Prosthetic sockets

fabrication using rapid

prototyping technology
P. Ng
P.S.V. Lee and
J.C.H. Goh
The authors
P. Ng, P.S.V. Lee and J.C.H. Goh are at the Department of
Orthopaedic Surgery, National University of Singapore
Keywords
Rapid prototyping, Computer-aided manufacturing
Abstract
The traditional way of making a prosthetic socket is by
draping a heated thermoplastic sheet over the positive
mould, or by applying layers of woven materials together
with acrylic resins over the positive mould. This process is
extremely labour intensive, and it usually takes two to
three days to make one socket. This paper presents the
development of a prosthetics Computer-Aided-
Manufacturing (CAM) system that utilises Rapid
Prototyping (RP) technology. The system reduces the
socket making time from days to less than 4 h. Clinical and
biomechanical studies are conducted to evaluate the
comfort and t of the new socket during gait. Preliminary
investigation of the new socket shows that its functional
characteristics are very similar to that of a traditional
socket
Electronic access
The research register for this journal is available at
http://www/emeraldinsight.com./research_registers
The current issue and full text archive of this journal is
available at
http://www/emeraldinsight.com/1355-2546.htm
Introduction
Rapid prototyping (RP) technology has
changed the way products are being designed
and manufactured. Being rst introduced in
1987, the technology has made a quantum
leap in all aspects over the past 10 years. The
building speed has been drastically increased,
parts dimensional accuracy has been
signicantly improved, and a wide variety of
new building materials have also been
identied. With these improvements, many
new uses for RP have been discovered.
Medical application is perhaps one of the
most mentioned new uses of RP in recent
years. Applications of RP in the medical eld
cover a variety of areas, like pre-operative
planning, surgical rehearsal, making of
prosthetic and orthotic devices, etc. While
some of these applications are still in their
infancies, others have already shown
promising results, as in the making of RP
prosthetic sockets.
Prosthetic socket fabrication and rapid
prototyping
The traditional way of creating a prosthetic
socket is a labour intensive process. Three
stages can be observed, namely,
measurement, rectication and fabrication.
The physical measurements of the amputees
stump are recorded in detail and a plaster
wrap cast is taken (Figure 1a). A positive
mould is subsequently created by lling the
wrap cast with plaster of Paris (Figure 1b).
Rectications to the plaster mould begin by
taking measurements of the mould, and
comparing them with those previously taken
on the stump. Based on information from past
experience, the mould is further altered by
removing or adding plaster to it. The aim is to
create a comfortable weight-bearing socket
that enables the soft tissue of the stump to be
compressed at pressure tolerant areas, and
relieved at pressure intolerant areas. Once
rectication is completed, draping clear
plastics over the positive plaster mould creates
a check socket (Figure 1c). Preliminary tting
is carried out using the clear check socket,
Rapid Prototyping Journal
Volume 8 Number 1 2002 pp. 5359
q MCB UP Limited ISSN 1355-2546
DOI 10.1108/13552540210413310
The authors wish to acknowledge the help provided
by the staff of the Prosthetic and Orthotic
Laboratory and the Gait Analysis Laboratory,
National University of Singapore.
53
where minor shape adjustment can be
achieved by reheating the socket (Figure 1d).
Once a good t has been assured, a denitive
socket is fabricated. This is accomplished by
re-creating a positive plaster mould from the
check socket. Depending on the nal material
of the socket, it can be vacuum formed by
draping heated thermoplastics over the
positive mould or by applying layers of woven
materials together with acrylic resins over the
positive mould. Upon completion of the
socket, the other components of the articial
limb are incorporated.
Different rapid prototyping technologies
have been investigated recently for prosthetic
socket fabrication. The technology has shown
enormous potential in eliminating the tedious
and time-consuming steps involved in
traditional socket fabrication. In a cost benet
analysis conducted by Freeman and
Wontorcik, 1998, two trans-tibial check
sockets were fabricated using
Stereolithography (SLA-250/40). The
traditional casting process was eliminated; the
sockets were fabricated directly from digitised
data. The time taken to fabricate the two
check sockets were 58 and 26h. Though these
gures can be reduced with faster systems like
the SLA-350/10 or SLA500/40, the real
concern with Stereolithography is the hefty
capital and manufacturing cost. Freeman et al.
also highlighted that the photopolymer
resins used in Stereolithography might be
strong enough for making check sockets, but
would be too brittle and weak for denitive
sockets.
To investigate the possibility of making
denitive sockets using RP technology, a
feasibility study was conducted by Tan et al.
(1998). The Fused Deposition Modelling
(FDM) process was chosen for the study
because of its minimum post-processing
requirement, and the superior mechanical
properties of the building material (P301
polyamide). The positive mould of a trans-
tibial socket was created and scanned using
the Digibot 3D laser digitising system
(Digibotics Inc, USA). The acquired shape is
represented by a mass of spatial points,
known as the points-cloud. These points were
reconstructed in 3D and a solid model was
generated using a CAD software. The solid
model was then converted to STL model/
format for rapid prototyping.
The physical and fabrication data of the
FDM socket is summarised in Table I. The
wall thickness of the socket varied from 3 mm
at the distal end to 7mmat the proximal brim,
to ensure rigidity at the proximal end for
supracondylar suspension. Gait analysis
showed that there were insignicant
deviations of gait parameters when
comparing the FDM socket to the
traditional socket that was created based on
the same positive mould. The investigation
demonstrated that FDM could be used to
produce prosthetic socket with functional
characteristics similar to that of a traditional
one. However, Tan et al. (1998) remarked
that the major drawback with the FDM
method was its long building time and high
manufacturing cost.
The ability to create freeform surfaces and
hidden features makes rapid prototyping an
ideal technology for automated socket
fabrication. However, to make SLA, FDM, or
any other commercial RP systems a practical
Figure 1 Measurement and rectication stage
Table I Fabrication data of the FDM socket
Height (mm) 228
Wall thickness (mm) 37
Weight (g) 332
Fabrication time (h) 29
a
a
Note: Using 3D Modelerw
Source: (Tan et al., 1998)
Prosthetic sockets fabrication using rapid prototyping technology
P. Ng, P.S.V. Lee and J.C.H. Goh
Rapid Prototyping Journal
Volume 8 Number 1 2002 5359
54
means of fabricating denitive sockets, the
manufacturing time and cost must be
signicantly reduced. As this may be difcult
to accomplish at the present moment, a
customised RP system specically designed
for socket fabrication is a pertinent
alternative.
Overview of rapid manufacturing
machine (RMM)
A customised system named the Rapid
Manufacturing Machine (RMM) was
developed for automated socket fabrication.
The requirements of the RMM were stated
below to ensure quick and economical
production of denitive sockets:
.
Most of the building materials used by
current RP systems are expensive and
may not be suitable for denitive sockets.
Therefore, the building material for
RMM must be inexpensive and must
meet the strength requirement of
denitive sockets.
.
Commercial RP systems usually have
precision and accuracy that is beyond the
requirement for socket fabrication; a
diametrical accuracy of ^ 1.0mm is
normally sufcient for prosthetic socket.
Therefore, the hardware and software
design of the RMM should make use of
the lower accuracy requirement to reduce
system cost and building time.
The working principle of the RMM is similar
to that of FDM, which dispenses semi-molten
material onto the machine table, layer by
layer, to form 3D objects. Polypropylene (PP)
is selected to be the building material for the
RMM because of its rigidity, strength and
resistance to fatigue; and most importantly, it
is inexpensive and has been used by
prosthetists for making above-knee and
below-knee prosthetic sockets.
There are two main components in the
RMM, namely the robotic system and the
dispenser. The schematic of the RMM is
shown in Figure 2.
The schematic of the dispenser is shown in
Figure 3. Polypropylene lament (4 mm) is
fed into the heating barrel. The lament
moves along the barrel, and is melted near the
exit, which is where the heating element is
mounted. The in-coming lament acts as a
piston to push the molten polypropylene out
of the nozzle.
The molten polypropylene forms a
continuous strand as it leaves the nozzle. The
strand is dispensed onto the machine table
according to the cross-sectional contour of the
stump. The second layer is laid in the similar
manner on top of the rst. The process
continues until the whole socket is built (see
Figure 4).
RMM design considerations
File format
Stereolithography Interface Specication, or
STL, is the most widely used format in rapid
prototyping (Georges and Chuck, 1996;
Chua and Leong, 1996). STL model is
formed by tessellation of the original model. It
is excellent for describing model with planar
surfaces, but when one turns to medical
applications, the limitations of STL models
become obvious (Dolenc and Makela, 1996).
The highly complex shape of a prosthetic
socket usually results in large le size when
tessellated. This in turn results in long data
verication and processing time, as in the
making of the FDM socket.
Figure 2 Schematic of RMM
Figure 3 Schematic of dispenser
Prosthetic sockets fabrication using rapid prototyping technology
P. Ng, P.S.V. Lee and J.C.H. Goh
Rapid Prototyping Journal
Volume 8 Number 1 2002 5359
55
As the STL format is apparently not a good
format for representing prosthetic sockets, the
RMM has discarded the format. At this
moment, the system accepts the following
data formats:
.
IGES (by Capod Systems/Digibotic).
.
Points cloud (by Digibotic).
Additional data lters can be developed if
required.
Nozzle
The nozzle used by the 3D Modeler FDM
machine (Stratasys Inc, USA) has a
relatively small diameter. Such conguration
enables the machine to build small and
delicate parts with good dimensional
accuracy. However, small nozzle diameter
has its trade-off in terms of the building speed.
This is particularly apparent when building
parts that are wider than the process road
width.
The road width of the FDM process ranges
from 0.254 to 2.54 mm, depending on the
diameter of the nozzle, the input speed of the
material and the extrusion head pressure. In
the case of making prosthetic sockets, the
required socket wall thickness is usually
thicker than the maximum road width. As it is
impossible to achieve the required wall
thickness in one pass, additional steps are
needed to form a thicker wall, as illustrated
in Figure 5.
However, if the road width is designed to be
the same as the required socket wall thickness,
fabrication time can be signicantly reduced.
To achieve this, the nozzle diameter has to be
increased accordingly. This modication is
justiable for this application as building
speed is the priority, and extreme precision in
the xy plane is normally not required. The
nozzle diameter used in the RMM is 3mm,
with a resultant road width or socket wall
thickness of 4 mm.
Figure 4 Various stages in the manufacturing of a trans-tibial prosthetic socket
Figure 5 Building a thick wall with narrow road width
Prosthetic sockets fabrication using rapid prototyping technology
P. Ng, P.S.V. Lee and J.C.H. Goh
Rapid Prototyping Journal
Volume 8 Number 1 2002 5359
56
Awider road width also increases the process
ability to build slanted surface without
external supports, as illustrated in Figure 6 at
the sockets patellar tendon-bearing (PTB)
region. Generating and building external
supports can be extremely tedious and time
consuming.
With a wider road width, the contact area
between layers is also increased. This gives a
better inter-layer bonding, which is an
important factor when socket strength is
concerned. The inter-layer bonding is further
enhanced with the nozzle pressing the current
layer against the previous, as shown in
Figure 7.
System accuracy
Besides manufacturing speed, socket accuracy
is also an important issue. An experiment was
conducted to determine the accuracy of the
system. Based on the normal working
diameter range, three cylinders of internal
diameters 60, 80 and 100 mm were
fabricated. The internal surfaces of these
cylinders were digitised by a 3D probe.
Figure 8 shows the surface prole of the
80 mm cylinder. The centre and diameter of
the circle will be determined by circle
tting.
There are many ways to t a circle, the
simplest way is perhaps the least square
method. An initial guess is made to
determine the centre (X
0
,Y
0
) and diameter
(D
0
). A range will be specied, say 2d, for
iteration. The rst iteration can be obtained
by offsetting the centre and diameter by d.
Therefore, the centre and diameter of the rst
circle is given by (X
0
2d,Y
0
2d) and (D
0
2d)
respectively. The root-mean-square (RMS)
error can be obtained by comparing the circle
with the actual prole. Three nested loops are
required to complete the iteration. The last
circle will be at centre (X
0
+d,Y
0
+d), with
diameter (D
0
+d ). The circle with the
minimum RMS error during iteration gives
the best t.
This algorithm may not give the most
accurate t, and is denitely not the most
efcient. However, the algorithm is easy to
implement, and the t is considered
reasonably accurate when the targeted system
accuracy is ^1.0 mmdiametrically. The circle
tting results of the three cylinders are given
in Table II.
Discussion and conclusion
To compare the RMM process with the FDM
process, the positive mould of the trans-tibial
socket that was used by Tan et al. (1998) in
their study was re-scanned using the Capod
AK/BK scanner (Capod Systems, Sweden).
Figure 6 Building slanted surface without external
support
Figure 7 Nozzle pressing current layer against the
previous layer
Figure 8 Circle tting results
Prosthetic sockets fabrication using rapid prototyping technology
P. Ng, P.S.V. Lee and J.C.H. Goh
Rapid Prototyping Journal
Volume 8 Number 1 2002 5359
57
The scanned data was exported to the
RMM for fabrication. A comparison of the
physical and fabrication data of the RMM and
FDM socket is summarised in Table III.
The RMM process proved to be more
superior in all respects, except that the
RMM socket was slightly heavier than the
FDM socket.
Due to better inter-layer bonding, the
RMM socket does not require any
increment of wall thickness at the proximal
end for rigidity. This simplies the tool
paths generation, and in turn, helps
reduce the data processing and manufacturing
time.
Clinical and biomechanical studies were
also conducted to evaluate the comfort and t
of the RMM socket during gait. A prosthetic
socket for a volunteer trans-tibial amputee
subject was manufactured using the RMM
(see Figure 9). The time taken to build the
socket was 3.5 h, and the complete prosthesis
weighed approximately 1.6 kg, which was
about 0.3 kg heavier than the traditional
prosthesis.
The comparison of the average temporal-
distance data of the stump wearing the RMM
and traditional socket is summarised in Table
IV. Preliminary investigation of the RMM
socket showed that its functional
characteristics were very similar to that of a
traditional socket.
Current/future work
Prosthetic sockets are commonly reported to
fail as a result of local bending and buckling at
the distal region, where the pylon is connected
(Wevers and Durance, 1987). Therefore, to
ensure that the RMM socket is safe for
patients usage, the prosthesis is currently
undergoing the ISO 10328 principal structure
tests. More clinical and biomechanical studies
will also be conducted in the future, to
evaluate the comfort and t of the RMM
socket during gait.
Table II Accuracy of the RMM
60 mm 80 mm 100 mm
Diameter 59.723 80.036 99.757
Centre (114.741,73.674) (111.321,71.879) (113.453,89.570)
RMS error 0.205 0.161 0.257
Deviation 2 0.277 0.036 2 0.243
Table III Fabrication data of the RMM and FDM socket
RMM FDM
Height (mm) 230 228
Wall thickness (mm) 4 37
Weight (g) 426 332
Fabrication time (h) 31/4 29
Table IV Biomechanical evaluation of the RMM socket
Trad RMM
Cadence (steps/min) 91 86
Walking speed (m/s) 0.85 0.81
Stride time (s) 1.32 1.39
Step time (s) 0.64 0.68
Single support (s) 0.34 0.38
Double support (s) 0.54 0.54
Stride length (m) 1.13 1.11
Step length (m) 0.61 0.59
Stance (percent) 66.67 65.83
Figure 9 A trans-tibial amputee with the RMM socket
Prosthetic sockets fabrication using rapid prototyping technology
P. Ng, P.S.V. Lee and J.C.H. Goh
Rapid Prototyping Journal
Volume 8 Number 1 2002 5359
58
References
Chua, C.K. and Leong, K.F. (1996), Rapid Prototyping:
Principles and Applications in Manufacturing, John
Wiley & Sons, Inc, Singapore.
Dolenc, A. and Ma kela , I. (1996), Rapid prototyping from
a computer scientists point-of-view, Rapid
Prototyping Journal, Vol. 2 No. 2, pp. 18-25.
Freeman, D. and Wontorcik, L. (1998), Stereolithography
and prosthetic test socket manufacture: a cost/
benet analysis, Journal of Prosthetics and
Orthotics, Vol. 10 No. 1, pp. 17-20.
Georges, M.F. and Chuck, K. (1996), Accuracy issues in
CAD to RP translations, Rapid Prototyping Journal,
Vol. 2 No. 2, pp. 4-17.
Tan, K.C., Lee, V.S., Tam, K.F. and Lye, S.L. (1998),
Automation of prosthetic socket design and
fabrication using CAD/CAM and RP technique, in,
Proceedings of the 1st National Symposium on
Prosthetics and Orthotics, pp. 19-22.
Wevers, H.W. and Durance, J.P. (1987), Dynamic testing
of below-knee prosthesis: assembly and
components, Prosthetics and Orthotics
International, Vol. 11, pp. 117-23.
Prosthetic sockets fabrication using rapid prototyping technology
P. Ng, P.S.V. Lee and J.C.H. Goh
Rapid Prototyping Journal
Volume 8 Number 1 2002 5359
59