3D Printed Child Ankle Foot Orthosis (AFO) Design and Testing

1
Derrick,Horner,Le,Walters2016
Project Report
ENSC29: 3D Printed Child Ankle Foot Orthosis (AFO)
Updated 04/20/2016
Introduction:
Prosthetics and orthotics are oftentimes necessary for a variety of patients, but can also be very costly for
both patients and their hospitals. The manufacturing processes for some of these devices require a
significant amount of time and resources,which can be inconvenient for every party involved. Shriners
Hospital for Children has expressed a need to improve the process of designing and constructing child
orthotics, involving materials research and 3D printing design to minimize time, materials, and steps in
the fabricating process.
Objective:
The main goal for this project was to assist Shriners Hospital and its leading orthoptist, Dr. Peter
Springs, with providing high-quality orthoses to children within an optimized time-frame. The main
purpose of our project is to create a basic design of an Ankle Foot Orthosis (AFO) using materials that
meet the strength, flexibility, and durability requirements for patients. Creating a simple, easily printed
AFO with the best composition and geometry to meet strength and comfort requirements is our top
priority. To avoid the delays caused by trial and error, production, and shipping, 3D printed designs will
be developed to reduce testing and development time and to quickly enable use by patients. In addition,
severaltypes of printer filaments will be researched and an optimal material that meets strength and
functionality requirements while minimizing cost will be determined.
Background:
The current process for those in need of an AFO brace is a timely and costly manner due primarily to time
consuming stages of AFO fabrication. A 3D-printed rapid prototyping process for these braces is the next
step for reducing cost and time.
There are currently a couple major projects delivering 3D printed prosthetics to children, however there
are very few for braces,as these bring in less profit for large orthotic and prosthetic manufacturing
companies. The current process for Shriners Hospital to produce an AFO consists of severalinvolved
steps.
First the patient's leg and foot are scanned and the digital 3D model is edited by the orthotist. This model
is then sent to Portland, Oregon where a physical positive mold of the foot is carved out of foam. The
brace itself is then fabricated by vacuum forming a sheet of plastic (usually polypropylene or a copolymer
of such) over the positive mold. After fabrication is completed, it is sent back to Shriners Hospital in
Spokane or any other Shriners location around the country. The patient is then fitted with the brace,and
finishing touches such as trimming and flaring are added.
In total this process can take as many as three weeks,and cost as much as $2000 for the patient.

2
Requirements:
These are a list of this project’s requirements, as instructed by Dr. Springs and the project advisor Patrick
Nowacki:
Customer:
Produce a children's leg brace via 3D printer that has equivalent or better specifications compared to the
current molded process.
Engineering:
1. Produced pieces must have consistently comparable strength to the current manufacturing
process,using Proteus NaturalCo-Polymer Polypropylene.
2. Produced pieces must be printable up to the size of the statistically tallest child that qualifies to
have an orthotic brace made by Shriners hospital, between 18 and 24 inches.
3. The thickness of the brace must be as thin as possible, preferably between 1.5 and 5
millimeters.
4. Lifespan of product must reach a maximum of three years under normal use.
5. Produce a comprehensive visual output of strength characteristics for simple comparison between
3D printed and current brace design.
Marketing:
6. Product must have equivalent design and strength characteristics to the current designs.
7. Product must be as thin and light as possible, for ease of use by the customer.
8. Production process must be simpler than the current process,and preferably of similar or lesser
price.
Constraints:
9. Must involve 3D printing
10. Recommended material of polypropylene
11. Constraints limited to printer available
Budget:
Funding for the AFO research project came from two sources. The largest portion of the funds was
granted from Gonzaga's Engineering Department due to the longevity of this research project. The printer
was purchased to be used by the university for years to come, as research and production will continue
after the preliminary project has finished. Gonzaga's funding equated to around $20,000 and the
remaining portion of funds of an undisclosed amount came from KEEN. This research project received
funding for the 3D printer, the associated printing materials, and any modifications made to the printer.
The project was budgeted to fit within the funding source but also to maximize the printer’s ability and
functionality for its intended purpose. Numerous printers had been considered in order to determine the
best printer for the task at hand.
3DP1000 Large Format 3D Printer with Second Head for Dual Head Print
 ~$23,000
 purchased for school use, funded by engineering departments
Materials
 3kg Polypropylene filament: $50/kg $150

3
 3kg Carbon Fiber PLA: $66/kg $198
 3kg PLA: $27/kg $81
 3kg PETG: $48/kg $144
 2lbs PCTPE Nylon: $38/lb $76
Other
 20 pieces of polypropylene Water-Jet cut $40
Total Material Budget Requested: $689
Liaison Collaboration:
There were severalinstances in which face-to-face collaboration between this group and Dr. Springs was
necessary. Two visits to Shriners Hospital consisted of witnessing how Dr. Springs’ Spectra 3D Scanner,
which is manufactured by Vorum, as well as taking a tour of the space where he modifies his patients’
AFOs by hand. Samples of both copolymer and homopolymer polypropylene to use in comparison to
current materials were also received during these visits. Dr. Springs took severaltrips to Gonzaga
University as well, mainly to discuss updated designs and the overall progress of the project – specifically
on the software side – as well as to discuss the crucial forces an AFO experiences during normal use.
3D Printing Procedure:
From 3D scan to the final print, the process is pretty simple but required constant monitoring at first.
After receiving the 3D scan from Dr. Springs which was created from a cast of the patient’s foot, the scan
is uploaded into MeshMixer, a free AutoDesk 3D modeling software. The reason MeshMixer was used as
opposed to SolidWorks was because 3D scans generate numerous faces on the 3D model to the order of
104
which requires heavy computational power to render, power that exceeds the limits of SolidWorks.
In MeshMixer, the scan is received as a 2D surface that must be extruded to a specified thickness that
creates the 3D model for the printer. The newly-created 3D model is exported to Simplify, a 3D modeling
software used specifically in conjunction with 3D printers such as the 3DP100. The model can be literally
be dropped into a 3D generated print bed displayed in the program and then shifted or moved for ease of
access during the print.
Within the specific settings of Simplify, the nozzle and bed temperatures can be set to best print the
designated material. Other features such as a raft and support material can also be adjusted in order to
support the odd shape of the bottom of the leg brace. A full infill of the model is used so that the leg brace
is a solid plastic part. After finalizing the print settings, the file is converted to a .gcode file that can be
read by the 3D printer.
Once the print is initialized and the nozzle and bed are set to the correct temperature,a small skirt is
printed around the model to help the 3D print technician determine if the print head needs to be minutely
lowered or raised to create the best bead and adhesion to the bed. This concludes the print technician's
involvement as the next step is to wait for the print to be completed (~14-16 hours depending on print
speed).
Settings such as temperatures,raft,support material and skirt within Simplify can be preset after
determining the best settings for the desired material to reduce labor and overall print time.

4
Material Testing:
Specifics:
Our project consisted of testing and/or printing with five different materials: polylactic acid (PLA),
carbon fiber reinforced PLA (CFPLA),3D printed polypropylene (3DP PP),polyethylene terephthalate
glycol (PETG), and plasticized copolyamide thermoplastic elastomer (PCTPE) nylon.
Two different experiments were completed for most of these materials. These tests were a tensile strength
test using an MTS Model 45 Electromechanical Monotonic Load Frame tester,and a bending fatigue test.
For the tensile strength tests,ten sample specimens of each material were printed, each about 110 mm ×
5.5 mm × 4 mm. Ten solid polypropylene samples of the same size were also tested, and were water-jet
cut from a sheet of material that was received from Dr. Springs. Test rates for PLA and PETG were 0.008
in/min. For both solid polypropylene and 3DP PP,the test rate was 0.02 in/min, and for CFPLA,the test
rate was 0.012 in/min.
The machine used to test bending fatigue was able to count the number cycles it applied to a sample of
material until failure. The settings used for these tests were 0.9 witness and a threshold of 90%. This
meant the machine would stop once the sample reached 90% of its original strength.
Data Analysis:
Each tensile test gave us the length that each sample was elongated in millimeters, along with the
corresponding force in Newtons that was applied to the sample. Data was taken every second until the
sample failed, at which point the test was stopped. Using this raw data, the associated stress and strain
values for each data point was found for every test. From here least squares regression analysis was
applied and using Excel’s solver add-in, the estimate of Young’s Modulus for each test was calculated.
Averaging these estimates gave an overall value for modulus for each material. Maximum tensile
strengths were also calculated and averaged for every material tested.
Comparison of Results:
At the beginning of our project, it was stated that our goal was to find a 3D-printer filament with material
properties within ± 5% of the corresponding properties for solid polypropylene. This objective was
changed for three reasons: incomparable data,printability of certain materials, and time constraints.
As the following table shows, none of the properties of our selected materials reached the stated goal of ±
5%. This necessitated a rating system in which each material was scored according to their average values
in five categories: Young's Modulus, tensile strength, fatigue results (number of cycles until failure),
printability and repeatability, and cost per kilogram. Each category had a highest possible score of 5, and
each material was ranked based on the order in which their numbers fell.

5
Table 1: Material ratings based on five critical categories.
It can be seen from Table 1 that PLA achieved the highest overall score,followed by PETG, CFPLA,
nylon, and 3DP PP,respectively. PLA scored at least a 4 in each category,and objectively it was by far
the best material tested. However,this table did not take into consideration severalsubjective categories
that were equally important, and therefore the decision on PLA as the one and only material choice could
not be easily made.
It can also be seen that no fatigue tests were completed for 3DP PP and nylon. While this affected the
final scores of these materials, it did not change which ones were either ruled out or recommended. 3DP
PP was ruled out because of its inability to be reliably printed using our printer, and so its material
properties could be deemed irrelevant. Nylon was ruled out because of extreme flexibility and low
modulus.
Figure 0: Fatigue Testing Comparison
Material Modulus (MPa) Score Tensile Strength (MPa) Score Fatigue Result Score Printability/Repeatability Score Cost (per kg) Score Overall Score
Solid PP 1507.72 --- 20.04 --- infinite --- N/A --- $14.00 --- ---
PLA 3930.26 4 42.66 5 13080 4
No warping, sticks to bed
well, layers stick very well to
each other, fast print speed,
very repeatable
5 (tie) $27.00 5 4.6
CFPLA 6501.44 5 39.52 4 2620 3
well, layers do not stick
together well, medium print
speed, difficult to load, not
repeatable (inconclusive)
2 $66.00 2 3.2
3DP PP 1316.54 2 16.89 1 N/A
Lots of warping, does not
stick to bed, layers stick
together reasonably well,
very slow print speed, not
repeatable
1 $50.00 3 1.75
PETG 2269.70 3 34.14 3 (tie) 400000 5
well, layers stick very well to
each other, fast print speed,
very repeatable
5 (tie) $48.00 4 4
PCTPE Nylon 73.04 1 34.79 3 (tie) N/A
Requires baking before
and/or after printing
3
$83.79
($38.00 per
pound)
1 2

6
Figure 1 shows stress vs. strain curves of each material, plotted together for easy comparison. Each curve
represents a single tensile test done on one sample of each type of filament. It is shown in this graph how
quickly CFPLA failed in relation to its amount of elongation under stress,a sign of how unyielding it
was.
Figure 1: Stress vs. Strain comparison of tested materials.
Both solid polypropylene and 3DP PP elongated much farther than the other materials, at the expense of
having low tensile strengths and modulus numbers. This translated to AFOs made of these materials
possibly having very limited capacities in terms of giving patients the support their legs needed.
It was clear after finishing the tensile tests that PETGwas the best compromise between 3DP PP's
flexibility and CFPLA's stiffness,and it can be seen from Figure 3 that this was the case. However,like
PLA,there were other factors involved that prevented PETGfrom receiving an overall recommendation.
Figure 2 shows the calculated results of Young's Modulus for every test, color coded by type of material.
CFPLA has the highest modulus by a significant margin, followed by PLA and PETG, respectively. 3DP
PP and solid polypropylene, however, are very similar in terms of this particular property. Unfortunately,
for reasons that will be explained further on in this report, 3DP PP was not a viable option for this project.

7
Figure 2: Young’s Modulus comparison of tested materials.
PLA:
While PLA had the highest tensile strength (42.66 MPa),there were concerns about its overall brittleness
and long-term viability. Brittleness is an issue for this application because if the AFO should fail, a brittle
product would more likely break all the way through and become completely useless to the patient.
Brittleness would also lead to breaks that have sharp edges and points, which could lead to patient injury.
However,its impressive tensile strength and modulus numbers minimize the risk of such incidents,
especially when compared to the other materials.
PLA is also known to be susceptible to biodegradation due to moisture adsorption and UV degradation,
which is cause for concern when the goal was to create a product that could sustain three years of normal
use. But PLA's fantastic test results for flexibility and strength, as witnessed by its solid values for both
modulus (3930.26 MPa) and the fatigue test (13,080 cycles),make it a very compelling alternative to the
current polypropylene product.
Carbon Fiber PLA:
It was assumed that CFPLA would be the optimum material choice for this project, considering carbon
fiber’s extreme rigidity and strength. However,when carbon fiber is blended with PLA,it becomes an
exceedingly brittle material with a tensile strength that is actually less than normal PLA. It also scored
poorly when it came to the fatigue tests,leading to a realistic assumption that a full AFO would not last
the required amount of time under normal demands.
Its high average modulus of 6501.44 MPa was 2571.18 MPa greater than the next highest value, which
belonged to PLA. This difference was greater than the modulus value of PETG and showed how rigid the
material was. However,its complete lack of flexibility could possibly lead to patient discomfort, as Dr.
Springs explained that some flexibility is oftentimes desired.
Figure 1 gives a great visual representation of how brittle CFPLA was compared to PLA. It can be seen
that the point of failure for this particular CFPLA sample occurred around 0.011% elongation, compared
to about 0.021% elongation for PLA. The sharper drop off in the curve for CFPLA in contrast to PLA

8
illustrates how the carbon fiber sample failed at a much more rapid pace,and with little warning, a strong
indication of its extreme brittleness.
There were severalprintability issues with CFPLA as well, most notably that the layers would not stick
together as the extruder moved farther from the bed. There were severalhypotheses for this, including
that the extruder was not at a high enough temperature,the baby-step was set too high, and that the
extruder fans should have been turned off during printing. Overall, the printability results were
inconclusive, but the severe brittleness and lack of flexibility of CFPLA lead to it ultimately being ruled
out as an option.
PETG:
With the second highest average score of 4 out of 5, PETGalso came off as a potential candidate for
recommendation. While its greatest trait was the ease with which it could print products reliably (it was
basically identical to PLA in this regard), it also showed great compromise between strength, flexibility,
and lack of brittleness.
While the PLA samples always broke cleanly during the tensile tests,the PETGsamples would break
cleanly only along the outside layers, and then would stretch considerably until it failed completely. This
suggested that if an AFO made of PETGwere to fail under intense stress,the likelihood of it breaking
cleanly all the way through is smaller than PLA,which means the patient might be able to still use it for a
limited time until needing a replacement, and it would reduce the chance of the patient getting injured by
broken sections of the AFO.
Evidence of this behavior can be seen in Figure 3, which shows one test of a PETGsample during the
tensile test. After the initial point of failure at about 0.023% elongation, the curve begins to plateau again,
an indication of the sample not breaking all the way through and elongating further.
Figure 3: Stress vs. strain curve of PETG
However,Dr. Springs was reluctant to recommend PETGfor his own use, as he had had experience with
the material, and explained that it had a tendency to split after long-term use. This was a trait that could
not be tested with the given resources,so it could not be objectively researched. Nevertheless,this
observation by Dr. Springs did prevent PETGto be chosen as the ideal material over PLA.
0
5
10
15
20
25
30
35
0 0.005 0.01 0.015 0.02 0.025 0.03 0.035
Stress(MPa)
Strain (%)
PETG Stress vs. Strain
Stress-Strain Modulus of Elasticity

9
3D Printed Polypropylene:
As stated above, 3DP PP was taken out of consideration not because of its material properties, but
because it was not feasible to produce a working product with the printer used for this project. The reason
3DP PP could not be printed reliably was the fact that its glass transition temperature Tg was well below
room temperature (-10°C to 0°C depending on sources). Each material’s Tg value is the temperature at
which each one will adhere to the printer bed, and our printer’s bed did not have the ability to be set
below room temperature. This caused 3DP PP to curl away from the bed surface during printing which
lead to severe warping of the printed objects. An attempt was made to create a full AFO, but it was unable
to pass more than a few layers before the print was no longer workable.
There are a few companies in the Spokane area that could possibly outsource the printing of
polypropylene products. However,approaching these companies was larger than the scope of this
particular project, and it is unlikely that Shriners would invest the time and resources necessary to pursue
such a venture for its own patients.
PCTPE Nylon:
The modulus and strength results for nylon given in Table 1 were found from the values posted on the
manufacturer’s (Taulman 3D) website for PCTPE nylon, and not actualtest results from this project. The
reason for this was because there was not enough time to print and test samples for both tensile testing
and fatigue testing, and only one AFO could be printed within this limited timeframe. However,using the
data found online, it is clear that nylon’s strength and flexibility properties would not be appropriate for
use with an AFO. It’s modulus of 73.04 MPa was only 5.5% of the material with the next lowest
modulus, 3DP PP,at 1316.54 MPa on average. This meant there would be considerable flex in the AFO
when a given stress was applied in any direction compared to the other materials, giving very limited
amount of support to the patient.
It should be noted that while PCTPE nylon was used for the AFO prototype, further research indicated
that 618 nylon would have been more appropriate to use, given its superior modulus over PCTPE nylon.
The PCTPE nylon that was available to use during the course of the project had been ordered by the
project advisor, and when it was realized that 618 nylon would have been the better material, there was
not enough time left to order it and complete the necessary prints. According to Taulman 3D’s website,
618 nylon has a modulus of 152.99 MPa and a tensile strength of 31.54 MPa,numbers that are still
significantly lower than those of the other tested materials.
Risk/Implementation Issues:
For this project we were mainly focused on the viability of a 3D printed AFO, and not necessarily the
long term use. Therefore,our material choice may not be the most optimal with regards to strength,
weight, durability, and cost. Since the current properties of heat-wrapped polypropylene have been proven
to be sufficient Shriner's patients' needs, any material that can match its strength and durability
characteristics would have worked. However finding a material that can be printed easily that falls into
this category is difficult if not impossible. Based on the data we accumulated, all of our materials were
ruled out either due to inability to print reliably, or due to incompatible strength characteristics. This was
despite adding more materials that were suggested to us later in the process by 3d printing professionals.
We were also worried about the brace failing while a patient is using it, directly harming the patient from
the broken brace or resulting in failure that could lead to a fall, possibly causing a serious health hazard or
fatality, so we opted for a combination of MTS machine tests and bending fatigue tests. Unfortunately

10
this system was not perfect,and results from a long term trial period would be a good idea before
implementing any of our results on an unsuspecting patient. Our initial plan was to test the braces on
ourselves, but due to the medical system's bureaucracy,we were not able to implement this as an option.
While our printer was compatible with a wide range of materials that had a likelihood of satisfying our
requirements, requiring additional parts or modifications to the printer did prove to be an issue. Several
aspects of the printer had to be replaced during the short time we had it and simple modifications were
made to improve our ability to print some of our materials. We limited our modifications to surface
treatments of the bed to improve sticking, but were unable to get a final result with polypropylene due to
the properties of the material while printing. One of the things that unexpectedly took us the longest time
to fix, was the issues we had with software. After spending over a month researching programs it took
our printer instillation technician to help us find the program we needed.
Research:
Preliminary research on the tensile strength of 3D printed PLA was conducted by a student at Gonzaga
University, Luke Blanchart, prior to this project. The research was done using the same MTS machine
and accompanying software as the one used for this project. Supervised by the head of the university’s
mechanical engineering department, Dr. Patrick Ferro, the data was granted to us to by Mr. Blanchart in
order to aid us in our own research.
Additional Internet research was also completed before testing commenced. Research on AFOs and the
reinforcement of such devices using composite materials including PLA and CFPLA was done by Dr.
Kenneth Dalgarno and Dr. Javier Munguia of the School of Mechanical and Systems Engineering at
Newcastle University (Ankle Foot Orthotics Optimization by Means of Composite Reinforcement of Free-
FormStructures,2013). Their research included strength testing of both materials which involved the use
of a 3-point bend test, a test procedure that was not available for this particular project.
Cost Analysis:
One of the largest of areas of focus is the time and cost reduction that is made by 3D printing the AFOs in
comparison to having the model vacuum heat-molded by a separate branch of Shriner's Hospital. This
analysis will only contain the steps beyond the casting of the patient's leg and 3D scan of said cast
because both processes contain these steps.
Time:
The current heat molding process of producing the AFO starts with the 3D scan being sent to a separate
branch of Shiner's Hospital in Portland, Oregon where a foam positive mold is created by a CNC foam
cutter. This foam mold is then used to simulate the patients leg as a heated piece of solid co-polymer
polypropylene is tightly wrapped around the positive mold and the vacuum sealed. The excess material is
cut off to create the hole in the front of the brace for the placement of the leg. This final piece is then sent
back to Shriner's Hospital for final, last minute adjustments. The entire process can take up to 2 weeks
depending on availability of the staff. The setup of the 3D model in Meshmixer and then Simplify with
presets can take a matter of minutes. With the 3D print taking up to 16 hours on a larger model, the 3D
printing process can cut production from a few weeks to a single day.
Cost:
With limited resources from the hospital it is difficult to pinpoint the total savings in dollar amounts by
3D printing the AFO but many assumptions can be made that point to the cost savings from 3D printing.

11
 3D printing only uses the required material unlike the heat molded process that has excess
material wasted after trimming away from leg hole (Each AFO uses about 0.5kg of material).
 3D printing is all in-house and doesn’t require shipping the AFO from branch to branch.
 3D printing is not labor intensive. After setting up the model in software,3D printing only
requires monitoring of the print for the first few layers to check adhesion to the print bed.
 3D printing materials are relatively cheap and will continue to decrease in price with the growing
interest in the technology (current pricing of material listed below).
o Cheap PLA 3D printed prototypes can be made to test fit with the patient before using a
more expensive material for the final product.
Final Deliverables:
Based on what Dr. Springs originally requested, our final deliverable will be the full scale AFOs we
printed. Some more additions we are going to present are our material research,from each of the
materials we tried to make an AFO from, as well as our personal recommendation based on our research
and experience with printing. We are also including a comprehensive video on how to get the scan to a
printable file type, and recommending settings on the printer for each material.
Conclusion:
Based on our results, while we feelwe did prove to Dr. Springs that a 3D printable ankle foot orthosis is
possible to print physically, we have insufficient data to recommend any materials we tested. There is
however significant viability in this project for future years. Outlined in our initial proposal was the fact
that further progress could have been possible were it not for our limited timeframe. Not only will the art
of 3D printing evolve with advancements in materials and technology, but our research will be used in
future years for related projects. Shriners Hospital has already agreed to liaison for more projects and we
have several directions in which those projects could take using our initial results.
Future Projects:
One of the test procedures that this project did not have the opportunity of undertaking was the
application of the 3D printed AFOs in field tests. This prevented elaborate research into the long-term
viability of each material, such as normal wear and tear,the effect of humidity and UV radiation, and
damage caused by walking on rough surfaces. While jumping through all the hoops necessary in order to
allow an actualpatient to test a 3D printed AFO prototype may not be the best use of time, an experiment
that could model such a durability test would be an interesting project for future groups to carry out.
Examining patients’ gait cycles is another area of research that can branch from the findings of this
project. Analyzing the areas of maximum stress on an AFO could lead to innovative geometries such as
structural reinforcements, varying thicknesses, and combinations of materials that are only possible with

12
3D printed parts. Dr. Springs has access to software which can diagnose how a certain patient walks and
can from there, the best geometry for the patient’s own AFO can be determined.
While the outsourcing of printed polypropylene AFOs was not within the scope of this project, it is
possible that future groups could look into such measures. This way, a full scale 3DP PP AFO could be
compared to the solid polypropylene model, and durability and strength tests could be successfully
completed.
As 3D printing technology advances,so too will the potential for more research into this topic. It is
entirely possible that new 3D filament materials will be produced in the next few years,materials that
could combine the best attributes of those that were tested for this project, while having superior
durability characteristics. There could also be nonindustrial 3D printers created in the near future that are
capable of printing with polypropylene, therefore outsourcing the manufacturing process would not be
necessary.

13
ENSC29–ProjectPlan

3D Printed Child Ankle Foot Orthosis (AFO) Design and Testing

Recommended

Recommended

More Related Content

What's hot

What's hot (13)

Similar to 3D Printed Child Ankle Foot Orthosis (AFO) Design and Testing

Similar to 3D Printed Child Ankle Foot Orthosis (AFO) Design and Testing (20)

3D Printed Child Ankle Foot Orthosis (AFO) Design and Testing