Load cell design

4.1.1 Strain gauge and its application in load cell

Strain gauge is a device used to measure strain through the change in electrical resistance, first invented by Edward E. Simmons and Arthur C. Ruge in 1938 [43].

Though strain can also be measured either by change in electrical capacitance, inductance, or by optical method, these methods’ sensitivity to dynamic strain, mounting requirements, and circuit complexity have limited their application [44].

Therefore, resistive-type strain gauges remain the most popular way to measure strain for their versatility. The most common type of strain gauge, the metallic foil-type, consists of a grid of metal wire filament (the resistor) of about 0.025 mm thickness, supported by a thin film of epoxy resin. By attaching the gauge to object with suitable adhesive, such as cyanoacrylate, the foil will deform with the surface of the attached object, causing change of its electrical resistance. The resistance change is usually measured with Wheatstone bridge, as shown in Figure 4-2, and is related to the strain by the quantity, known as the gauge factor (the GF), defined in equation (38):

R R R R withstand force directly, a structure with predictable strain related to force is needed when applying strain gauge on load cell. By measuring the strain of the structure, force can be obtained with known strain-stress and then stress-force relation.

Figure 4-1 Schematic of Wheatstone bridge circuit, where Rs are electrical resistances, VIN and VOUT are excitation voltage and output voltage; A, B, C, and D are nodes [44].

4.1.2 Design target for load cell

Since we set our maximum assistive force to be 100 N, our load cell needs to be able to withstand about 200 N of tensile force if we set a safety factor of 2. Furthermore, the load cell needs to be small and easy to install due to the limited space in our device.

In order to sense the change in force more accurately, we set a secondary criteria for the load cell that it should be able to measure force at a 0.5 N resolution level.

4.1.3 Structure design of load cell

As a result of needs for measuring heavy load while space is limited, sus304 steel is used to build the structure of the load cell for its high tensile strength. The load cell is designed to be compatible with straddle cable, which is used to connect wearer’ limbs to our device. Since straddle cable mainly transfers tensional force, we design our load cell to be a pull-type load cell. As shown in Figure 4-2, the load cell has a 6 mm hole in the middle, enabling it to be installed directly with a M6 screw. The two holes with groove are slightly larger than the head of straddle cable for climbing bike, which makes it easy to install by directly putting the cable head into the hole as shown in Figure 4-3. Since the geometry is complicated for theoretical model, we use finite element simulation to verify our design .

Figure 4-2 The structure of our load cell (dimension in mm)

Figure 4-3 Load cell design for straddle cable head installation

4.1.4 Finite element method simulation result

There are two main tasks in our FE analysis. One is to estimate whether the structure is able to withstand 200 N without yielding, and the other is to find the position for the strain gauge and to estimate whether the strain is adequate. Since our load cell is symmetrical about two axes as shown in Figure 4-4(a), we build a quartered model of our load cell for FE simulation, as shown in 4-4(b). In order to apply force more accurately, we build a fillet with radius of 0.5 mm on the corner where cable head contacts our load cell, which generates a small surface for force to apply. Based on the geometrical symmetry of our load cell, we assume the two cross sections marked in Figure 4-4(b) to be fixed boundaries.

Figure 4-4 (a) The load cell is symmetrical about two axes. (b) A quartered model of our load cell for FE simulation.

As shown in Figure 4-5, because there is a contact angle of 33.23° between the contact point on load cell and the direction of tensile force, a y-component force of 2/3 intensity will be generated on the contact point if we pull the cable along the x-direction.

We then start our FE simulation with the two fixed boundaries and the total force exerted on the fillet surface. During the simulation, a set of external forces of 100 N x-component and 65.5 N y-x-component force, which causes 200 N tensile force in cable, is used. The main criterion of our simulation is that the maximum stress should not exceed 200 MPa, which is approximately the yielding stress of steel. The simulation results are shown in Figure 4-6. Figures 4-6(a)-(c) show the result with the first set of external force. The maximum stress is 125 MPa, which is 37.5% smaller than our criterion, showing that our load cell is capable of withstanding 20 kgw force. The maximum strain on the upper surface occurs at 14 mm away from the center of load cell along longitudinal direction, generating elongation of 2.04 μm, which is too small for a metallic foil-type strain gauge to have a fine resolution. Therefore, strain gauges with much higher GF, like semi-conductor type strain gauges, are needed to meet our requirement. As a result, KSN-2-120-E3-16 strain gauge, made by Kyowa, boasting GF of -105, are used.

Figure 4-5 Diagram of contact angle

Figure 4-6 Simulation result of 100 N x-component and 65.5 N y-component force applied.

(a) of stress (b) of strain (c) 1D plot of strain along the line marked in (b).