) spread (mm)trial # 1 3.0 5.0 3.5 9.0 3.0 5.0

trial # 2 3.0 6.0 4.0 7.0 2.0 4.0

trial # 3 4.0 7.0 3.0 8.0 4.0 6.0

trial # 4 4.0 7.0 3.5 7.5 3.0 5.0

trial # 5 3.5 8.0 3.5 8.0 3.5 5.0

trial # 6 3.5 5.0 2.0 8.0 3.5 6.0

trial # 7 4.0 7.0 3.5 6.0 4.0 7.0

trial # 8 3.5 7.0 3.0 7.0 3.0 6.0

trial # 9 4.0 6.0 4.0 8.0 4.0 6.0

trial # 10 3.0 7.0 3.5 7.0 3.0 5.0

mean 3.55 6.50 3.35 7.55 3.30 5.50

std. dev. 0.44 0.97 0.58 0.83 0.63 0.85

Note: Measurements are taken to the nearest 0.5 mm.

Powder Deposition Width vs. Nozzle Length

0

2

4

6

8

10

2 6 10

Nozzle Length (mm)

Width

Spread

Figure 4.7: Powder deposition width vs. nozzle lengths.

Page 79: Transporte Neumatico Thesis

66

Table 4.12: Powder path width and spread for various nozzle diameters.

NOZZLE DIAMETER (mm)1.5 3.0

width (mm) spread (mm) width (mm) spread (mm)trial # 1 3.5 9.0 5.0 8.0

trial # 2 4.0 7.0 4.5 7.0

trial # 3 3.0 8.0 5.0 7.0

trial # 4 3.5 7.5 5.0 7.0

trial # 5 3.5 8.0 4.5 6.0

trial # 6 2.0 8.0 5.5 7.0

trial # 7 3.5 6.0 5.5 7.0

trial # 8 3.0 7.0 5.0 8.0

trial # 9 4.0 8.0 5.0 7.0

trial # 10 3.5 7.0 5.5 8.0

mean 3.35 7.55 5.05 7.20

std. dev. 0.58 0.83 0.37 0.63

Note: Measurements are taken to the nearest 0.5 mm.

Powder Deposition Width vs. Nozzle

Diameter

0

2

4

6

8

10

1.5 3.0

Nozzle Diameter (mm)

Width

Spread

Figure 4.8: Powder deposition width vs. nozzle diameter.

Page 80: Transporte Neumatico Thesis

67

Table 4.13: Powder path width and spread for various nozzle heights.

NOZZLE HEIGHT (mm)5.0 7.0 9.0

width (mm) spread (mm) width (mm) spread (mm) width (mm) spread (mm)trial # 1 3.5 9.0 3.5 11.0 4.5 8.0

trial # 2 4.0 7.0 4.0 9.0 5.0 9.0

trial # 3 3.0 8.0 3.5 12.0 4.0 11.0

trial # 4 3.5 7.5 3.0 13.0 4.0 12.0

trial # 5 3.5 8.0 3.5 13.0 4.5 13.0

trial # 6 2.0 8.0 4.0 10.0 4.0 12.0

trial # 7 3.5 6.0 3.0 12.0 5.0 11.0

trial # 8 3.0 7.0 3.5 13.0 5.0 12.0

trial # 9 4.0 8.0 3.5 10.0 4.5 12.0

trial # 10 3.5 7.0 4.0 14.0 4.5 13.0

mean 3.35 7.55 3.55 11.70 4.50 11.30

std. dev. 0.58 0.83 0.37 1.64 0.41 1.64

Note: Measurements are taken to the nearest 0.5 mm.

Powder Deposition Width vs. Nozzle Height

0

2

4

6

8

10

12

14

5 7 9

Nozzle Height (mm)

Width

Spread

Figure 4.9: Powder deposition width vs. nozzle height.

Page 81: Transporte Neumatico Thesis

68

Table 4.14: Powder path width and spread for various slider collection system speeds.

Slider Collection System Speed (mm/s)192.0 286.0 385.0

width (mm) spread (mm) width (mm) spread (mm) width (mm) spread (mm)trial # 1 2.5 13.0 3.5 9.0 2.5 6.0

trial # 2 2.5 10.0 4.0 7.0 3.0 7.0

trial # 3 3.0 11.0 3.0 8.0 2.5 6.0

trial # 4 3.0 12.0 3.5 7.5 2.5 8.0

trial # 5 2.5 11.0 3.5 8.0 2.0 6.0

trial # 6 2.5 11.0 2.0 8.0 3.0 8.0

trial # 7 3.0 11.0 3.5 6.0 2.5 7.0

trial # 8 3.0 12.5 3.0 7.0 2.5 6.0

trial # 9 2.5 12.0 4.0 8.0 3.0 7.0

trial # 10 3.0 11.0 3.5 7.0 2.5 8.0

mean 2.75 11.45 3.35 7.55 2.60 6.90

std. dev. 0.26 0.90 0.58 0.83 0.32 0.88

Note: Measurements are taken to the nearest 0.5 mm.

Powder Deposition Width vs. Slider

Collection System Speed

0

2

4

6

8

10

12

14

192 289 385

Velocity (mm/s)

Width

Spread

Figure 4.10: Powder deposition width vs. slider collection system speed.

Page 82: Transporte Neumatico Thesis

69

Table 4.15: Powder path width and spread for various air pressures.

Air Pressure (KPa)17.2 25.9 34.5

width (mm) spread (mm) width (mm) spread (mm) width (mm) spread (mm)trial # 1 3.5 9.0 2.5 15.0 4.0 20.0

trial # 2 4.0 7.0 2.0 14.0 4.5 20.0

trial # 3 3.0 8.0 2.5 16.0 4.0 20.0

trial # 4 3.5 7.5 2.0 15.0 4.5 20.0

trial # 5 3.5 8.0 2.0 13.0 5.0 20.0

trial # 6 2.0 8.0 2.5 13.0 4.5 20.0

trial # 7 3.5 6.0 2.5 16.0 4.0 20.0

trial # 8 3.0 7.0 2.0 15.0 5.0 20.0

trial # 9 4.0 8.0 2.5 14.0 5.5 20.0

trial # 10 3.5 7.0 2.5 17.0 4.0 20.0

mean 3.35 7.55 2.30 14.80 4.50 20.00

std. dev. 0.58 0.83 0.26 1.32 0.53 0.00

Note: Measurements are taken to the nearest 0.5 mm and the maximum width of the deposition areais 20.0 mm which was equaled or exceeded each case by the 34.5 KPa air pressure.

Powder Deposition Width vs. Air Pressure

02468

1012141618

20

17.2 25.9 34.5

Air Pressure (KPa)

Width

Spread

Figure 4.11: Powder deposition width vs. air pressure.

Page 83: Transporte Neumatico Thesis

70

The results show that a powder path of 3 to 4 mm can be consistently deposited.

The average width for all values not including the wider diameter nozzle is 3.4 mm with a

standard deviation of 0.84. The total spread of the powder changed dramatically with the

various parameters. The average value was 10.63 mm with a standard deviation of 4.55.

Changing the nozzle length had little effect on the width and spread of the powder

path other than slightly reducing the width as the length increased. Enlarging the nozzle

diameter caused the main width to increase but had little effect on the overall spread of

the powder. In future systems, a wider nozzle could be used in areas where the tolerance

of the position of the powder is relaxed. Increasing the height of the nozzle from the

deposition surface had minimal effects on the main width of powder but caused

significant widening of the overall spread. Increasing the speed of the slider collection

system had minimal effect on the main width but caused the overall spread to reduce.

This is because the longer time the nozzle spends over a particular area, the more force is

applied to the underlying powder which causes the powder to spread further.

Varying the air pressure effected the deposited powder path more than any other

of the parameters. Changing the air pressure causes a change in the volumetric flow rate

and hence the force in which the powders impact the collection surface. No linear trend

was observed on the main powder width but the higher air pressures caused a significant

increase in the total spread of the powder. This shows that higher air pressures should be

Page 84: Transporte Neumatico Thesis

71

avoided to reduce the force the powder has on the underlying bed and to make sure that a

transition to lean-phase conveying does not occur.

4.4 MULTIPLE POWDER MIXING

The main benefit of powder deposition with the experimental apparatus is not its ability

to accurately and rapidly deposit a powder layer, but rather its ability to deposit

selectively heterogeneous material compositions throughout each layer. A series of

experiments were therefore performed to analyze the effects that system parameter

changes had on the mixing of two powders being deposited simultaneously. The

parameters tested were the nozzle’s vertical height, diameter, and length. Observations

were made on the general mixing capabilities and any patterns observed were recorded.

As before, the default experimental parameter settings were set as shown in Tbl. 4.10.

The experiments were performed by creating two equal powder plugs, red and

blue, in two of the parallel polyurethane tubes. The air pressure from the two air sources

were set equal and directed through the manifold to each polyurethane tube. The powder

plugs were then set in motion and reached the nozzle simultaneously, where upon they

were deposited on the traversing slider collection system.

With the original nozzle setup (length equal to 6.0 mm, height equal to 5.0 mm,

and diameter equal to 1.5 mm), the mixing was poor and somewhat segregated (Fig. 4.12).

Mixing occurred in the main width, but there were distinct areas of red and blue rather

Page 85: Transporte Neumatico Thesis

72

than a uniform blend. The different powders seemed to be divided by the centerline of

the cross-section. The spread was completely divided by color on each side of the

centerline.

blue side

red side

(mm)

(a) (b)

Figure 4.12: Default parameter setup with a height of 5.0 mm;

(a) original image, (b) annotated image.

Elevating the nozzle height to 7.0 mm produced greater mixing in the main powder

width but the powder types were still somewhat segregated by the centerline (Fig 4.13).

The mixing was not thorough for there are distinct areas of bi-modal colors rather than a

uniform blend. Raising the nozzle height to 9.0 mm produced a greater mixing (Fig. 4.14).

The main width appeared thoroughly mixed with no spatial patterns evident. However,

the mixing remained macroscopic rather than microscopic. In particular, there were areas

of red and blue powder rather than blended powder particles. Increasing the length of the

nozzle produced greater mixing at the center of the main powder width but the powders

were still somewhat segregated by the centerline (Fig 4.15). Increasing the nozzle

Page 86: Transporte Neumatico Thesis

73

diameter improved mixing in the main width, but uniform mixing was still not achieved

(Fig. 4.16).

blue side

red side

(mm)

(a) (b)

Figure 4.13: Nozzle height of 7.0 mm; (a) original image, (b) annotated image.

blue side

red side

(mm)

(a) (b)

Figure 4.14: Nozzle height of 9.0 mm; original image (a), annotated image (b).

Page 87: Transporte Neumatico Thesis

74

blue side

red side

(mm)

(a) (b)

Figure 4.15: Default parameter setup with nozzle length of 10.0 mm;

(a) original image , (b) annotated image.

(mm)

(a) (b)

Figure 4.16: Default parameter setup with nozzle diameter of 3.0 mm;

(a) original image, (b) annotated image.

Uniform mixing was not achieved for any setting. Future systems should

therefore consider developing a flow analysis of the powder passing through the nozzle to

better stimulate thorough mixing of the multiple powders. Also, it was not possible to

analyze the proportions of powder deposited with various volumetric flow rates because

the differences of volumetric flow rates achieved by this experimental device are on the

order of ten percent. Future systems should therefore also consider maintaining a

Page 88: Transporte Neumatico Thesis

75

constant volumetric flow rate for each powder supply line and control the deposited

material composition by proportioning the amount of powder which each nozzle receives.

4.5 MULTIPLE MATERIAL TRANSITIONS

In order to facilitate truly continuous heterogeneous material composition, it is important

that the deposition system is capable of changing the material composition on the fly.

The experimental apparatus provides this capability. In particular, it can currently mix

two powders on the fly, and if automated, it would also be able to change this

composition on the fly to enable a point-by-point layer design.

Experiments were therefore conducted with the experimental apparatus to examine

its transitional powder mixing capabilities. In these experiments it was not possible to

measure quantitative data to accurately illustrate the transitions because of the lack of

automated control within the experimental apparatus. However, transition from one

powder color to the next was tested. The experiment was set up identically to that

described in Section 4.4 with one exception: Instead of setting the plugs in motion

simultaneously, one was sent slightly ahead of the other. The air supplied to the leading

powder plug was ceased as the trailing powder plug approached the nozzle. This resulted

in a transition from one material type to the next. Figure 4.17 illustrates a transition from

red to green powder colors.

Page 89: Transporte Neumatico Thesis

76

red green

(mm)

(a) (b)

Figure 4.17: Transition from red to green powder, (a) original image, (b) annotated image.

These transitions are obviously severely error prone with the current experimental

setup. A more precise transition would require timing to the fraction of a second to

provide the transitions desired. A future system would require automation to control the

timing and would need to be calibrated to determine the transient stages required for

initiating and stopping powder flow.

4.6 OBSERVATIONS

The experimental results and analysis demonstrate that depositing continuously

heterogeneous powder composition with the experimental apparatus is feasible. Many

beneficial deposition qualities were illustrated by the experimental apparatus. However,

several deficiencies were also apparent. These deficiencies need to be overcome before

deposition by plug-phase pneumatic conveying can be integrated into SLS, 3DP, or FPM.

The results illustrate that powder can be transported through the tubing system

and deposited with control. Conveying at lower air pressures resulted in lower particle

Page 90: Transporte Neumatico Thesis

77

speeds which lend to this control. The results also show the potential for selectively

depositing continuously heterogeneous material compositions. Multiple powders were

directed to the tubing and deposited simultaneously through a single nozzle. Other

results demonstrate the capabilities of selectively altering the material composition of the

deposited powder on the fly.

Several deficiencies of the experimental apparatus also became apparent. First,

the range of operational volumetric flow rates is limited, varying only 10% among the

various setups. This is unfortunate because until this problem is overcome, it will not be

possible to achieve precise heterogeneous material deposition varying under computer

control. Also, three powders were unable to be simultaneously deposited. This was

because the pressure drops in the parallel tubing lines were not equal and the air bypassed

one of the lines. A third controlled pressure input would remedy this problem.

However, the basic feasibility of deposition of continuously heterogeneous

material composition by plug-phase pneumatic conveying has been demonstrated. The

next step, in addition to developing a more precise powder mixing control, is therefore to

explore how this concept can be integrated with SLS, 3DP, or FPM. Hence, Chapter 5

will present a conceptual design of a proposed larger scale system based on the results

and analysis of the performance of the experimental apparatus.

Page 91: Transporte Neumatico Thesis

78

CHAPTER 5

MULTIPLE POWDER DEPOSITION SYSTEM

5.1 INTRODUCTION

The results and analysis of the performance of the experimental apparatus illustrate that

depositing continuously heterogeneous powder layers is feasible for Selective Laser

Sintering (SLS), Three Dimensional Printing (3DP), and Freeform Powder Molding

(FPM) with a plug-phase pneumatic conveying system. In its present form, the

experimental apparatus is not suited for implementation into SLS, 3DP, or FPM. It lacks

the automation and the speed necessary for satisfactory operation. However, its

operating principles can be used as the basis for the design of a larger scale system to

provide selectively heterogeneous material fabrication potential.

The results from testing of the experimental apparatus are used for developing a

conceptual design of the multiple powder deposition (MPD) system (Fig. 5.1). Similar to

the experimental apparatus, the MPD system conveys three powders, each through its

own tubing system, to a nozzle where they selectively merge prior to deposition. Drastic

improvements are made by automating the powder plug creation and by depositing with

an array of nozzles to help reduce the deposition time. The array of nozzles scans the

build area in raster fashion rather than having the build area move under a stationary

nozzle.

Page 92: Transporte Neumatico Thesis

79

Powderprint head

Powder storageand loadingchambers

Supplytubing

MotionsystemDischarge

tubing

Figure 5.1: Multiple Powder Deposition System.

Included in this chapter are overviews of the major functional components

contained in the conceptual design of the MPD system, a review of the software and

control requirements, and the powder properties that most effect the conveying of

various powders. An analysis of the MPD system’s expected performance in depositing

a single layer of powder is estimated based on the results discussed in Chapter 4.

Page 93: Transporte Neumatico Thesis

80

5.2 NOZZLE ARRAY AND POWDER DEPOSITION HEAD

Depositing each powder layer with a single nozzle system would take the ‘rapid’ out of

‘rapid prototyping’. To aid in minimizing the deposition time required for each layer, an

array of independently operating nozzles is proposed. These nozzles are similar in

design to that used in the experimental apparatus with some minor modifications. The

nozzle array is contained within the powder deposition head and scans the build area in

unison during the powder deposition cycle.

The general form of the nozzle remains unchanged from that discussed in

Section 3.1.2 except for a few minor design modifications (Fig. 5.2). In order to better

accommodate the numerous tubes and fittings into a small area, vertical nozzle extensions

should be used. They should taper into the nozzle without severe bending. Based upon

results discussed in Chapter 4, a nozzle length of 10.0 mm is recommended. This length

minimized the deviation of the powder path and produced greater mixing of multiple

powders over the shorter lengths. This mixing was far from uniform though. There were

still significant areas of segregated powder types. A more accurate analysis of the flow

patterns observed within the nozzle will be needed. This flow analysis should be applied

to the design of a new nozzle which will allow for the optimization of multiple powder

mixing. Areas that should be addressed are the optimal angle for merging the nozzle

extensions and alternative shapes and sizes of the nozzle mixing passageway.

Page 94: Transporte Neumatico Thesis

81

Figure 5.2: Proposed deposition nozzle.

Future nozzles will furthermore have to be fabricated by a process other than

FDM. In particular, a process is needed that produces precise, smooth surfaces inside

the nozzle to minimize its friction with the powder. The surface quality of the nozzle is

currently the weak point in the tubing system. A glass blown nozzle is recommended for

its improved surface quality, ability to fabricate a variety of shapes, and its transparency

for visualizing internal flow patterns.

A space of 18.0 mm is required between the center points of adjacent nozzles

because of the robust size of the fittings required for connecting the nozzle extensions to

the tubing. The powder deposition width of each nozzle is on the order of 3 mm, so there

will be a significant void between each nozzle. To help minimize this void, a second row

of nozzles is used. Each nozzle is placed in the center of the pair of nozzles in the row in

Page 95: Transporte Neumatico Thesis

82

front at distance of 18.0 mm behind. A pair of rows each containing five and four,

respectively, nozzles is used.

As discussed previously, plug-phase conveying does not supply a continuous

deposition of powder. To obtain a continuous deposition, a second pair of nozzle rows

will be used. The second row pair will operate 180 degrees out of phase with the other

row pair and they will be aligned directly behind the first row pair at a distance of 18.0

mm (Fig. 5.3).

18.0

18.0

9.0

Primary nozzlerow pair

Secondary nozzlerow pair

18.0 18.0 18.0

18.0

18.0

9.0 9.0 9.0

dimensions in mm 100.0 mm

100.0 mm Fastaxis

Slowaxis

Figure 5.3: Nozzle array positioning.

Finally, a powder deposition head (Fig. 5.4) will be needed to house the array of

nozzles and transport them in unison with the motion system. In the future system, a

100.0 mm by 100.0 mm area head would contain the array of nozzles shown in Fig. 5.3.

Page 96: Transporte Neumatico Thesis

83

The open top accommodates the tubing lines which supply powder to the nozzles. The

circular openings permit it to slide on the motion system poles, and bearings are used to

reduce sliding friction during motion. The large opening shown is used for the exit of the

discharge tubing.

Opening fordischarge tubing

exit

Openings for slidingon the motionsystem poles

Figure 5.4: Powder deposition head.

5.3 MOTION SYSTEM

To deposit a complete layer of powder, the nozzle array must scan the entire build area.

Unlike the motion system used for the experimental apparatus, the MPD system moves

the nozzles over the build area rather than moving the build surface underneath a

stationary nozzle. This is the more feasible approach for there is greater difficulty in

moving the larger inertia powder bed with the same speed and accuracy.

Page 97: Transporte Neumatico Thesis

84

The motion requirement for the MPD system is linear motion with constant

velocity in two dimensions, namely x and y. Changes in the vertical height will be

provided by the lowering and raising of the powder bed, as is presently done with both

SLS and 3DP. Numerous mechanical systems can easily accommodate these design

requirements. The motion system designed for the MPD system operates with

principles similar to that of an ink jet printer. One axis of travel is fast. During travel on

this axis, the nozzles deposit powder onto the powder bed surface. The second axis of

travel is slower. Travel along this axis provides a small increment in position for the

nozzles to make a new pass along the fast axis over the powder bed.

The motion system designed for the MPD system is a simple series of pulleys,

sliders, and stepper motors (Fig. 5.5). Along the fast axis, a pair of slider poles is used to

guide the powder deposition head and maintain its stability. Connected to each side of

the deposition head is a V-belt. The deposition head is propelled along the axis in the

requisite direction by a stepper motor attached to one of the pulleys. The velocity

requirements are to be determined based upon the flow analysis, but should be on the

order of 250 mm/s. The slow axis operates with similar principles. A pair of stepper

motor driven pulley systems is attached to each side of the fast axis pulley system. Each

motor is controlled through a position feedback to aid in positioning control. After the

completion of each pass along the fast axis, the slower axis increments an appropriate

Page 98: Transporte Neumatico Thesis

85

Fast Axis

Slow axis

Figure 5.5: Proposed motion system for the Multiple Powder Deposition System.

distance. After deposition of each powder layer, the deposition head and slider poles can

be positioned as to not interfere with either the laser scanning or binder spreading.

The motion pattern required is similar to a simple raster pattern with some

modifications required due to the spacing of the nozzles. After a single pass along the

fast axis, 6 mm voids will be left in between the paths of powder. The slow axis will

increment the deposition head over 3 mm (the width of each powder path) and another

deposition pass will be made. This results in 3 mm voids between the powder paths.

After another 3 mm increment along the slow axis and deposition pass, a complete area of

Page 99: Transporte Neumatico Thesis

86

powder is deposited. This area is 81 mm wide. Before making the next deposition pass,

the deposition head needs to be moved 82.5 mm along the slow axis and the pattern is

repeated until a complete layer is deposited. A portion of this motion sequence is shown

in figure 5.6. The motion system modeled can deposit a powder bed 300 mm by 250 mm.

Slow axis

Fast axis

Figure 5.6: Portion of motion pattern for powder deposition.

5.4 POWDER PLUG CREATION & DELIVERY

An automated plug creation system similar to the commercial systems illustrated in

Fig. 2.7 is used to feed powder to the nozzles. A timer operated valve switches the air

flow between the loading bin and the tubing. When air enters the bin, powder is forced

out the bottom and enters the tubing. After the desired powder plug length of is created,

the air flow is switched to the tubing system. This air propels the powder plug through

the tubing system. The use of alternating air valves creates equally spaced powder plugs

spaced by equal lengths of air within the tubing (Fig. 5.7).

Page 100: Transporte Neumatico Thesis

87

Figure 5.7: Powder plugs passing through tubing.

In the MPD system, each nozzle requires independent delivery of 3 powders.

There are a total of 18 nozzles, thus 54 plug loading systems are needed. Figure 5.1 of

the complete MPD system only shows a portion of these. Each plug loading system

(Fig. 5.8) operates as described for the commercial system. These containers are

relatively large and are therefore stored “remotely” while connected by tubing to the

system.

Pressurized air input

Timer operatedsolenoid valve

Powder loading bin

Figure 5.8: Plug loading system, 1 of 54 required.

The tubing system connecting the plug loading system to each nozzle needs to

have a low coefficient of friction and be flexible. During the motion of the deposition

head, the distance from the plug loading systems to the nozzle array will fluctuate

Page 101: Transporte Neumatico Thesis

88

tremendously. Great care needs to be made in the routing of the tubing as to not cause

bends which will disrupt the flow of powder.

Calibration of the flow rates for various plug length and air pressure combinations

is required for each material and each of the 54 delivery tubes. Proportioning the powder

composition by adjusting the flow rate of powder through the tubing system is not

recommended for several reasons. As shown in testing of the experimental apparatus,

flow rates within the operating range of the system may only be different on the order of

10%. This would severely restrict the material compositions available. Also, changing

flow rates within each tube may not be achievable on the fly. For instance, the gradual

increasing of the volumetric flow rate of an individual plug may not be achieved because

the plugs traveling behind it will prevent an air flow rate adjustment from reaching the

leading plug.

A simple solution to the problem would be to convey the powder in each tubing

system at identical flow rates and to control the flow of powder entering the nozzle

through the nozzle extensions. A valve system needs to be designed and tested to

accomplish this need. The valve system should be able to direct the proportion of

powder desired to enter the nozzle through the tubing extensions and to divert the

remainder of the powder or conveying air into the discharge tubing (Fig. 5.9). Controlling

the powder flow in this manner will prevent stalling of the powder plugs within the

tubing and spraying of the powder bed with pressurized air.

Page 102: Transporte Neumatico Thesis

89

Discharge tubing

Powder delivery tube Nozzle extension

Unwanted powder portionand conveying air

Powder portion used fordesired material contribution

Proportioningvalve system

Figure 5.9: Proposed operation of proportioning valve for powder composition adjustment.

5.5 SOFTWARE & CONTROL

Current CAD software allows for specifying only discrete material regions. In order to

design for heterogeneous material blends, the current state of software needs to be

improved to include continuously heterogeneous material composition in three

dimensions. Through this software the material composition at each point of the three-

dimensional part needs to be determined to develop the control of the MPD system.

The software might break the three-dimensional part into a finite element grid with

the material composition specified for each node. The spacing of the grid will be

determined by the tolerance achievable by the MPD system. From this node data, the

necessary control parameters for the MPD system can be generated.

Page 103: Transporte Neumatico Thesis

90

The most difficult area of control is the array of proportioning valves. 162 valve

systems need to be independently controlled based on the nozzle position over the build

area and the proportion of air or powder to pass into the nozzle extension. A

programmable logic controller (PLC) can handle the control of both the valve system and

motion system. The loading and conveying of plugs of powder can be done continuously

or only when the powder from the individual tube is to be deposited. The latter method

would save tremendously on material costs.

5.6 POWDER PROPERTY EFFECTS

The deposition experiments performed thus far have only used one powder,

polycarbonate, dyed different colors to simulate the deposition of multiple materials.

The ability to fabricate parts of multiple colors would only produce cosmetic

improvements. To capture all of the benefits of continuously heterogeneous material

deposition, a variety of materials with diverse properties should be able to be deposited.

Mainwaring [Mainwaring93] summarizes the effects that material properties have

on conveying. No one property dictates a material’s suitability for conveying. The

properties that influence material’s flow the most are its particle size distribution and

particle shape. A wide particle size distribution is more problematic than fine powders,

specifically those with an average size below 100 µm, as is used in SLS and 3DP.

Spherical particle shapes are also recommended. Particle shapes that naturally do not

Page 104: Transporte Neumatico Thesis

91

flow well have difficulty being conveyed. These include cylindrical and disk shaped

particles.

The use of plug-phase conveying helps to initiate the flow of particles that

normally will not flow. Even if a product will not flow naturally in plug form, short

plugs can be artificially created to permit reliable conveying [Mainwaring93].

Guidelines suggest that the particle sizes be at least three times smaller than the

inner diameter of the tubing transporting them. Thus for the MPD system, particle sizes

should be maintained under 500 µm. All of the materials commercially available for use in

SLS satisfy these requirements and should be able to be deposited in the MPD system

with no modifications. Sample SLS material properties are shown in Tbl. 5.1, 5.2, and

5.3.

Table 5.1: Laserlite™ LN4010 Nylon Compound Properties [DTM94a].

General Properties Value Test MethodSpecific Gravity, 20ºC 1.04 g/cm3 ASTM D792Moisture Absorption, 20ºC, 1.0% ASTM D570 65% relative humidityPowder Tap Density 0.58 g/cm3 ASTM D4164Volume Average Particle Size 120 microns laser diffractionParticle Size Range, 90% 60-250 microns laser diffraction

Table 5.2: Laserlite™ LNF5000 Nylon Compound Properties [DTM94b].

General Properties Value Test MethodSpecific Gravity, 20ºC 1.04 g/cm3 ASTM D792Moisture Absorption, 20ºC, 1.0% ASTM D570 65% relative humidityPowder Tap Density 0.55 g/cm3 ASTM D4164Volume Average Particle Size 50 microns laser diffractionParticle Size Range, 90% 15-90 microns laser diffraction

Page 105: Transporte Neumatico Thesis

92

Table 5.3: Laserlite™ LWX2010 Wax Compound Properties [DTM94d].

General Properties Value Test MethodSpecific Gravity, 20ºC 1.04 g/cm3 ASTM D792Powder Tap Density 0.55 g/cm3 ASTM D4164Volume Average Particle Size 105 microns laser diffractionParticle Size Range, 90% 25-235 microns laser diffraction

5.7 PERFORMANCE ANALYSIS

While the complexity of the MPD system is much greater than that of the experimental

apparatus, both deposit powders with similar operating principles. Thus, the

performance expected for the MPD system should be accurately predicted from the

testing and analysis of the experimental apparatus. The most important parameters that

will affect the MPD system’s acceptance into the solid freeform fabrication industry are

the time required for depositing each layer and the spatial increment at which the material

composition may be specified and met by the deposition system.

The results of the experimental apparatus predict that a volumetric flow rate on

the order of 1.5×103 mm3/s can be expected for each nozzle. One might therefore expect

to completely deposit a powder layer of 10” by 10” (254 mm by 254 mm) with a

thickness of 0.008” (0.2 mm) with the MPD system in approximately one second. The

actual time may be slightly higher due to acceleration and deceleration of the motion

system. On par, the total deposition time, however, should remain below five seconds.

This is slower than the deposition time of the conventional roller distribution method, due

Page 106: Transporte Neumatico Thesis

93

mainly to the added complexity of depositing continuously heterogeneous material

composition.

The spatial interval achievable for the specified powder composition may vary in

all three dimensions. The 3 mm width of the powder path dictates a spacing of roughly

3.0 mm in the direction of the slow axis. The composition of the material deposited may

be accurately controlled at a spacing of roughly 3 mm. The composition of the material in

between the 3 mm interval will be determined by the nearest node.

The spatial interval in the vertical direction will be on the order of the powder

layer thickness assuming that the powder does not “bleed” into the underlying powder

bed and a compaction method is used to level the surface. Material composition in

between the specified positions will be equal to the nearest node.

Lastly, the spatial interval relies on the transient capabilities of the control

system. Calibration is necessary to accurately predict the transient patterns of the

control system relative to the powder deposition. The control points in the deposition

system will be the proposed proportioning valves. Having these control points located

near the deposition points (nozzles) should help to minimize residual effects of

commencing and ceasing flow.

The limiting factor in minimizing the nodal spacing is the width in which the

powder may be deposited. Deposition in the direction of the head motion can be

accurately controlled to a much tighter interval.

Page 107: Transporte Neumatico Thesis

94

5.8 DESIGN SUMMARY

The proposed conceptual design of the MPD system applies the principles developed

and tested with the experimental device into a larger scale operation. This system should

be capable of depositing powder layers with continuously heterogeneous material

composition. It may selectively deposit three powder types through an array of nozzles

that as a whole are capable of providing continuous powder deposition.

An automated plug loading system is proposed. The powder plugs are created

and transported through the supply tubing similar to the commercial systems previously

discussed (Section 2.4). Identical volumetric flow rates should be used for each supply

tubing and the control of the deposited powder composition may be met through a

proportioning valve system.

An array of 18 nozzles, with 9 operating simultaneously, will be used to accept

the powder from the supply tubing, uniformly mix each of three powder components,

and deposit the final powder composition onto the build area. The nozzle array is

necessary to help minimize the deposition time for each layer. The powder print head,

which contains the nozzle array, will scan the build area in an altered raster pattern

(Fig. 5.6). The motion is provided by a system similar to an ink-jet pulley system.

The MPD system’s acceptance into the SFF industry will rely on the deposition

time for each layer and the tolerance at which the powder can be deposited on a point-to-

point basis. Based on the results of the experimental apparatus, deposition times can be

Page 108: Transporte Neumatico Thesis

95

expected to be on the order of two seconds. Deposition tolerances will vary significantly

in each dimension. The use of plug-phase conveying helps to minimize deposition error

by providing deposition at low particle speeds and with a minimal amount of air

combined in the flow.

Further development of this conceptual deposition system requires much testing,

analysis, and calibration. However, based on the performance of the experimental

apparatus and that estimated for the proposed MPD system, plug-phase pneumatic

conveying combined with a raster motion is clearly a feasible method for the deposition of

continuously heterogeneous material composition for use in SLS, 3DP, or FPM.

Page 109: Transporte Neumatico Thesis

96

CHAPTER 6

CONCLUSION

This thesis has explored the fabrication of continuously heterogeneous parts in physical

rapid prototyping. It was achieved by developing a new method for depositing powder

layers for use in Selective Laser Sintering, Three-Dimensional Printing, and Freeform

Powder Molding. It presents a foundation to which other research should be built upon.

This chapter concludes the thesis by outlining the contributions made by this thesis and

suggestions for future work.

6.1 CONCLUDING REMARKS

There is a definite need for fabricating parts of continuously heterogeneous material

composition. When achieved, engineers may then begin to design at the microscopic level

with endless control over final part properties. Because they use additive build

processes, some solid freeform technologies, in particular SLS, 3DP, and FPM, can

readily adapt to selectively heterogeneous material blends.

Many SFF systems have fabricated using dual material build processes. These are

used primarily for support generation, but have expanded to include methods of

controlling the microstructure of fabricated parts. In order to adapt SLS, 3DP, and FPM

to the capability of continuously heterogeneous material composition, a new method of

Page 110: Transporte Neumatico Thesis

97

powder deposition needs to be devised. Plug-phase pneumatic conveying is a viable

choice.

This thesis explored the feasibility of implementing a plug-phase pneumatic

conveying powder delivery system for three-dimensional continuously heterogeneous

material composition for potential use in SLS, 3DP, or FPM. It was done through the

design, fabrication, and testing of an experimental apparatus. The experimental apparatus

used plug-phase conveying to transport three materials through a tubing system to a

deposition nozzle where they were selectively mixed prior to deposition.

Results show that continuously heterogeneous material composition is feasible

through the deposition methods of the experimental apparatus. These results are used to

model a conceptual design of a larger scale deposition system. Further development of

this larger scale system should be the next stage in future research.

6.2 CONTRIBUTIONS

This thesis addresses the need for fabricating continuously heterogeneous material parts

in SFF by developing a method of powder deposition that can accommodate changing

material composition on the fly. The final results were limited by the simplicity of the

experimental apparatus, but they demonstrate that plug-phase pneumatic conveying is a

feasible approach for depositing continuously heterogeneous powder layers.

Page 111: Transporte Neumatico Thesis

98

6.3 RECOMMENDATIONS FOR FUTURE WORK

This thesis builds a foundation for further research in fabricating parts of continuously

heterogeneous material composition. A number of issues remain before a fully functional

prototype system can be built:

• Development of a flow analysis and redesign of the nozzle. The nozzle used in the

experimental apparatus definitely had its limitations. Further research should focus

on analyzing the flow conditions that take place within the nozzle to achieve uniform

mixing. This will include exploring other nozzle shapes, sizes and adding static mixing

in the nozzle passageway. This analysis should be applied to a redesign of the nozzle

which optimizes its mixing and deposition potential.

• Method of proportioning powder composition. Because of the limited range of

operational flow rates achievable, an alternative method of varying composition needs

to be developed. This method needs to be able to selectively allow certain quantities

of each powder into the nozzle and be able to adjust the composition rapidly.

• Transient effects in the switching of powder composition. Once the methods of

proportioning are developed the transient effects of switching powder composition

needs to be further analyzed. These results need to be incorporated into the control

of the proportioning system.

• Deposition onto underlying powder bed. Testing needs to be performed to determine

the force at which the underlying powder bed becomes disturbed and adjust the flow

Page 112: Transporte Neumatico Thesis

99

correspondingly. The underlying powder bed should be compacted and heated to

simulate the conditions present in SLS, 3DP, and FPM.

• Exploring other powder alternatives. This research focused on only one powder,

polycarbonate. Assumptions concerning the flow of other powders are made. These

assumptions need to be validated and the range of material and their relevant

properties need to be explored for potential implementation into this deposition

system.

• Height measurements. The height of the deposited powder needs to be accurately

measured. This could be done by shadowing as is used in measuring contact angles.

The path height could be projected onto a screen by a light source. The scaling factor

can be determined by calibration through the projection of an object of known height.

• Compaction methods. The powder layers deposited may be of lower density than

desired. Methods of compaction and leveling each layer need to be explored.

• Adjusting other parameters within SLS, 3DP, and FPM to adapt to heterogeneous

material composition. These fabrication systems currently use only homogeneous

material blends. Changing material properties throughout the part may also result in

changing fabrication techniques. For example, in SLS the time and intensity at which

the laser must pass over each point may become variable along with the material

composition.

Page 113: Transporte Neumatico Thesis

100

• Development of software capable of heterogeneous material composition. Current

CAD software allows for specifying only discrete material regions. In order to design

for heterogeneous material blend, the current state of software needs to be improved.

The heterogeneous geometry needs to be converted into a form for controlling the

parameters of the deposition system.

Page 114: Transporte Neumatico Thesis

101

REFERENCES

[Agarwala95] M. Agarwala, R. van Weeren, R. Vaidyanathan, A.Bandyopadhyay, G. Carrasquillo, V. Jamalabad, N. Langrana, A.Safari, S. Garofalin, S. Danforth, J. Burlew, R. Donaldson, P.Whalen, and C. Ballard, “Structural Ceramics by FusedDeposition of Ceramics,” Proceedings Solid FreeformFabrication Symposium, August 7-9 1995, University of Texas,Austin Texas, http://www.caip.rutgers.edu/sff/Austin_paper_presented.html.

[Albright51] C. Albright J. Holden, H. Simmons, and L. Schmidt, “PressureDrop in Flow of Dense Coal-Air Mixtures,” Industrial andEngineering Chemistry, vol. 43, no. 8, August 1951, pp. 1837-1840.

[Ashley95] S. Ashley, “Rapid Prototyping is Coming of Age,” MechanicalEngineering, vol. 117, no. 7, July 1995, pp. 62-68.

[Badrinarayan92] B. Badrinarayan and J. Barlow, “Metal Parts From SelectiveLaser Sintering of Metal-Polymer Powders,” Proceedings SolidFreeform Fabrication Symposium, August 3-5 1992, Universityof Texas, Austin Texas, pp. 141- 146.

[Beck92] J. Beck, F. Prinz, D. Siewiorek, and L. Weiss, “ManufacturingMechatronics Using Thermal Spray Shape Deposition,”Proceedings Solid Freeform Fabrication Symposium, August 3-51992, University of Texas, Austin Texas, pp. 272- 279.

[Behrendt95] U. Behrendt and M. Shellabear, “The EOS Rapid PrototypingConcept,” Computers in Industry, vol. 28, 1995, pp. 57-61.

[Bourell92] D. Bourell, H. Marcus, J. Barlow, and J. Beaman, “SelectiveLaser Sintering of Metals and Ceramics,” The InternationalJournal of Powder Metallurgy, vol. 28, no.4, 1992, pp. 369-381.

[Brown70] R. Brown and J. Richards, Principles of Powder Mechanics, P.Danchwerts (ed.), Pergamon Press, 1970, pp. 1-12.

Page 115: Transporte Neumatico Thesis

102

[Burns93] M. Burns, Automated Fabrication - Improving Productivity inManufacturing, PTR Prentice Hall, 1993.

[Carter93] W. Carter and M. Jones, “Direct Laser Sintering of Metals,”Proceedings Solid Freeform Fabrication Symposium, August 9-11 1993, University of Texas, Austin Texas, pp. 51- 59.

[Cima92] M. Cima, A. Laudner, S. Khanuja, and E. Sachs, “MicrostructuralElements of Components Derived from Three DimensionalPrinting,” Proceedings Solid Freeform Fabrication Symposium,August 3-5 1992, University of Texas, Austin Texas, pp. 220-227.

[Deckard93] L. Deckard and T. Claar, “Fabrication of Ceramic and MetalMatrix Composites From Selective Laser Sintered CeramicPreforms,” Proceedings Solid Freeform Fabrication Symposium,August 9-11 1993, University of Texas, Austin Texas, pp. 215-222.

[Dickson78] A. Dickson, B. Skews, and R. Marcus, “Plug Phase Conveying,”Proceedings of Pneumotransport 4, paper D6, Organized byBHRA Fluid Engineering, Cranfield, Bedford, UK, June 1978.

[DTM94a] Laserlite™ LN4010 Nylon Compound product data sheet, DTMCorp., Austin, Texas, 1994.

[DTM94b] Laserlite™ LNF5000 Nylon Compound product data sheet,DTM Corp., Austin, Texas, 1994.

[DTM94c] Laserlite™ LPC3000 Polycarbonate Compound product datasheet, DTM Corp., Austin, Texas, 1994.

[DTM94d] Laserlite™ LWX2010 Wax Compound product data sheet, DTMCorp., Austin, Texas, 1994.

[FDM96] FDM users group meeting, William Priedeman, Stratasys, Inc.,June 24-26, 1996. Alexandria MN.

[Flain72] R. Flain, “Pneumatic Conveying: How the System is Matched tothe Materials,” Process Engineering, November 1972, pp. 88-90.

Page 116: Transporte Neumatico Thesis

103

[Geldart90] D. Geldart, “Powder Processing - The Overall View,” in:Principles of Powder Technology, M. Rhodes (ed.), John Wileyand Sons, 1990, pp. 1-8.

[Greul95] M. Greul, T. Pintat, and M. Greulich, “Rapid Prototyping ofFunctional Metallic Parts,” Computers in Industry, vol. 28, 1995,pp. 23-28.

[Heinzel85] J. Heinzel and C. Hertz, “Ink-Jet Printing,” in: Advances inElectronics and Electron physics, P. Hawkes(ed.), AcademicPress, Inc., vol. 65, 1985, pp. 91-171.

[Konrad80] K. Konrad, D. Harrison, R. Nedderman, and J. Davidson,“Prediction of the Pressure Drop for Horizontal Dense PhasePneumatic Conveying of Particles,” Proceedings ofPneumotransport 5, paper E1, Organized by BHRA FluidEngineering, Cranfield, Bedford, UK, April 1980.

[Konrad81] K. Konrad, “Dense Phase Pneumatic Conveying of Particles,”Ph.D. Dissertation, University of Cambridge, UK, 1981.

[Konrad86] K. Konrad, “Dense-Phase Pneumatic Conveying: A Review,”Powder Technology, vol. 49, 1986, pp. 1-35.

[Lakshminarayan92] U. Lakshminarayan and H. Marcus, “An Experimental Study ofthe Relationship between Microstructure and MechanicalProperties of a Ceramic Fabricated by Selective Laser Sintering,”Proceedings Solid Freeform Fabrication Symposium, August 3-51992, University of Texas, Austin Texas, pp. 44-53.

[Lee93] S. Lee, E. Sachs, and M. Cima, “Powder Layer PositionAccuracy in Powder-Based Rapid Prototyping,” ProceedingsSolid Freeform Fabrication Symposium, August 9-11, 1993,University of Texas, Austin Texas, pp. 223- 236.

[Lippert66] A. Lippert, “Pneumatic Conveyance of Solids at HighConcentrations,” Chemie-Ingenieur-Technik, vol. 38, no. 3, 1966,pp. 350-355.

Page 117: Transporte Neumatico Thesis

104

[Mainwaring93] N. Mainwaring, “Characterization of Materials for PneumaticConveying,” American Ceramic Society Bulletin, vol. 72, no. 8,August 1993, pp. 63-71.

[Marcus90] R. Marcus, Pneumatic Conveying of Solids, Chapman and Hall,1990.

[Melvin91] L. Melvin and J. Beaman, “The Electrostatic Application ofPowder for Selective Laser Sintering,” Proceedings SolidFreeform Fabrication Symposium, 1991, University of Texas,Austin Texas, pp. 171- 177.

[Michaels92] S. Michaels, E. Sachs, and M. Cima, “Metal Parts Generation ByThree Dimensional Printing,” Proceedings Solid FreeformFabrication Symposium, August 3-5 1992, University of Texas,Austin Texas, pp. 244- 250.

[Pegna95] J. Pegna, “Exploratory Investigation of Layered FabricationApplied to Construction Automation,” Design EngineeringTechnical Conferences, vol. 1, ASME 1995, pp. 219-226.

[Prabhu93] G. Prabhu and D. Bourell, “Supersolidus Liquid Phase SelectiveLaser Sintering of Prealloyed Bronze Powder,” Proceedings SolidFreeform Fabrication Symposium, August 9-11 1993, Universityof Texas, Austin Texas, pp. 317- 324.

[Rock95] S. Rock and C. Gilman, “A New SFF Process for Functional PartRapid Prototyping and Manufacturing: Freeform PowderMolding,” Proceedings Solid Freeform Fabrication Symposium,August 7-9 1995, University of Texas, Austin Texas.

[Sachs92] E. Sachs, M. Cima, P. Williams, D. Brancazio, and J. Cornie,“Three Dimensional Printing: Rapid Tooling and PrototypesDirectly from a CAD Model,” Journal of Engineering forIndustry, vol. 114, November 1992, 481-488.

[Sadler49] A. Sadler, “Gas-Solids Fluidizing for Transport,” ChemicalEngineering, vol. 56, no. 5, 1949, pg. 110.

Page 118: Transporte Neumatico Thesis

105

[Sanders96] Sanders Prototype, Inc., May 1996, http://www.sanders-prototype.com/techdesc.htm.

[SFF Group95] Solid Freeform Fabrication Group, University of Texas, atAustin, 1995, http://sffoffice.utexas.edu.

[Sindel94] M. Sindel, T. Pintat, M. Gruel, O. Nyrhila, and C. Wilkening,“Direct Laser Sintering of Metals and Metal Melt Infiltration forNear Net Shape Fabrication of Components,” Proceedings SolidFreeform Fabrication Symposium, August 8-10 1994, Universityof Texas, Austin Texas, pp. 94- 101.

[Tobin93] J. Tobin, B. Badrinarayan, J. Barlow, J. Beaman, and D. Bourell,“Indirect Metal Composite Part Manufacture Using the SLSProcess,” Proceedings Solid Freeform Fabrication Symposium,August 9-11 1993, University of Texas, Austin Texas, pp. 303-307.

[Vail92] N. Vail and J. Barlow, “Ceramic Structures by Selective LaserSintering of Microencapsulated, Finely Divided CeramicMaterials,” Proceedings Solid Freeform Fabrication Symposium,August 3-5 1992, University of Texas, Austin Texas, pp. 124-130.

[Van der Schueren95] B. Van der Schueren and J. Kruth, “Powder Deposition inSelective Metal Powder Sintering,” Rapid Prototyping Journal,vol. 1, no. 3, 1995, pp. 23- 31.

[Yoo93] J. Yoo, M. Cima, S. Khanuja, and E. Sachs, “Structural CeramicComponents by 3D Printing,” Proceedings Solid FreeformFabrication Symposium, August 9-11 1993, University of Texas,Austin Texas, pp. 40-50 .

Page 119: Transporte Neumatico Thesis

106

APPENDIX A

BILL OF MATERIALSFOR

EXPERIMENTAL APPARATUS

PART SOURCE CATALOG # COST QUANTITY TOTAL COST

manifold Cole Parmer H-06464-82 $6.00/each 1 $6.00luer adapter Cole Parmer H-06359-17 $8.00/25 4 $8.00

luer lock Cole Parmer H-06359-67 $8.00/25 4 $8.00screw clamp Cole Parmer H-06833-10 $10.25/3 3 $10.25

str. micro-fitting Cole Parmer H-06173-00 $41.00/3 3 $41.00reducer fitting Cole Parmer H-06391-70 $19.43/each 6 $58.29

Teflon PFA tubing Cole Parmer H-06375-01 $27.00/25 ft. $27.00Polyurethane tubing Cole Parmer H-06423-01 $20.75/50 ft. $20.75

instrument cart Cole Parmer MO02-K31 $309.50 1 $309.50power receptacle Cole Parmer AO-09 $52.00 1 $52.00

motor C and H DCGM7505 $29.95 1 $29.95motor control kit Marlin P. Jones 4057-MD $15.95 1 $15.95AC/DC converter Radio Shack 273-1653B $21.99 1 $21.99

board True Value plywood $5.00 1 $5.00tape Walmart Scotch Mailing

Tape$1.99 1 $1.99

nozzle self-fabricated 1nozzle support self-fabricated 1

spool self-fabricated 1slider platform self-fabricated 1

slider tray self-fabricated 1funnel self-fabricated 4

circuit support self-fabricated 1main support stand scrap metal 1

track groove scrap metal 1eye hook spare 1

thread spare 1powder DTM Laserlite™

LPC3000Polycarbonate

TOTAL $615.67

Page 120: Transporte Neumatico Thesis

107

APPENDIX B

PARTS LISTFOR

EXPERIMENTAL APPARATUS

PART QUANTITY REFERENCE CODEbase board 1 A

track groove 1 Bmain support stand 1 C

main air line 1 1 D1main air line 2 1 D2

luer fitting 5 Eluer lock 5 Fmanifold 1 G

middle tubing 3 Hscrew clamps 3 Ireducer fitting 3 JTeflon tubing 3 K

straight micro-fitting 3 Lnozzle 1 M

nozzle support 1 Nmotor 1 O

motor spool 1 Pspeed control circuit 1 Q

circuit support 1 Rslider top multiple S

slider bottom 1 Tadhesive surface multiple U

eye hook 1 Vfunnel 4 W

AC/DC converter 1 X

Page 121: Transporte Neumatico Thesis

108

APPENDIX C

SUPPLEMENTAL DRAWINGSOF

EXPERIMENTAL APPARATUS

Page 122: Transporte Neumatico Thesis

109

Quantity : 1

Reference Code : A

Catlog Number :

Material : Plywood

Part : Base Board Source : True Value

Note: All dimensions in mm.

Figure A.1: Base board.

Page 123: Transporte Neumatico Thesis

110

Quantity : 1

Reference Code : B

Catlog Number :

Material : Sheet Metal

Part : Track Groove Source :

Note: All dimensions in mm.

Figure A.2: Track Groove

Page 124: Transporte Neumatico Thesis

111

Quantity : 1

Reference Code : C

Catlog Number :

Material : Sheet Metal

Part : Main Support Stand Source :

Note: All dimensions in mm.

Figure A.3: Main support stand.

Page 125: Transporte Neumatico Thesis

112

Quantity : 1

Reference Code : D1

Catlog Number : H-06375-01

Material : Polyurethane

Part : Main Air Line 1 Source : Cole Parmer

Note: All dimensions in mm.

Figure A.4: Main air line 1.

Page 126: Transporte Neumatico Thesis

113

Quantity : 1

Reference Code : D2

Catlog Number : H-06375-01

Material : Polyurethane

Part : Main Air Line 1 Source : Cole Parmer

Note: All dimensions in mm.

Figure A.5: Main air line 2.

Page 127: Transporte Neumatico Thesis

114

Quantity : 5

Reference Code : E

Catlog Number : H-06359-17

Material : Plastic

Part : Luer Fitting Source : Cole Parmer

Note: All dimensions in mm.

Figure A.6: Luer fitting.

Page 128: Transporte Neumatico Thesis

115

Quantity : 5

Reference Code : F

Catlog Number : H-06359-67

Material : Plastic

Part : Luer Lock Source : Cole Parmer

Note: All dimensions in mm.

Figure A.7: Luer lock.

Page 129: Transporte Neumatico Thesis

116

Quantity : 3

Reference Code : H

Catlog Number : H-06423-01

Material : Polyurethane

Part : Middle Tubing Source : Cole Parmer

Note: All dimensions in mm.

Figure A.8: Middle tubing.

Page 130: Transporte Neumatico Thesis

117

Quantity : 3

Reference Code : I

Catlog Number : H-06833-10

Material : Plastic

Part : Screw Clamps Source : Cole Parmer

Note: All dimensions in mm.

Figure A.9: Screw Clamp.

Page 131: Transporte Neumatico Thesis

118

Quantity : 3

Reference Code : J

Catlog Number : H-06391-70

Material : PTFE

Part : Reucer Fitting Source : Cole Parmer

Note: All dimensions in mm.

Figure A.10: Reducer fitting.

Page 132: Transporte Neumatico Thesis

119

Quantity : 3

Reference Code : K

Catlog Number : H-006375-01

Material : Teflon PFA

Part : Teflon Tubing Source : Cole Parmer

Note: All dimensions in mm.

Figure A.11: Teflon tubing.

Page 133: Transporte Neumatico Thesis

120

Quantity : 3

Reference Code : L

Catlog Number : H-06173-00

Material : Plastic

Part : Straight Micro-Fitting Source : Cole Parmer

Note: All dimensions in mm.

Figure A.12: Straight micro-fitting.

Page 134: Transporte Neumatico Thesis

121

Quantity : 1

Reference Code : M

Catlog Number :

Material : ABS plastic

Part : Nozzle Source :

Note: All dimensions in mm.

Figure A.13: Nozzle.

Page 135: Transporte Neumatico Thesis

122

Quantity : 1

Reference Code : N

Catlog Number :

Material : ABS plastic

Part : Nozzle Support Source :

Note: All dimensions in mm.

Figure A.14: Nozzle support.

Page 136: Transporte Neumatico Thesis

123

Quantity : 1

Reference Code : P

Catlog Number :

Material : ABS plastic

Part : Motor Spool Source :

Note: All dimensions in mm.

Figure A.15: Motor spool.

Page 137: Transporte Neumatico Thesis

124

Quantity : 1

Reference Code : R

Catlog Number :

Material : ABS plastic

Part : Circuit Support Source :

Note: All dimensions in mm.

Figure A.16: Circuit Support.

Page 138: Transporte Neumatico Thesis

125

Quantity : 6

Reference Code : S

Catlog Number :

Material : ABS plastic

Part : Slider Top Source :

Note: All dimensions in mm.

Figure A.17: Slider Top.

Page 139: Transporte Neumatico Thesis

126

Quantity : 1

Reference Code : T

Catlog Number :

Material : ABS plastic

Part : Slider Bottom Source :

Note: All dimensions in mm.

Figure A.18: Slider bottom.

Page 140: Transporte Neumatico Thesis

127

Quantity : 1

Reference Code : U

Catlog Number : Scotch Mailing Tape

Material :

Part : Adhesive Surface Source : Walmart

Note: All dimensions in mm.

Figure A.19: Adhesive surface.

Page 141: Transporte Neumatico Thesis

128

Quantity : 1

Reference Code : V

Catlog Number :

Material : Steel

Part : Eye Hook Source :

Note: All dimensions in mm.

Figure A.20: Eye Hook.

Page 142: Transporte Neumatico Thesis

129

Quantity : 4

Reference Code : W

Catlog Number :

Material : ABS plastic

Part : Funnel Source :

Note: All dimensions in mm.

Figure A.21: Funnel.

Page 143: Transporte Neumatico Thesis

130

Figure A.22: Layout of fixtures attached to base board.

Page 144: Transporte Neumatico Thesis

131

Figure A.23: Layout of fixtures attached to main support stand.

Page 145: Transporte Neumatico Thesis

132

Figure A.24: Exploded manifold assembly.

Page 146: Transporte Neumatico Thesis

133

Figure A.25: Exploded nozzle assembly.

Page 147: Transporte Neumatico Thesis

134

Figure A.26: Exploded motor assembly.

Page 148: Transporte Neumatico Thesis

135

APPENDIX D

SUPPLEMENTAL DRAWINGSOF

MULTIPLE POWDER DEPOSITION SYSTEM

Page 149: Transporte Neumatico Thesis

136

Figure A.27: Powder print head.

Page 150: Transporte Neumatico Thesis

137

Figure A.28: Proposed nozzle.

Page 151: Transporte Neumatico Thesis

138

Figure A.29: Motion system.

Page 152: Transporte Neumatico Thesis

139

Figure A.30: Powder loading system.

Page 153: Transporte Neumatico Thesis

140

VITA

Shawn Fitzgerald was born in Pompton Plains, New Jersey on October 2, 1972. He

attended St. James High School in Carney’s Point, New Jersey. He obtained his Bachelor

of Science degree in Mechanical Engineering from Virginia Tech in May of 1994. After an

eight month internship with the Delaware River and Bay Authority in New Castle,

Delaware, Shawn returned to Virginia Tech to receive his Master of Science degree in

Mechanical Engineering in July 1996. He is working for the Defense Systems and

Electronics group of Texas Instruments located in the greater Dallas area.