e-ISSN: 2146 - 9067

International Journal of Automotive

Engineering and Technologies

journal homepage:

https://dergipark.org.tr/en/pub/ijaet

Original Research Article

External torque sensor design providing wireless and real-time

data customized for drivetrain
Sedat Tarakçı1, * , Oğuzhan Aldemir2 , Turan Solmaz3 , Efe Işık4
1, 2, 3, 4
Tirsan Kardan A.S, Tirsan Kardan R&D Center, M anisa, Turkey.
1
İzmir Institute of Technology, Department of M echanical Engineering, İzmir, Turkey.
ARTICLE INFO ABSTRACT
Orcid Numbers In this study, a torque sensor design that can be used in both laboratory and field
1. 0000-0002-9125-5919 tests simulating road conditions of drivetrain of the vehicles is explained.
2. 0000-0002-3471-5034 Mechanical design of the sensor has been carried out via finite element method
3. 0000-0001-9166-9241
4. 0000-0001-7657-317X (FEM). Material of the torque sensor and the relevant heat treatment were
Doi: 10.18245ijaet.982530 determined. Electronic measurement circuit design and its setup have been
* Corresponding author carried out including amplifier, filter, ADC and data transmission features which
s.tarakci@tirsankardan.com.tr are suited for full bridge connection. Verification and calibration tests have been
Received: Aug 13, 2021 performed with respect to DIN-51309. As a result of the studies, it was
Accepted: Apr 14, 2022
determined that the measurement accuracy of the system in the torque range of
Published: 01 May 2022
250 ÷ 2500 Nm was 0.5%. The effect of noise-induced deviations was found
Published by Editorial Board
Members of IJAET 0.05% for the precision value. Consequently, a verified universal torque sensor
© T his article is distributed by has been developed for any drivetrain application under torsional effect with the
T urk Journal Park System under given specifications. It is hoped that this sensor might become a common feature
the CC 4.0 terms and conditions.
on any and every drivetrain for advanced monitoring of the drive components.
Keywords: Torque, driveshaft, data acquisition, calibration

1. Introduction and automotive sensors continue to grow, torque

sensors have still limited usage [2]. Torque
Automotive industry has increasingly become
sensor is a device to measure the torque on a
more of a sensing and measurement venue. Cars
rotating system [3]. Driveshaft and other
are turning into sensing and computing
platforms. Naturally, more of the automotive drivetrain members are one of the measureme nt
interest for torque sensors.
components are interconnected, and hence
Main reasons for torque measurement are to
communicating with both on-board computers
control mechanical power for industrial reasons
and the drivers. These interconnected vehicle
and to improve design studies. Enginee r ing
components are being easily monitored over
measurements play a significant role in the
their streamed data on statistics, health
product development process. Examining the
monitoring, tracking etc basis [1].
parts that make up the vehicle both provides
As computer controlled car systems have gained
information about product performance and
importance, different sensors have an important
provides input for future designs.
role of providing information on the controllers.
Drivetrain members are subjected to verifica tio n
Although, the sensor numbers are getting higher
19 International Journal of Automotive Engineering and Technologies, IJAET 11 (1) 18-27

tests prior to design release [4]. Torque is one of has been done for this force measuring element
the important variable that is measured in order so that it can work in the outdoor environme nt
to understand the limits of the product. This [9].
knowledge provides an opportunity to Niedworok presents an innovative approach to
characterize the mechanical limits of the product measure and monitor the dynamic torque
[5]. changes in drive systems of the vehicles during
A driveshaft may be overdesigned or selected their natural operation [10].
improperly according to the type of vehicle. Sanponpute et al. designed a torque measuring
This situation affects driveshaft’s costs and device by using a loadcell. They verified their
vehicle’s weight in negative manner. An study by finite element analysis and
opportunity to collect data of transmitted torque experimental results [11].
of the driveshaft gives a chance to improve the In this study, a torque sensor design that can
design or to choose a suitable product [6]. provide input for the development processes of
As with all parts on the vehicle, the power drivetrains in vehicles, can be used in tests
transmission parts are also open to development. simulating road conditions, and collects torque
Power transmission parts are labored depending data in fixed or rotating positions is explained.
on the vehicle's transmission gear ratio, road The sensor, which is the subject of this study, is
conditions or vehicle trailer load. The said designed to be resistant to external effects and
stresses are observed as a high torque effect on impermeable to use in road conditions. Torque
these parts. At this point, the monitoring of the data collected by the sensor in real time can be
torque effect emerges as a necessity during the read wirelessly by mobile devices or computer.
operation in order to ensure appropriate and 2. Mechanical Design
acceptable design.
Within the scope of this need, a study by [Joel Torque is one of most important physical
C. Schnackenberg under the name of "Wheel parameter for the drivetrain. The power
Load Transducer"] is a torque measureme nt transmission capacity of the drivetrain depends
research. In the study carried out in 2002, the on the torque that can be transmitted. Therefore,
power transmission is made with the help of an design of driveshaft is done according to the
adapter connected to the vehicle wheel hub. capacity of torque. When torque is applied,
Thanks to this adapter, forces coming from tension occurs on the driveshaft and the
different directions to the wheel hub are resulting stress determines the capacity of
detected with the help of beams designed on driveshaft [12]. If the applied torque is out of the
different axes on the adapter [7]. limits, the driveshaft may be damaged. If the
Werner Nold's 2002 patent study "Recorder for torque is kept applying, that influences the
Measuring applied forces has a rotationa lly product life [13].
symmetric body with measurement legs The stress is calculated as follows.
arranged between holes on which strain gages TR
𝜏𝑚𝑎𝑥 = J
(1)
are placed to measured shear forces, which are
intensitive to load placement" The mechanica l 𝜋𝑑4
J= (2)
system, which is symmetrically connected, 32

designed to measure at the same time, and the As shown in Figure 1 when the torque T apply
mechanical system consisting of specially to a bar, twisting occurs and angle of twist (θ)
designed measuring points and directions are can be calculated by equation (3) [14, 15].
shown. In this study, forces in different axes are TL
𝜃= J (3)
detected and recorded [8].
In the patent study titled “Torque Sensor Having Metallic materials are flexible according to their
Sealing Membrance” by Michael Grandl et al. in working principle. When the torque is applied
2017, the basic torque measurement body and an metallic material, elastic strain is occurred.
inner flange design with the force input are Strain gauges can measure amount of strain on
made. Attenuated mounting sections are the metallic material so stress can be calculated.
designed for sensors that generate output signals In this case, if the torque is applied to a
in mechanical design. In addition, sealing work cylindrical geometry, amount of stress and
International Journal of Automotive Engineering and Technologies, IJAET 11 (1) 18-27 20

torque can be calculated. between two points, the strain gage is attached
to the circular surfaces as follows [3].

Figure 4. Strain gage application [17]

The purpose of the torque sensor is to measure
torque over torque-transmitting parts such as the
driveshaft. The measurement is taken by
adhering strain gages to the torque sensor. At the
same time, this sensor can be mounted to the
component to be measured and the torque of that
Figure 1. Torque applied to a bar [15] component can be measured. The geometry of
the part to be measured is designed according to
When a rod is pulled, an elongation of ∆A
the strain gage working mechanism. Strain gage
occurs. Thus, L extends from its original length
measurement mechanism for torque
to L+∆L.
measurement is the formation of minus and plus
The ratio of this elongation (or shortening) ∆L
orthogonal stress distribution as in figure 3. The
to the original length L is called strain and is
orthogonal stress distribution should be as in
denoted by the sign ε (epsilon) [16]. Figure 2
Figure 5. As indicated in Figure 4, measureme nt
shows an elongated rod.
is taken by connecting strain gages. The
geometry of the part to be measured is designed
to obtain a stress distribution like in figure 3.

Figure 2. Rod under deformation [16]

When torque is applied to cylindrical geometry,
as like in figure 3, a homogeneous field of
orthogonal stresses is occurred. The stress can
calculate the following formula [17].

Figure 5. Orthogonal stress distribution on a pipe

Design& finite element analysis (FEA) studies
were continued until desired stress distributio n
has been obtained. The part in figure 6 was
designed according to further design & analysis
studies. The analysis results show in figure 7
that stress distribution is similar to the desired
one and homogeneous stress distribution was
obtained.
According to the finite element results, the
design in figure 6 was chosen. The maximum
stress values that will occur on the part under the
nominal torque are limited to 30% of the tensile
strength of the part and the part dimensions have
Figure 3. Stress distribution [3] been determined accordingly the design analysis
2τr cycle.
𝜎= (4)
𝜋𝑅4 4340 materials were selected to produce the
Since the strain gage measures the change design. According to ISO 18265, the hardness
21 International Journal of Automotive Engineering and Technologies, IJAET 11 (1) 18-27

value was selected according to the desired 3. Electronic Design

tensile strength from the conversion table for Suitable mechanical design has been completed
low alloy steels [18]. The selected hardness for observing the strain flows clearly. In order to
value was measured as in figure 8 by taking a measure the torque data, strain gauge
section from the heat treated geometry. The positioning have to be on strain flows direction.
resultant hardness value was between HRC 34 For this reason, strain gauges were placed at 45
to 35. degrees to the axis of the mechanical design part
and 90 degrees between themselves. The
locations and axial positions were determined
and marked before to make the instrumenta tio n
correct. Figure 9 shows the electronic
application of strain gages.

Figure 6. Final design of torque sensor

Figure 9. Half bridge application [19]
A printed circuit was designed for half bridge
measurement method. Half bridge applicatio n
includes two single strain gauge for torque
measurement. There are two 1000 Ohm resistors
parallel to strain gauges on the PCB. Resistors
will be a reference for strain gauges and there
will occur potential voltage difference [20]. This
voltage is a very low level signal. The signal is
amplified by a commercial amplifier before
signal processing. Amplified signal is processed
on the microcontroller and converted into a
digital signal [21]. Digital signals, which are
Figure 7. FEA results
interpreted by mathematical operations on the
microcontroller, are transferred to the wireless
communication section of the PCB. Wirelessly
broadcast data is made available to
communication devices such as hand termina ls,
smart phones, tablets or computers. Bluetooth
communication method was used to transmit
data. A bluetooth module which uses SPP (serial
port profile) was choosen. SPP imitates the
RS232 serial connection and is used to send and
receive data between two devices [22]. In the
circuit diagram, the microcontroller supplied
digital data from its TX port to the RX port of
the bluetooth module. Transmitted data is
Figure 8. Hardness measurement after heat treatment broadcasted wirelessly.
International Journal of Automotive Engineering and Technologies, IJAET 11 (1) 18-27 22

The PCB’s schematic and printing design was the full bridge connection, four strain gages
made on Proteus Professional software. Printing have been glued on the driveshaft axis at 45°
design and prototype of PCB is shown the figure angle. Connections with electronic circuit have
10. The 50 mW power required by the PCB is been carried out. The test setup is shown in
provided by lithium- ion batteries. Figure 12.

Figure 10. PCB design and prototype

The resistance of strain gage changes depending Figure 11. Comparison graph difference between half
and full bridge
on the length of conductive wires inside the
strain gauge. The change of the resistance
affects the voltage of branches and creates a
potential voltage difference. In the half bridge
method described, the values of the 1000 ohm
resistors were used as a reference. For this
reason, the voltage difference is quite narrow
and contains noises [21].
Full bridge method was applied as second
approach to torque measurement. Here, Figure 12. Test setup
decreasing the noise from collected data and to Similarly, to the symmetry of the region where
improve the sensitivity were aimed [23]. Two the strain gages had been applied in the torque
strain gauges of the same ohm value with sensor, a new set of strain gages were adhered
reference resistors were used instead reference and prepared for data collection with verified &
resistors used in half bridge method. Resistors, calibrated commercial data acquisition board.
holes and paths on the PCB were updated and it The application can be seen in figure 13.
has been suited for full bridge application.
A more stable signal was obtained with the full
bridge application. The processes followed by
the new generated signal are the same as the half
bridge method. However, the sensitive voltage
difference created by the full bridge changes the
noise levels of the collected data. In figure 11,
the measurement results obtained with full
bridge and half bridge methods. It was clearly
seen that the full bridge was more stable with
low noise level.
Figure 13. Strain gages dedicated for commercial board
4. Verification Tests
According to the test plan, the system was
The Torque sensor, which is developed for preloaded 3 times and then let the system has
calibration and verification tasks, has been been stabilized. Afterwards, calibration has
subjected to a branch of tests according to a been started. Torque level was gradually
verification plan. DIN 51309 standard was used increased from 0 to 2500 Nm for 500 Nm steps
as a reference [24]. in clockwise direction. The value read from the
The torque sensor is mounted between the sensor has been recorded at each step.
driveshaft and the test rig. In accordance with Afterwards, torque was applied in counter
23 International Journal of Automotive Engineering and Technologies, IJAET 11 (1) 18-27

clockwise direction from 0 to 2500 Nm for 500 Similarly, the calibration of the torque sensor
Nm steps. Applied torque values and the raw was also performed with the strain gages
values read from the sensor are given in Table 1. associated with commercial data acquisitio n
Table 1 Calibration values board. In Table 2 the applied torque values and
Applied Sensor Data related voltage values can be seen.
Torque (Nm) (Unit less) Table 2 Calibration values of commercial system
-2500 9611500 Applied Calibration
-2000 9240500 Torque (Nm) values (mV/V)
-1500 8868250 -2500 1.421
-1000 8495250 -2000 1.179
-500 8122250 -1500 0.933
0 7748500 -1000 0.675
500 7374750 -500 0.409
1000 7001750 0 0.152
1500 6629000 500 -0.121
2000 6257500 1000 -0.383
2500 5872500 1500 -0.644
2000 -0.894
When a graph was drawn with these values,
2500 -1.139
information such as the equation and linearity of
the curve has been obtained. In the graph shown A graph was drawn with the collected values
in Figure 14, it is seen that the second order term and information such as the equation and
is very small and negligible, and the calibratio n linearity of the curve has been obtained. In the
curve is linear. graph shown in Figure 15, it is seen that the
second order term is very small and negligib le,
and the calibration curve is linear.

Figure 14. Calibration graph

Figure 15. Calibration graph of commercial system
The torque sensor has been calibrated by
After the calibration, verification tests were
introducing the sensor data obtained during the
started. In the validation tests, intermediate steps
calibration procedure to the developed software.
are also added. According to the test scenario,
A linear equation is defined for calibratio n
torque level was gradually increased from 0 to
between each calibration interval.
2500 Nm for 250 Nm steps in clockwise
Even though figure 14 shows that a curve with
high linearity is achieved in calibratio n direction. There are very short delays between
steps regarding the continuous procedure [24].
procedure, calibration equations are constituted
It was also observed that the sensor was
between each applied torque intervals. The fact
stabilized quickly. After reaching the value of
that the torque sensor software allows to be
2500 Nm, it was decreased to 0 Nm with 250
defined calibration equation from multip le
Nm steps. Afterwards, 250 Nm steps were taken
points increases the system accuracy and
up to 2500 Nm in counterclockwise direction.
minimizes the problems related to linear ity.
After reaching 2500 Nm, it has been decreased
Calibration formula is given in equation (5).
to 0 again with 250 Nm steps. Collected data
s−𝑆𝑟𝑖
𝑇𝑟 = (𝑆 (5) from the sensor during the test is shown in figure
𝑟 (𝑖+1) −𝑆𝑟𝑖 )÷(𝑇𝑎 (𝑖+1) −𝑇𝑎𝑖 )
16.
𝑓𝑜𝑟 𝑖 = 1 𝑡𝑜 𝑛, 𝑛 ∈ 𝑍 + Simultaneously, torque verification tests have
International Journal of Automotive Engineering and Technologies, IJAET 11 (1) 18-27 24

been performed with commercial system with during the test.

the strain gages attached to the symmetr ica l
arms of the sensor. The collected data from both
systems were compared. Difference between the
value read from the sensor and the applied
torque value were compared according to
equation (6) for each point.

Figure 17. Torque measurement with commercial system

verification test

Figure 16. Torque sensor verification tes t

𝑇𝑟𝑖 −𝑇𝑎 𝑖
%𝑎𝑐𝑐𝑢𝑟𝑎𝑐𝑦 = 𝑥100 (6)
𝑇𝑎 𝑖
𝑓𝑜𝑟 𝑖 = 1 𝑡𝑜 𝑛, 𝑛 ∈ 𝑍 +
It is seen that there is a deviation of less than
0.5% between them.
Precision of the sensor was calculated for full
bridge strain gage arrangement. A specific
torque value was applied for a certain time
period. The applied torque value was compared
with the highest and the lowest value read by the
sensor for that time interval. The ratio provides
the precision level of the sensor. The deviation
is caused by noise-induced effects. Figure 11
shows the deviations of the sensor for a limited
time period. Equation (7) represents the
approach to calculate the noise level. The effect
of noise-induced deviations was found 0.05%.
𝑇𝑟,𝑚𝑎𝑥 −𝑇𝑟 ,𝑚𝑖𝑛
%𝑝𝑟𝑒𝑐𝑖𝑠𝑖𝑜𝑛 = 𝑥100 (7)
𝑇𝑎 𝑖
+
𝑓𝑜𝑟 𝑖 = 1 𝑡𝑜 𝑛, 𝑛 ∈ 𝑍
The graph of the data collected by the
commercial system is given in figure 17. Since
the data acquisition frequencies of the
commercial system and the system developed Figure 18. Test conditions
within the scope of this study are different from
each other, the x-axis of the graph cannot be When the verification tests were completed, a
unified and both data cannot be shown in the validation test was also performed for the design
same graph. & FEA cycle of the system. A test has been
As a result, a comparison and verification study performed to determine the stress distributio n
has been carried out both with the system and stress levels in order to understand relation
developed within this study, with a commercia l between the FE analyzes and the real conditions
system and with the torque values applied by the and to determine the strength level of the torque
verified test device. Figure 18 shows that data sensor. A strain gage was attached to a suitable
was read from all three systems simultaneo us ly point on the sensor to collect strain data. Figure
25 International Journal of Automotive Engineering and Technologies, IJAET 11 (1) 18-27

19 shows the application. The information about the model is given in

figure 21.
The stress value of the region where the strain
gage was attached was found as 128 MPa at a
torque of 3500 Nm in the finite element model.
When the two models were compared, deviation
of 8% has been seen.
5.Conclusion
Prototype of the developed torque sensor has
been produced regarding to design and analysis
studies. Likewise, related electronic circuit
design and production has been carried out. It
Figure 19. Design validation test was observed that accuracy errors and noise
From this point, strain data were collected again levels decreased with the full bridge applicatio n.
with a commercial system. Torque was applied The fact that the torque sensor software allows
to be defined calibration equation from multip le
up to 3500 Nm both clockwise and
counterclockwise with gradual increments. The points increases the system accuracy and
collected strain values were converted into minimizes the problems related to linear ity.
equivalent stress values and given in figure 20. Calibration of the sensor has been performed on
a test device which is calibrated in an accredited
institution. In the subsequent verification tests,
the accuracy of the system was demonstrated by
comparing the measurements performed with
the developed system both with the test rig and
with a commercial system simultaneously.
The measurement accuracy of the system in the
torque range of 250 ÷ 2500 Nm was calculated
as 0.5%. The effect of noise-induced deviatio ns
was found 0.05% as the precision value.
In the test carried out for the design validatio n
of the system, it was determined that there was
Figure 20. Equivalent stress graph
an 8% deviation between the finite element
model and the prototype. The model was
considered suitable as it was within the 10%
margin of error accepted for the finite element
method.
Consequently, a verified torque sensor has been
developed for any application under torsional
effect with the given specifications.
Acknowledgement
We would like to acknowledge the support
Figure 21. Finite element results of the torque sensor provided by Tirsan Kardan R & D center and
test center during our experimental work.
When the collected data were examined, the
stress value of the region where the strain gage CRediT authorship contribution statement
was attached was measured as 118 MPa at a Sedat Tarakçı: Conceptualizatio n,
torque of 3500 Nm. methodology, writing-original draft, validatio n,
The region where the strain gage was attached investigation, formal analysis, Project
was determined on the finite element model. administration. Oğuzhan Aldemir: Validatio n,
The stress occurred at this point under the torque writing- original draft, Data Curation. Turan
of 3500 Nm has been figured out on the model. Solmaz: Methodology, investigation. Efe Işık:
International Journal of Automotive Engineering and Technologies, IJAET 11 (1) 18-27 26

conceptualization, supervision, Writing - 8. N. Werner, “Recorder for measuring

Review & Editing applied forces has a rotationally symetric body
with measurement legs arranged between
Declaration of Competing Interest
through holes on which strain gages are placed
The authors declare that they have no known to measure shear forced, wgich are insensitive to
competing financial interests or personal load placement”, DE10055933A1, 2002.
relationships that could have appeared to 9. M. Grandl, “Torque sensor having a
influence the work reported in this paper. sealing membrance”, CN109661566A, 2019.
Nomenclature 10. A. Niedworok, “Dynamic Torque
Measurement of Drive Transmission in
𝜏𝑚𝑎𝑥 Maximum shear stress Vehicles Being in Motion Using Strain Gauges
T Torque Measurement Technique and Wireless Data
R Radius Transmission”, 16th European Conference on
J Polar moment of inertia Power Electronics and Applications Conference
L Length Book,1-8, 2014.
𝜎 Normal Stress 11. T. Sanponpute, C. Watthong, “Prototype
6. References of Load Cell Application in Torque
1. O. Aldemir, S. Tarakçı, T. Solmaz, E. Measurement”, IMEKO 2010, 285-289, 2010.
Işık, “External system design for real time 12. B. S. H. Tonin, Y. He, N. Ye, H. P.
monitoring of temperature and torque data on Chew, A. Fok, “Effects of tightening torque on
the driveshaft”, 10th International Automotive screw stress and formation of implant-abutme nt
Technologies Congress Book, 769-776, 2020. microgaps: A finite element analysis”, The
2. G. Persson, O. Persson,” Torque Sensor Journal of Prosthetic Dentistry, 2021.
for Automotive Applications”, Lund Univers ity, 13. J. Bai, X. Wu, F. Gao, H. Li, “Analys is
2015. of powertrain loading dynamic characteristics
3. M. Hilal Muftah, S. Mohamed Haris, K. and the effects on fatigue damage”, Applied
Petroczki and E. Awad Khidir, “An Improved Sciences, Volume 7, Issue 10,1-15, 2017.
Strain Gauge-Based Dynamic Torque 14. R. G. Budynas, J. K. Nisbett, “Shigle y's
Measurement Method”, International Journal of Mechanical Engineering Design_TextBook ”
Circuits, Systems and Signal Processing, Mcgraw-Hill, 2011.
Volume 7, Issue 1, 66-73, 2013. 15. L. Mendoza and E. Yuen, “Diseño de un
4. D. Danesin, C. Girardin, A. Sorniotti, A. prototipo de máquina trituradora de botellas
Morgando, M. Velardocchia, “Driveline layout PET”, Peru Technology University, 2019.
influence on four-wheel drive dynamics”, SAE 16. https://www.ni.com/en-
Transactions, Section 6: Journal of passenger tr/innovations/white-papers/07/measuring-
cars: Mechanical systems, Volume 113, 534- strain-with-strain- gages.html,13/8/21.
541, 2004. 17. Y. Chen, “Development of highly
5. C. Chen, T. Ma, H. Jin, Y. Wu, Z. Hou, magnetostrictive composites for applications in
F. Li, “Torque and rotational speed sensor based magnetomechanical torque sensors”, Iowa State
on resistance and capacitive grating for University, 1999.
rotational shaft of mechanical systems”, 18. ISO 18265 Metallic materia ls
Mechanical Systems and Signal Processing, Conversion of hardness values.
Volume 142, 106737, 2020. 19. https://www.ni.com/en-
6. P. Bingham, S. Theodossiades, T. tr/innovations/white-papers/07/measuring-
Saunders, E. Grant, R. Daubney, “A study on strain-with-strain- gages.html, 1.8.21.
automotive drivetrain transient response to 20. D.M. Ştefănescu, “Wheatstone Bridge -
‘clutch abuse’ events”, Proceeding of the The Basic Circuit for Strain Gauge Force
Institution of Mechanical Engineers, Part D: Transducers. In: Handbook of Force
Journal of Automobile Engineering, Volume Transducers.”, Springer, 2011.
230, Issue 10, 1403-1416, 2015. 21. P. R. Nagarajan, B. George, V. J. Kumar,
7. J. C. Schnackenberg, “Wheel load “A Linearizing Digitizer for Wheatstone Bridge
transducer”, US6439063B1, 2002. Based Signal Conditioning of Resistive
27 International Journal of Automotive Engineering and Technologies, IJAET 11 (1) 18-27

Sensors”, IEEE Sensors Journal, Volume 17,

Issue 6, 1696 – 1705, 2017
22. P. Tawadros, M. Awadallah, P. Walker,
N. Zhang, “A low-cost bluetooth torque sensor
for vehicle jerk and transient torque
measurement”, Proceedings of the Institution of
Mechanical Engineers, Part D: Journal of
Automobile Engineering, Volume 234, Issue 2-
3, 2020
23. D. Marioli, P. Rolla, A. Taroni, “Strain
gauge transducers: an evaluation of accuracy
limits”, Measurement, Volume 10, Issue 3, 98-
104, 1992.
24. DIN 51309 Materials testing machines -
Calibration of static torque measuring devices.