Vous êtes sur la page 1sur 9

# Measurement 44 (2011) 611619

## Contents lists available at ScienceDirect

Measurement
journal homepage: www.elsevier.com/locate/measurement

## Investigations on machined metal surfaces through the stylus type

and optical 3D instruments and their mathematical modeling with the
help of statistical techniques
P. Demircioglu a, M.N. Durakbasa b,
a

Adnan Menderes University, Faculty of Engineering, Department of Mechanical Engineering, 09010 Aydin, Turkey
Vienna University of Technology, Institute for Production Engineering and Laser Technology, Department of Interchangeable
Manufacturing and Industrial Metrology, Karlsplatz 13/311, 1040 Wien, Vienna, Austria

a r t i c l e

i n f o

Article history:
Accepted 14 December 2010
Available online 19 January 2011
Keywords:
Surface metrology
Surface roughness
Comparative study
Stylus and 3D optical methods
Statistical analysis

a b s t r a c t
The measurement of roughness on machined metal surfaces is of considerable importance
to manufacturing industries as the roughness of a surface has a signicant inuence on its
quality and function of products. In this paper, an experimental approach for surface
roughness measurement has been based on the comparison of roughness values taken
from the stylus and optical type instruments on the machined metal surfaces (turning,
grinding and milling) is presented.
Following this experimental study, all measured surface roughness parameters have
been analyzed by using Statistical Package for Social Science (SPSS 15.0) statistically and
mathematical models for the two most important and commonly used roughness parameters Ra and Rz have been developed so that Ra = Ra (F, P, C) and Rz = Rz (F, P, C, M), whereas F
expresses feed, P periodicity, C contrast and M the type of material. The statistical results
from numerous tests showed that there has been a correlation between the surface roughness and the properties of the surface topography and there have been slight differences
among three measurement instruments on machined metal surfaces in this experimental
study.

1. Introduction
Surface roughness plays a vital role in determining the
desired quality of a machined metal surface for todays
engineering industry. The quality of assessment of engineering surfaces with respect to their functional and optical properties for different loading conditions is inuenced
by roughness parameters characterizing basically the surface microtopography [1]. It is traditionally dened by
two parameters: arithmetical mean deviation of the assessed prole Ra and average maximum height of assessed
prole Rz as they are one of the most commonly used and
accepted by researchers and in industry as well. Surface
Corresponding author.
doi:10.1016/j.measurement.2010.12.001

## roughness inspection is one of the essential quality control

processes carried out to ensure that manufactured parts
conform to specied standards [2].
The surface parameter used to evaluate surface roughness in this experimental study is the roughness average
(Ra), the most widely used parameter for surface texture.
The roughness average is the area between the roughness
prole and its central line, or the integral of the absolute
value of the roughness prole height over the sampling
length. Determination of Ra is normally computed by the
software but can be derived using the following formula:

Ra

1
lr

lr

jzxjdx

## where is z(x) is the prole deviation from the mean line

and lr is the sampling length.

612

## P. Demircioglu, M.N. Durakbasa / Measurement 44 (2011) 611619

The more common parameter for roughness is maximum height of prole (Rz). Rz is calculated by measuring
the vertical distance from the highest peak to the lowest
valley within ve sampling lengths, then averaging these
distances. Rz averages only the ve highest peaks and the
ve deepest valleys therefore extremes have a much
greater inuence on the nal value.
The goal of this research work is to obtain mathematical
models of Ra and Rz estimating the coefcients of the linear
equation, involving a few independent variables (feed in
mm, periodicity, type of material, contrasting, type of production process, etc.) with an analysis of variance (ANOVA)
and regression analysis.
2. Surface topography techniques for the comparative
study
Conventionally, surface roughness measurement has
been performed by using a stylus instrument [3,4]. When
a stylus traverses a surface, the vertical motion of the stylus is converted by way of a pick-up into an electrical signal. The pick-up is generally a linear variable differential
transducer (LVDT). The electric signal is amplied and processed or converted into a digital signal via an A/D converter and then analysed using a computer. A schematic
diagram of such a system is shown in Fig. 1.
In the diagram shown, the stylus is held stationary
while the specimen surface is moved in a raster scan using
precision X, Y-tables. The movement of the table is controlled via a computer, allowing numerous combinations
of area size and data sample spacing to be selected [6].
The stylus measurement method is a contact type, the
main drawback of which is that the loaded stylus can damage or scratch the surface being measured, especially on
soft surfaces [7]. The transducer and stylus tips are often
fragile, hence the instrument must be applied in a fairly
vibration free environment. Consequently, this direct con-

## tact measurement method is not suitable to be used on a

test object undergoing a machining process simultaneously [8].
These two main drawbacks of the stylus-based surface
measurement technique make it necessary to develop
non-contact optical methods that can be used for in-process measurement and the measurement of soft surfaces
[9]. Stylus and optical type prolometers are, in a sense,
different, that is; while traditional stylus method is used
for height information, the optical method refers to areal.
Optical technique as a complement to the stylus instrument, a non-destructive and non-contact method, appears
to be a suitable alternative for carrying out measurement
of surface quality including surface roughness [10].
New breakthroughs by the instrumentations have been
made in recent years, to establish high-tech instruments
which can acquire a 3D surface structure of the precisely
machined surfaces to fulll the requirements for the application in industrial environment. In this experimental
study, both surface measuring systems will be examined
with great many practical applications [11].
The measurements of optical systems were carried out
by two different instruments in this experimental study.
One is a new non-contact optical surface characterization
technique called focus variation used by the innite focus
microscope (IFM) to build true color 3D images of surfaces
and microscopic structures. Its operating principle combines the small depth of focus of an optical system with
vertical scanning to provide topographical and color information from the variation of focus. The system delivers
dense measurements over large areas with a density of 2
Mio 25 Mio measurement points and a high vertical resolution up to 20 nm [12].
This non-destructive method utilizes coaxial white light
which is provided by a light source delivered through a
beam splitter to a series of selectable, innity-corrected,
high-Numerical Aperture (N.A.) objectives contained in a

Fig. 1. Schematic diagram illustrating the major constituents of a stylus-type of surface texture measuring instrument [5].

613

## Fig. 2. Schematic visualization of the focus-variation technology [12].

six-place nosepiece. The specimens reected light is projected through the beam splitter onto a color digital sensor
(Fig. 2).
The other one is a scanning type confocal laser microscope with high resolution, high contrast and drastically
enhanced resolution in light axis direction via confocal optics. Confocal optics are designed to have almost innite
small depth of focus. Thats why, not only the variation of
focus, but also the absolute maximum value is analysed
[10]. The scanning type confocal laser microscope targets
laser beam at a very small spot with objective lens and
scans over the specimen in XY directions. It then captures
a light from specimen with detector and outputs the image
of specimen on monitor [13]. A schematic diagram of such
a system is illustrated in Fig. 3.
In this experimental study, measurements by stylus
type instrument were performed with Form Talysurf Intra
50, which is skidless and can be used for waviness, prole
and other parameters such as material ratio with absolute
condence in the measurement results and measurements
by optical 3D surface measurement devices, the innite
focus microscope and the confocal laser scanning
microscope.
3. Conditions of experiment
In this experimental study, the measurements were
made with commercially available Form Talysurf Intra 50
by Taylor Hobson GmbH for the tactile surface evaluation
with a high resolution, in 1.0 mm range 16 nm. By its software ultra it is possible to analyze and monitor operations. It was used 60 mm stylus arm length, 2 lm radius
conisphere diamond stylus tip size, 1 mN force
(speed = 1 mm/s) and Gaussian lter in all measurements
by the stylus instrument [2,14,15]. Each measurement

## Fig. 3. Schematic representation of the principle of a scanning type

confocal laser microscope [13].

## was redone six times. During the process of measuring,

the cutoff length was taken 0.8 mm and the sampling
length 4 mm according to the ISO standards. The analyses
were then performed for all specimens for several roughness parameters which are Ra and Rz parameters giving
us much of the idea [16,17]. Altogether 90 experiments
were conducted in order to allow performing ANOVA and
regression analysis using SPSS 15, capable of predicting
precise surface roughness. The measurements were taken
with commercially available instruments.
4. Experimental work
The precise measurement and evaluation processes of
this research work have been carried out in Vienna University of Technology, Interchangeable Manufacturing and
Industrial Metrology, Nanometrology Laboratory. The
experiments were done by preparing at specimens with
different machining processes such as turning, grinding
and milling. Fifteen at samples with periodic and random
proles of different roughness value classes, having shiny
and browned surfaces were used in this experimental
work. The samples were classied into two groups as periodic and random surface proles in this research work. The
reason for having two different groups of proles is that
the prole is an important concern when compared the devices. Six consecutive measurements were taken for each
condition. The same alignment system was applied for all
specimens. Obtained results from both, stylus and optical
surface measuring instruments for the surface evaluation
under laboratory conditions are presented in this paper
[10].

614

## ments with nine specimens. By using the same alignment

system, six successive measurements were made on nine
samples processed with face turning by means of three
instruments. Face turning steel shiny sample 1 has periodic
surface proles as shown in Fig. 4. The diagrams of roughness prole belonging to the sample taken from the stylus
type device and two optical instruments were obtained as
shown in Fig. 5.
Comparisons of the roughness values belonging to the
at specimens taken from the tactile system and two optical systems are shown in Fig. 6, where o1 is the abbreviation of the innite focus microscope, c is the abbreviation
of stylus instrument and o2 is the abbreviation of the confocal laser scanning microscope. The at specimens having
periodic surface proles have also been evaluated in terms
of RSm as well as Ra and Rz parameters according to ISO
4288: 1998 [2].
For this sample, it was taken comparable results for
mean values of Ra parameters. (For the stylus system,

## 4.1. Specimens with periodic surface proles

Three face turning aluminum shiny specimens, three
face turning steel shiny samples and three face turning
steel browned workpieces with periodic surface proles
were measured by three instruments. Comparability for
periodic surfaces will be veried as a result of the experi-

260
255
250
245
240
235
230
225
220
215
210

2)
(o

2)
M
ea
n

(o

2)
3

(o

2)
1

(o

(c
6

(c

)
(c
2

7,000
6,000
5,000
4,000

Ra

3,000

Rz

2,000
1,000

6
(
M o1)
ea
n
(o
1)
1
(c
)
2
(c
)
3
(c
)
4
(c
)
5
(c
)
6
(c
)
M
ea
n
(c
)
1
(o
2)
2
(o
2)
3
(o
2)
4
(o
2)
5
(o
2)
6
(o
2
M
ea )
n
(o
2)

1)
(o

(o

1)
5

1)

(o

(o

(o
3

1)

0,000

1)

## Roughness Values (m)

ea
n

(o

1)

1)
(o
5

(o
3

(o

1)

RSm (m)

1)

Roughness values ( m)

Fig. 5. Diagrams of roughness prole belonging to face turning steel shiny sample 1 respectively obtained them from the contact stylus measuring system
(on the right), the innite focus microscope (in the middle) and the confocal laser scanning microscope (on the left) [10].

Fig. 6. Comparisons of the roughness values belonging to face turning steel shiny sample 1 taken from the contact stylus and two optical measuring
systems in terms of the parameter RSm, Ra and Rz [10].

615

## P. Demircioglu, M.N. Durakbasa / Measurement 44 (2011) 611619

1.031 lm; for the innite focus microscope, 1.113 lm; for
the confocal laser scanning microscope, 1.111 lm were taken). The standard deviation of the data belonging to the
face turning steel shiny sample 1 is also similar. It was also
taken compatible results for RSm and Rz values according to
the ISO standards.

## 4.2. Specimens with random surface proles

Two surface grinding steel browned samples, two
peripheral milling steel browned samples and two front
milling steel browned samples with random proles were

used in order to strengthen this study. In this group of proles there will be couples with roughness class difference
of one. Comparability for six specimens with random surfaces will be veried as a result of the experiments. The
diagrams of roughness prole and values belonging to
peripheral milling steel browned sample 1 (Fig. 7) taken
from both systems were obtained as shown in Fig. 8. Comparisons of roughness values taken from both measurement systems in terms of parameters Ra and Rz are given
in Fig. 9.
For the peripheral milling steel browned sample 1, it
was taken comparable results for mean values of Ra parameters (For the stylus system, 1.031 lm; for the innite focus microscope, 1.113 lm; for the confocal laser scanning
microscope, 1.111 lm were taken.). It was also taken compatible results for Rz values according to the ISO standards.

5. Experimental results

## During this experimental work, all measured surface

roughness parameters have been analysed by using SPSS
15 statistically. Data for each test were statistically analyzed by the help of regression analysis, which is the most
popular empirical methods. The regression analysis determines which factors and interactions are signicant.

7,000
6,000
5,000
4,000

Ra

3,000

Rz

2,000
1,000

1
ea )
n
(o
1)
1
(c
)
2
(c
)
3
(c
)
4
(c
)
5
(c
)
6
(c
)
M
ea
n
(c
)
1
(o
2)
2
(o
2)
3
(o
2)
4
(o
2)
5
(o
2)
6
(o
2
M
ea )
n
(o
2)

)
(o
1

(o

1)
(o

(o
1

(o
1

(o
2

1)

0,000

## Roughness Values (m)

Fig. 8. Diagrams of roughness prole belonging to peripheral milling steel browned sample 1 respectively obtained them from the contact stylus measuring
system (a), the innite focus microscope (b) and the confocal laser scanning microscope (c) [10].

Fig. 9. Comparisons of the roughness values belonging to peripheral milling steel browned sample 1 taken from the contact stylus and two optical
measuring systems in terms of the parameter Ra and Rz [10].

616

## A One-way Analysis of variance (Oneway ANOVA) has

also been used (a = 0.05) to test the signicant difference
between measurement systems. When the Oneway ANOVA has been applied so as to test the equality of three
instruments at one time by using variances (feed in mm,
periodicity, type of material, contrasting, type of production process, etc.), a comparison of them was done employing a Post-Hoc test to identify which groups were
signicantly different from others assuming a 95% of condence level. These tests showed that there has been a
slight difference between the contact stylus surface measurement system and two non-contact optical measurement systems on engineering surfaces statistically.
In Figs. 10 and 11, the distribution curves (Frequency
Histograms) of the descriptive statistical results in terms
of Ra and Rz parameters were presented.
Finally, Linear Regression of Ra and Rz parameters estimated the coefcients of the linear equation, involving a few

## independent variables (feed in mm, periodicity, type of

material, contrasting, type of production process, etc.), that
best predicted the value of the dependent variable. And
two mathematical models giving the values of Ra and Rz
roughness parameters have been established in terms of
feed in mm, the periodicity, contrasting and type of material
as a result of the statistical analyses as shown below.
According to Table 1, Ra depends on periodicity, the
type of material, contrast, the type of production processes,
feed and the type of machine as its independent variables
but feed, periodicity and contrast predicted the value of the
Ra parameter in a best way. The linear equation of Ra
parameter is given below. In Fig. 12, the distribution curves
of Ra with descriptive statistic values for measurement
data were presented.

## whereas F feed, P periodicity and C contrasting.

Fig. 10. The distribution curves (Frequency Histograms) of the descriptive statistical results belonging to the measurements of face turning aluminum shiny
sample 1, face turning steel shiny sample 1 and face turning steel browned sample 1 in terms of Ra and Rz parameters.

Fig. 11. The distribution curves (Frequency Histograms) of the descriptive statistical results belonging to the measurements of peripheral milling steel
browned sample 1, peripheral milling steel browned sample 2, front milling steel browned sample 1, front milling steel browned sample 2 in terms of Ra
and Rz parameters.

617

## The measurement results for 15 at specimens were

very consistent with each other. This showed that there
was an inner accuracy of the three measurement systems.
Furthermore, it could be seen that measured values were
comparable and the differences between the methods
were small. The measurement results taken from the tactile instrument in terms of Ra parameters turned out to
be extremely well-matched with those taken from two
optical instruments, because Ra value refers to a mean value. However Rz value refers to the height of a prole,
between the minimum and maximum points of the prole.
Rz is calculated by measuring the vertical distance from the

## According to Table 2, Rz depends on periodicity, the type

of material, contrast, the type of production processes, feed
and the type of machine as its independent variables but
feed, periodicity, contrast and the type of material predicted the value of the Ra parameter in a best way. The linear equation of Rz parameter is given below. In Fig. 13, the
distribution curves of Rz with descriptive statistic values
for measurement data were presented.
Rz 3:258 38:059  F 4:495  P 0:906  M  1:500  C

## whereas F feed, P periodicity, C contrasting and M type of

material.

Table 1
The coefcients of the linear equation, involving a few independent variables (feed in mm, periodicity, type of material, contrast, type of production process,
etc.), that best predicted the value of the dependent variable Ra.
Model

(Constant)
Periodic surface proles
The type of material
Contrast
The type of production processes
Feed
The type of machine

Unstandardized coefcients

Standardized coefcients

Std. Error

Beta

1.314
1.279
0.059
0.188
0.024
9.097
0.025

0.244
0.283
0.087
0.087
0.106
0.124
0.036

0.234
0.010
0.037
0.011
1.097
0.008

Sig.

5.396
4.525
0.680
2.166
0.227
73,633
0.687

0.000
0.000
0.497
0.031
0.821
0.000
0.493

Fig. 12. The distribution curves of Ra with descriptive statistic values for measurement data [10].

Table 2
The coefcients of the linear equation, involving a few independent variables (feed in mm, periodicity, type of material, contrast, type of production process,
etc.), that best predicted the value of the dependent variable Rz.
Model

(Constant)
Periodic surface proles
The type of material
Contrast
The type of production processes
Feed
The type of machine

## Dependent variable: Rz.

Unstandardized coefcients

Standardized coefcients

Sig.

Std. error

Beta

Std. error

3.258
4.495
.906
1.500
.067
38.059
.051

.960
1.113
.342
.342
.419
.487
.142

.194
.036
.070
.007
1.082
.004

3.395
4.038
2.648
4.386
.159
78.203
.357

.001
.000
.009
.000
.874
.000
.721

618

## P. Demircioglu, M.N. Durakbasa / Measurement 44 (2011) 611619

Fig. 13. The distribution curves of Rz with descriptive statistic values for measurement data [10].

9
8
7
6
5

Sa
Ra

4
3
2
1

15

14

13

12

11

10

## The no of measured samples with flat surfaces

Fig. 14. Comparisons of Ra and Sa parameters belonging to the samples having at surfaces taken from the contact stylus measurement system and two
non-contact optical systems [10].

## highest peak to the lowest valley within the sampling

lengths, then averaging these distances. Rz averages only
the few highest peaks and the deepest valleys, therefore
extremes have a greater inuence on the nal value. This
is the reason why the measurement results of the samples
taken from the non-contact instruments in terms of Rz
parameters came out slightly higher than the contact
instrument. The differences between three measuring
instruments were small because these 15 at samples have
ordinary engineering surfaces.
In this study, the 3D surface parameters were also measured. For evaluation of surface roughness as the most
popular 3D surface parameter Sa was chosen for comparison. Comparisons of Ra and Sa parameters belonging to
the samples having at surfaces taken from the contact
stylus measurement system and two non-contact optical
systems are shown in Fig. 14. As it can be seen from gure,
measurement results for Sa turned out to be quite similar
to those for Ra parameter, because Ra is the 2D counterpart
of the 3D descriptor Sa. Both Ra and Sa reect the arithmetic
mean of the absolute values of the surface point departures
from the mean plane within the sampling area [18].

6. Conclusion
This paper is an experimental study of the roughness
analyses of the conventionally machined samples with at
surfaces in order to compare the data obtained from the
contact stylus measurement device with two non-contact
optical surface measurement instruments and the capabilities of three measurement systems were noted in terms of
their similarities and differences. The major disadvantage
of using a stylus instrument is that it requires direct physical contact, which limits the measuring speed. In addition,
the instrument readings are based on a limited number of
line samplings, which may not represent the real characteristics of the surface. This kind of deviation may cause
serious errors in the surface quality assessment especially
when the surface prole is periodic.
It is observed that three devices are giving comparable
results if the surface has a good reection value, is not very
ne machined surface with a periodic prole and not
ruined or scratched. According to the presumptions, the
problem with the ne machined samples having periodic
proles is that contact system can not detect the extreme

## values of the surface and give a result due to general ne

prole. But optical systems can detect precise value of
the prole so that the results are getting scattered. At this
point, when the inadequacy of stylus was compared to a
light beam, it has been noted that stylus was inefcient because of its geometrical form. Of course no man building
stylus can reach the thickness of a light beam in todays
technology.
The next step of this research work is to compare spherical specimens to understand how the contact stylus and
the optical measurement instruments react to spherical
samples in terms of their geometrical shapes. That is, contact stylus surface measurement instrument has some limitations by measuring samples with spherical surfaces.
Further works could be compared the data of the 3D
roughness parameters obtained from the stylus type prolometer with those from the optical systems. 3D surface
parameters give more precise picture of the surface, so it
is possible to evaluate more precise surface roughness
parameters according to used machining technique.
Although surface topography has been studied in great detail in terms of 2D parameters, a comprehensive comparison on the nature of 3D surface topography generated by
different machining styles at different machining conditions is still missing. Additionally, the relationship between 3D surface topography and functional aspects is
poorly understood. Because of this reason, the following
step of this paper is to analyze the 3D surface topography
of the machined surfaces and compare the measurement
systems in terms of 3D surface roughness parameters.
References
[1] B. Dhanasekar, N. Krishna Mohan, Basanta Bhaduri, B. Ramamoorthy,
Evaluation of surface roughness based on monochromatic speckle
correlation using image processing, Precision Eng. 32 (2008) 196
206.
[2] ISO 4288: 1996, Geometrical Product Specications (GPS) Surface
Texture: Prole Method Rules and Procedures for the Assessment
of Surface Texture. TCor 1: 1998.

619

[3] D.J. Whitehouse, Stylus contact method for surface metrology in the
ascendancy, Meas. Cont. 31 (2) (1998) 4850.
[4] D.J. Whitehouse, Handbook of Surface and Nanometrology, Institute
of Physics, IOP Publishing, London, 2003.
[5] L. Blunt, K.J. Stout, Three-dimensional Surface Topography. Penton
Press, London, 2000, ISBN: 1857180267, p. 320.
[6] H.J. Pahk, K.J. Stout, L. Blunt, A comparative study on the threedimensional surface topography for the polished surface of femoral
[7] M.N. Durakbasa, P.H. Osanna, P. Aksoy (Demircioglu),
Characterization and Analysis of High-Precise Surface Structures by
Contact and Non-contact Methods, The First Conference ERIN,
Faculty of Mechanical Engineering in Bratislava, Slovakia, April 25
26, 2007, p. 61.
[8] C.J. Tay, S.H. Wang, C. Quan, H.M. Shang, In situ surface roughness
measurement using a laser scattering method, Optics Commun. 218
(2003) 110.
[9] U. Persson, Surface roughness measurement on machined surfaces
using angular speckle correlation, J. Mater. Process. Technol. 180
(2006) 233238.
[10] P. Aksoy (Demircioglu), Evaluation of High Precision Surface
Structures by Contact Stylus and Non-Contact Optical Methods,
PhD Thesis. Vienna University of Technology, Vienna, Austria, 2008.
[11] M.N. Durakbasa, P.H. Osanna, P. Aksoy (Demircioglu), L. Kruter,
Contact and Non-contact Measurement and Analysis of the Surface
of High Precision Workpieces, The 12th International Conference on
Metrology and Properties of Engineering Surfaces, July 8th10th,
Rzeszow, Poland, 2009.
[12] R. Danzl, F. Helmli, S. Scherer, Focus Variation A New Technology
for High Resolution Optical 3D Surface Metrology, The 10th
International Conference of the Slovenian Society for NonDestructive Testing, Application of Contemporary Non-Destructive
Testing in Engineering, September 13, Ljubljana, Slovenia, 2009.
[13] http://www.leedsmicro.com/LEXT.asp.
[14] ISO 3274: 1996 Geometrical Product Specication (GPS) Surface
texture: Prole method Nominal characteristics of contact stylus
instruments. Tcor 1: 1998.
[15] ISO 4287 DAM 1: 2007; Geometrical Product Specication (GPS)
Surface Texture: Prole Method Terms, Denitions and Surface
Texture Parameters; Amendment 1: Peak Count Parameter.
[16] H.C. Beck, R. Pleul, Methods for Selecting Function-Relevant Surface
Characteristics: TEQ Technologietransfer & Qualittssicherung
GmbH,
Chemnitz,
Institut
fr
Management
und
Fertigungsmesstechnik.
[17] P. Pawlus, Digitisation of surface topography measurement results,
Measurement 40 (2007) 672686.
[18] D.J. Whitehouse, Handbook of Surface and Nanometrology, Institute
of Physics, Bristol, (published in 1994), 2003, ISBN: 0-7503-05835, 2/
1150.