Final Technical Report

December 2016 Technical Report
3D Bioprinter Senior Design - LabFab
Ryan Herbst and Alison Below
The George Washington University
Mechanical and Aerospace Engineering Laboratory - MAE 4199
December 19 2016

Table of Contents
I. Alpha Model ………………………………………………………...………………………………………….….……………….. 1-6
A. Testing …………………………………………….……...………………………………………….….……………….. 1-2
B. Improvements ……………………………….…….....……………………………………….….…..…………….. 2-5
1. Hardware …………………………………...………………………………………….….….…………….. 2-5
a) Tubing …………………………………………………………………………………………….…. 2
b) Internal Extrusion ……………………………………………………………………………. 2-3
c) ATX Power Supply Mounting …………………………………………………………. 4
d) Sealed Enclosure ……………………………………………………………………………... 4-5
e) Firmware Updates ………………………………………………………………………………. 5
2. CAD Work …………………………………...………………………………………….….………………….. 5
II. Beta Model ………………..………………..….………………………………………………………………..….…………….. 7-15
A. Design Description ……………………………….….………………………………………….….………………. 8-15
1. Enclosure ……………………………….…...………………………………………….….……………... 8-10
2. Actuation …………………………….……...…………………………………………..…………….. 10-12
3. Extrusion ……………………………….……...………………………………….….…………..…….. 12-13
a) Internal …………………………………….……...…………………....…..….……………….. 12
b) External ……………………………….……...…………………………….……………….. 12-13
4. Z Axis ……………………………………………...……...………………………….….……………….. 13-14
a) Printing Bed ………………………….……...…………………………….….…………… 13-14
b) LED Grid ……………………………….……...…………………………….….……….……….. 14
5. Firmware and Electronics ……………………………….……………………….….………….. 14-15
B. Commercialization Efforts ……………………………….……...……………...………….……..……….. 15-16
III. Related Documents ..……………….……….………………………………….…………………….….………..……….. 16-17
A. Bill of Materials ………………………………………..….……...……………………….…………….……….. 16-17
B. Knowledge Transfer …………………………………….……...…………………………….….……….……….. 17

1
I. Alpha Model
A. Testing
Work with the completed Alpha model began with calibration and extrusion tuning testing. The
first focus was to establish basic functionality using a consistent material. Toothpaste was
chosen. Despite being more viscous than most biomaterials, toothpaste was successful in the
process to refine layer heights, printing speeds and extrusion rates.
It was decided to use toothpaste to write out “LabFab” to film and feature a successful and
recognizable print for the Kickstarter video. A custom Gcode toolpath was implemented to
enable the printing of letters only a single layer thick. The task was easier said than done, but
after dozens of hours of tuning printing parameters, tweaking the custom Gcode, the LabFab
was able to write its name with excellent precision. The extrusion rates required constant
scrutiny during testing because the LabFab would sometimes extrude the viscous toothpaste
after missing deposition locations, or would over-extrude towards the end of the print, resulting
in ugly lines that were much thicker than the starting lines. Adjusting the PID constants within
the firmware to achieve the proper layer height for the tuned extrusion rates required
additional work. Too low, the nozzle would drag on the surface of the bed and too high, the
extruded material would meander or weave instead of laying down a straight line.
Figure 1. Printed “LabFab” with toothpaste
The next test was to use the newly designed microencapsulation printing head, discussed below.
Polyethylene Glycol (PEG) dyed with a fluorescent color was extruded surrounding a chemical
crosslinking agent, creating a tube structure within the PEG. The result from the
microencapsulation printing was a network of lines that hardened into hollow networks upon
crosslinking. The purpose was to print vascular networks that simulate veins. This experiment
Figure 2 (a) Mid-printing with microencapsulation technique; (b) Final print

2
was performed with the assistance of Tao, one of Dr. Zhang’s research masters candidates, who
was seeking to investigate vascular flow with the use of our prints. He was integral both in the
design of the microencapsulation extruder block and the process of identifying and perfecting a
print material for experimentation.
The final test was working with photocrosslinking materials. The LED grid was implemented as a
component in the Alpha to satisfy a design requirement from Dr. Nathan Castro. He believed
exposing a solution of inactive PEG, the correct activating compound, and a photo-initializing
agent to 420 nm visible blue light would enable photocrosslinking. Experimentation began with
using various solution concentrations of PEG and the varying types of active compounds. By
mass, the same concentration of photo-initializer was used. A battery of experiments were
conducted to determine the effectiveness of LabFab’s photocrosslinking capabilities. However,
the LED grid was not nearly as effective as hoped. Further investigation into what is preventing
the desired crosslinking, is required. Potentially, the LEDs could simply be lacking enough in
intensity in the 4207 7nm spectrum, a different photo initializer should be tested, or additional
unknown factor(s) could be in play. It may be possible that the borosilicate glass is refracting or
interfering with the LED light. This experimental process took multiple weeks and participation
of Dr. Nathan Castro internationally, as well as many researchers within Dr. Zhang’s laboratory.
Overall it was learned that multi-material extrusion requires lots of tweaking as speed, layer
height, extrusion height, extrusion rates, number of materials to be extruded and nozzle
diameters were all unique between materials of different viscosities.
B. Improvements
1. Hardware
a) Tubing
It was discovered that the originally selected tubing used for transporting the printing
material from the syringe to the internal extruder block was too stretchy and required
too much head volume to prime the lines (4.8mm inner diameter). Stronger tubing of a
smaller 1.6mm inner diameter was chosen. This allowed for less material waste.
b) Internal Extrusion
Consequently, due to the design alteration of the tubing, the internal extruder block
needed to be redesigned. This is due to the fact that smaller facets needed to be used.
The facets are the hardware components that attach and secure a connection between
the tubing and extruder block channels.
Initially, the new internal extruder was 3D printed with a high-end FDM printer, in the
3D printing laboratory in the SEH B2 level. The results were acceptable, however upon
testing, tiny gaps between layers existed, causing the block to leak. Different efforts
were made to mediate this shortcoming, such as using an acetone chemical treatment
to seal any imperfections. Without success, it was ultimately decided to have the
internal extruder 3D printed by a SLA printer.

3
The result of designing and printing new extrusion blocks were incredible. Gaps no
longer existed between the layers, allowing for the material to travel without issue. The
block additionally aesthetically was more pleasing. The part was completely clear
throughout with no sign of any defects or gap between layers. Figure 3 shows the new
extruder blocks.
Figure 3. New SLA printed extruder blocks, two single channels (left) and one single and
one 2-to-1 channel block (right)
A microencapsulation extrusion block, as discussed in section (a), was also created. This
extrusion type is key in creating prints that simulate vascular networks. The block can
be seen in Figure 4.
Figure 4. Microencapsulation extruder block

4
c) ATX Power Supply Mounting
Using the existing side panel, an opening with mounting holes was made with a laser
cutter. This allowed the power supply to be mounted onto the the body of the printer
and allow external access to the power supply switch
and AC connection. Originally, the power supply unit
was placed on the LabFab floor, unattached and able to
be moved to enable easier testing power and access.
Throughout the tests, the need to have the power
supply unit fixed was nixed, and it was decided to
install it permanently (almost) to the LabFab itself.
The original right hand side panel was removed from
the printer and recut in the laser cutter. Mounting four
mounting holes (seen left) were incorporated to allow
the printer to hover on the inside of the printer. Not
only did this design decision increase amount of space
inside the printer, but it allowed the cooling fan to
exhaust into the atmosphere instead of the enclosure.
The power supply was additionally hidden from user
view by a makeshift shroud.
Figure 5. Power supply mount
d) Sealed Enclosure
As testing progressed, the constant need to have access to the interior of the printer
subsided. Subsequently, all body panels were installed. Essentially, during this time,
LabFab progressed from a testbench to a working prototype. Although sealing each of
the panels was not necessary for the testing and experimentation, it was successful in
closing a design parameter we had not achieved last Spring.
Figure 6 shows the change from Alpha model in Spring 2016 to fully sealed model.
Figure 6. Alpha Model Spring 2016, unsealed (right), to current, sealed (left)

5
e) Firmware Updates
Throughout the various phases of testing, tuning, and experimentation, the LabFab’s
firmware was updated to improve motor control acceleration values, extrusion rates,
the location of the print bed in three space, control of the cooling fan for the RUMBA
board, current values, and more. The Marlin firmware is in a perpetual state of
improvement, and such updates are necessary to improve functionality of the LabFab. In
the future, more current updates of the firmware, such as vRC6 should be in use instead
of older Marlin version. These newer Marlin versions may be buggy, but also may
implement features and updates that the LabFab would benefit from.
2. CAD Work
After various improvements were made, it was necessary to finalize and complete the
SOLIDWORKS renderings of the Alpha. This included incorporating components such as the
LED grid, power supply, power supply shroud and miscellaneous details such as hardware.
Finalizing and updating the CAD was the final step taken in formally finishing the Alpha
LabFab model. Although it can perpetually be improved, its design is complete, both in real
life and in SOLIDWORKS. Final CAD models with all LabFab components are displayed in
Figure 7.
Figure 7: Final Alpha Model in CAD
The final improvements, testing, and experimentation done to the Alpha model of the LabFab mark its
transition from our hands to Dr. Zhang’s laboratory. We have finalized and exceeded all of her original
design parameters, as given to us approximately one year ago by her previous lab manager, Dr. Nathan
Castro. Table 1 shows the achievement chart of all design parameters and their completion level.

6
Design Requirements Completed Spring 2016 Completed Fall 2016 (current)
Qualitative Parameters
1 No deeper than 2 feet Met Met
2 Glass stage with 8”3
build volume Met Met
3 Multimaterial dual extrusion Met Met and exceeded
4 Resolution of at least 200 microns Met and exceeded Met and exceeded
5 Cost less than off-the-shelf
bioprinters ($10,00+)
Met and exceeded Met and exceeded
Quantitative Parameters
1 Overhead gantry travel along y-axis Met Met
2 Y-axis carriage mounted on gantry,
moves about x-axis
Met Met
3 Stage moves about z-axis Met Met
4 Electronic controller implementing
open-source firmware/software
Met Met
5 Extruding motors outside enclosure Met Met
6 Closed frame Work in progress Met
7 Plexiglass viewing windows and front
access door
Work in progress Met and exceeded
8 Gasket silicon seals Work in progress Met
9 Photoinitializer LEDS beneath glass
stage
Met Met
10 Top mounted UV lighting and UV
blocking film
Not achieved Met and exceeded
Table 1. Final Design Parameter Analysis

7
II. Beta Model
The Beta Model was almost a complete rehaul from the Alpha version. Each component except for the
internal extrusion system and the majority of firmware and electronics were redesigned and improved
upon. Table 2 belows details each component of the printer and whether it has changed in cost, size,
weight, or complexity from the Alpha model.
Component Subsystem Changed?
Cost Size Weight Complexity
Enclosure Framing YES Same Smaller Less -
Door YES More Larger Same Much better design
Panels YES Less Smaller Less Slide in and out, easier to
install
Sealant YES More - - Never in Alpha model
Actuation X-Axis YES Less Smaller Less -
Y-Axis YES Less Smaller Less -
Y-Axis passive YES More Smaller Less Previous design was
inadequate and not thought
out
Z-Axis YES Less Taller More -
Extrusion Internal NO - - - -
Extrusion YES Less Less Less Two bay extrusion only
Z Axis Bed YES Less Less Less No need for spacer block
LED Grid YES - - - Taken out of design
Firmware &
Electronics
Controller NO - - - -
Power Supply NO - - - -
UV Lamp YES - - - Taken out of design
Capacitive Sensor YES - - - Resistor ladder to voltage
divider
Firmware NO - - - -
Table 2. Alpha to Beta Design Improvements and Changes
The pricing between Alpha and Beta model of the LabFab improved through the redesign. Detailed in
the supplementary Bill of Materials for both models, the Alpha totals to $4084.00 and the Beta totals to
$2980.26. This price difference accounts for shipping of parts and lack of educational pricing for the Beta
model, and is mainly a difference due to improved design and stronger engineering knowledge and
principles. Testing and experimentation with the Alpha gave light to what was unnecessary, what could
be changed, and what could be improved.

8
In addition, design changes were implemented as to achieve the goal of creating a commercializable
product - one that is easily assembled, lightweight, inexpensive, yet effective. Various design decisions
that were selected for the alpha model were realized during both the build and testing phases, and were
able to be corrected in the Beta design. These reasons will be addressed below in the design description
section.
A. Design Description
1. Enclosure
To reduce weight, cost and complexity, the enclosure was fully redesigned. For the Beta
model the frame was designed again entirely out of 1” extruded aluminum bars, sourced
from vendor 80/20. The enclosure hardware was selected mainly from the 80/20 vendor, as
a vast amount of accessories are available for extruded aluminum rods, lessening the
reliance on custom-made components.
One of the first goals was to find a cheaper and lighter way to enclose the printer. Instead of
using sheets of steel that are screwed on with 72 facets, High Density Polyethylene panels
will slide into place along the slots of the extruded aluminum. To ensure a good seal and
proper fit, gasket strips will be inserted in the gap of the extruded aluminum. One added
benefit to the removable panels, they can be easily replaced for additional windows,
different materials and have custom alterations that would have been impossible with the
stainless steel panels without having access to a laser cutter.
Figure 8 below shows the enclosure of the Beta model from an isometric and top view.
Figure 8. Beta model enclosure
With the new enclosure design, cheaper and simpler to install mounting plates could be
mounted on the exterior of the printer using the externally exposed mounting slots. Using
external hardware avoided the need for custom cuts to be made on the enclosure panels.

9
On the Alpha, the door functions adequately, but had room for improvement. Instead of a
thin door design, the door for the Beta is fully framed with extruded aluminum. Four pieces
of extruded aluminum are miter cut at 45 degree angles and screwed together as a rigid
frame. This prevents the door from bending and guarantees a solid seal when the door is
closed. Similar to the enclosure, the door also borrows the the enclosures sliding in gasketed
panel, and external mounting hardware. This design also accommodates gasketing along the
entire outer surface of the door and frame, for a completely sealed door-jam when closed.
Figure 9. Opened (left) and closed (door) Beta Model door
To address concerns of damage caused to the rumba controller by falling and/or making
contact with a metallic surface, a custom controller mount was designed. Intended to be 3D
printed, it is lightweight, simple and allows the system controller to be securely mounted to
the enclosure. Additionally, the mount also addresses overheating concerns with an
integrated fan mount. Although unlikely, it is possible that the system controller might
overheat.
Figure 10. Mount with RUMBA controller (right), mount without controller (left)
A new shroud was designed to isolate the power supply and system controller from the
interior of the printer. Designed to fully enclose this hardware, the only interior opening is a

10
small gasket for wires to connect to the motors and sensors. User access to the power
supply and system controller is intended to be exclusively from the exterior of the printer.
Just like with the Alpha, the Beta allows the power supply to be secured to the side panel, as
well as the system controller, allowing user access to the USB connection, power switch and
AC connector.
Figure 11. Shroud designed to hide power supply and RUMBA from user view
2. Actuation
To reduce cost, complexity and exterior volume, the custom linear guide rods from PBC
linear were chosen with integrated stepping motors. In the Alpha, an effort to reduce costs
backfired and it resulted in requiring more hardware and larger dimensions than necessary.
In addition to buying separate motors and guide rod systems, couplers and motor mounts
also needed to be purchased. With the integrated systems, each guide rod system is roughly
1.2 inches shorter, with the exact same print volume. As a direct result, the Beta offers the
same functionality in a smaller package.

11
Figure 12. Similar actuation systems implemented in Alpha (left) and Beta (right)
To further reduce complexity and exterior volume, a new passive bearing system was
designed. Testing the Alpha model, it was determined that the passive rail bearing design
proved to be over-engineered, bulky and unnecessary for the relatively slow printing speeds
with multi-material extrusion. For the Beta, a simple roller-on-a-rail system offers the same
functionality with a smaller form factor.
Figure 13. New passive rail for y-axis Beta model
On the Alpha, solid aluminum parts were milled out of aluminum to couple the x and y axis
systems into a single XY structure. The aluminum proved to be difficult to manufacture and
it was determined that cheaper and lighter parts with the same functionality, could be 3D
printed.

12
Figure 14. New 3D printed spacer blocks in Beta model
3. Extrusion
a) Internal
The internal extrusion blocks designed and implemented as of Spring 2016 proved faulty
over the course of testing. The delrin block manufactured and installed originally was
inadequate when the inner tubing of the printer was redesigned to be smaller.
Connections could not be made between the channels and the tubing exit. For this
reason and to implement various experiments of Dr. Zhang’s researchers, new internal
extruder blocks were professional Stereolithography (SLA) printed, with a contracted
service. Such is discussed further and pictured in Section IB1(f). Due to these vast
improvements, no further changes were implemented for the Beta model.
Figure 15. New internal extruder blocks for Alpha Model to be repurposed in Beta Model
b) External
The External Extrusion system for the Alpha was a resounding success. However in
practice, the 3D printed system was larger and more bulky than necessary.

13
Additionally, testing determined that although useful to have a dedicated single and
double carriage for extrusion, the single bay and corresponding stepping motor was
entirely redundant. A new external extrusion system was designed for the Beta with
only a double bay carriage. By shaving off unnecessary volume and component
positioning the Beta external extruder is approximately 40% the volume of the original
with 100% the functionality of the Alpha external extruder. Figure 16 shows the
difference between the original external extrusion system and the new external
extrusion system.
Figure 16. Alpha model (left) and Beta model (right) external extrusion system
Figure 17 shows additional angles of the extrusion system.
Figure 17. External extrusion system redesign
4. Z Axis
a) Printing Bed
Late in the design process of the Alpha, a 1” spacer plate was required to allow the
printing head to have full range of access to the build plate. For the Beta, the bed was

14
simply extended. Such can be seen in Figure 18, showing the difference between the
original bed and the new bed design for the Beta model.
Figure 18. Alpha (left) versus Beta (right) bed design from bird’s eye view excluding
enclosure. Note aluminum spacer block on Alpha model right hand side.
During testing, it was found that having additional overhead room inside the printer
enclosure would improve accessibility. To accommodate, the Beta has an extra 1” of
overhead space.
b) LED Grid
The LED Grid system was not implemented in the Beta model of the LabFab. During
testing, it was determined the biomaterials used were mainly chemically crosslinked.
Moreover, the photocrosslinking capabilities of the LED grid were seemingly
unsuccessful in various experiments. For this reason, and due to immense amount of
time spent individually soldering the LED sectionals together, the LED grid was not
implemented. If a Beta program partner identifies a need for photocrosslinking by
means of this grid, the component will be redesigned in terms of complexity and time
spent manufacturing it. It is likely that future versions will abandon the underneath-the-
entire-bed approach for a more traditional overhead photocrosslinker, however a
design with area effect should be considered.
5. Firmware and Electronics
The firmware and electronics from Alpha to Beta model were largely unchanged. The
firmware, power supply unit, and RUMBA controller board did not change. The UV lamp
fixture as a whole was taken out of the Beta model, to reduce costs and complexity. If the
customer/Beta model program partner did not require the use of the lamp for sterilization

15
or curing purposes, its installation would only lead to unnecessary safety precautions. The
capacitive sensor would only be changed so that the current resistor ladder that lowers
output voltage from its signal line to end stop switch input on the RUMBA board would be
implemented via a 12V to 5V voltage divider. These two changes were the only ones made
in the firmware and electronics component sections for the Beta LabFab model.
The final comparison between Alpha and Beta models is pictured in Figure 19.
Figure 19. Alpha (left) vs. Beta (right)
B. Commercialization Efforts
Commercialization efforts for the LabFab began in Spring 2016, upon winning the GWU School
of Engineering and Applied Science Pelton Award for Outstanding Senior Project, in which the
LabFab was deemed patentable and commercially viable by a panel of academic and
professional judges. Nathan Castro, our Senior Design mentor in Spring 2016, as well as his
business partner Benjamin Holmes, and ourselves decided to pursue transitioning the Alpha
LabFab into a commercializable product. Thus, the efforts for the Beta model were born.
Various meetings occurred over the course of this semester to begin the commercialization
process. The team met with contacts from both Ben and Nathan to discuss the possibility of a
Kickstarter campaign, to work on Beta redesign, to develop industry and educational partners,
and to develop a business plan. The Kickstarter was planned to be a means to gain capital to
implement a Beta program, where the team would build and deploy a small batch of Beta
models to various research and educational institutions. Their feedback/suggestions and use of
the LabFab would identify if the product was satisfying a need and could become a viable
product, or if it was not necessary and would not be successful. At the beginning of the
semester the plan for commercialization was as follows in Table 3:

16
Date Goal
Early Fall Write script, edit, and film Kickstarter video
Mid Fall Redesign beta model. Create all final CAD drawing and establish a price
point, including shipping costs. Consider assembly processes and
modularity in terms of customer needs.
December Have final video editing complete. Create Kickstarter page content and
identify “kick-backs,” tiered prizes and services that contributors receive
after donating
Mid December Reach out to existing 3D printing websites, magazines, and communities.
Advertise to and identify these as a means to show off and extend
awareness of campaign. Similarly, establish Beta model candidates from
key partners: use Dr. Zhang’s conference contacts and Ben/Nathan
knowledge of industry to find these participants.
January Launch campaign for a 60 day period
March Evaluate success or failure. If success, pursue building Beta models as
quickly as possible and deliver to intended partners. If failure, reach out
to Dr. Zhang to attempt to obtain a provisional patent through GWU for
external and internal extrusion systems (the main novel components of
the printer)
Table 3. Commercialization Plan
As of December 13, the plan is close to planned course. The team is in the process of identifying
key advertisers and Beta program partners, and intend to begin the Kickstarter campaign mid-
January to February.
III. Related Documents
A. Bill of Materials
Both the Alpha and Beta Bill of Materials have been included with this document. Such was
done as a formal cost comparison was required as part of the commercialization process, and
for the intents of this course. Finalizing the Alpha BOM meant including all small components
not previously reported on in the May 2016 report, including some hardware, 3D-printed parts,
and additional features and changes made over the summer and into this semester.
The efforts to create and formalize a BOM for the Beta version was done to give a base price
point for the Kickstarter. This was necessary in setting a goal for the Kickstarter. If the goal for
the Kickstarter is implement a 5 printer Beta testing program, and each Beta printer cost a total
of $5,000 to source parts, build, ship, and install, the Kickstarter would need to ask for at least
$30,000 to get the program up and running, neglecting all labor costs. Getting a final quote for
the Beta version would also allow for ease in computing how much an entire printer could be
built for, sold for, and shipped for, if the Kickstarter was to get funded and the product
commercialized. The Beta version has various tabs for each different part vendor and their
proposed quotes.

17
Both Bill of Materials are included as Microsoft Office Excel Documents as a supplement to this
report.
B. Knowledge Transfer
In order to prepare Professor Zhang’s laboratory for our graduation, it was necessary to leave
her and her laboratory researchers with the necessary information required to operate the
LabFab Alpha model. The Knowledge Transfer folder provided to her included the Alpha Bill of
Materials, firmware, all SOLIDWORKS CAD drawings, and a Operations Manual. The Operations
Manual is a document containing basic design explanation, troubleshooting, safety measures,
tips/tricks for use, and information on how to open and view the attached documents. The
Knowledge Transfer PDF document is provided as a supplement to this report.

Final Technical Report

More Related Content

Similar to Final Technical Report (20)

Recently uploaded (20)

Final Technical Report