29

Determination of the ability of an FDM type additive manufacturing process using

the Cp index

Determinación de la habilidad de un proceso de manufactura aditiva del tipo FDM

mediante el índice Cp
Baltazar-Lopez, Servando*a

a
Tecnológico Nacional de México - Tecnológico de Estudios Superiores de Valle de Bravo • LWH-9758-2024 •
0009-0005-7449-8803

CONAHCYT classification DOI: https://doi.org/10.35429/H.2024.16.29.38

Area: Engineering
Field: Engineering
Discipline: Industrial engineering
Sub-discipline: Quality control

Key Handbooks

The main contributions of this research are focused on the knowledge of the scope of additive manufacturing as a reliable
manufacturing process of components and industrial products with high quality standards. The basis of this focuses on the
relationship between the dispersion of a manufacturing process and the tolerance of the specifications given in the design
of a part. The key aspects to understand are the effect that even in disruptive processes, such as additive manufacturing,
the use of the right tool for manufacturing has on the statistical stability of a manufacturing process. The main conclusions
of this investigation are that the ability of the machine under analysis to produce parts within specification is poor for a
fine tolerance of 10 tenths of a millimeter above and below the nominal value. In addition to the above, the mean shows a
bias towards the lower limit of the specification, so a way to focus the process independently of reducing the scatter should
be sought. It should be noted that the thickness of the filament used in the adjustment to make the parts was 0.254 mm, a
fraction of an inch. If it is considered that the layers are superimposed one after the other without overlapping, to generate
a 10 mm high cube, the machine would have to deposit an inaccurate amount of layers, therefore, it is inferred that designing
a part in fractions of an inch could increase the process skill rates. So far the author does not have any grants, nor has he
generated any citations in the last year. The author of this research work is assigned to the Tecnológico Nacional de México,
campus Valle de Bravo, being this a public Institution of Higher Education of the Government of the State of Mexico.

Citation: Baltazar-Lopez, Servando. Determination of the ability of an FDM type additive manufacturing process using the
Cp index. 29-38. ECORFAN.

* [email protected]

Handbook shelf URL: https://www.ecorfan.org/handbooks.php

ISBN 978-607-8948-54-3/©2009 The Authors. Published by ECORFAN-Mexico, S.C. for

its Holding Mexico on behalf of Handbook HAIESMA. This is an open access chapter
under the CC BY-NC-ND license [http://creativecommons.org/licenses/by-nc-nd/4.0/]

Peer Review under the responsibility of the Scientific Committee MARVID®- in

contribution to the scientific, technological and innovation Peer Review Process by training
Human Resources for the continuity in the Critical Analysis of International Research.
 30

Abstract

The present project analyzes the Cp index of machine ability for an Additive Manufacturing process by
measuring a sample of 30 parts manufactured by the FDM method. The analysis of the parts is carried
out using Minitab® software for two tolerance classes, fine and medium, obtaining an insufficient result
for both. This, however, may be due to an adjustment parameter in units of the English system, used in
parts of the international system.

Determination of the ability of an FDM type additive manufacturing process using

the Cp index

Objectives Methodology Contribution

Determine the Cp index of A sample of 30 pieces was The analysis showed that the
an additive manufacturing measured and statistical process is not capable of
process. analysis was performed. producing with consistency
within a fine tolerance and
adjustments to the process are
required.

Dimensional tolerances, quality control, standard deviation

Resumen

El presente proyecto analiza el índice Cp de habilidad de máquina para un proceso de Manufactura

Aditiva mediante la medició n de una muestra de 30 piezas fabricadas por el método FDM. El análisis
de las piezas se lleva a cabo mediante el uso del software Minitab® para dos clases de tolerancia, fina y
media, obteniendo un resultado insuficiente para ambas. Esto, sin embargo, puede deberse a un parámetro
de ajuste en unidades del sistema inglés, usado en piezas del sistema internacional.

Determinación de la habilidad de un proceso de manufactura aditiva del tipo FDM mediante el

índice Cp

Objetivos Metodología Contribución

Determinar el índice Se midió una muestra de El análisis mostró que el

Cp de un proceso de 30 piezas y se llevó a cabo proceso no es hábil para
el análisis estadístico producir con consistencia
manufactura aditiva. dentro de una tolerancia fina,
por lo que se requiere realizar
ajustes al mismo.

Tolerancias dimensionales, control de calidad, desviación estándar

 31

Introduction

Additive manufacturing, also known as 3D printing, is a manufacturing process by adding layers one on
top of the other with a small deviation in position from the previous one to, after depositing a certain
number of layers, form a volume from a CAD file. Among the advantages is that designers do not have
to worry about how the parts will be manufactured, only about structure and function. Customisation, the
MA does not require moulds, so different products can be produced in one batch. In addition, MA has
the property of small batches (Huang, Wang, & He, 2015). A recent application example is that in 2016
General Electric reportedly employed additive manufacturing to manufacture 25 thousand LEAP motor
nozzles (Cotteleer, 2014). By 2018, additive manufacturing was used 50% for prototyping and 50% for
final products, forcing quality control in processes to become increasingly important. Some processes,
for example, have on-site sensors and cameras to monitor and analyse quality in real time (Kim, Lin, &
Tseng, 2018).

According to their characteristics, manufacturing processes can be classified into five categories,
subtractive, additive, joining, splitting and transformation (Abdulhameed, Abdulrahman, Ameen, &
Hammad Mian, 2019). These are classified in DIN ISO 8580 (ISO, 2020) as follows:

1. Preformed
2. Formed
3. Cutting
4. Assembly
5. Coatings
6. Modification of material properties

Although this classification refers to conventional processes, additive manufacturing processes

can be classified as preforming processes, which consist of producing a rigid body from a shapeless
material in granules or powder form. MA processes fall into this classification, except that in some cases
the material is not in granules or powder, but in filament form. There is also an international standard
that provides specific terms for AM processes, ASTM F2792-12a (ASTM, 2013), which provides a
structure for grouping current and future AM processes, which are:

1. Material extrusion
2. Binder injection
3. Powder bed fusion
4. Vat polymerisation
5. Direct energy deposition
6. Material injection
7. Sheet rolling

This standard allows for much greater clarity when referring to the different AM processes for
educational or commercial purposes. Within this same standard, direct reference is made to the process
to be used in this work, FDM®, fused deposition modeling, which is defined as a material extrusion
process used to make thermoplastic parts by extrusion and hot deposition of materials layer by layer, a
term that denotes machines built by Stratasys, Inc.

The use of this manufacturing technology has made remarkable advances in the variety of parts
that can be produced with complex geometries. With the MA processes, the specific materials they
employ, as well as their particular properties, new GD&T standards need to be implemented (Udroiu &
Braga, 2020).

The present project evaluates the ability to consistently produce parts with specific dimensions
using the additive manufacturing process MFD of the TESVB reverse engineering laboratory machine,
using the Cp and Cpk indices, which compare the dispersion of the processes with the design tolerance,
thus establishing more or less acceptable levels of dispersion according to the needs of use of the parts.
Considering a normal distribution in the sample, which will allow us to establish that, using a Gaussian
bell amplitude of six standard 6σ deviations, we will have control over 99.74% of the population (Lind,
Marchal, & Wathen, 2015).
 32

Since the parts manufactured for the project have no specific use, as they are manufactured for
the sole purpose of the study, there is no dimensional requirement from a customer, therefore the
dimensional tolerances are established in accordance with ISO 2768-1 ‘General tolerances. Tolerances
for linear and angular dimensions without indication of individual tolerance’.

Problematic

The step from manufacturing prototypes to end-use parts implies that the manufacturing processes, both
subtractive and additive, which is the object of analysis of this work, have statistical stability in obtaining
the characteristics of the product, which guarantees a level of compliance in the quality of the product
and thus satisfies the requirements of the client. This work will focus on the geometric dimensions of
parts manufactured using the FDM MA process on the Strasasys® uPrint SE® machine in the Industrial
Engineering laboratory of the Tecnológico de Estudios Superiores de Valle de Bravo.

As has also been mentioned, this equipment has not been conceived to manufacture final parts
and its use in the institution has been for merely didactic activities and, in some cases, for prototyping,
but in no case as equipment for the production of final parts. Because of this, there is no tolerance to be
met defined by the requirements of any customer. However, for the analysis, the aim is to define a unit
design, i.e. 10 mm per feature, wherever possible, to simplify the analysis. The tolerances to these
dimensions are precisely what the statistical analysis is intended to determine, using the potential process
capability parameter Cp and actual process capability Cpk, which takes into account the bias in the data
as a result of the process adjustment.

Since there is no given specification for manufacturing the parts, a unit-dimension design, i.e. a
three-dimensional part, namely a die, 10 millimetres per side, will be used as the starting point. This
choice of geometry and dimensions is intended to facilitate the calculations and thus to focus on the
results and the statistical behaviour of the calculations. Tolerances are assigned according to the
international standard ISO 2768-1 General Tolerances - Part 1: Tolerances for linear and angular
dimensions without individual tolerance indications. When performing the test, a potential process
capability index Cp ≥ 1.67 and an actual process capability index Cpk ≥ 1.33 are expected.

The dimensions will be evaluated in the different forms of material deposition, in order to
determine the effect they have on the dimensional stability of the parts.

Justification

The present work aims to understand the ability of the additive manufacturing machine in the laboratory
at TESVB and, although it belongs to the MA processes, it is not, nor will it be used to manufacture
consumer parts and, although it could be said that the results obtained from the analysis could be
extrapolated to processes in industry, the reality is that what is obtained from this analysis could be
extrapolated to processes in industry, the reality is that what emerges here will be for purely academic
and didactic purposes, as the data obtained, as well as serving to develop the scientific interest of the
students who participate in it, will also serve to enrich the programmes of subjects such as statistics and
statistical quality control within our institution.

In the same way, this work aims to be a spearhead for similar projects with different conventional
equipment, but whose ability is not known, and although it is of no interest for accessing markets, as the
objective of having these machines is not to sell anything, it is interesting and important to know through
practice the potential of equipment that, in other environments, would form part of productive processes.

However, it is worth exploring the potential of AM from our small but significant trench, as a
reality in today's industry, an industry of goods with intrinsic disposable value, in which environment
AM and its flexibility fits perfectly. For, while it is not the panacea of new manufacturing, it does
represent an important alternative to the needs of societies increasingly focused on experiences and less
on possessions.

Thus, putting the scientific knowledge of statistical process control into practice in a tangible
process will be of great value to the undersigned and, moreover, desirably to the industrial engineering
community, first and foremost, to the entire TESVB.
 33

Target

To determine the capability of the Stratasys® uPrint SE® additive manufacturing machine by the Fused
Deposition Modelling method of the TESVB Industrial Engineering laboratory to manufacture
dimensionally stable cubic parts when evaluated by statistical analysis using the process capability
indicators Cp and Cpk.

Methodology

This is a quantitative investigation carried out by producing a sample of size n=30, with a confidence
interval of 95% and a relative error of 10% for the Cpk factor. The parts shall have a design of unit
dimensions (10 mm per side), in the form of a cube, to print them and evaluate their dimensions in the
distance between the different faces, this will give an idea of the ability of the machine to reproduce the
dimensional characteristics and their relationship with the position of the deposited material in layers,
establishing, if any, a difference between the dimensions obtained along the deposition or tangent to it.
The data shall be evaluated according to the following statistics:

Cp = Tolerance/Dispersion (1)

In which the tolerance is given by:

𝑇𝑜𝑙𝑒𝑟𝑎𝑛𝑐𝑒 = 𝐿𝑆𝐸 − 𝐿𝐼𝐸 (2)

Where:

𝐿𝑆𝐸 = 𝑈𝑝𝑝𝑒𝑟 𝑠𝑝𝑒𝑐𝑖𝑓𝑖𝑐𝑎𝑡𝑖𝑜𝑛 𝑙𝑖𝑚𝑖𝑡

𝐿𝐼𝐸 = 𝐿𝑜𝑤𝑒𝑟 𝑠𝑝𝑒𝑐𝑖𝑓𝑖𝑐𝑎𝑡𝑖𝑜𝑛 𝑙𝑖𝑚𝑖𝑡

While the dispersion is given by:

𝐷𝑖𝑠𝑝𝑒𝑟𝑠𝑖𝑜𝑛 = 6𝜎

Where:

𝜎 = 𝑃𝑟𝑜𝑐𝑒𝑠𝑠 𝑠𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛

Development

For the present work, parts were printed on a Stratasys® uPrint SE Plus® machine (see Figure 1) using
ABSplus P430® as construction material, without support material, given the geometry of the parts.

Figure 1 Impresora Stratasys uPrint SE Plus

 34

The design of the parts corresponds to cubes of 10 mm per side (see Figure 2). This design was
defined in such a way as to have longitudinal dimensions by which the machine's ability to manufacture
parts with consistency in these dimensions could be checked. In addition, two radii were added, one
internal and one external, in order to have the possibility of eventually adding these dimensions to the
study. In this work the study of these dimensions was not carried out, only the linear ones.

Figure 2 Screenshot of the design of the part. STL file

In order to carry out an ability study it is necessary to establish tolerances to the nominal
measurement. As the part under study does not belong to any assembly, nor is it requested by any
customer, only the nominal value is available. Therefore, the international standard ISO 2768-1:1989
Tolerances for linear and angular dimensions without indication of individual tolerances was used. This
standard provides guidance for setting specification limits according to four tolerance classes: fine,
medium, coarse and very coarse (see Table 1). For the present study, the fine and medium tolerances
were chosen whose tolerances, for the range in which the nominal dimension falls, are ±0.1 mm and ±0.2
mm, respectively (ISO, 1989).

Values in millimetres
Tolerance class Permissible deviations for basic dimension ranges
Designation Description 0.5 a More More Over 30 More than More than Over Over
3 than 3 than 6 and up to 120 and 400 and 1000 and 2000 and
and up to and up to 120 up to 400 up to 1000 up to up to
6 30 2000 4000
f fina ±0.05 ±0.05 ±0.1 ±0.15 ±0.2 ±0.3 ±0.5 ---
m half ±0.1 ±0.4 ±0.2 ±0.3 ±0.5 ±0.8 ±1.2 ±2
c broad ±0.2 ±0.3 ±0.5 ±0.8 ±1.2 ±2 ±3 ±4
v very large --- ±0.5 ±1 ±1.5 ±2.5 ±4 ±6 ±8

Table 1 Permissible deviations for linear dimensions, except broken edges

Source: ISO 2768-1:1989

Measurements were made with a Moore & Wright digital micrometer with a measuring range of
0 - 25 mm and a resolution of 0.001 mm. This instrument is suitable for the study, since the resolution of
the measuring instrument must be less than 10% of the defined tolerances (Preissler, Rosenberger, &
Notni, 2017).
𝑖𝑀𝑒𝑎𝑠𝑢𝑟𝑒𝑚𝑒𝑛𝑡 𝑢𝑛𝑐𝑒𝑟𝑡𝑎𝑖𝑛𝑡𝑦
0.1 ≥
𝑚𝑎𝑛𝑢𝑓𝑎𝑐𝑡𝑢𝑟𝑖𝑛𝑔 𝑡𝑜𝑙𝑒𝑟𝑎𝑛𝑐𝑒

The sample size is 30 pieces, which allows a confidence level between 90% and 95% with a ratio
of the lower confidence limit and the Cp of 80% (Franklin, 1999), according to the following formula
(see Table 2).

𝜒2 (𝛼)
𝐶𝑝𝑙𝑖𝑐 = 𝐶̂𝑝 √ 𝑛−1
𝑛−1
 35

Desired ratio de 𝑪𝒌𝒍 ̂

𝒑 ⁄𝑪𝒑 Confidence level for the lower confidence limit
90% 95% 99%
0.80 n = 24 n = 36 n = 67
0.8034 0.8011 0.8002
0.85 n = 40 n = 63 n = 120
0.8503 0.8509 0.8504
0.90 n = 87 n = 139 n = 269
0.9001 0.9003 0.9000

Table 2 Comparison of an approximate sample size for the Cp formula compared to the exact size
(obtained iteratively)
Source: (Franklin, 1999)

The formula used for the calculation of the Cp index is as follows:

𝐿𝑆𝐸−𝐿𝐼𝐸
𝐶𝑝 =
6𝑆

Where:

LSE = Upper Specification Limit

LIE = Lower Specification Limit
S = Standard deviation of the sample, obtained from the following formula:

∑𝑛
𝑖=1(𝑥𝑖 −𝑥̅ )
2
𝑆=√ 𝑛−1

The settings for the manufacture of the parts were carried out according to the following table
(see Table 3):

Setting parameter Value

Filament thickness 0.254 mm
Filling density High-density hollow
Print direction Axes of the radii perpendicular to the plane of the printing plate
Supporting filler SMART
STL file scale 1:1
Orientation 180° (Y axis) by positioning the figure such that the base is the face with the rounded edges
(top view)

Table 3 Manufacturing parameters

Results

It should be noted that all parts evaluated are smaller than the nominal value of the design, which implies
that a calculation of the Cpk index is unnecessary, as it would be out of control. Furthermore, the Cp
index values for both fine and medium tolerance classes indicate an out-of-control process. For fine
tolerance 10±0.1 mm the Cp index is 0.77, which is equivalent to a probability of out-of-specification
parts of 130013 parts per million (see Figure 3). This is largely due, in addition to the dispersion, to the
displacement of the data towards the lower specification limit, obtaining out-of-tolerance parts even
below this limit.

For medium tolerance (see Figure 4) the Cp index is 1.53, equivalent to a probability of out-of-
tolerance parts of 305 parts per million. By opening the tolerance, machine performance improves
significantly. However, the shift of the values towards the lower limit still needs to be corrected.
 36

Figure 3 Capability analysis with Cp indices for fine tolerance

Figure 4 Capability analysis with Cp indices for mean tolerance

Conclusions

The ability of the machine under analysis to produce parts within specification is poor when dealing with
a fine tolerance with 20 tenths of a millimetre tolerance, this ability improves substantially when the
tolerance is opened to double, to have a total of 40 tenths of a millimetre total tolerances. However, an
adjustment must be made to centre the process, because if the analysis of the Cpk index, which considers
the process bias, were to be carried out, it would have a very poor result for the reason mentioned above.
It should be noted that the thickness of the filament used in the adjustment for the production of the parts
was 0.254 mm, this is a fraction of an inch, if it is considered that the layers are superimposed one after
the other without overlapping, to generate a 10 mm high cube, the machine should deposit an inaccurate
amount of layers, therefore, it is inferred that designing a part in fractions of an inch could increase the
process skill indexes.

Statements

This work analyses the relationship between the tolerance specified in the design of a part and the
dispersion of the process that produces it, by measuring a sample and obtaining the arithmetic mean and
standard deviation, and by obtaining the process skill index Cp , being clear that, being a sample, this
index represents only an estimate, whose confidence level depends, to a large extent, on the sample size
chosen, which has been n = 30.
 37

The results show that the process is not stable for a fine tolerance of 0.20 mm between the lower
and upper specification limits, and that the mean and all individual values show a trend below the nominal
value of 10 mm. This highlights the need to focus the process to bring the values closer to the nominal
specification. An observation made during the study is that one of the construction parameters of the
parts, the layer thickness, is 0.254 mm, which is equal to 0.01 inch. This measurement makes sense since
the machine is of American manufacture. This is the hypothesis for further studies in which either tools
are available that allow layer thicknesses in fractions of a metre, or parts are designed with dimensions
in fractions of an inch. This should increase the value of the Cp index, i.e. the process capability. This
hypothesis is put forward and remains to be demonstrated in subsequent projects.

Conflict of interest

The authors declare that they have no conflict of interest. They have no financial interests or personal
relationships that could have influenced this book.

Authors' contribution

Baltazar-López, Servando: Contributed to the formulation of the research project and the design of the
methodology. Design of the parts and determination of the specification, manufacture and measurement
of the parts, as well as the statistical analysis and writing of the final manuscript.

Availability of data and materials

All data and information presented in this research are available in the references.

Funding

This research did not require funding.

Acknowledgements

The author gratefully acknowledges the support of the Industrial Engineering academic programme of
the Tecnológico de Estudios Superiores de Valle de Bravo for the facilities to carry out the project with
the Additive Manufacturing equipment and materials of the Reverse Engineering and Dimensional
Metrology laboratory.

Abbreviations

FDM Fuse Deposition Modelin

DIN Duetsches Institut für Normung
ISO International Organization for Standardization
LSE Upper Specification Limit
LIE Lower Specification Limit
MA Additive Manufacturing

References

Background.

Huang, T., Wang, S., & He, K. (2015). Quality Control for Fused Deposition Modeling Based Additive
Manufacturing: Current Research and Future Trends. The First International Conference on Reliability
Systems Engineering, 1-6. doi:doi: 10.1109/ICRSE.2015.7366500

Kim, H., Lin, Y., & Tseng, T.-L. B. (2018). A review on quality control in additive manufacturing. Rapid
Prototyping Journal.
 38

Basics

Abdulhameed, O., Abdulrahman, A.-A., Ameen, W., & Hammad Mian, S. (2019). Additive
manufacturing: Challenges, trends, and applications. Advances in Mechanical Engineering.

Alafaghani, A., & Qattawi, A. (2018). Investigating the effect of fused deposition modeling processing
parameters using Taguchi design of experiment method.

Bellini, A., & Güçer, S. (2003). Mechanical characterization of parts fabricated using fused deposition
modeling. Rapid Prototyping Journal.

Franklin, L. A. (1999). Sample size determination for lower confidence limits for estimating process
capability indices. Terre Haute, IN, USA: Pergamon.

Lind, D. A., Marchal, W. G., & Wathen, S. A. (2015). Statistics Applied to Business and Economics.
Mexico: McGraw-Hill/Interamericana Editores S. A. de C. V.

MacDonald, E., & Wicker, R. (2016). Multiprocess 3D printing for increasing component functionality.
Science.

Support

ASTM (2013). Standard Terminology for Standard Terminology for Additive Manufacturing.

ISO (1989). ISO 2768-1 General Tolerances- Part 1: Tolerances for linear and angular dimensions
without individual tolerance indications. ISO.

ISO, D. (2020). DIN 8580:2020-01 Fertigungsverfahren - Begriffe, Einteilung. Germany.

Discussions

Kumar, N., Jain, P. J., Tandon, P., & Pandey, P. M. (2018). The effect of process parameters on tensile
behavior of 3D printed flexible parts of ethylene vinyl acetate (EVA). Journal of Manufacturing
Processes.

Preissler, M., Rosenberger, M., & Notni, G. (2017). AN INVESTIGATION FOR PROCESS
CAPABILITY IN ADDITIVE MANUFACTURING. 59th Ilmenau Scientific Colloquium.