Page 72 - ATZ11 November 2019 Professional

RESEARCH GEARBOXES

or angular error, small transverse forces will result despite the
compensating coupling [2]. The resulting transverse force-related
falsifications of the torque are mainly caused by bending of the
torque sensor and extra bearing forces in the gearbox.
In order to avoid friction and falsifications, high-quality torque
sensors operate without bearing of the measuring shaft. This
means that the measuring shaft rotates freely in a narrow air gap
between rotor and stator. This results into two requirements:
– The compensating elements (couplings) must not be too flexi-
ble. Otherwise the sensor shaft could sag down by its own
weight. This would result in friction and abrasion between sen-
sor shaft and stator [2].
– There will also be friction and wear between the sensor shaft
and the stator if the axle offset and angular misalignment are
too large. To reach the highest possible measuring quality, it is
necessary to use an adjustment concept in addition to suitable
compensation elements.
FIGURE 3 shows the structure of an adjustable measuring slide
developed. The adjustment concept enables fine adjustment of
the motor/brake position to the sensor as well as the pre-adjusted FIGURE 3 Universal measuring slid (© UAS South Westphalia)
drive sensor unit or brake sensor unit to the gear shaft. The axle
offset and the angular failer are fine-adjusted in x- and y-direction
via adjusting screws. The setting process is divided into the fol-
lowing steps: than 0.1 mm [3]. If higher alignment qualities have to be
– First, the torque and speed sensor (A) is fixed to the intermedi- achieved, an appropriate measuring system [2, 4] is
ate plate (B) of the measuring slide. The axis of the sensor builds recommended.
the reference for the alignment of the drive unit (C) or the brake – The same options are available for aligning the drive or brake
as well as the intermediate plate (B) to the gear. sensor unit with the shaft of the gear. The alignment in z-direc-
– Second, screws (1) can be used to adjust the motor/brake posi- tion and the rotation around the y-axis can be done by the
tion in the z-direction and the rotation about the y-axis. Screws screws (3). The screws (4) can be used to move the drive or
(2) move the motor/brake unit in the y-direction and rotate it brake sensor unit in y-direction and rotate it about the z-axis.
around the z-axis.
– A fine adjustment in x-direction is not necessary, as possible
4 MEASUREMENT DEVIATION
errors will be compensated by opening and closing the interme-
AND CALIBRATION CONCEPT
diate coupling.
– In the simplest case, the screws (1) and (2) are manipulated To interpret test results it is always necessary to know the maxi-
until the coupling hole and the sensor shaft can be pushed mum possible measurement deviations. To estimate the devia-
together without resistance. Thus the axis offset in y- and tions, the sources of the errors and their effects have to be known.
z-direction can be set with an accuracy of ≤ 0.2 mm and the Usually, the maximum deviation is given as percentage of the nom-
angular error around the y- and z-axis with an accuracy of ≤ 0.5°. inal value of the sensor. This means, for example, that a relative
Without tools, the human eye cannot detect structures smaller deviation of 3 % for a torque sensor with 0.5 Nm nominal torque
guarantees an absolute error of ≤ 0.0015 Nm. But if this sensor
is used at a test torque of 0.1 Nm, the guaranteed absolute error
is also 0.0015 Nm. This results in a maximum relative deviation
of 15 % for 0.1 Nm.
Compared to the torque measurements, the deviations of the
speed measurements are relatively small and can usually be
ignored. Against this background, it becomes clear that especially
a calibration of the torque sensors in the correct measuring range
seems to be very reasonable. The concept bases on a centrally
mounted disc to which a cord with test weight is attached. The
testing torque (T_Test) results from the test- and cordmass (m_
(test mass)+m_cord ), the gravitational acceleration (g), the effec-
tive lever disc radius (r_disc) and the lever disc bearing friction
(T_(bearing friction)). The resulting bearing disc friction shall be
added when lifting the test mass and subtracted when lowering.
FIGURE 4 shows an example of the torque curves during lifting
and lowering of a test mass of approximately 0.5 kg. The differ-
ence results in a frictional torque of the lever disc of 0.00503 Nm
FIGURE 2 Test setup with coaxial gearbox (© UAS South Westphalia) and a mean torque value of 0.49847 Nm.
66