To see how the lateral diameter of the HC-DEA bubble affects the force output of the system; 10, 12.5 and 15mm diameter HC-DEAs (see fig below) were investigated. Frediani et al (2014b) initially proposed a HC-DEA tactile display, a bubble of effectively 21mm in diameter was created. The size of this bubble is too large as to be placed on multiple fingertips of the hand simultaneously, and therefore is unsuitable for a multi-finger tactile display system. To change the lateral dimensions of the HC- DEA bubble the internal diameter of the laser cut internal support frame needs to be altered. See next figure (6-1).
Figure 6-9 A sliced front and isometric view illustration of a triple layer HC-DEA, with the frame diameter/ HC-DEA bubble diameter labelled.
Changing the internal diameter of the support frame will also affect two other aspects of the HC-DEA. These are the overlapping electrode area on the active membrane and the volume of coupling gel within the bubble. Changing the overlapping electrode area is relatively trivial, as the laser cut stencil used to apply the electrodes is easily changed to the diameters tested. The volume of gel, on the other hand, has to be calculated to provide the same stretching of the membrane across the comparative samples, as varying prestretch amounts of the material will affect actuator performance, as discussed by Koh et al (2011b). The diameters of bubbles to be tested in this work are 10, 12.5 and 15mm. To produce a 12.5mm diameter bubble with a single layer membrane, 0.8ml of coupling gel was used. 0.8ml was rather arbitrarily
passive and active membranes being in the relaxed elastic state. If we make the assumption that the bubble is two spherical caps, the following calculation was used to ensure each different diameter bubble has comparable stretches:
𝑉𝑐𝑎𝑝 = 1
6 𝜋ℎ(3𝑎'+ ℎ')
The previous equation gives the volume of a spherical cap. Considering our HC-DEA is formed of two spherical caps, Vcap is multiplied by 2 to get the total coupling gel volume Vgel. Considering this formula and that a 12.5mm diameter bubble has a vertical height diameter of ~10mm and contains 0.8ml of coupling gel, the following values are obtained for 10mm and 15mm diameter bubbles:
Table 6-1 Table of gel volumes HC-DEAs at 10, 12.5 and 15mm in diameter.
Ø (mm) A (mm) h (mm) Vcap (ml) Vgel (ml)
10.0 0.500 0.4 0.191 0.382
12.5 0.625 0.5 0.375 0.750
15.0 0.750 0.6 0.643 1.286
Considering the accuracy of injecting the gel into the cavity formed by the vacuum in the manufacturing process, +-0.05ml being the minimum increment of the syringes used, 0.4, 0.8 and 1.3ml of coupling gel will be used for bubbles of 10, 12.5 and 15mm respectively.
6.4.1 Results pertaining to change in lateral diameter on force output
To investigate how the lateral diameter of the HC-DEA affects the uniaxial force performance, the tests described in section 6.2.3 with the HC-DEAs are described in 6.1.1. A 4kV voltage was applied across the actuators relating to 10, 12.5 and 15mm in lateral bubble diameters with corresponding gel volumes of 0.4, 0.8 and 1.2ml accordingly. A load cell was then placed upon the deformed passive surface of the actuator, once a contact force of 0.05N was recorded by the load cell, the voltage was switched to 0V and force values were recorded at a rate of 100Hz. It was decided to only present the force data pertaining to the first 1000ms as ~95% of the full force was typically applied in this time period. 300% prestretch was applied to membranes before forming into bubble configuration. The data pertains to 5 samples at each diameter size.
coupling gel. Although not seen at this scale, this could be a potential issue at larger diameter HC-DEAs. Below is a graph showing the force delta (F(n+1)-F(n)) for the tests performed, and showing the rate of change of force applied over the first 400ms.
Figure 6-11 Average change of force between samples collected by load cell at the different diameter actuators tested over the first 400ms. For each diameter, five specimens were tested.
The peak rate of force application appears to happen within the first ~25ms for 12.5 and 10mm actuators, whereas this is shifted to ~40ms for the larger 15mm diameter bubble. Potentially due to the increased mass of coupling gel. An interesting feature is the double peak in the small diameter actuators. A potential cause for this is that the same cylindrical 12mm in diameter loadcell indenter was used for all the tests performed, and the interaction of the passive membranes for differently sized actuators could potentially give varying rate of force application profiles. One way to test this would be to use load cell indenter parameters that are proportional to the actuator diameter being tested.
To test how the force is applied to the loadcell due to a linear ramping signal, the test described in 6.2.2 is performed on the differently sized actuators. Below are the results from 5 samples for each of the different diameters, again with the initial prestretch of 300%.
Figure 6-12 Results of the voltage induced force test outlined in 6.2.2 for different diameter actuators. 5 samples were tested at each diameter and a fitting line drawn. Error bars represent 2 standard deviations from the mean (95% confidence level).
In the above figure (6-17), the performance increase of the larger diameter actuators can be seen. As expected there is a slight curve to voltage induced force, as identified by Fedianni et al (2014b). As described in 6.2.2, the voltage was decreased linearly from 4kV to 0.7kV (linear operating range of circuitry) over a period of 6.5 seconds. Considering that this actuator is to be placed within the width of the fingertip (<20mm), we opt to compromise between the force output and lateral diameter by using the 12.5mm as the diameter actuator. An important factor to consider is the size of the supporting frame and the plastic enclosure of the actuator, as this currently requires a further lateral width across the fingertip. Therefore, to increase the force performance of the actuator, the multi-layer approach described in chapter 4 will be tested using 300% pre-stretched membranes with a diameter of 12.5mm.