Surface Roughness Measurement
Waris Khan
IIST Aerospace Department
( January 20, 2017)
Surface roughness is critically important engineering parameter as it can affect the part accuracy, reduce
the post processing costs and improve the functionality of the parts.This paper presents an experimental
technique to measure the variation of surface roughness on different material machined by a particular process
and also its variation with the machining process. In the experiment conducted, specimens used are mild steel
and aluminum, both of which are machined by milling and grinding. To make it more intensive, roughness
was measured for various cut-off length and bandwidths. The instrument used for the measurement was
Taylor Hobson surface profiler which provided a resolution of 16nm. The results are focused on roughness
and texture surface parameters for horizontal surfaces.
1 Introduction where the lower roughness limit is not enough and a
higher roughness limit is necessary such as in main-
Roughness is a measure of the texture of a taining a lubricating oil film or in case of patterned
surface. It is quantified by the vertical deviations surfaces which have a definite pattern or surface tex-
of a real surface from its ideal form. If these de- ture and have a particular roughness range for the
viations are large, the surface is rough; if they are surface which is helpful in controlling the wettability
small the surface is smooth. Roughness is typically of surface.
considered to be the high frequency, short wavelength
component of a measured surface.Roughness plays an
important role in determining how a real object will 2 State Of The Art
interact with its environment. Rough surfaces usu-
ally wear more quickly and have higher friction co- Roughness may be measured using contact
efficients than smooth surfaces. Roughness is often or non-contact methods. Contact methods involve
a good predictor of the performance of a mechani- dragging a measurement stylus across the surface;
cal component, since irregularities in the surface may these instruments include profilometers. Non-contact
form nucleation sites for cracks or corrosion. Al- methods include interferometry, confocal microscopy,
though roughness is usually undesirable, it is difficult electrical capacitance and electron microscopy.
and expensive to control in manufacturing. This of- Non-contact measurements have an advan-
ten results in a trade-off between the manufacturing tage as they can be used even when the surface is
cost of a component and its performance in applica- fine or narrow, or soft. Also the speed of measure-
tion. ment can be chosen conveniently on the basis of the
It depends on the type of material as well type of non-contact measurement. Whereas contact
as the machining processes used. It is desirable to measurements generally involve a stylus which move
minimize surface roughness, as it is often associated along a surface and have a range of speed for accurate
with friction and wear and tear. It is also seen that measurements hence they are more reliable[1].
smoother and harder surfaces have greater strength, Laser scattering is a non-contact measure-
higher load bearing capacities and longer tool life. ment where the collimated laser beam is directed to
Failure due to fatigue often begins at sharper corners a beam splitter by a polarizer. One of the beams ob-
due to high stress concentration1. Hence in order tained from the beam splitter is reflected onto a half
to produce better components and instruments, it is wave plate after which is focused onto the specimen
necessary to consider surface finishing process for a surface by a combined doublet lens. On striking the
given material and application. surface, the polarized laser beam undergoes scatter-
Smoothness of bearings and piston movement ing and specular reflection, which are then gathered
in engines, the drag characteristics of an airfoil at by a measuring lens and then directed onto photo
high Reynolds number, pipe friction, etc. all vary diode array. The second beam from the beam split-
with surface roughness. There are some systems ter is focused perpendicularly onto the surface along
1
the optical axis of the measuring lens, due to which stylus ranges between 0.1-50 mm. The instrument
the reflected beam returns along the same path and is provides skidless contact measurement and vertical
then directed onto the photo diode. The photo diodes movement of the stylus is recorded by an inductance
receive two maximum intensity beams at predeter- pickup (displacement of a magnetic core in a coil in-
mined spots(which correspond to normally reflected duces emf, which is amplified and then processed).
light and the specular reflected light) which ensures In order to filter off the waviness and vibrations, a
that the surface isplaced at the focal plane of the gaussian filter is used.
measuring lens. Now the angle variation of the scat-
tered light can be judged by reading the intensities as
captured by the photodiode,which in turn can used 4 Theory
to calculate surface roughness.This technique shows
Surface irregularities comprises of waviness
accurate Ra readings for 0.005µ to 6µ range[2].
and surface roughness. Waviness refers to the higher
The laser speckle method works on Fraun-
wavelength surface irregularities which generally oc-
hoffer diffraction model in which a collimated beam
curs due to inaccuracies of machining, which may be
is directed onto the test surface and uses the inten-
lack of straightness of guideways of the tool post, mis-
sity image developed on the back focal plane of the
alignment errors or non-linear feed motion. Whereas
measuring lens to find surface roughness[2]. Atomic
surface roughness comprises the finer irregularities of
force microscope makes use of molecular attraction
smaller wavelength caused due to tool vibrations, fric-
or repulsion of a nano-sized probe brought very close
tion or inappropriate tool feed rate[4]. Hence in order
to the surface, hence making it a contact measure-
to measure roughness, it is necessary to use cut-off
ment technique, with very high resolution. Confo-
length which determines both the sampling length as
cal microscopy is widely used for surface profiling.
well as waviness cut-off wavelength. A lower cut-off
Scanning electron microscope cannot give quantita-
wavelength is also provided by the instrument in or-
tive measurements directly without stereomicroscopy.
der to remove higher frequency vibration which are
Other techniques are tracer type profilo-
not required for the measurement. The ratio of the
gram which is a slight modification to inductance
higher to the lower cut-off wavelength set as the band-
based contact stylus profilers in which the stylus is
width of the filter used. The radius of probe also acts
pivots with a mirror. A collimated beam of light is di-
a mechanical filter and thus lesser the radius, finer
rected onto the mirror which upon refleecting strikes
the measurement[5].
a revolving drum with light sensitive film. Hence as
Measurement can be skid or skidless. The
the tracer moves along the surface, a magnified pro-
main advantage of the skid is it acts as a mechanical
jection of the surface is obtained on the film[3].
filter and sets a datum for measurement. Whereas
in skidless measurements, stick-slip effect can disturb
the probe movement, and the datum is set by the ma-
3 Experimental Setup chine. When the surface texture is visible to naked
eye, roughness should be measured perpendicular to
Fig 1:Taylor Hobson Profiler
lay. However, when the direction of lay cannot be as-
certained, roughness should be measured in 2 perpen-
dicular directions and the direction which provides a
higher roughness value should be considered.
Quantitative measurement of roughness
uses a mean line as a reference, which is an imagi-
nary line dividing the surface profile in such a way
that within the sampling length, the total area above
and below this imaginary line are the same. Some
important roughness parameters are listed below.
1. Ra : Average roughness.
Pn
|yi |
Ra = i=1
n
where yi represents the vertical distance of the
surface profile from the mean line for the ith
data point, summed over n divisions of the sam-
Taylor Hobson surface profiler is the mea- pling length.
suring instrument used for this experiment. It uses
2um radius diamond tip stylus which provides a res- 2. Rq :Root mean square roughness (within a sam-
olution of 16 nm or better. Traverse length of the pling length).
2
q Pn
|yi |2 5. Now proceed to Analysis where the cutoff
Rq = i=1
n
length (set according to the type machining
following same notations. process used to obtain the surface) and band-
width are set and hence the roughness of the
3. Rv :Maximum valley depth (in evaluation surface is obtained.
length).
6. This is repeated for different cutoff length and
bandwidth.
Rv = minyi
4. Rp :Maximum peak height (in evaluation
length).
Rp = maxyi
5. Rt :Maximum height of profile (in evaluation
length).
Rt = Rv + Rp
6. Rz(DIN,ASM E) :Average of the distance between
the highest peak and lowest valley in each
sampling length, evaluated over the evaluation
length.
1
Ps
Rz(DIN,ASM E) = s i=1 Rti
where where s is the number of sampling
lengths, and Rti is Rt for the ith sampling
length.
7. Rz(JIS) :Japanese Industrial Standard for
Rz ,based on the five highest peaks and low-
est valleys over the entire sampling length.
1
P5
Rz(JIS) = 5 i=1 Rpi − Rvi Fig 2 :Average surface roughness is plotted against
bandwidth for different cut-off lengths, materials with
where Rpi and Rvi are the highest peak and milled or grinded surfaces.
lowest valley respectively within a sampling
length.
6 Result and Inferences
5 Procedure Grinding gives better surface finish when com-
pared to milling . The Fig 2 also shows that varia-
1. The surface of the specimen should be cleaned. tion in roughness in a grinding process is much lesser
when compared to a milling process. The roughness
2. The direction of lay is identified and accordingly
value in a grinded steel surface is under 1µ for most
specimen is placed under the probe.
cutoff lengths and bandwidth. However for a milled
3. Inputs are fed into the profiler then probe is surface, the roughness varies from 6 − 15µ, which
positioned on the surface and appropriate load shows that the surface roughness is process depen-
is applied. dent.If we compare milled surface of steel and alu-
minum,again a clear distinction is seen in the range
4. Then through Measure and Analysis the data of values of surface roughness as aluminum is much
length (traversing length), run-up length and softer than steel and can be machined more easily.
measurement speed are set to 15mm, 0.3mm Other roughness parameters such as Rz(DIN,ASM E)
and 0.5 ms respectively. The stylus now moves ,Rz(JIS) and root mean squared roughness follow a
across the lay and the profile data points are similar trend. When we check the variation of rough-
fed to the software. ness with bandwidth, it is seen that with an increase
3
in bandwidth, the surface roughness reaches a con- the cutoff values were given only for a particular pro-
verging limit.If the cutoff length is lower then this cess without any consideration of the material.
limit can be reached for a lower bandwidth, partic-
ularly seen in aluminium milled surface. This oc-
curs because the surface roughness mainly depends 7 Conclusion
on the amplitudes of a certain dominant wavelengths,
hence only if that wavelength is seen to come un- In this experiment, it was seen that the sur-
der the bandwidth, surface roughness values will con- face roughness is higher for milled steel surface than
verge. Similar reasoning could be associated to con- milled aluminum surface and also higher for milled
vergence of roughness values at lower cutoff lengths steel surfaces than grinded steel surfaces. Hence can
(sampling length) for a given bandwidth. This wave- be concluded that surface roughness is both process
length can be reached at lower bandwidths of lower and material dependent. Surface roughness also de-
cutoff lengths or higher bandwidths in case of higher pends on the cutoff length and bandwidth chosen.
cutoff length. Steel doesn’t show any particular con- The correct range of bandwidth and cutoff length
verging limit, which could mean that a different range can be decided by checking convergence of roughness
of measurement is needed for milled steel surfaces as within the range.
References
[1] Jain, R.K, Engineering Metrology 20th edition, Khanna publishers,New Delhi,638666
[2] http://www.olympus-ims.com/en/knowledge/m etrology/roughness/
[3] C.J Tay, , S.H Wang, C Quan, H.M Shang, In situ surface roughness measurement using a laser
scattering method, Elsevier, Optics Communications, Volume 218, Issues 13, pp. 110.
[4] Whitehouse, David (2012). Surfaces and their Measurement. Boston: Butterworth-Heinemann. ISBN
978-0080972015.
[5] Shahabi, H.H.,Ratnam, M.M., Simulation and measurement of surface roughness via grey scale image
of tool in finish turning, Precision Engineering, Volume 43, January 2016, Pages 146-153.
4
8 Appendix
Fig 3:Roughness Profile for mild steel with cutoff length 0.8 and bandwidth 30
Fig 4:Roughness Profile for aluminium with cutoff length 0.8 and bandwidth 10
