University of Arkansas, Fayetteville

ScholarWorks@UARK
Mechanical Engineering Undergraduate Honors
Mechanical Engineering
Theses

5-2016

Low-Cost Inkjet Process for 3-D Printing

Christopher T. Schmitt
University of Arkansas, Fayetteville

Follow this and additional works at: http://scholarworks.uark.edu/meeguht

Part of the Electro-Mechanical Systems Commons, and the Manufacturing Commons

Recommended Citation
Schmitt, Christopher T., "Low-Cost Inkjet Process for 3-D Printing" (2016). Mechanical Engineering Undergraduate Honors Theses. 54.
http://scholarworks.uark.edu/meeguht/54

This Thesis is brought to you for free and open access by the Mechanical Engineering at ScholarWorks@UARK. It has been accepted for inclusion in
Mechanical Engineering Undergraduate Honors Theses by an authorized administrator of ScholarWorks@UARK. For more information, please
contact scholar@uark.edu, ccmiddle@uark.edu.
 Abstract

This paper presents a low-cost 3-D inkjet process design. Current commercially

available 3-D inkjet printers (e.g., Objet Connex) that are capable of multi-material printing are

expensive and out-of-reach of most consumers. While fused deposition modeling and other

extrusion-based hybrid processes for multi-material printing are becoming available, these

techniques are slow and relatively low-resolution. This research attempts to fill the gap by

presenting a simple and low-cost solution to bring desktop 3-D inkjet to the community by

utilizing components from an Epson inkjet printer. The mechanical system was designed with

maximum utilization of the Epson carriage system with the addition of a third axis. A 3rd party

ink supply system was used in replacement of the standard ink cartridges and an integrated UV

LED curing system was developed. Finally, a simple software control system was implemented

using Arduino. Costs were managed by exploiting the original driver electronics, power supply,

and X-stage motor and feedback mechanism.

ii
 Acknowledgements

First and foremost, I would like to thank Dr. Wenchao Zhou for introducing me to this

exciting field of study. His dedication as a teacher and commitment to the project has helped

and inspired me tremendously. Dr. Zhou’s team of undergraduate and graduate students at

the AM3 lab has been a great work environment for advice, support, and comradery.

Exposure to the many diverse projects at the lab has taught me immensely, as well as given

me some great friends.

I would also like to thank Danny Mora for his assistance with mechanical design, and

Garrett Urban for his experience with Arduino. Without them, and their many hours of help,

this project would not be where it is today. Many thanks also go to Zachary Zelenka and Joe

Moquin, for their electrical and software help, especially with regards to the UV LED modules

and voltage divider. Their knowledge and consultation was extremely valuable to me.

Finally, many thanks to the Honors College (including Walton Family Foundation) and

Arkansas Department of Higher Education for their generous scholarships. The Fellowship

and Governor’s Scholarship have made it possible for me to focus on my studies without any

financial burden. It has been both an honor and a humbling experience to live and learn here

at the University of Arkansas.

iii
 Contents

Abstract ......................................................................................................................................................................... ii
Acknowledgements................................................................................................................................................ iii
Figures ........................................................................................................................................................................... v
Tables .......................................................................................................................................................................... vii
Chapter 1. Introduction .......................................................................................................................................... 1
1.1 Background of Additive Manufacturing ............................................................................................. 1
1.2 Current Capabilities .................................................................................................................................... 3
1.3 Open Source Revolution ............................................................................................................................ 4
1.4 Motivation for Study ................................................................................................................................... 6
Chapter 2. Design Overview ................................................................................................................................. 9
2.1 Epson’s System ........................................................................................................................................... 10
2.1.1 WF30 Specifications ........................................................................................................................ 10
2.1.2 WF30 Mechanical System ............................................................................................................. 11
2.1.3 WF30 Electrical System & Print Head ..................................................................................... 14
2.2 Mechanical System.................................................................................................................................... 18
2.2.1 X-Y Stage ............................................................................................................................................... 18
2.2.2 Z-stage ................................................................................................................................................... 21
2.2.3 Mechanical Endstops....................................................................................................................... 23
2.2.4 Ink Supply System ............................................................................................................................ 24
2.3 Electrical System ....................................................................................................................................... 25
2.3.1 Arduino MEGA 2560 & RAMPS 1.4............................................................................................ 26
2.3.2 Voltage Divider .................................................................................................................................. 26
2.3.3 Photo-curing System ....................................................................................................................... 30
2.4 Software ........................................................................................................................................................ 33
Chapter 3. Results & Discussion ...................................................................................................................... 35
3.1 Results ............................................................................................................................................................ 35
3.2 Discussion ..................................................................................................................................................... 38
Chapter 4. Conclusion & Future Work .......................................................................................................... 39
4.1 Conclusion .................................................................................................................................................... 39
4.2 Future Work ................................................................................................................................................ 40
Appendix A. Software Code ............................................................................................................................... 42
Appendix B. Optocoupler Circuit Design ..................................................................................................... 43
References ................................................................................................................................................................ 44

iv
 Figures

Figure 1-1. CAD model of teapot, showing smooth surface features [2]…………………………………….2

Figure 1-2. Teapot printed using Stratasys uPrint SE Plus, showing horizontal layering…………..3

Figure 1-3. Helmets printed using Connex3 in full-color photopolymer material [6]…………………4

Figure 1-4. Tower of Pisa, printed on Ultimaker 2, surface artifacts exist on tower's loggia………5

Figure 1-5. Piezoelectric Inkjet, PZT represented in red [14]……………………………………………………7

Figure 1-6. Thermal Inkjet, heater represented in red [14]……………………………………………………...7

Figure 2-1. WF30 print head nozzle layout [20]……………………………………………………………………11

Figure 2-2. WF30 printer mechanism diagram [20]……………………………………………………………...12

Figure 2-3. WF30 Waste Ink System [20]……………………………………………………………………………...13

Figure 2-4. WF30 main board………………………………………………………………………………………………14

Figure 2-5. Print head assembly – Nozzle Plate (top), CN5 & CN7 connectors (bottom right)….15

Figure 2-6. Analog Drive Waveform of WF30 [19]……. ………………………………………………………….15

Figure 2-7. CN5 Circuit snippet, CHA and CHB may be located [23] ……………………………………….16

Figure 2-8. CN7 Circuit snippet, LAT, NCHG, SI1, SI2, and SI3 may be located [23] ………………….17

Figure 2-9. Motors and sensors in WF30 [20] ……………………………………………………………………17

Figure 2-10. Epson “main frame assembly” (upside down), used as X-stage [20] ……………………18

Figure 2-11. Carriage mounts left and right, designed with the freedom of AM………………………19

Figure 2-12. X-Y stage assembled, NEMA 17 motor and drive shaft (bottom right) ………………20

Figure 2-13. Carriage mount (left) connected to timing belt on its underside………………………...21

Figure 2-14. Z-stage design: base mount (in red), build plate frame (in blue) ………………………21

Figure 2-15. Z-stage installed into printer……………………………………………………………………………22

v
Figure 2-16. X-Y-Z stages of printer with build plate…………………………………..…………………………23

Figure 2-17. Mechanical Endstop…………………………………………………………………………………….23

Figure 2-18. Epson CISS, supply lines exit the top of the cartridge top units (left) to a major

reservoir for extended printing sessions (right) ……………………………………………………….. 25

Figure 2-19. PF motor signal reading from oscilloscope………………………………………………………..27

Figure 2-20. Voltage Divider Circuit Diagram……………………………………………………………………28

Figure 2-21. Voltage divider into side of RAMPS – black (GND), yellow (A10) ………………………29

Figure 2-22. Voltage divider breadboard configuration……………………………………………………..29

Figure 2-23. Left: UV photo-curing circuit, Right: lumped element model [24] ………………………30

Figure 2-24. Breadboard test of UV photo-curing circuit [24] ……………………………………………….31

Figure 2-25. UV LED PCB design [24] ……………………………………………………………………………….32

Figure 2-26. UV LED mount design…………………………………………………………………………………...32

Figure 2-27. UV LED mounts installed………………………………………………………………………………….32

Figure 2-28. UV LED on print head……………………………………………………………………………………….32

Figure 3-1. Nozzle check pattern printed……………………………………………………………………………...36

Figure 3-2. Nozzle check pattern, according to Epson [20] ……………………………………………………36

Figure 3-3. Rectangle printed………………………………………………………………………………………………37

Figure B-1. Proposed twin optocoupler design for isolating Epson/Arduino circuits………………43

vi
 Tables

Table 2-1. Arduino Analog Signal Pins for RAMPS Endstops…………………………………………………24

vii
 Chapter One.

Introduction

1.1 Background of Additive Manufacturing

Commonly known as 3D printing, additive manufacturing (AM) is a process that has

become increasingly popular in the recent years. Its capability to turn a computerized virtual

model into existence with relative speed and ease is revolutionary. There are a wide range

of AM techniques available, including stereolithography (SLA), fused deposition modeling

(FDM), selective laser sintering (SLS), and inkjet (binder jetting or material jetting), to name

a few. AM processes share the similarity that they are layer-based manufacturing methods,

in which three-dimensional parts are built through the sequential deposition of two-

dimensional layers of material. These layers are literally cross-sections of the final part –

and as layers become increasingly thin, the 3-D print will appear to more closely represent

the original computer aided design (CAD) model (Figure 1-1). The final physical part (Figure

1-2) can thus be considered an approximation of the initial virtual data [1].

1
 Figure 1-1. CAD model of teapot, showing smooth surface features [2]

There are numerous benefits to 3D printing that make it an attractive tool for

designers. AM’s layer-based approach allows for the creation of complex geometries that

would otherwise be impossible or very difficult and time-consuming using traditional

subtractive manufacturing approaches. Also, AM is a digital fabrication process, meaning

exceedingly customized parts could be manufactured relatively economically, in a highly

automated environment. Additionally, AM has the capability for rapid prototyping –

designers can prototype and redesign with enhanced speed and freedom. Finally, AM has

the potential to allow anyone to design, build, and create – consolidating what used to be the

need for a large workspace and a wide variety of tools into a single instrument that can

become an intuitive extension of the imagination.

2
 Figure 1-2. Teapot printed using Stratasys uPrint SE Plus, showing horizontal layering

1.2 Current Capabilities

Since the invention of the first commercial SLA machine in 1986 by Chuck Hull (who

went on to found 3DSystems) [3], current capabilities of rapid prototyping has increased

tremendously. From improvements in speed, build volume, cost, materials and part

strength, AM machine are starting to close the gap with traditional methods, such as injection

molding, in terms of final part accuracy, quality, and end-usability. Direct Metal Printing

(DMP) is a technology that allows one to create metal and ceramic parts – that has proved

useful in the medical, dental, and aerospace industries [4-5].

The Objet Connex series by Stratasys are currently the only printers on the market

capable of multiple material printing in high-resolution. In addition, these printers can print

in multiple colors simultaneously. A sample print from a Connex3 is shown below in Figure

1-3.

3
 Figure 1-3. Helmets printed using Connex3 in full-color photopolymer material [6]

Build resolution of an Objet 500 Connex3, the series flagship, for example, is 600 DPI in the

X- and Y-axes and 1600 DPI in the Z-axis. Layer thickness is down to 16-micron resolution,

for smooth surface features and minimal artifacts. Using inkjet print heads, the printer is

capable of using a wide variety of materials with properties that are rigid, flexible, thermal

resistant, and even translucent [7]. A preventative price tag of over a quarter-million USD,

however, limits this powerful technology from many potential consumers [8].

1.3 Open Source Revolution

Previously, 3D printing has been largely confined to businesses, research institutions,

and design professionals. Recently, however, the market for consumer-levels printers has

grown considerably with the help of a strong open-source community. The RepRap printer

is an open-source community project that is dedicated to the development of a low-cost self-

replicating printer [9]. Here, enthusiasts can communicate, post instructions, and offer

improvements on printer designs. Primarily FDM printer information is found on the

RepRap site, due to the technology’s inherent simplicity and low-cost. MakerBot (now

4
acquired by Stratasys) spawned from the RepRap project, and is focused on producing low-

cost FDM printers. The MakerBot Replicator printer uses a 0.4mm nozzle with 100-micron

layer resolution at a cost of USD$2899 [10]. The MakerBot, and other similar FDM printers

such as Ultimaker, are optimized for depositing polylactic acid (PLA) and acrylonitrile

butadiene styrene (ABS). These thermoplastics are easy to extrude, are available in a wide

variety of colors, and are relatively inexpensive. Figure 1-4 shows the capability of an

Ultimaker 2 FDM printer.

Figure 1-4. Tower of Pisa, printed on Ultimaker 2, surface artifacts exist on tower's loggia

However, unlike their more expensive counterparts, these printers are not capable of multi-

material printing. A printer called the Fab@Home was created to address this issue. Like

RepRap, Fab@Home is also entirely open-source, and was the first multi-material printer

5
available [11]. This project has since seemed to have fallen by the wayside, as their website

is no longer in service [12].

1.4 Motivation for Study

The ultimate goal is to make the 3-D printer as simple-to-use and commonplace as an

appliance. Therefore, it needs to be affordable to everyone, while still maintaining a level of

performance and precision that rivals the expensive industrial printers of today. Low-cost

FDM printers are widely available, however FDM performance is inherently slow because it

“requires material to be plotted in a point-wise, vector fashion that involves many changes

in direction” [1]. A solution, then, is to use inkjet technology, particularly drop-on-demand

(DOD) inkjet technology, because it is fast, capable of multi-material, and high-resolution.

The two most common types of inkjet technology available are piezoelectric inkjet

and thermal inkjet (also known as bubble jet). In a piezoelectric inkjet print head, the

piezoelectric element sits atop a thin diaphragm that is positioned above the nozzle. One of

the most common piezoelectric materials is lead zirconate titanate (PZT), a man-made

ceramic. Its efficiency in reception and transmission makes it an ideal material for

transducers, converting an electrical signal into mechanical vibrations [13]. In an inkjet print

head, a voltage is applied to the PZT, causing the diaphragm to bend upwards, drawing in ink

from a reservoir located above the mechanism. Subsequently, an opposite voltage is applied,

causing the diaphragm to invert rapidly, expelling a single droplet of ink [14], as illustrated

in Figure 1-5.

6
 Figure 1-5. Piezoelectric Inkjet, PZT represented in red [14]

A thermal inkjet print head shares similar ink plumbing, however a small heating element

sits in place of the PZT, as shown in Figure 1-6. This heater (essentially a power resistor),

vaporizes the ink generating a pressure increase, thus causing a droplet to be ejected [15].

Figure 1-6. Thermal Inkjet, heater represented in red [14]

The issue with this thermal inkjet process is that the heaters have a short lifetime, due the

high electrical currents necessary to run the heating elements [14, 16]. Manufacturers try to

solve this issue by packaging the print head as part of the cartridge, in order to ensure that

it is replaced regularly. An additional issue is that thermal inkjet can’t be used with heat-

sensitive materials [14]. The thermal inkjet process can be found in nearly all consumer-

7
level desktop office (2-D) printers from companies like Hewlett-Packard, Lexmark, and

Canon. The exception is Epson.

Epson first introduced their "Micro Piezo” technology in the early 1990s [17]. The

piezoelectric inkjet process does not alter the chemical composition of the jetted material

and has the ability to be finely tuned in order to produce adequate pressure for droplet

generation [15]. Additionally, PZT print heads have a low operating temperature, increasing

reliability and allowing for the print head to be integrated into the machine [14]. Epson has

used this technology in their entire lineup of desktop office printers since 1993 with the

release of the Epson Stylus 800 [17]. Utilizing the print head and components in these

relatively inexpensive machines, one could build a 3-D inkjet printer at a fraction of the price

of current offerings. This paper presents a low-cost inkjet process for additive

manufacturing, part of a larger and ongoing effort to eventually bring the desktop 3-D printer

to more homes around the world.

8
 Chapter Two.

Design Overview

To create something that creates is an exceptional task. In order to make the 3-D

printer like an appliance, it should be simple to use and unobtrusive in the user’s life –

subtle and overt. Furthermore, to develop a product for the mass-market presents a

significant challenge because it must be desirable and affordable. Therefore, maximum

utilization of off-the-shelf components is crucial in order to manage costs. The printer was

made with this design philosophy in mind. Parts such as the print head, driver electronics,

power supply, and X-stage motor and feedback mechanism were all sourced from an Epson

printer and are discussed in Section 2.1. Section 2.2 details the mechanical system that was built

around these components, providing three axes of motion, as well as an ink supply system for the

print head. Section 2.3 presents the open-source Arduino platform that was chosen for

controlling the stepper motors, the voltage divider for reading Epson’s signals, and the integrated

photo-curing system. Finally, Section 2.4 outlines the software ideology, as well as some basic

calculations for keeping the X- and Y-stage in sync using Arduino. Arduino’s vibrant

community of users, especially amongst “RepRappers”, made it an ideal platform for this

9
project. This project may very well be the first inkjet RepRap incorporating a Micro Piezo print

head [18].

2.1 Epson’s System

The desktop inkjet printer chosen for this project was the Epson WorkForce 30

(WF30). The relatively large amount of information available on this printer was the main

reason for choosing this model at the time. A similar project, the MultiFab, conducted at the

Massachusetts Institute of Technology, used the WF30 print head because of its high

performance and low cost [19]. In addition, a service manual was obtained for the WF30,

aiding substantially in the understanding of the printer’s mechanical and electrical systems.

2.1.1 WF30 Specifications

The WF30 uses drop-on-demand piezoelectric ink jet technology developed by

Epson. The print head is Epson’s F3-3 Mach Turbo II, containing 540 nozzles, divided into 3

rows of 180 nozzles each. The print head accepts 5 ink cartridges: cyan, magenta, yellow,

and two cartridges of black (CMYK). The first and second row are both devoted to black,

while the third row is subdivided equally into 3 nozzle banks for cyan, magenta, and yellow

(60 nozzles each), shown in Figure 2-1 [20]. The print head can eject droplets of 6 pl, 13 pl,

or 26 pl of color ink and 13 pl or 26 pl droplets of black ink [21].

10
 Figure 2-1. WF30 print head nozzle layout [20]

2.1.2 WF30 Mechanical System

The WF30 shares the same printer mechanism with the Epson Stylus C110, C120,

D120, Office B30, T30, T33, and ME Office 70 [20]. The printer may be divided up into the

print head, carriage mechanism, paper loading mechanism, paper feed mechanism, and ink

system, as shown in Figure 2-2.

11
 Figure 2-2. WF30 printer mechanism diagram [20]

Similar to most 2-D printers, the carriage unit with print head traverses along a

metal main frame assembly, driven by a timing belt and DC motor (CR motor). In the

transverse direction, paper feeds in through a series of rollers, driven by another DC motor

(PF motor). During the printer initialization sequence, nozzle priming, and after printing

completes, the carriage unit travels back to its “home position” above the waste ink system.

This is where it is stored when the printer is off. The waste ink system (Figure 2-3)

contains a deployable wiper, rubber cap, pump, and waste ink chamber.

12
 Figure 2-3. WF30 Waste Ink System [20]

This system is surprisingly intricate. As the carriage unit travels back to the waste ink

system, the PF motor reverses direction, causing the wiper to raise by a cam. This rubber

wiper, essentially a squeegee, cleans the nozzle plate of residual ink. When the carriage

reaches the end of its travel, it bumps the cap slider, thus pulling the rubber cap up into the

nozzle plate, forming a seal. The PF motor then spins in a rapid state, driving the pump

unit, which removes any waste ink through a mesh screen in the cap and into a reservoir

underneath. This act of drawing ink from the channels and applying suction to the nozzle

plate helps to mitigate the PZT print head susceptibility to entrapped air [22]. Finally, the

carriage lock lever raises, securing the carriage in its home position. This means that the

PF motor serves two functions – to feed paper past the print head and to drive the waste

ink system via a clever clutch mechanism.

13
2.1.3 WF30 Electrical System & Print Head

The main circuit board of the WF30 (Figure 2-4) is where the analog drive waveform

for the PZTs is created, printer interfaces with the PC, and is the central hub for all of the

printer’s electrical components.

Figure 2-4. WF30 main board

The main ribbon cables CN6, CN7, and CN5 (each with 17, 17, and 18 leads, respectively)

connect the main board to the carriage unit, with CN5 and CN7 connecting directly to the

print head (Figure 2-5). CN6 splits at the carriage unit and goes to the CR Scale (optical

sensor for X-stage feedback) and the Head FFC Connector for reading the ink cartridges.

Other important nomenclature on the main board: CN8 connects to CR motor, CN9 to PF

motor, CN10 to PF encoder sensor, and CN3 to paper empty (PE) sensor.

14
 Figure 2-5. Print head assembly – Nozzle Plate (top), CN5 & CN7 connectors (bottom right)

The print head utilizes a trapezoidal analog waveform for driving the PZTs and a

series of digital signals for nozzle selection and droplet size. One period of the analog

waveform (Figure 2-6) contains four sub-waveforms. Combinations of these sub-waveforms

results in the three droplet sizes possible: 6, 13, and 26 pl [19].

Figure 2-6. Analog Drive Waveform of WF30 [19]

15
In order to control printing, there are eight digital signals: NCHG, LAT, CHA, CHB, SI1, SI2,

SI3, and CK. Their corresponding pins on CN5 and CN7 may be seen in Figures 2-7 and 2-8,

respectively. NCHG determines whether the analog waveform activated the ink chambers.

CHA, CHB, and LAT show the start of a sub-waveform. SI1, SI2, and SI3 are responsible for

specifying drop size. Finally, CK is the signal of the clock [19]. This information about

Epson’s electrical signals was sourced from Joyce G. Kwan.

Figure 2-7. CN5 Circuit snippet, CHA and CHB may be located [23]

16
 Figure 2-8. CN7 Circuit snippet, LAT, NCHG, SI1, SI2, and SI3 may be located [23]

The WF30 is powered by a 42VDC power supply. The high voltage is necessary for

driving the PZTs. As mentioned briefly in section 2.1.2, there are two 42V DC motors: the CR

and PF motor. The CR motor has a photo interrupter sensor with a resolution of 180

pulse/inch [20] for use as a feedback mechanism on the carriage unit’s whereabouts.

Similarly, the PF motor has a rotary photo interrupter sensor for tracking motor movement.

The paper empty (PE) sensor is another photo interrupter for detecting the top and bottom

edge of sheets of paper. These components are pointed out in Figure 2-9.

Figure 2-9. Motors and sensors in WF30 [20]

17
2.2 Mechanical System

The idea was for maximum utilization of the Epson components in order to minimize

costs. The entire “main frame assembly”, containing the carriage unit with print head, CR

motor, and CR Encoder Sensor was taken from the Epson printer. In addition, the power

supply and main board unit were utilized. The 3-D printer was designed according to a

Cartesian coordinate system, with the three axes situated perpendicular to one another.

2.2.1 X-Y Stage

In Epson’s system, the printing surface is fed under the fixed print carriage.

Conversely, for this 3-D printer, the print carriage should instead be moved over the fixed

printing surface (build plate). With this idea of relative motion in mind, the entire main

frame assembly was lifted from the Epson printer. It was unmodified with the intention that

it would be the ideal X-stage, since motor and feedback mechanism were already integrated

(CR motor and CR Scale, respectively). The main frame may be seen in the Figure 2-10 (CR

motor and CR Scale removed).

Figure 2-10. Epson “main frame assembly” (upside down), used as X-stage [20]

18
The main frame (hereinafter referred to as the carriage) assembly is made out of sheet-metal,

thus subject to flex, as it was clearly not intended by Epson to be moved about. Several

iterations of 3-D printed carriage mounts (Figure 2-11) were created so that the carriage

assembly could slide transversely along some smooth rod in the Y-direction using press-fit

bearings. These brackets take on somewhat odd shapes in order to integrate with the

carriage assembly and as well as reduce flex and distribute load via four linear bearings (two

per mount).

Figure 2-11. Carriage mounts left and right, designed with the freedom of AM

Belt connection points were positioned on the bottom side of both of the carriage

mounts. Since the carriage assembly is relatively heavy, driving it from both sides was
19
necessary. Commonly found in RepRap printers, GT2 pulleys and timing belts were used to

create the Y-stage, driven by a NEMA 17 stepper motor. Figure 2-12 shows the fully

assembled X-Y stage.

Figure 2-12. X-Y stage assembled, NEMA 17 motor and drive shaft (bottom right)

The NEMA 17 stepper motor connects to a drive shaft via a short closed-loop timing belt.

The drive shaft then connects to the underside of both carriage mounts (Figure 2-13) via

timing belt, thus powering the carriage from both ends. This design results in smooth Y-

stage travel.

20
 Figure 2-13. Carriage mount (left) connected to timing belt on its underside

2.2.2 Z-stage

The Z-stage was inspired by the cantilevered Ultimaker 2 design. It is completely

decoupled from the X-Y stage (Figure 2-14). This allows for some freedom in its placement

underneath the X-Y stage, as well as an ability to be utilized in future 3-D printer designs.

Figure 2-14. Z-stage design: base mount (in red), build plate frame (in blue)

A base mount for the Z stepper motor was 3-D printed that supports two smooth rods. The

base mount attaches to a base plate of medium-density fiberboard (MDF) by 4 screws, one

21
at each corner. The motor was directly connected to a lead screw via a segment of vinyl

tubing. In order to keep costs down, the X-Y stage was mounted above the Z-stage using 2

in. x 4 in. lumber. The entire assembly was then mounted to a MDF base plate. The printer

assembly can be seen below in Figure 2-15. An 8 in. x 8 in. MDF build plate (Figure 2-16) sits

atop the build plate frame (in blue). As the X-Y stage travels over the build plate, the Z-stage

drops down layer-by-layer.

Figure 2-15. Z-stage installed into printer

22
 Figure 2-16. X-Y-Z stages of printer with build plate

2.2.3 Mechanical Endstops

Basic mechanical switches were fitted at each end of the Y-stage motion, as a

preventative safety stop mechanism to prevent the motor from burning out in the event the

carriage assembly over-travels (Figure 2-17).

Figure 2-17. Mechanical Endstop

23
These mechanical endstops were packaged with the RAMPS 1.4 and are commonly found

on most RepRap printers. The endstops have three pins: S (signal), + (positive), and –

(ground). The + pin accepts 5V from the Arduino. The S pin outputs 0V (low) or 5V (high)

if triggered. RAMPS 1.4 makes connecting endstops easy by providing S, +, and – male

connector pins on the board. For reference, Table 2-1 lists the analog pins on Arduino that

RAMPS uses. For purposes of this project, only two endstops for Ymin and Ymax were

utilized.

Table 2-1. Arduino Analog Signal Pins for RAMPS Endstops

Xmin Xmax Ymin Ymax Zmin Zmax

Arduino Pin: 3 2 14 15 18 19

2.2.4 Ink Supply System

A 3rd party Continuous Ink Supply System (CISS) was obtained. The CISS effectively

tricks the Epson into thinking it is a genuine cartridge by using the same cartridge contact

modules of a genuine cartridge. It replaces the individual cyan, magenta, yellow, and two

black cartridges with an entire packaged unit. The CISS has a much larger ink reservoir than

a standard cartridge – supply lines extend from the tops of the cartridge sub-units to a major

reservoir. The major reservoir is easy to fill with different inks/materials (Figure 2-18). In

addition, the CISS has a reset button that tells the printer that it is refilled, whether or not

there is actually ink present. This is useful for testing printer operations without actually

jetting material. After loading ink into the CISS, air in the supply lines must be drawn out

24
using a syringe, simultaneously filling the cartridge sub-units with ink. The major reservoir

must be positioned approximately the same elevation as the cartridges print head. If the

major reservoir is positioned too high, unwanted ink may drip from the print head when not

in operation. If it is too low, the PZTs may have trouble drawing in ink. The CISS may be

seen installed into the printer in Figure 2-16.

Figure 2-18. Epson CISS, supply lines exit the top of the cartridge top units (left) to a major

reservoir for extended printing sessions (right)

2.3 Electrical System

In order to manage costs, the main board unit from the WF30 was used because it

already contains the driver electronics to activate the print head. This eliminates the need

to design a custom board that is capable of producing the trapezoidal analog waveform and

digital signals that control printing. In addition, the main board integrates with a PC via USB,

allowing one to easily send image slices from computer to printer (further explained in

Section 2.4).

25
2.3.1 Arduino MEGA 2560 & RAMPS 1.4

The Arduino MEGA 2560 microcontroller and RAMPS 1.4 (RepRap Arduino Mega

Pololu Shield) combination was chosen based on its open-source and popularity among the

RepRap community. An inexpensive 400W, 12V standard PC power supply was sourced for

powering the RAMPS. The RAMPS provides an interface between the microcontroller and

stepper motors. NEMA 17 stepper motors were used for the Y- and Z-stages.

2.3.2 Voltage Divider

One design challenge was getting the X- and Y-stages to move in sync. Since the X-

stage is operated by the Epson main board and the Y-stage by the Arduino, the Arduino had

to somehow be able to read and interpret the signals of the Epson. Therefore, the objective

was to figure out how to be able to read the signals of the PF motor. This is because, in

Epson’s system, the PF motor can be thought of as their Y-stage motor (the carriage holds the

print head as it travels along the X-direction, as paper feeds transversely in the Y-direction).

If the PF motor’s signals could be read, the Arduino could interpret that information, and

then, in turn, drive the Y-stage stepper motor. The reasoning behind not using the PF motor

as the Y-stage motor is because returning the carriage assembly to its starting position

between “pages” is necessary for layered printing.

An issue, however, is the large voltage differential that exists between the Epson and

Arduino. Since the PF motor operates at a voltage of 42V and the Arduino MEGA 2560’s

analog pin can only accept 5V maximum, a solution had to be developed to either (1)

electrically isolate the two systems or (2) reduce this voltage differential without affecting

26
either system. Either way, it had to do-so at a frequency that could keep pace with Epson’s

X-stage movement.

One of the proposed solutions was to use an optocoupler to electrically isolate

Epson’s circuit from the Arduino. A twin 4N35 optocoupler circuit was designed that

would allow the Arduino to read the PF motor’s movement and direction (Appendix B).

However, testing revealed the optocoupler to be unreliable, possibly due to its reliance on

an LED and phototransistor, and thus this solution was abandoned.

Another solution was a voltage divider. A capture of the signal sent to the PF motor

may be seen in Figure 2-19, where Vin,max = 45V and Vin,min = 17V, approximately.

Figure 2-19. PF motor signal reading from oscilloscope

27
To keep maximum output voltage of the voltage divider under 5V and avoid high current

flow into the Arduino, resistance values R1=220kΩ and R2=10kΩ were used, based off the

voltage divider equation below.

,-
𝑉"#$,&'( = 𝑉*+,&'( (1)
,. /,-

Following Equation 1, Vout,max = 1.96V and similarly, Vout,min = 0.74V, both sufficiently less than

5V, yet separated enough to be distinguishable using Arduino’s analogRead() command. A

circuit diagram of the voltage divider may be seen in Figure 2-20.

Figure 2-20. Voltage Divider Circuit Diagram

When connected, RAMPS commands nearly all of the Arduino’s analog pins. Fortunately, it

does not use pin A9, A10, A11, and A12. This voltage divider design used A10 as the analog

input from the Epson. Since the RAMPS does not have pass-through connections for these

28
pins, the wire has to be wrapped around the RAMP’s support pin and smashed between the

RAMPS and Arduino for a secure connection, as shown more clearly in Figure 2-21 and 2-22.

Figure 2-21. Voltage divider into side of RAMPS – black (GND), yellow (A10)

Figure 2-22. Voltage divider breadboard configuration

This configuration produced excellent results, with Arduino’s Serial Monitor outputting a

clean “0” (when PF motor is feeding paper) and “361” when it is in idle (on a scale of 0 to

29
1023). This large difference in the “analogRead()” inputs allowed for code that was

relatively tolerant. This means that if there were ever any analog signal less than “200”, for

example, the Y-stepper was instructed to move (because 0 is less than 200). If the signal was

over “200” (since 361 is greater than 200) then the Y-stepper was told to hold position. The

code for the voltage reading may be found in Appendix A.

2.3.3 Photo-curing System

Similar to the large-scale industrial 3-D inkjet printers, photo-curable polymer was

the logical material to use in this setup. Therefore, a photo-curing system had to be

developed. UV LEDs seemed to be the most reliable and economical choice for this project.

Two banks of 6 LEDs at 390 nm wavelength, connected in parallel with a bipolar junction

transistor (BC547) and 67 Ω resistor, were designed. The circuit of the photo-curing system

may be seen in Figure 2-23, designed for 12V DC input from the PC power supply.

Figure 2-23. Left: UV photo-curing circuit, Right: lumped element model [24]

30
 The 390 nm wavelength UV LEDs were selected based on their similarity with the

405 nm violet laser found in the FormLabs 1+ printer [25]. Using OrCAD PSpice Designer

voltage simulation and sweeping the resistance (R7) required in Figure 2-23, the 67 Ω

value for the resistor was calculated. Sweeping the resistance allowed for the optimal

resistor to be selected for the circuit. After successfully testing the circuit on a breadboard

and finding acceptable penetration and curing of FormLabs’ methacrylate photopolymer

resin (viscosity 850-900 cP, Figure 2-24 [25]), the circuit was created on an Accurate A427

CNC PCB router, located in John A. White, Jr. Engineering Hall, University of Arkansas. This

allowed for a robust and compact design that could easily be mounted on the underside of

the print head.

Figure 2-24. Breadboard test of UV photo-curing circuit [24]

The PCB have power, ground, and signal pins on both sides for a design that integrates

willingly with many future 3-D printer designs (Figure 2-25). A simple mount was created

31
(Figure 2-26) that gives the modules the proper angle towards the print area when glued to

the print head (Figure 2-27 and 2-28).

Figure 2-25. UV LED PCB design [24] Figure 2-26. UV LED mount design

Figure 2-27. UV LED mounts installed Figure 2-28. UV LED on print head

32
2.4 Software

In order to simplify the systems necessary to develop this 3-D printer, the idea was

to “trick” the printer into thinking it was doing nothing different from its original intention.

In addition, the main goal was to develop a machine that can print 2-D layers of JPEG

images over-and-over again to develop a 3-D object. The process for taking a 3-D model

and converting it into slices and then into JPEG images that could be sent to the printer was

of less-importance to the author. As a proof of concept, the author’s intentions were to load

JPEGs of various geometries into Microsoft Word and feed those images directly to the

Epson main board. Each JPEG “slice” would be on its own page in Word, with multiple

pages (a.k.a. a “ream”) stacking to form the 3-D object. This creates a straightforward

procedure for the user: Open → File → Print.

The Epson PF motor moves approximately 1mm for each “step” during a high

quality printing sequence. This was determined from some basic experiments in which a

ballpoint pen was held lightly above some paper the PF motor was feeding during a 2-D

test print. The markings measured consistently to be 1mm in length. The NEMA 17 motor

used for the Y-stage completes a full-rotation from 12,800 steps in Arduino IDE, equating to

35.55 steps/degree or 0.0281 degrees/step, as shown below in the following calculation.

𝑠𝑡𝑒𝑝𝑠 1 𝑟𝑜𝑡𝑎𝑡𝑖𝑜𝑛 𝑠𝑡𝑒𝑝𝑠 𝑑𝑒𝑔.

12800 × = 35.55 𝑜𝑟 0.0281
𝑟𝑜𝑡𝑎𝑡𝑖𝑜𝑛 360 𝑑𝑒𝑔. 𝑑𝑒𝑔. 𝑠𝑡𝑒𝑝

In order to move the Y-stage 1mm as well, the 22.4mm diameter pulley needs to rotate

2.5578 degrees, as calculated by the following.

33
 𝑥
1𝑚𝑚 𝑜𝑓 𝑡𝑟𝑎𝑣𝑒𝑙 = 2𝜋 ∙ 22.4𝑚𝑚 ∙
360

∴ 𝑥 = 2.5578 𝑑𝑒𝑔.

Multiplying this number by steps/degree of the NEMA 17, this equates to approximately 91

steps in Arduino code, as shown by the following.

𝑠𝑡𝑒𝑝𝑠
35.55 ×2.5578 𝑑𝑒𝑔. = 90.93 𝑠𝑡𝑒𝑝𝑠
𝑑𝑒𝑔.

In other words, 91 steps on the stepper motor equates to 1mm of travel on the Y-stage. A

way to think of how this works is: (1) the Epson main board sends a signal to the CR motor

telling the X-stage to make a pass, during which the PF motor remains idle; (2) at the end of

the pass, the PF motor is instructed to move 1mm; (3) simultaneously with Step 2, the

Arduino reads and interprets the PF motor’s signal, thus driving the Y-stage. An important

note is that the print head can print in both directions, thus the PF motor will move once

for each pass of the X-stage (meaning, once for a pass and once again for a return pass).

This process repeats until the carriage reaches the end of the Y-stage, which is

approximately 2 full rotations of the pulley (which is 25,600 steps on the motor or

approximately 281mm of travel). The software was coded in Arduino IDE and may be

found in Appendix A.

34
 Chapter Three.

Results & Discussion

3.1 Results

Shortly after the printer had been fully designed and built, and before significant

testing was conducted, an unfortunate issue with cartridge recognition arose. The printer

suddenly refused to communicate with the CISS as well as genuine Epson cartridges,

halting any ability to print. A second WF30 had also been purchased for spare parts in case

an event such as this occurred. However, within just a week of using the spare parts from

the backup printer, the same issue occurred, despite careful disassembly and procedures to

eliminate electrostatic discharge. After discovering of similar issues with other users of the

WF30, the author has hypothesized that this issue may be due to planned obsolescence by

Epson. The author has proposed some solutions in Section 4.2 for future work.

Despite this ill-fated circumstance, several photographs were obtained before the

cartridge failure, indicating some positive results. Figure 3-1 shows a photo of a nozzle

check pattern conducted. The gaps in the lines suggest many of the nozzle were still

clogged after an attempt at using Objet Cleaning Fluid to test jetting higher-viscosity

material. The nozzles and ink reservoirs were in the process of being cleaned with

35
isopropyl alcohol (IPA), which is the reason for the faded prints. Figure 3-1 is the result of

IPA mixed with Objet Cleaning Fluid and some residual black ink.

Figure 3-1. Nozzle check pattern printed

Figure 3-2. Nozzle check pattern, according to Epson [20]

36
Shortly after, in order to continue purging the nozzles, a rectangle was printed (Figure 3-3).

The result appears to be a gradient; however this is due to the IPA diluting the residual

inks.

Figure 3-3. Rectangle printed

These results suggest that the Arduino is interpreting Epson’s signals correctly from

the PF motor. The X-stage and Y-stage are in sync. A video captured by the author shows

the printer reaching the end of the Y-stage, returning back to its initial position at the

completion of a “page”, dropping the Z-stage, and continuing on the next page.

37
3.2 Discussion

Due to the cartridge malfunction, testing of the printer ceased until the issue could be

resolved. The plan was to test the printer using Stratasys materials: VeroWhitePlus RGD835

(Part No. OBJ-04054) photopolymer resin, since it is used in the Objet30 3-D inkjet printer,

available in the University of Arkansas, Dept. of Biomedical Engineering. Therefore, the

purchasing and testing of this material was not followed-through with as it costs USD$300

for two 1kg cartridges, with no guarantee of compatibility. The author agrees that this high

material price seemingly defeats the purpose of a low-cost printer.

Ideally, a custom in-house photopolymer material would be created for this printer,

taking into account jettability and curability with the UV LED system. A dimensionless

number called the Ohnesorge number (Oh) is one way, to some extent, determine if a

material is jettable. The Ohnesorge is related to the Reynolds and Weber numbers [15],

denoted by

RS T
𝑂ℎ = = [26]
,S UVWX

where μ is the dynamic viscosity, ρ is the density, σ is the surface tension, and Lc is the

characteristic length. If a liquid’s Oh is between 0.1 and 1, it is potentially jettable [26].

The author recognizes that simply sourcing a material that works with the printer is

a very important aspect, whether or not it follows the low-cost philosophy. This matter will

have to be left as future work.

38
 Chapter Four.

Conclusion & Future Work

4.1 Conclusion

First and foremost, this project attempted to design and build something that doesn’t

exist, with minimal experience and a low-budget. While the project hit an unfortunate wall

just as the printer was nearing completion, there are several contributions that were made.

This paper has presented a genuinely low-cost method for inkjet 3-D printing that has

promise and potential. The author believes the hardware in desktop inkjet printers is

entirely capable of becoming multi-material manufacturing devices, and that the prices of

printers like the Stratasys Object Connex series could be significantly reduced with the

introduction of a low-cost competitor.

This project also leaves behind hardware that can be utilized in future printer designs.

The Y- and Z-stages were designed with the ability to be easily modified or taken from the

printer for use on new iterations. The simple, modular design also allows for significant

adjustment to be made, especially with regards to build volume.

From a learning perspective, this project has provided the author with a rather

extensive knowledge of the operations and mechanisms of the WF30 that are not readily

39
available on the Internet. The idiosyncrasies of the printer were discovered through many

hours of dissection and tinkering. Additionally, the knowledge acquired about Arduino and

RAMPS has inspired the author to use these devices in future projects. Although there is still

much work to be done and much that could be improved, this paper has laid the foundation

for future development. It is meant to be a comprehensive documentation of this project’s

history and current status for use as a reference for the next student willing to take on the

challenge.

4.2 Future Work

As with any research, there is always more to be done. Firstly, the author suggests

abandoning the WF30 printer, as the hardware is nearly 8 years old and parts are becoming

increasingly difficult and expensive to acquire. For reasons unknown, a used WF30 on the

Internet can be double or more of its original MSRP. In addition, if the hypothesis that

planned obsolescence killed the WF30 is true, then it should most definitely be avoided. A

much newer, inexpensive Epson printer model would be the best choice for hardware

moving forward, as print head technology has presumably changed very little. Something

like the Expression Home XP-420, one of Epson’s budget printers (approximately USD$50

currently) would be a viable option. The main downside might be the many features that

even low-cost printers of today have – such as touchscreens and Wi-Fi. The WF30 benefits

by not having these distractions from its focused intent. They simply add complexity to the

circuitry that will have to be dissected and understood. Nevertheless, printer functions are

likely quite similar, allowing the carriage from the new printer to be mounted on this frame

and similar code to be used to read its signals. The author admits that this project relied

40
heavily on Epson’s functions for driving the print head, X-stage, and software from PC to

printer. If reliance on the Epson main board was eliminated, then theoretically the WF30

hardware should still be adequate. This brings up the second suggestion – taking over the

print head entirely via Arduino. This process is already underway, with Garrett Urban

spearheading this operation. A more complete understanding of the printer electronics and

circuitry is necessary for this. The high voltage driving waveform that the WF30 main board

creates would have to be simulated with Arduino. The ability to alter the waveform, though,

would allow for precise tuning of the PZTs to the jetted material in question. This flexibility

would allow for multi-material printing. As mentioned in Section 3.2, a custom

photopolymer formulated for the print head could help optimize the printing process.

Multiple print heads could even be added to the carriage for increased build speed and

coverage. Complete control of the print head would provide the ultimate freedom and truly

is the future of this project.

41
 Appendix A

Software Code

The main code used for printer operation, provided below:

_________________________________________________________________________________________________

#include <Stepper.h>
#define Y_ENABLE_PIN 56
#define Y_MIN_PIN 14
#define Y_MAX_PIN 15

Stepper X_STEPPER(12800, 54, 55); //Z-stepper motor, using X pins for convenience
Stepper Y_STEPPER(12800, 60, 61); //Y-stepper motor

const int Y_Trigger_Input = 10;//analog input from Epson for Y-stage movement

int stepCount = 0; //initial=0, number of steps the motor has taken

int pageCount = 0; //initial=0, number of pages the printer has completed
const int maxStep = 120; //printer defined, number of steps Y-stage can fit
(absolute max is ~280)
const int layers = 50; //user defined, layers (in z-direction)

void setup() {
Serial.begin(9600); // initialize the serial port:
pinMode(Y_ENABLE_PIN, OUTPUT);

Y_STEPPER.setSpeed(150); //Speed (rpm)

while (pageCount < layers){

Serial.println(analogRead(Y_Trigger_Input)); //prints the voltage read to the
serial monitor
delay(10);

int VoltageRead = analogRead(Y_Trigger_Input);

if (VoltageRead < 200){ //this is on a 0 to 1023 scale, voltage @ Epson idle
=~361
Y_STEPPER.step(91); //should be 91 for regular printing
Serial.print("steps:");
Serial.println(stepCount);
stepCount++;
delay(100);
}

if (stepCount == maxStep){
Y_STEPPER.step(-91*stepCount); //returns the Y-stage to initial position
stepCount = 0; //resets stepCount
pageCount++;
X_STEPPER.step(12800);
Serial.print("Completed Page: ");
Serial.println(pageCount);
delay(100);
}
}

42
 Appendix B

Optocoupler Circuit Design

Figure B-1. Proposed twin optocoupler design for isolating Epson/Arduino circuits

43
 References

1. Gibson, Ian, David Rosen, and Brent Stucker. Additive Manufacturing Technologies
3D Printing, Rapid Prototyping and Direct Digital Manufacturing. New York, NY:
Springer, 2015. Print.

2. Schmitt, Christopher T. Homework 3: Bezier Surface Modeling. PDF. MEEG 491V -

Additive Manufacturing, 2015.

3. C.W. Hull, Apparatus for production of three-dimensional objects by

stereolithography. 11 Mar. 1986: U.S. Patent No. 4,575,330.
4. "ProX™ 200 Dental." 3D Systems. Web. 29 Mar. 2016.
<http://www.3dsystems.com/3d-printers/production/prox-200-dental>.

5. Kellner, Tomas. "World's First Plant to Print Jet Engine Nozzles in Mass Production -
GE Reports." GE Reports. 15 July 2014. Web. 29 Mar. 2016.
<http://www.gereports.com/post/91763815095/worlds-first-plant-to-print-jet-
engine-nozzles-in/>.

6. "Stratasys Redefines Design and Manufacturing with the New Objet500 Connex3
Color Multi-Material 3D Printer." Stratasys Blog. 27 Jan. 2014. Web. 15 Apr. 2016.
<http://blog.stratasys.com/2014/01/27/objet500-connex3-color-multi-material-
3d-printer/>.

7. "Objet500 and Objet350 Connex3." Stratasys. Web. 29 Mar. 2016.

<http://www.stratasys.com/3d-printers/production-series/connex3-systems>.

8. Sevenson, Brittney. "Stratasys Introduces Objet500 Connex1 & Connex2 Triple-

jetting Multi-material 3D Printers." 3DPrint.com. 08 Sept. 2014. Web. 29 Mar. 2016.
<http://3dprint.com/14431/objet500-connex1-connex2/>.

9. "RepRap." RepRap.org. Web. 11 Apr. 2016. <http://reprap.org/>.

10. "MakerBot Replicator Desktop 3D Printer." MakerBot. Web. 11 Apr. 2016.

<http://store.makerbot.com/replicator>.

11. "Fab@Home." Wikipedia. Wikimedia Foundation. Web. 11 Apr. 2016.

<https://en.wikipedia.org/wiki/Fab@Home>.

12. Fab@Home. Web. <www.fabathome.org/>.

13. "Material of the Month: Lead Zirconate Titanate." The Institute of Materials,
Minerals, and Mining. Materials World Magazine, 7 May 2012. Web. 15 Apr. 2016.
<http://www.iom3.org/materials-world-

44
 magazine/feature/2012/may/07/material-month-lead-zirconate-titanate>.

14. Gough, Jeff. "File - Print - Electronics." 27c3. Berlin. Web.

<https://media.ccc.de/v/27c3-4099-en-file_print_electronics#video&t=501>. 2011.

15. Zhou, Wenchao, and David W. Rosen. "Interface Dynamics in Inkjet Deposition."
Diss. 2014. Web.

16. "Why Thermal Printer Printheads Wear Out and How to Extend Printhead Life."
Rotech. 2014. Web. 15 Apr. 2016.
<http://www.rotechmachines.com/support/faq's/lists/buying-guide-list/why-
thermal-printer-printheads-wear-out-and-how-to-extend-printhead-life>.

17. "Epson Stylus 800 - The First Inkjet Printer Equipped with Micro Piezo Technology."
Epson. Web. 29 Mar. 2016.
<http://global.epson.com/company/corporate_history/milestone_products/24_styl
us800.html>.

18. "InkJet." RepRap.org. 7 Aug. 2015. Web. 15 Apr. 2016.

<http://reprap.org/wiki/InkJet>.

19. Kwan, Joyce G. "Design of Electronics for a High-resolution, Multi-material, and

Modular 3D Printer." Master’s Thesis. Massachusetts Institute of Technology, 2013.
Print.

20. Service Manual - EPSON Sylus C110/C120/D120/WorkForce 30/Stylus OFFICE

B30/T30/T33/ME OFFICE 70, Rev. C. Seiko Epson Corp. PDF.

21. Fukano, Takakazu. Ink Jet Printer. Seiko Epson Corporation, assignee. U.S. Patent
6,758,544. July 2004.

22. Seshimo, Tatsuya and Hikonosuke, Uwai. Ink Jet Printer and Printing System Using
the Same. Seiko Epson Corporation, assignee. European Patent 0933215A2. Apr.
1999.

23. Service Manual - EPSON Sylus C110. Seiko Epson Corp. PDF.

24. Ashley, Jonathan, Ted Goll, Joe Moquin, Dominic Rose, Christopher Schmitt, Jonathan
Stewart, and Zachary Zelenka. Additive Manufacturing Undergraduate Final Project -
3D INKJET Printer. University of Arkansas. MEEG 491V. PDF. 2015.

25. "Desktop SLA 3D Printing Tech Specs." Formlabs. Web. 11 Apr. 2016.
<http://formlabs.com/products/3d-printers/tech-specs/>.

26. "Ohnesorge Number." Wikipedia. Wikimedia Foundation. Web. 11 Apr. 2016.

<https://en.wikipedia.org/wiki/Ohnesorge_number>.
45