15.08.2016
The influence of a varying temperature on the measurement of different quantities is generally known. In addition, precision transducers for the measurement of torques also show a measurable dependence on the relative humidity. In PTB’s Solid Mechanics Department, the effects have been thoroughly examined, whereby not only the sensors, but also other parts of the chain of traceability – the measuring amplifiers and the bridge standards applied for their calibration – were used. The results will be of interest especially for metrology institutes and accredited laboratories as well as manufacturers of the respective measuring instruments.
For the electrical measurement of the mechanical quantities of force, torque or pressure, sensors are often used which are based on strain gauges. These strain gauges are installed on the sensor bodies – mostly glued – and electrically linked to a Wheatstone bridge circuit. A precision amplifier supplies the circuit with a defined supply voltage. The mechanical quantity acting upon the sensor distorts the sensor, and the strain gauges experience a small change in resistance which results in an output voltage in the bridge. The relationship between the output voltage and the supply voltage is measured and indicated by the amplifier as the stress ratio. So-called bridge standards create a well-known stress ratio and help to trace the measurement amplifier to the scale of the electric unit of "voltage".
In contrast to force transducers, torque transducers are – for constructive reasons – usually not hermetically closed, i.e. the bridge circuit is in contact with the ambient air. In this way, the influences of the relative humidity and of the temperature of the ambient air are greater than in the case of hermetically closed transducers, a fact which was already known in the preparatory phase of the examinations. Now the effects have been examined in detail and published. A first part (published in [1]) dealt with the quantification of influences of the above-mentioned quantities on the individual measuring instruments. Thereby, bridge standards, measuring amplifiers and torque transducers were examined, whereby the latter were not loaded; so the main issue was the influence on the zero signal of the sensor. For these measurements, the instruments were individually exposed to varying temperatures and humidity values in a climatic cabinet. The bridge standards turned out to be very stable. The measuring amplifiers generally showed the linear dependence of the measured signal change on the temperature, humidity and stress ratio, whereby only one quantity was varied at a time. The effects became visible in the signals relatively quickly. The individual amplifiers – although they were all of the same type – revealed differences from one to another. A possible reason for this might be the degree of scatter in the parameters of the components used. In the case of the zero signals of the torque transducers, the time component played an essential role, as it sometimes took several hours until the display stabilized after the conditions had been modified. Thereby, both the distributions and the orders of magnitude of the effects differed greatly. As to the distributions, an asymptotic warm-up and also an overshooting to the original value could be observed – besides a rapid adjustment to a new value. The distributions were described by means of a rheological model.
Whereas the influences on the zero signal of a strain gauge transducer usually do not pose a problem (exception: permanent measurement without the possibility of unload the transducer), it is not that easy to take into account the sensitivity change of a transducer. In a second phase, the sensitivity variation of different transducers at variable ambient conditions was examined, evaluated and published in [2]. A 20 N•m torque standard machine was used for the measurement according to the principle of the direct deadweight effect (supported lever-mass system) which was equipped with a specially adapted climatic cabinet which generated the desired adjustable ambient conditions (Figure 1) specifically for the torque transducer.
Figure 1: 20 N•m torque standard machine: 1 – Load mass, 2 – Lever, 3 – Air bearing,
4 – Torque transducer (partly covered), 5 – Climatic cabinet with temperature and humidity control
The measurement process was determined as follows: The standard measuring facility repeatedly and continuously, in a time rhythm of several minutes, pre-loaded the transducer with 100 % of its nominal torque – and then carried out a measurement sequence with zero signal, 50 %, 100 %, 50 % again and zero signal with a subsequent waiting time. During this process which could carry on for several days, the temperature and the relative air humidity were modified successively in the climatic cabinet between 18 °C and 40 °C and between 37 % r. h. and 70 % r. h., respectively. Thereby, the signals of the torque transducer were recorded and after that, the signal changes were calculated and the data was evaluated. Here also, the respective transducer showed typical behaviour which can be described as asymptotic warm-up (Figure 2) or overshoot (Figure 3).
Figure 2: Signal changes of a 5 N•m torque transducer for temperature (red curve) or for relative
humidity (blue curve) modifications
Figure 3: Signal changes of a 10 N•m torque transducer for temperature (red curve) or for relative
humidity (blue curve) modifications
The effects can be both positive and negative, and the order of magnitude can also differ within a certain transducer type. The values seem to attest to a much bigger influence of temperature as compared to humidity. However, one should not ignore the fact that the temperature in laboratories can be kept constant quite well at ±1 K, whereas humidity variations may well reach ±10 % r. h. Then the humidity variations can also have a significant influence on precision measurements.
The procedures and instruments used allow a sufficient number of insensitive instruments or influences to be determined and, if necessary, the improvement of measurement results by applying corrections.
Bibliography:
[1] K. M. Khaled, D. Röske, A. E. Abuelezz, M. G. Elsherbiny: Humidity and temperature effects on torque transducers, bridge calibration unit and amplifiers, Measurement, Volume 74, October 2015, Pages 31-42, 
doi:10.1016/j.measurement.2015.07.007
[2] K. M. Khaled, D. Röske, A. E. Abuelezz, M. G. Elsherbiny: The influence of temperature and humidity on the sensitivity of torque transducers, Measurement, Volume 94, December 2016, Pages 186-200, doi:10.1016/j.measurement.2016.07.028
Contact person:
Dirk Röske, Department 1.2, WG 1.21, e-mail: 
dirk.roeske(at)ptb.de