euspen’s 19th International Conference &
Exhibition, Bilbao, ES, June 2019
On-machine optical surface topography measurement sensor based on focus
Subbareddy Darukumalli 1,2, Teguh Santoso 1, Wahyudin P. Syam1, Franz Helmli2 and Richard Leach1
1Manufacturing Metrology Team, University of Nottingham, NG8 1BB Nottingham, UK
2Alicona Imaging GmbH, Dr. Auner Straße 21a, 8074 Raaba, Austria
Wireless technology is becoming one of the key technologies that allows an increase in the productivity of a manufacturing process.
In this paper, a compact on-machine areal surface texture and form measurement sensor based on focus variation is presented.
Considering the feasibility of operating in real-time inside a production machine, the current prototype has been designed with the
dimensions of 80 mm (diameter) × 200 mm (length) and the typical time for a single measurement is less than 20 s. The instrument
design is presented. The performance of the on-machine sensor has been compared with measurement results acquired by a
benchtop focus variation instrument. Additionally, the current prototype real-time wireless signal strength has been studied in a
milling machine and the results are presented.
In-process metrology, focus variation, surface texture, surface form
New manufacturing trends in precision manufacturing require
in-process quality control (QC) of parts for every manufacturing
step to control tolerance stack-up. The reduction of feature
dimensions requires tighter manufacturing tolerances up to the
sub-micrometre level. These features require appropriate
measuring instruments for QC .
In-process monitoring can produce the highest profit gains, in
terms of reduced capital costs, increased automation, reduced
human error and, most importantly, less scrap and rework. Due
to the complex nature of this manufacturing stage, in-process
monitoring is the least exploited and understood QC monitoring
stage. Recently, on-machine measuring instruments are
becoming an option on state-of-the-art machine tools .
Non-contact optical areal measuring instruments are
increasing in versatility. In contrast to traditional tactile
measuring instruments, non-contact optical instruments have a
number of advantages, such as relatively high measurement
speeds, no surface contact and small feature accessibility .
Non-contact areal topography measuring instruments based
on focus variation (FV) have become increasingly popular .
Optical measuring instruments based on FV are less sensitive to
environmental noise (for example, vibration) that are commonly
found in a production environment. A customized optical
measuring instrument based on FV is a potential method to be
used for on-machine and in-process QC in production .
In the context of the newly evolving Industry 4.0 framework
, which includes the use of cyber-physical systems, wireless
technology is becoming one of the key technologies. This allows
an increase in the productivity of a manufacturing process with
automation and wireless monitoring.
In this paper, we present a brief overview of the FV method
(Section 2), the design concept of the on-machine FV optical
instrument (Section 3), measurements on a calibrated artefact
(Section 4) and wireless performance of the prototype in a real-
time milling machine (Section 5) and finally conclusions (Section
2. Focus variation
The FV method combines the small depth of focus of a
microscopic optical system with vertical scanning to provide
topographical and colour information from the variation of focus
[4, 7]. The main component of the FV instrument is a precision
optical unit containing various lenses. With different microscopy
objectives, measurements can be obtained with different
resolutions. The FV method can also deliver colour information
with full depth of field, which is mapped to the 3D points.
3. Design concept
Figure 1. A conceptual diagram of the on-machine FV sensor.
Considering the feasibility of the sensor for operating inside a
production machine (on-machine, see  for definitions), the
current prototype has been designed with the dimensions of
80 mm diameter × 200 mm length. It has a field of view of (1.7 ×
1.7) mm with a 10× objective magnification lens. The instrument
has a maximum scanning range of 20 mm with a 20 nm vertical
resolution. The measuring time for single measurements is less
than 20 s.
Figure 1 shows the conceptual diagram of the developed on-
machine optical surface topography sensor. It has firstly, a
precision single-axis linear motion stage and its motion
controller, secondly, an optical setup and illumination system,
and finally, computing electronics and WiFi system.
Figure 2. (a). Optical setup mounted on the axis (top) the optical
system is in-line with the linear guides (bottom). (b). Optical setup
mounted to the axis with a distance using linear guides (top) the optical
system is not in-line with the linear guides (bottom).
As shown in Figure 2 (a), the optical setup is mounted on the
axis in between the linear guides to reduce the Abbe error. The
positioning of the precision linear stage is controlled by using a
motion controller with CAN bus (controller area network)
protocol peer-to-peer communication. During vertical scanning,
image data is continuously captured and each image position is
read from a high accuracy linear encoder.
The current prototype design is in Figure 3. The optical setup
of the prototype is made up of a high-speed USB 3.0 interface
CMOS sensor with specially designed optics and a 10×
microscope objective lens, which is mounted on the axis. A ring
light is mounted around the objective lens for illumination of the
The captured image data is processed on a mini computer
built-in the sensor. Measurement settings and results are
monitored using a wireless display or the machine integrated
display. The sensor is powered using a rechargeable battery or
with a direct power connection. For further detailed evaluation,
the measurements data is moved to a remote workstation or
server via a WiFi connection or alternatively via a cable.
The sensor, thanks to its compact design, can be mounted (see
Figure 3 mounting threads on the top) into various types of
manufacturing machines (for example, additive manufacturing
machine, micro-scale milling and micro-scale injection
moulding) with a specific adapter or mount (e.g., in CNC
machine: HSK-A63, SK-40, HSK-A100 and in micro milling
machine: HSK-E40, HSK-A63, HSK-E32).
Figure 3. Current prototype design.
With the wireless communication and rechargeable battery,
the sensor can be installed into an automatic tool changer
system on a milling machine and clamped by its spindle to
perform on-machine measurements. Measurements can be
performed during (in-line) or after (in-process) the
manufacturing process and the instrument can be operated
automatically or manually using a wired or wireless connection.
The sensor is enclosed by a specially designed case to protect is
from contamination during the machining operation. However,
when the case is open during the measurement, the objective is
exposed to the contamination.
An example of a step by step working process of the sensor in
a closed-loop milling production machine is described as follows.
1. A milling machine is turned on and a milling process is in
operation on a part.
2. The automatic tool changer of the machine will pick the
sensor from its tool magazine.
3. The sensor is positioned on the part and measures the
desired part’s surface. This is done by using the milling
machine software that can control the tool changing
4. The captured areal surface measurement data is processed
locally on minicomputer built-in the sensor.
5. The areal surface measurement results can be used for QC
of the part, for optimizing the future milling processes and/
or for compensating the error of the milling on the current
6. Once the measurement is finished, the sensor is placed
back into the tool magazine. Where, the battery will be
recharged using a built-in wired/ wireless charging outlet.
4. Measurements of a calibrated artefact
A calibration artefact (Alicona IF-Calibration tool) with a step
height of 1 mm (DKD calibration value is 0.9999 mm ±
0.0001 mm at k = 2) was used as a measurement artefact.
Alicona software was used for the areal evaluation, and an
Alicona InfiniteFocusSL instrument  was used as the reference
benchtop FV instrument.
Figure 4. Step height measurement with the 10× objective and with
0.0001 mm vertical resolution.
Table1. Step height measurement.
The measured areal data of the step height artefact can be
seen in Figure 4. After the measurement, a levelling on the lower
surface and a profile extraction was performed. These profiles
were used to fit two parallel lines by a least-squares method. The
step height measurement results are shown in Table 1. From
Table 1, the measurement result shows a 20 nm deviation from
the reference value (which is still within the uncertainty of the
reference value). However, the uncertainty of the prototype
measurement is six times higher than that of benchtop FV
instrument and fortunately, we have managed to obtain a step
height measurement within the reference uncertainty in the
initial prototype evaluations.
The instrument measurement noise is calculated by using the
subtraction of two repeated measurements of the calibration
artefact (difference method ) and by using the proprietary
estimation of the measurement repeatability value provided by
the software (estimated repeatability method) .
Table2. Instrument measurement noise estimation.
The measured areal data using a 10× objective magnification
with 300 nm vertical resolution has been used for the
instrument measurement noise estimation results shown in
Table 2, which shows that the prototype instrument
measurement noise is approximately three times higher than
the reference benchtop FV instrument. This is because the optics
on the prototype sensor are not yet corrected for aberrations
and distortion. A large reduction in measurement noise of the
prototype sensor is expected to be comparable with the
benchtop system after implementing a full optical calibration.
5. Real-time wireless performance monitoring of the prototype
in a milling machine
To test the signal strength performance of the WiFi
connection, a prototype setup has been placed in the machine
chamber of HURCO VMX24t 5-axis milling machine. The WiFi
signal strength was monitored using a commercial iSSIDer Wifi
scanner application running on a laptop.
Table3. WiFi prototype monitoring in a 5-axis milling machine.
As shown in Table 3, the prototype has been tested in three
different scenarios: the machining chamber door was open, was
closed when the machine was not operating and while the
machine is in operation. From the results in Table 3, no
significant network delay was reported. In all the three
scenarios, with an average of 36.35 Mbps uploading speed,
1.71 Mbps downloading speed and average of -25.3 dBm
(decibel-milliwatt) signal strength are obtained. The WiFi
network data throughput speed and signal strength are
calculated using the IEEE 802.11n WLAN standard .
In this paper, a focus variation based prototype sensor for on-
machine areal surface topography measurement is presented.
Initial measurement accuracy evaluations with a step height
artefact measurement show deviations of 20 nm with respect to
the reference value. Due to the lack of flatness deviation
adjustment, optical adjustment for aberration and distortion,
the instrument measurement noise is approximately three times
higher than that of the benchtop FV instrument. Consistency in
the wireless network uploading and downloading speed and
signal strength shows no noticeable network delay for data
transfer during a milling process. Future work will be to correct
for flatness deviation, optical aberrations and distortion, and a
comprehensive test of the developed sensor and quantification
of measurement uncertainty of the results.
This work has received funding from the European Union
Horizon 2020 Marie Skłodowska Curie ITN projects PAM^2 and
MICROMAN under the grant agreement numbers 721383 and
 Gao W, Haitjema H, Fang F Z, Leach R K, Cheung C F, Savio E and
Linares JM 2019 On-machine and in-process surface metrology for
precision manufacturing. Annals CIRP. (to be published)
 Zhao F, Xu X and Xie S 2009 Computer aided inspection planning-the
state of the art. Computers in Industry, pp 453–466.
 Piska M and Metelkova J 2014 On the comparison of contact and
non-contact evaluations of a machined surface. MM Science
Journal, pp 476—480.
 Danzl R, Helmli F and Scherer S 2011 Focus Variation–a robust
technology for high resolution optical 3D surface metrology.
Strojniški Vestnik – Journal of Mechanical Engineering, pp 245–256.
0.99997 ± 0.003
1.00012 ± 0.0005
0.99999 ± 0.00001
 Danzl R, Lankmair T, Calvez A and Helmi F 2017 Robot solutions for
automated 3D surface measurement in production. 18th
International Congress of Metrology, Paris, France.
 Lars G, Arno K, David R, Paul M, Christoph M, Sebastian S, Dania D,
Singh L, Julie K and Matthew S 2015 A discussion of qualifications
and skills in the factory of the future: a German and American
perspective. VDI/ASME Industry 4, pp 1-28.
 Duboust N, Ghadbeigi H, Pinna C, Soberanis S A, Collis A, Scaife R,
and Kerrigan K 2017 An optical method for measuring surface
roughness of machined carbon fibre-reinforced plastic composites.
Journal of Composite Materials, pp 289-302.
 ISO/CD 25178-700 2018. Geometrical product specification (GPS) --
Surface texture, Areal -- Part 700: Calibration, adjustment and
verification of areal topography measuring instruments
(International Organization for Standardization working document).
 Danzl R, Helmli F and Scherer S 2007 Automatic measurement of
calibration standards with arrays of hemi-spherical calottes. 11th
International Conference on Metrology and Properties of
 IEEE 802.11n-2009 - IEEE Standard for Information technology--
Local and metropolitan area networks-- Specific requirements-- Part
11: Wireless LAN Medium Access Control (MAC)and Physical Layer
(PHY) Specifications Amendment 5: Enhancements for Higher
Throughput. (Institute of Electrical and Electronic Engineers)