TORQUE TRANSDUCER PROJECT IN AENG 260

Group 4. Tess F. Bueno Ruel G. Peneyra INTRODUCTION Strain gauge has been an important tool since its invention way back in 1930. The resistance strain gage is, in itself, a transducer, since it converts mechanical deformation, or strain, into a corresponding electrical signal. Beginning, however, with the initial motivation for its invention, the strain gage has always been applied not only to the measurement of strain for stress analysis purposes, but also as the sensing element in numerous types of transducers for measuring other mechanical variables such as force, torque, and pressure. OBJECTIVE This project aimed to develop, evaluate and characterize a simple yet effective torque transducer for small devices.

REVIEW OF LITERATURE Note: The following literature was extracted from the technical note "Strain Gage Based Transducers" by the Measurements Group, Inc., and was published in 1988. Torque Transducers Although a less frequent requirement than force measurement, it is sometimes necessary to measure the torque in a mechanical system, and a strain-gage-based transducer is usually the preferred means for doing so. In the case of a circular cross-section shaft subjected to pure torque, there are equal shear stresses in both the longitudinal and circumferential directions, everywhere on the shaft surface. The shear strains corresponding to these stresses cannot be measured by longitudinally and circumferentially oriented strain gages, however, because a single strain gage is sensitive to only normal strains along its axis, and is totally insensitive to shear strains. But, from elementary mechanics of materials, it is known that pure shear

produces plus and minus normal strains of equal magnitude at angles of plus and minus 45 from the shear direction. It is also known from mechanics principles that the difference between the two principal normal strains is equal to the maximum shear strain. Thus, the shear strains, which are proportional to the applied torque, can be measured with a pair of strain gages installed on the shaft surface and oriented at 45 on either side of a line parallel to the shaft axis. Most commonly, two pairs of such gages are mounted on the shaft, as shown below, to maximize the output signal and to provide for cancellation of extraneous signal components due to bending or direct stress ("Strain Gage Based Transducers", Measurements Group, Inc., 1988).

Following are the design relationships for a circular cross section torque transducer:

DESIGN

OF THE

IMPROVISED TORQUE TRANSDUCER

The improvised torque transducer is a flat aluminum bar measuring __ x ___ x ____. The material was chosen for its inherent characteristics of twisting at small torsional load capacity. Two rosettes were fixed at the surface of the bar, one on each face located at both longitudinal and traverse axes. Strain gages on each rosette are oriented at 45 on either side of a line parallel to the longitudinal axis. The figure below illustrates the position and circuit of the full bridge torque transducer configuration.

Vin

gage 1 gage 2

Vou t

gage 4 gage 3

Front

Back

Figure 2. Illustration of the full bridge torque transducer

---------------------- Insert actual picture of the transducer here in horizontal plane Figure 3. Actual torque transducer Design analysis For any direction of torque, gages 1 and 3 are strained in tension, and gages 2 and 4 in compression; and the strains due solely to the applied torque are all equal in magnitude. When the gages are connected in the bridge circuit as shown, the resistance changes in every pair of adjacent bridge arms are opposite in sign, while those in opposite arms have the same sign. As a result, the bridge output, is four times that of a single gage. On the other hand, the thermal outputs of all four gages are the same in sign and in nominal magnitude (if at the same temperature). Since the same resistance changes in all arms of the bridge have no effect on the state of bridge balance, the strain gage thermal output is canceled within the circuit. Such cancellation also occurs for strains due to a purely axial load on the shaft, and for the same reason. However, if the bar is subjected to a bending moment in any plane, the resulting strains in plane opposed gages (1 and 3, and 2 and 4) are always numerically equal but opposite in sign. These are likewise canceled because the corresponding resistance changes appear in opposite pairs of bridge arms. Evaluation of the Improvised Torque Transducer The evaluation of the improvised torque transducer was carried out in bending, tension, compression, and torsional load application. In all tests, the Vin terminals were connected to power supply while the Vout terminals were

connected to the signal amplifier set at x 100 range. One end of the transducer was fixed on the bench table while loads were applied at the other end. For tension, bending, and axial, loads of 100 g to 900 g at 100 g increment were applied. For the torsion test, an arm with a length of 20 cm was secured perpendicularly at the free end of the transducer to serve as arm. Same load magnitudes and increment were applied at the free end of the arm to cause torsional deformation on the flat bar. Insert Images Here Figure 4. Evaluation setup and test of the transduer Results of the Evaluation of the Improvised Torque Transducer Based on the results of the evaluation made, the torque transducer or the strain gages doesnt respond on bending, tensile and compression loading. In both increasing and decreasing load application, the reading remains constant, hence, the transducer was found unaffected by these types of load. In the torsion test, however, the transducer responds very well. Table 1 shows the raw data during the trials of the device. Table 1. Voltage Readout, Vo, during the experimental test Trial 2 Loa Torqu Trial 1 d e, N- Ascend Descend Avera Ascendi Descend Avera g m ing ing ge ng ing ge 0.096 0 0 0.093 0.095 0.094 0.096 0.097 5 0.196 0.102 100 2 0.099 0.101 0.1 0.102 0.103 5 0.392 0.105 0.108 200 4 0.105 0.106 5 0.108 0.109 5 0.588 0.111 0.114 300 6 0.111 0.112 5 0.114 0.115 5 0.784 0.120 400 8 0.118 0.118 0.118 0.12 0.121 5 500 0.981 0.122 0.124 0.123 0.127 0.127 0.127 1.177 600 2 0.128 0.13 0.129 0.133 0.133 0.133 1.373 0.134 700 4 0.134 0.135 5 0.139 0.139 0.139 1.569 0.140 800 6 0.14 0.141 5 0.145 0.145 0.145

Overall Averag e 0.0952 5 0.1012 5 0.107 0.113 0.1192 5 0.125 0.131 0.1367 5 0.1427 5

1.765 900 8 0.148 Arm Length = 20 cm

0.148

0.148

0.152

0.152

0.152

0.15

Graph of the resulting data shows linear relationship between the voltage reading and the actual torque applied

Figure 5. Graph of the torque transducer characteristic curve showing both trial projections Resolving for the function that will predict the torque applied as function of Vo resulted to the equation: Torque, N-m = 32.63(Vo) - 3.102 where 99.96 percent of the variation can be explained by the model. The succeeding figure shows the characteristic equation and the linearity curve of the transducers response to torsional load.

Figure 6. Graph and equation of the torque transducer characteristic curve POST ANALYSIS The improvised torque transducer performs as expected base on the weight limits defined (0-900 g). The linearity is well fitted at the defined range. However, few other observations are noticeable:
1. That the two trials have differences subject to varying Vout reading

in the no load condition. However, the incremental increase in Vout was found to be almost constant. Deeper analysis may show that the material property of aluminum with regards to its elasticity has some problem as the material deformation and returning back to its original shape is a time function. Hence, to revert back to its original zero value, the aluminum bar must be allowed to rest for quite sometimes to remove all unnecessary or undesired stresses created by the previous loading.
2. These applications of Wheatstone bridge arithmetic for augmenting

the torsional output signal while canceling unwanted signal components require that the gages be precisely positioned and oriented. The centers of all four gages must lie on the same flat section; they must be oriented along 45 helices on the flat surface;

and the centers of each pair of opposed gages must lie on a common reference. The gage installation task can be simplified considerably by employing special torque gages, each of which contains two grids (on a common backing) precisely oriented at plus and minus 45 from the gage axis.

SUMMARY AND RECOMMENDATION Based on the results of the study, the improvised transducer was found effective with well fitted linearity curve at torsional loads of 0 2 N-m. Furthermore, the results showed that the strain gage configuration and arrangement was able to cancel the bending, compression and tensile loads applied to it. It was further suggested that: 1. a validation study should be performed for higher torque application to the transducer
2. the arrangement and positioning of the strain gages should be

taken into serious consideration in order to cancel all undesirable signals.