Thesis Submitted in Partial Fulfillment of the Requirements for the Degree of Master in Science in Textile Engineering
The Swedish School of Textiles
2013-06-02 Report no. 2013.14.2
Pressure sensitive textiles for integration in saddle pads
Therese Engvall
2
Acknowledgements
During this thesis I came across many problems and questions, which would not have been solved without the help and assistance from the following:
Lena Berglin - who has always been supportive and encouraging throughout the whole process.
Emanuel Gunnarsson - thank you for your assistance and sharing your
knowledge. Many of the experiments would not have been possible without you.
Hanna Lindholm - the weaving process would not have been as efficient without your assistance and knowledge.
Also, I would like to thank my family and friends and give my special thanks to Kim Fasting and Karin Rundqvist who have been helpful and supportive through thick and thin.
Borås, May 2013
Therese Engvall
3
Abstract
In this thesis, capacitive textile pressure sensors have been developed. The sensors were meant to be integrated into saddle pads and be able to measure the pressure between the saddle and horse. The aim of the thesis was to create a theoretical and practical based map on how a textile pressure sensor can be made. Capacitance was found to be the most suitable pressure sensitive technique to be implemented in a textile structure. The project was divided into two cycles, where the first cycle consisted of laminating capacitive textile pressure sensors of readymade fabrics in different thicknesses and sizes. After testing the pressure sensitivity of these laminates, it was concluded that a thin fabric with some compressibility was sufficient for making a textile capacitive pressure sensor. However, the area cannot be too small. The second cycle consisted of weaving capacitive pressure sensors as three layer structures. The pressure sensitivity of the sensors and the effect of moisture were tested. The results showed that most of the woven sensors were able to sense a 50g change in weight even after a 700g load was put on. The moisture and water tests showed that the pressure sensors must be protected from water and moisture. It was also discovered that there is a lack of knowledge in how textile structures and fibres behave under compression and release. Models of how textiles behave during pressure are needed to do correct transformations between compression and pressure and predict how the textile will behave during different pressures.
Keywords: Textile pressure sensor, saddle pad, saddle pressure, capacitive pressure sensor, press sensitive principles.
4
Popular abstract
Horses are sensitive animals and a poorly fitted saddle will cause the horse pain and eventually back problems. A saddle with good fit on the other hand enhances the performance of both the rider and horse. In this thesis textile pressure sensors have been developed. Textiles are usually thin, soft and flexible and a piece of fabric is already commonly used between the saddle and horse's back - a saddle pad. Therefore, textile pressure sensors would be suitable for integration into saddle pads as they would be able to measure the pressure between the saddle and horse. Textile pressure sensors were both laminated and woven and their pressure sensitivity were tested. Most of the sensors were able to sense a 50g change in pressure even when a 700g load already was on the sensor. The effect of water and high humidity on the sensors was also tested and the results showed that the sensor must be protected from water and moisture to function correctly. It was also found that more research must be made on how textile structures and fibres behave under compression and release.
5
Table of Contents
1. Introduction ... 9
1.1 Related products ... 10
1.2 Assumption ... 10
1.3 Aim ... 11
1.4 Research questions ... 11
2. Theory ... 12
2.1 Pressure ... 12
2.2 The press sensitive principles ... 12
2.2.1 Piezoresistive effect ... 12
2.2.2 Piezoelectric effect ... 13
2.2.3 Fibre optics ... 13
2.2.4 Capacitance ... 14
2.2.5 Other press sensitive principles ... 15
2.3 Related work ... 15
2.4 Choice of technique to work further on ... 17
3. Experimental ... 18
3.1 First cycle – materials ... 18
3.1.1 Fabrics ... 18
3.1.2 Adhesives ... 18
3.2 First cycle – methods ... 19
3.2.1 Sample production – lamination ... 19
3.2.2 Relative permittivity ... 19
3.2.3 Capacitance measurements ... 19
3.2.4 Pressure sensitivity measurements of laminated samples ... 20
3.3 Second cycle – materials ... 22
3.3.1 Yarns ... 22
3.4 Second cycle – methods ... 22
3.4.1 Sample production – weaving ... 22
3.4.2 Shielding fabrics production ... 24
3.4.3 Test of contact between the conductive layers ... 24
3.4.4 Capacitance measurement ... 25
3.4.5 Pressure sensitivity measurements of woven samples ... 25
3.4.6 Pressure sensitivity test of woven samples with woven shields ... 26
3.4.7 Effect of moisture ... 27
4. Results ... 28
6
4.1 First cycle ... 28
4.1.1 Laminated samples ... 28
4.1.2 Capacitance measurements ... 29
4.1.3 Pressure sensitivity measurements of laminated samples ... 30
4.2 Second cycle ... 31
4.2.1 Woven samples ... 31
4.2.2 Woven shield fabrics ... 33
4.2.3 Contact test ... 34
4.2.4 Capacitance measurement ... 34
4.2.5 Pressure sensitivity measurements of woven samples ... 34
4.2.6 Pressure sensitivity test of woven samples with the woven shields ... 37
4.2.7 Effect of moisture ... 39
5. Discussion ... 43
5.1 First cycle ... 43
5.1.1 Material discussion ... 43
5.1.2 Capacitance measurements ... 43
5.1.3 Pressure sensitivity measurements of laminated samples ... 43
5.2 Second cycle ... 44
5.2.1 Choice of technique ... 44
5.2.2 Choice of structures and materials ... 44
5.2.3 Contact test ... 45
5.2.4 Pressure sensitivity measurements of woven samples ... 46
5.2.5 Pressure sensitivity test of woven samples with the woven shields ... 48
5.2.6 Effect of moisture ... 49
6. Conclusion ... 51
7. Future work ... 53
References ... 54
Appendix A ... 56
Draft of the first setup of the loom – waffle weave ... 56
Appendix B ... 57
Draft of the second setup of the loom – waffle weave ... 57
Appendix C ... 58
Draft of the rep weave ... 58
Appendix D ... 59
Results of the woven samples ... 59
Appendix E ... 63
7 Results from the pressure sensitivity test of the woven samples with the woven
shields ... 63
Appendix F ... 67
Comparison between pressure measurements with and without the 38g weight ... 67
List of figures Figure 2.1. Left: the relationship between capacitance and spacing. Right: the relationship between capacitive impedance and spacing ... 15
Figure 3.1. The Agilent U1732A LCR meter and the sample holder ... 20
Figure 3.2. The circuit of the square wave generator ... 21
Figure 3.3. A picture showing the faraday's cage and the oscilloscope ... 21
Figure 3.4. A picture taken inside the faraday's cage, showing the circuit and the sample in the sample holder ... 21
Figure 3.5. A schematic picture of different amount of conductive weft in the top and bottom layer.Yellow is non conductive and black is conductive yarn. ... 23
Figure 3.6. The draft of the different woven shield fabrics. ...24
Figure 3.7. A schematic picture of the backside of an English saddle and a rough estimation of the dimensions ...26
Figure 4.1. Some of the laminated samples. The top three samples are the ones with single jersey fabric as middle layer. The middle ones are with 1:1 rib fabric and the four at the bottom are with the 4.4 mm thick spacer fabric ...28
Figure 4.2. The cross section of the laminated spacer fabrics. From the top: 3.3 mm spacer fabric, 4.4 mm spacer fabric and 6 mm spacer fabric ....29
Figure 4.3. The capacitance measured with Agilent 1732A was divided with the calculated capacitance and shows how much the measured capacitance differs from the calculated capacitance ...30
Figure 4.4. Sensitivity factor of the 2.5×2.5 cm samples ...30
Figure 4.5. Sensitivity factor of the 5×5 cm samples ...31
Figure 4.6. Sensitivity factor of the 10×10 cm samples ...31
Figure 4.7. A picture of all the woven samples ...32
Figure 4.8. A picture of the woven shield fabrics and the isolating fabric. 1: isolating fabric, 2: shielding fabric where every fifth yarn is conductive (sparse shield), 3: shielding fabric where every third yarn is conductive (middle shield), 4: shielding fabric where every second yarn is conductive (dense shield) ...33
Figure 4.9. Frequencies of some of the woven samples ...35
Figure 4.10. Sensitivity factor of some of the woven samples ...35
8 Figure 4.11. A close up of the sensitivity factor of some of the woven samples
at + 100g to - 100g ...36
Figure 4.12. Frequencies of sample 4, 5 and 6 ...36
Figure 4.13. Frequencies of sample 9, 12 and 13 ...37
Figure 4.14. Frequencies of sample number 8 with the different shields ...38
Figure 4.15. Frequencies of sample number 16 with the different shields ...38
Figure 4.16. Sensitivity factor of sample number 8 with and without the 38g weight. The most dense shield fabric was used ...39
Figure 4.17. Frequency of sample number 8 with and without the 38g weight The most dense shield fabric was used ...39
Figure 4.18. Frequency of sample 4 during water spraying test ...40
Figure 4.19. Frequency of sample 16 during water spraying test ...40
Figure 4.20. Sensitivity factor of sample 16 in dry, moist and wet conditions ..41
Figure 4.21. Sensitivity factor of sample 4 in dry and moist conditions ...41
Figure 4.22. Decreasing frequency during increasing relative humidity of sample 4 and 16 ...42
Figure A.1. The first draft of the loom - waffle weave. ...56
Figure B.1. The second draft of the loom - waffle weave. ...57
Figure C.1. The draft of the loom - rep weave. ...58
Figure D.1. Frequencies of the woven samples. ...59
Figure D.2. Sensitivity factor of the woven samples. ...60
Figure D.3. A close up of the sensitivity factor of the woven samples at + 100g to – 100g. ...61
Figure F.1. Sensitivity factor of sample number 4 with and without the 38g weight. The most dense shield fabric was used. ...67
Figure F.2. Frequency of sample number 4 with and without the 38g weight. The most dense shield fabric was used. ...67
Figure F.3. Sensitivity factor of sample number 11 with and without the 38g weight, the most dense shield fabric was used. ...68
Figure F.4. Frequency of sample number 11 with and without the 38g weight. The most dense shield fabric was used. ...68
List of tables Table 3.1. Fabrics ... 18
Table 3.2. Yarns ... 22
Table 3.3. Pressure sensitivity test sequence. ...26
Table 4.1. Measured and calculated capacitances of the laminated samples. ...29
Table 4.2. Woven samples ...32
Table 4.3. Woven samples’ geometry. ...33
Table 4.4. Capacitance of the woven samples. ...34
Table D.1. Sensitivity factor of the woven samples during pressure increase. ...62
Table D.2. Sensitivity factor of the woven samples during pressure decrease. ..62
9
1. Introduction
This thesis was a part of a research collaboration between The Swedish School of Textiles, Chalmers University of Technology and The University of Sydney. The goal of this collaboration is to solve problems in equine and equestrian science with the aid of smart textiles and theoretical and applied physics. The larger aim with this research group is to improve the welfare and performance of horses in two ways:
1. Be able to do research on horses without stressing them or causing them pain by integrating smart textiles into ordinary horse equipment familiar to the horse.
2. This developed equipment can then be used to solve problems and increase the performance and welfare of the horse.
Riding is one of the largest sports in Sweden [1] and it exist a little more than 350 000 horses in the country [2]. Horses are expensive animals and they are very valuable to their owners, both economically and emotionally. Horses are also sensitive animals as their only effective defence against predators is to flee. They are designed to have a wide field of vision and movable ears to localise sounds. It is in the horse’s nature to run away as soon as it hears a suspicious sound or sees a sudden movement in its surroundings. [3] Horses are therefore easily stressed when exposed to new environments and equipment. By integrating sensors and examination equipment into gear already known by the horse, research and veterinary examinations can be done without stressing the animal.
A poor fitted saddle will cause the horse pain and eventually back problems and even lameness. A saddle with good fit on the other hand enhances the
performance of the rider and horse. [4] In [5], a pressure sensor mat from Novel was used to investigate if there is a correlation between how well the saddle fits and the wellbeing of the horse. The horse’s health was proven to be related to the fit of the saddle.
To be able to measure the pressure between the horse’s back and the saddle, the pressure sensor must be soft and flexible. It must be able to shape itself around the horse’s back and it must be sufficiently thin to measure the pressure correctly. A sensor that is too thick would only be pressed into the soft surface of the saddle and horse’s back, giving incorrect readings due to the decreased load area and increase of measured pressure. [6] It is important that the pressure mat does not affect the fitting of the saddle; it has to be thin and flexible. In [7], a capacitive textile pressure sensor has been developed. However, it was 1.1 cm thick. Thus, it would be too thick to fit in a saddle pad and it may also affect the fitting of the saddle.
10 1.1 Related products
There are several companies providing pressure sensor mats for measuring saddle pressures. There are also companies providing large area, flexible pressure sensor mats with no specific field of application. These could probably be used for measuring the pressure between the saddle and horse. Though, none of them are made of textiles. Tactilus Equestrics® and Tekscan® both sell saddle pressure mapping products based on piezoresistive technology [8] [9]. Novel is another company providing a system for measuring the pressure between the saddle and the horse. Their pliance®-s system is wireless and is based on capacitive sensors.
[10] Xsensor and Pressure Profile Systems are two companies which produces flexible large area pressure sensor mats based on capacitive technology [11] [12].
In [13] a commercially available pressure sensor mat was used to test its validity and reliability in saddle fitting. The results showed that the pressure mapping system was valid, but only in highly standardised conditions. As the commercially available saddle pressure mapping systems are derived from devices measuring human pressures, such as in wheelchairs and bicycles, they are not optimal for measuring saddle pressures on horses. For saddle fitting on horses, the pressure measurement should be made during riding as well. The specific pressure mat in the study could only measure a maximum pressure of 40 kPa. This value was found to be too low during the measurements. However, the individual sensors were quite large (about 6×2 cm) and only measured the maximal pressure, not the average. It was also found that the shape and stiffness of the pressure mat could be improved. The rectangular mat had to be wrinkled and as it was sensitive to folding it affected the measurements. It was therefore concluded that a pressure sensor mat shaped as a saddle pad would simplify the positioning of it.
1.2 Assumption
Throughout the thesis this assumption has been used:
Textile pressure sensors can be made more comfortable than pressure sensors made of other materials.
Textile pressure sensors can be made flexible and soft and can be integrated into textile based equipment already familiar to the horse. Existing pressure sensors made of other materials can be hard and stiff. Incorporated into a saddle pad, they might cause the horse pain when the pressure from the saddle and rider press the hard sensors into the horse’s back. A textile pressure sensor may be made to have the same softness and flexibility as the saddle pad and therefore will not be noticeable by the horse.
Compared to the existing products for measuring saddle pressures, which are made of plastic materials and have quite hard structures, a textile structure would be more convenient. A textile pressure sensor can be made soft and flexible and
11 the sensor might be incorporated into a structure which is pressure relieving. The resolution of the pressure mapping might also be possible to do better in a textile structure compared to the conventional technique.
Textile pressure sensors could also be used for other purposes. One example is healthcare where the textile pressure sensor could be integrated in mattresses and wheelchairs to prevent pressure ulcers.
1.3 Aim
The aim of this thesis is to create a theoretical and practical based map of how a textile pressure sensor can be made.
1.4 Research questions
1. How can a textile pressure sensor be designed?
a. What thickness is needed?
b. Is the size influencing the sensitivity of the sensor?
c. How do sensors of different sizes and thicknesses respond to different pressures?
2. Multilayered textile structures could be an applicable textile technique for this purpose.
a. Is it possible to achieve a multi-layered pressure sensor in one process (e.g. knitting, weaving)?
b. Is more than one technique needed to create a functional textile pressure sensor?
3. How can the sensor be protected from disturbing signals and humidity?
12
2. Theory
In this chapter, pressure and the different principles to measure pressure are described and their pros and cons are presented. Related work of earlier studies of textile pressure sensors is also described.
2.1 Pressure
When the force acting on a surface is divided by the surface’s area, the quotient is the pressure. Pascal (Pa) is the SI unit of pressure and 1 Pa = 1 N/m2. One single Pa is a low pressure and in technical and industrial application it is common to use other units to measure pressure, such as atmosphere (atm) or Torr (1 Torr = 1 mmHg). [14]
Pressure can be measured in different ways and also in comparison to different pressures. When the pressure is measured relative to the normal atmospheric pressure, it is called gauge pressure. Absolute pressure sensors measure the pressure relative to a perfect vacuum. It is also possible to measure a differential pressure which is the difference between two absolute pressure sensors’ output.
[15]
A sensor is a device which can sense absolute values or changes in physical quantities, like pressure or temperature, and transform this information into a signal (usually electrical) which can be collected by a data collecting system. [16]
[15] A sensor usually does not work by itself; it needs other devices connected to it. Such devices could be actuators, signal processors and signal conditioners, and also data recorders and memory devices [14]. Pressure sensors are commonly quite complex and need to process the signal in several steps to convert the energy from the pressure to an electrical signal [14].
2.2 The press sensitive principles
There are several techniques available for measuring pressure. Here are the principles of them described and their advantages and disadvantages for making into a textile structure discussed.
2.2.1 Piezoresistive effect
Piezoresistive materials can be used in pressure measurements as the resistivity of the material changes when the dimensions of it are changed. They are made of semiconductor materials where a region of p-type material is diffused into a base of n-type material [17]. An applied pressure induces a strain in the piezoresistive material, causing it to change resistance. The change in resistance is dependent on the value of the applied pressure force. Thus, the change in pressure can be
measured by measuring the change in resistance. The measured electrical output is then converted into a value for the applied pressure force. [15]
13 Piezoresistive pressure sensors commonly have the advantage of having simple read-out technique, but they also suffer from disadvantages such as high
hysteresis, poor stability and they are affected by changes in humidity and temperature. [6]
2.2.2 Piezoelectric effect
The piezoelectric effect is when an electric charge is produced in a crystalline material which is under a mechanical load [14]. There are natural piezoelectric materials, such as quartz, synthetic ones like lithium sulphate and also polymers like poly(vinylidine fluoride) that have piezoelectric properties. Piezoelectric materials can therefore be used for pressure measurements due to that a voltage is produced when a force is placed on a piezoelectric material. The voltage can be measured and translated into the amplitude of the applied force. When a
mechanical force is applied to a piezoelectric material, the asymmetrical network of molecules is distorted. This deformation shifts the electrical charges and the positive and negative charges are therefore displaced within the material. The negative and positive charges are gathered on opposite sides of the surfaces of the material and they can therefore be measured as an output voltage by placing electrodes on the two opposite surfaces of the material. [17] With the two electrodes on each side of the piezoelectric material, it becomes like a capacitor where the piezoelectric material is the dielectric material between the electrodes [14]. The voltage output measured from a rectangular block is given by:
𝑘𝐹𝑑
𝐴 (2.1)
where k is the piezoelectric constant, F is the force applied on the material in g, the thickness of the material is d and A is the area of the material. [17]
The piezoelectric technology is not suitable for measuring static pressures due to that there is a leaking current. The measurement of a static pressure with a piezoelectric pressure sensor will decrease over time and eventually decrease down to zero. [6] Piezoelectric materials are only available in fibre form at research level and can just be bought as film or granulate.
2.2.3 Fibre optics
Fibre optical sensors are another way to measure pressure. They have the advantage of not being affected by electromagnetically induced noise, they produce no heat and they are not sensitive to electrically discharges. When the optical fibre is compressed, the refractive index is changed which in turn changes the intensity of the transmitted light. The difficulty with fibre optics is to ensure that the light entering the optical fibre is maximised. [17] [18] Common optical fibres of glass and plastic are stiff and fragile and thus, difficult to use in textile
14 production processes. Fibre optics also uses special equipment and instruments which are rather expensive.
2.2.4 Capacitance
Capacitance is measured between two isolated conductive objects and is a value of the objects’ ability to store charges. Two parallel conductive plates where one plate is positively charged (+q) and the other negatively charged (-q), is a
capacitor. It can be characterised by q which is the size of the charge on the plates and V, which is the positive potential difference between the plates. The
capacitance C of the capacitor is the ratio between the charge q and voltage V:
𝐶 =𝑞
𝑉 (2.2)
The size, shape, relative position of the conductive objects and the medium in between also influences the capacitance. The capacitance between two parallel plates can therefore be measured by:
𝐶 =𝜀0𝜀𝑟𝐴
𝑑 (2.3)
where ε0 is the electric permittivity of vacuum, εr is the relative permittivity of the dielectric medium, A is the area of the conductive surfaces and d is the distance between the parallel conductive surfaces. As can be seen in the formula, the capacitance is changed when the distance between the surfaces is changed.
Therefore, capacitors are used in pressure sensors. [14] [17]
The relationship between capacitance and spacing (d) is not linear; ε0, εr and A are constants, giving the relationship:
𝐶 =𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡𝑠
𝑑 (2.4)
Hence, capacitors with a small d are very sensitive to even small changes in spacing (see figure 2.1). This can be a problem if the capacitance is measured directly, but a linear output is received if capacitive impedance is measured (see figure 2.1):
𝑍 = 1
2𝜋𝑓𝐶 (2.5)
Z is impedance, f frequency and C is the capacitance. [19]
15
Figure 2.1. Left: the relationship between capacitance and spacing. Right: the relationship between capacitive impedance and spacing.
Farad (F) is the SI unit for capacitance and 1 F is an extremely large capacitance.
Therefore, capacitance is commonly measured in µF (=10-6), nF (=10-9) or pF (=10-12). [14]
The electric field lines between the plates in a parallel plate capacitor are straight and uniform, but at the edges the fields are not straight due to the fringing effect.
These fringing fields add capacitance to the total capacitance of the capacitor, but are not included in the formulas for calculating capacitance. As long as d is kept small in comparison to the area of the capacitor, the fringing effect is so small that it can be neglected. [19]
Compared to piezoresistive pressure sensors, capacitive sensors are in general less sensitive to changes in temperature and humidity [6]. The difference between the dielectric constants of dry air at 0°C and steam at 110°C is 0.00726. Thus, even if the temperature or relative humidity of the surrounding air changes the dielectric constant does not change much. A capacitive sensor must be protected from getting wet, as the dielectric constant of water is about 80 (depending on the water temperature) compared to air which has a dielectric constant of about 1. [19]
2.2.5 Other press sensitive principles
It is also possible to make pressure sensors using hydraulic and pneumatic technologies, though they are not suitable for making thin and flexible pressure sensors. [6]
2.3 Related work
Textile pressure sensors has been developed and tested in earlier studies with different techniques and results. Textile pressure sensors based on capacitive sensing has been tested in different ways in [7] [20] [21] [22]. Strips of
conductive fabric in rows and columns with a foam layer in between have been used in [21]. Though, the use of a foam layer introduces non-linearity in the function of pressure versus capacitance. Spacer fabric has been used in [7] and [20] to create textile capacitive pressure sensors. A conductive fabric was used as a common electrode on one side of the spacer. On the other side, small single
16 electrodes were embroidered with conductive thread to create an array of
capacitors. The spacer had high hysteresis and it had to be modelled to reduce the error induced from it. In [22], capacitive sensors were created between single yarns by coating nylon yarns with conductive and non-conductive materials and weaving them together.
Piezoresistivity has been used in [23] to create a partly textile pressure sensor. A resistive sensor fabric was placed between two toothed elastomeric plates. A pressure put on the top elastomeric plate will induce an in-plane strain in the fabric and change the resistivity of it. The pressure sensor can therefore translate the in-plane strain into the vertical pressure. Several hundreds of compressing cycles were needed to get a sensor with good repeatability.
Common optical fibres made of plastic or glass are stiff and fragile and are therefore not suitable for textile production processes. When they are integrated into textiles, the textile feeling is lost and the textile becomes rather stiff. In [18], flexible optical fibres made of thermoplastic silicone have been produced. Then the optical fibres were woven together with cotton into textiles. The produced optical fibres had poor light transmission, compared to ordinary optical fibres.
However, they were more flexible and the light was sufficient for short distances.
A 2×2 pressure sensor matrix was produced which sensed changes in pressure by the optical fibre’s change in cross section. Due to the elastomeric properties of the fibre the cross section change is reversible. The signal transmission element and the sensor are combined in one unit and thus, the sensor cannot be used for shape recognition or multi-touch sensing.
Polymeric piezoelectric film pressure sensors have been integrated in upholstery fabric, between the top fabric and foam, for car seats in [24] [25]. It was not possible to measure the pressure through change in charge of the piezoelectric film; instead phase variation of the resonant frequency was measured. This method was successful in measuring the induced stresses on the fabric.
In earlier studies of textile pressure sensors, especially capacitive sensors,
different spacer fabrics and foams have been used. These thicker structures proved to have high hysteresis and the function of pressure versus capacitance was not linear. Little research has been done on thinner textile pressure sensors, with an ordinary fabric as compressible layer. Creating a multilayered textile pressure sensor in one step has not been investigated thoroughly. In research of
multilayered pressure sensors, it is common to use readymade fabrics and laminating them together. Embroidery is also used. To produce a multilayered textile pressure sensor in one technique and one single step needs more research.
17 2.4 Choice of technique to work further on
Out of the studied techniques, capacitance was chosen to be used further in this project. Capacitance does not need any special equipment to be measured, nor is any special material (other than conductive and non-conductive materials) needed to create a capacitor. Capacitance measurements are not as sensitive to changes in temperature and humidity as piezoresistive measurements are. However,
capacitance has the disadvantage of being sensitive to noise and must be shielded from external electric fields.
18
3. Experimental
The thesis was divided into two cycles where the results from the first cycle were concluded into the second cycle. The first cycle consisted of creating capacitive pressure sensors of different sizes and thicknesses. This was done by laminating readymade fabrics of different thicknesses with a conductive fabric. The different samples’ pressure sensitivity were tested and compared with each other. This was done to get an indication of how thin the compressible middle layer could be. Out of the results from the first cycle, theories were created about how textile pressure sensors could be produced in a three layer structure in one step.
The second cycle consisted of creating new samples based on the theories from the first cycle. In this cycle, yarns were used instead of readymade fabrics and the samples were made in one single process. The samples were tested and evaluated for their pressure sensitivity.
3.1 First cycle – materials
The materials used in the first part of the thesis were readymade fabrics of
different types and thicknesses. Also an adhesive was used to laminate the fabrics together.
3.1.1 Fabrics
The fabrics used in the first samples were all made of polyester and had different thicknesses. Two ordinary weft knitted fabrics and three different warp knitted fabrics were used. A conductive fabric, Shieldex Medtex 180 +B, was also used for the first samples. All fabrics and their thicknesses can be seen in table 3.1.
Table 3.1. Fabrics
Name Type Material Thickness [mm]
Shieldex Medtex 180 +B Warp knitted Silver-plated nylon 0.47 Single jersey fabric Single jersey Polyester 0.44
1:1 rib fabric 1:1 rib Polyester 1.0
Spacer 1 Warp knitted spacer Polyester 3.3
Spacer 2 Warp knitted spacer Polyester 4.4
Spacer 3 Warp knitted spacer Polyester 6.0
The single jersey fabric and 1:1 rib fabric were knitted with the same yarn, polyester dtex 167/48/1.
3.1.2 Adhesives
During lamination, a thin polyamide web adhesive with an activation temperature of 85-95°C was used.
19 3.2 First cycle – methods
The methods in the first cycle consisted of manufacturing the laminated samples and the testing of the samples’ capacitance and pressure sensitivity.
3.2.1 Sample production – lamination
The fabric thickness was measured with a thickness measurement gauge tool, J100 Tasterform, on all fabrics. The conductive fabric, Shieldex Medtex 180, was then laminated onto both sides of the polyester fabrics with a polyamide web adhesive. All samples were laminated in a transfer press in 90°C for 45 seconds, two times; one time for each side of conductive fabric. The spacer fabrics were laminated as a larger piece and were cut in the different sample sizes. In the samples with the two thinner weft knitted fabrics, the middle layer was cut much bigger than the conductive fabric. This was done to prevent contact at the edges between the two conductive layers. Each fabric was made into samples of four different sizes: 1×1 cm, 2.5×2.5 cm, 5×5 cm and 10×10 cm.
A conductive thread was sewn onto each side of the conductive fabrics on all samples to simplify the connection to instruments. The samples were tested with the Agilent 34405A multimeter to make sure no contact occurred between the conductive layers.
3.2.2 Relative permittivity
The relative permittivity of the medium between the conductive surfaces affects the capacity. Therefore, the relative permittivities of the polyester fabrics were estimated with this method: each fabric was cut in 2.5×2.5 cm pieces and they were weighed and their volumes calculated. By means of the density of polyester, the weight of a solid polyester piece with the same volume was calculated. The mass of the fabric piece and the mass of the solid polyester piece were used to calculate the amount of air and polyester in the fabric. The relative permittivities of the fabrics were calculated with equation 3.1.
𝜀𝑟 = 𝜀𝑎𝑖𝑟 ∗ 𝑎𝑖𝑟 𝑎𝑚𝑜𝑢𝑛𝑡 + 𝜀𝑝𝑜𝑙𝑦𝑒𝑠𝑡𝑒𝑟 ∗ 𝑝𝑜𝑙𝑦𝑒𝑠𝑡𝑒𝑟 𝑎𝑚𝑜𝑢𝑛𝑡 (3.1) When the fabrics are compressed, the relationship between air and polyester amount will be changed as the volume changes. This will increase the amount of polyester in comparison to air, which will increase the relative permittivity.
3.2.3 Capacitance measurements
Before the capacitance measurements started with the laminated samples, the capacitance for ideal samples were calculated with equation 2.3. This ideal capacitance was used to compare with the measured capacitances.
20 3.2.3.1 Agilent U1732A LCR meter
The Agilent U1732A LCR meter was used to measure the capacitance of the different laminated samples at 1 kHz. The samples were fixed in a holder, which can be seen in figure 3.1. The holder was connected to the LCR meter and before the samples were fixed in the holder, the LCR meter was set to zero. When the value had stabilised it was recorded.
Figure 3.1. The Agilent U1732A LCR meter and the sample holder.
3.2.3.2 Agilent 34405A multimeter
The capacitance of the laminated samples was also measured with the Agilent 34405A multimeter. The sample holder was used here as well and the multimeter was set to zero before the sample was fixed in the holder. When the value had stabilised it was recorded.
3.2.4 Pressure sensitivity measurements of laminated samples An oscillator, which generates a square wave, was used to measure the samples change in capacitance when pressure was applied. The frequency of the output signal is determined by the capacitor, as can be seen in equation 3.2. A sketch of the circuit can be seen in figure 3.2. The circuit was built on a bread board and connected to an oscilloscope, Agilent DSO1012A, which in turn was connected to a computer. 10 000 measuring points of time and voltage were loaded into an Excel (Microsoft Corporation) document, which was used to calculate the frequency. The frequency is related to the capacitance through this formula:
𝑓 = 1
2 ln 3 𝑅3𝐶 (3.2)
21 Where f is the frequency, R3 is the resistor number 3 in the circuit and C is the capacitance. Equation 3.2 is from [26], in which it is derived.
Figure 3.2. The circuit of the square wave generator.
The circuit was kept in a faraday’s cage to shield it from noise and the samples were put in the same sample holder used in the capacitance measurements.
Pictures of the setup can be seen in figure 3.3 and 3.4.
Figure 3.3. A picture showing the faraday’s cage and the oscilloscope.
Figure 3.4. A picture taken inside the faraday’s cage, showing the circuit and the sample in the sample holder.
22 The 1×1 cm samples were too small to put any weights on them and were
therefore excluded from this test. The other samples were tested by first recording the frequency without pressure. This frequency was also used to calculate the capacitance of the samples, which were compared with the other capacitance measurements. A starting pressure was added, which was 603 Pa and equal for all samples. A square piece of hard plastic was used on each sample to distribute the pressure evenly and its weight was included in the starting pressure. Then weights of 50g, 20g and 6g were added separately and the frequency was measured
between each addition of pressure.
3.3 Second cycle – materials
The materials used in the second cycle were only yarns as the samples were woven.
3.3.1 Yarns
Several different yarns of different materials and quality have been used during the weaving process. The details of the different yarns are presented in table 3.2.
Table 3.2. Yarns
Name Tex number Material Type
Statex Shieldex 58 tex Silver plaited polyamide 2-ply, multifilament conductive yarn Trevira CS 73 tex Polyester 2-ply staple fibre yarn
Bouclé yarn 139 tex Polyester Bouclé, fancy yarn Velour yarn 204 tex Polyester Velour, fancy yarn
Interlaced yarn 135 tex Polyamide Multifilament, textured, interlaced Monofilament 51 tex Polypropylene (PP) Monofilament yarn
The Statex Shieldex yarn was used as warp and weft in the top and bottom layer for the woven samples. Trevira CS was used as warp and weft for the middle layer, but also as weft in the top and bottom layer. The bouclé, velour, interlaced and monofilament yarn were used as weft in the middle layer.
3.4 Second cycle – methods
The methods during the second cycle included a weaving process in which the samples were produced. Different shielding fabrics were also woven. The woven samples’ pressure sensitivity were tested and their capacitance measured. An experiment of the samples’ performance during moist and wet conditions was also performed. The effectiveness of the different woven shields was tested.
3.4.1 Sample production – weaving
After the first laminated samples had been made and tested, it was decided that the next samples should be a woven three layer structure. Two narrow warps were put up in a hand loom with double warp beams; one warp with the conductive fibres for the top and bottom layer and one warp with a polyester yarn for the middle layer. The thread count was the same for all layers and was 13 ends/cm.
23 The warp width was set to 4 cm for the top and bottom layer and 7 cm for the middle layer. The top and bottom layer were narrower than the middle layer to prevent the layers from coming in contact at the edges. The reed used during weaving was 65/100.
The weaving pattern for the top and bottom layer was plain weave. For the middle layer, waffle weave and rep weave were used. The three layers have to be
connected and bonded together in some way; otherwise three separated fabrics will be produced. In the first weave pattern that was used, the layers were connected by binding warp threads of the top and bottom layer to the middle layer. This was done during weft insertion for the middle layer. The warp threads of one shaft for each top and bottom layer were bonded to the middle layer by separated treadles. The draft of the setup of the loom for this weaving pattern can be seen in appendix A.
Binding the top and bottom layer with every forth warp thread to the middle layer gave too close binding points, which would only increase the risk for contact between the top and bottom layer. Therefore, the tie up of the binding treadles was changed after sample number 2. The new draft can be seen in appendix B. During weaving of sample number 4, it was realised that the binding points were better hidden if the warp threads were bonded during their long floats.
The binding pattern for the rep weave is shown in appendix C. It uses the same pattern for binding the layers as the waffle weave. The bonding points in the top and bottom layer were tried to be kept apart and about 9-13 picks were added between the bonding points in each layer.
The samples were woven about 7 cm long with the top and bottom layer about 1 cm shorter in each end. In most samples, Statex Shieldex was used as weft for the top and bottom layer. To test if less conductive weft is sufficient some samples were made with none or only two picks of conductive weft. A schematic picture of this is seen in figure 3.5.
Figure 3.5. A schematic picture of different amount of conductive weft in the top and bottom layer. Yellow is non conductive and black is conductive yarn.
24 3.4.2 Shielding fabrics production
The woven pressure sensors must be protected from surrounding electric fields during pressure measurements. Therefore, shielding fabrics were woven. Three different shields were woven with different densities of conductive yarn. Also, an isolating fabric was woven to isolate the pressure sensor from the shield fabric.
The warp used for the shielding and isolating fabrics was the same as for the pressure sensor. The isolating fabric was woven with only non conductive yarns.
The fabrics were woven about 30 cm long to be able to be folded around the woven pressure sensor samples. The weaving patterns for the different shields can be seen in figure 3.6.
Figure 3.6. The draft of the different woven shield fabrics.
3.4.3 Test of contact between the conductive layers
When the woven samples were finished and cut down from the loom, a short piece of Statex Shiedltex yarn was sewn to each side of the samples to simplify the connection to instruments. To test if there was contact between the conductive layers in the samples they were connected to the Agilent 34405A multimeter.
Weights with a total mass of 2100g were put on the samples and the samples showing contact were noted and did not move on to the following tests.
25 3.4.4 Capacitance measurement
The Agilent U1732A LCR meter was used to measure the capacitance of the different woven samples at 1 kHz. The test was performed in the same way as for the laminated samples.
3.4.5 Pressure sensitivity measurements of woven samples
The pressure sensitivity measurements of the woven samples were performed with the same type of circuit as in the test of the laminated samples, but this was
soldered together on a circuit card. The circuit was connected to the Agilent 34405A multimeter, which in turn was connected to a computer. The multimeter measured the frequency directly and it was recorded by the computer about every second. The samples were put in the same sample holder used in the capacitance measurements. A faraday’s cage was not used; instead the sample holder was packaged in a thin sheet of plastic and aluminium foil which was connected to ground.
To decide which weights to use in the pressure sensitivity test, a rough estimation of the pressure between the saddle, including a rider, and horse’s back was made.
The dimensions of an ordinary English saddle’s backside were roughly estimated and an estimation of the contact area of the saddle was calculated. A schematic picture of the saddle’s backside can be seen in figure 3.7. The total contact area was 0.183 m2. The weight of a saddle varies a lot depending on type and material, but a typical English saddle made of leather, with stirrups, saddle pad and girth weighs about 6-10 kg. In this case 6 kg was used. The weight of the rider was set to 70 kg. The pressure induced by the tightened saddle girth was not included. The pressure was calculated with this formula:
𝑃 =𝑚𝑔
𝐴 = 76∗9.82
0.183 = 4078 𝑃𝑎 (3.3)
The weight needed was calculated by using the mean value of the woven samples’
areas, which was 0.00147 m2. The weight needed:
𝑚 =𝑃𝐴
𝑔 = 4078 ∗0.001465
9.82 = 609𝑔 (3.4)
This calculated pressure was added to simulate the pressure from the saddle and rider and then smaller weights were added to simulate pressure changes. The exact weights used were 38g, 568g, 100g, 50g and 20g. The first 38g was a plastic plate used to distribute the pressure evenly over the sample area. The multimeter measured the frequency constantly and weights were added and removed with about 15 seconds in between. The test sequence is presented in table 3.3.
26
Figure 3.7. A schematic picture of the backside of an English saddle and a rough estimation of the dimensions.
Table 3.3. Pressure sensitivity test sequence.
Weight Time [s]
No weight 1-15
+ 38g 16-30
+ 568g 31-45
+ 100g 46-60
+ 50g 61-75
+ 20g 76-90
- 20g 91-105
- 50g 106-120
- 100g 121-135
- 568g 136-150
- 38g 151-165
3.4.6 Pressure sensitivity test of woven samples with woven shields The test of the woven shields was performed in the same way as the pressure sensitivity measurements of the woven samples, except for that the sample holder could not be used. Instead, the shield and isolating fabric were folded around the woven sample and it was put flat on the table. The cables from the sample to the circuit were shielded too as they also were sensitive to noise. The shield fabric and the cable shield were connected to ground.
The plastic plate used in the test of the pressure sensitivity of the woven samples was too big to be used. Instead a smaller jar, filled with water to weigh 38g, was used. The jar’s surface was flat and large enough to cover the whole sample. The test sequence was the same as in the pressure sensitivity test for the woven samples with the aluminium foil shield. As the 38g plastic plate weight could not be used any longer, a few samples were also tested without it to find how the samples reacted to the immediate large increase/decrease in pressure.
27 3.4.7 Effect of moisture
An experiment with two different woven samples during moist and wet conditions was performed to test the effect of moisture on the pressure sensors. The test was made in three steps: the first step was to test the sample dry, in the next test was a humidifier used to increase the relative humidity and in the last test a spray bottle was used to spray water on the sample.
The tests were made in the same faraday’s cage that was used during the pressure sensitivity test of the laminated samples. The sample holder from the tests of the laminated samples was also used. The same soldered circuit used in the pressure sensitivity test of the woven samples was used and it was connected to the Agilent 34405A multimeter, which in turn was connected to a computer. The multimeter measured the frequency every second which was recorded by the computer.
The dry tests were performed in a relative humidity of about 40% and a
temperature around 20°C. The frequency was measured for about 1 minute before 38g, 568g and 50g weights were added and removed with about 15 seconds in between. The starting 38g weight was a hard piece of plastic used to distribute the pressure evenly over the sample surface.
The test with increased relative humidity also started by measuring the frequency for 1 minute before the mist from the humidifier was added. The mist from the humidifier was blown into the faraday’s cage by a fan placed in one side of the cage. A tube with a funnel in each end guided the mist from the humidifier to the fan. The relative humidity was measured with a simple hygrometer placed inside the cage. The relative humidity was increased up to 90% inside the cage, then the humidifier and fan was turned off and the shutter was closed in front of the fan.
The frequency was measured for about 1.5 minutes before the weights were added and removed. The same weights and interval as in the dry test were used.
The last test started with a dry sample and the frequency was measured for 1 minute. Then the sample was sprayed with water from a spray bottle once, from about 1 dm distance. After 5 minutes, the weights were added and removed with 15 seconds in between.
28
4. Results
The results from the different experiments and tests are presented in this chapter.
4.1 First cycle
In the first cycle, the laminated samples were manufactured and their pressure sensitivity and capacitance were measured.
4.1.1 Laminated samples
Some of the samples are presented in a picture in figure 4.1. In figure 4.2 the cross sections of the spacer laminates are shown. Four different sizes were made: 1×1 cm, 2.5×2.5 cm, 5×5 cm and 10×10 cm. Unfortunately, there were always contact between the layers of the 10×10 cm single jersey sample, even though several samples were made.
Figure 4.1. Some of the laminated samples. The top three samples are the ones with single jersey fabric as middle layer. The middle ones are with 1:1 rib fabric and the four at the bottom are with
the 4.4 mm thick spacer fabric.
29
Figure 4.2. The cross section of the laminated spacer fabrics. From the top: 3.3 mm spacer fabric, 4.4 mm spacer fabric and 6 mm spacer fabric.
4.1.2 Capacitance measurements
In table 4.1, the results of the measured and calculated capacitances of the laminated samples are presented. In figure 4.3, the capacitance measured with Agilent U1732A is divided with the calculated capacitance with equation 2.3 to see how much they differ.
Table 4.1. Measured and calculated capacitances of the laminated samples.
Sample Calculated C
[pF]
Agilent 1732A [pF]
Agilent 34405A [pF]
Square wave generator [pF]
Single jersey 1×1 cm 2.79 4.0 4 36.59
Single jersey 2.5×2.5 cm 17.41 20.3 21 56.49
Single jersey 5×5 cm 69.63 73.8 80 115.32
1:1 rib 1×1 cm 1.19 2.6 3 34.82
1:1 rib 2.5×2.5 cm 7.45 11.2 13 43.85
1:1 rib 5×5 cm 29.82 38.1 46 77.58
1:1 rib 10×10 cm 119.27 137.1 159 192.69
3.3 mm spacer 1×1 cm 0.30 1.4 2 32.51
3.3 mm spacer 2.5×2.5 cm 1.88 4.1 5 35.24
3.3 mm spacer 5×5 cm 7.51 12.9 13 46.89
3.3 mm spacer 10×10 cm 30.02 47.1 48 98.72
4.4 mm spacer 1×1 cm 0.23 0.9 2 30.89
4.4 mm spacer 2.5×2.5 cm 1.45 3.1 4 34.61
4.4 mm spacer 5×5 cm 5.80 9.3 12 41.78
4.4 mm spacer 10×10 cm 23.18 31.1 32 80.18
6 mm spacer 1×1 cm 0.16 0.8 2 29.69
6 mm spacer 2.5×2.5 cm 0.97 2.9 4 34.20
6 mm spacer 5×5 cm 3.88 7.4 8 40.13
6 mm spacer 10×10 cm 15.52 23.6 29 71.79
30
Figure 4.3. The capacitance measured with Agilent 1732A was divided with the calculated capacitance and shows how much the measured capacitance differs from the calculated
capacitance.
4.1.3 Pressure sensitivity measurements of laminated samples The measured frequencies were used to compare the samples. By dividing the difference in frequency with the difference in pressure, between every added weight, a sensitivity factor was calculated. This was used to compare the samples and see how sensitive they were to the added pressures and also if they were consistent in their response. In figures 4.4-4.6 graphs of the sensitivity factor is presented. If a sample would have been equally sensitive to the added pressures, the graph would be a straight line. The frequency decreases when the samples are compressed, thus the sensitivity factor becomes negative.
Figure 4.4. Sensitivity factor of the 2.5×2.5 cm samples.
0.00 1.00 2.00 3.00 4.00 5.00 6.00
1×1 cm 2.5×2.5 cm 5×5 cm 10×10 cm
Agilent 1732A, C/Calculated, C
Single jersey fabric 1:1 rib fabric 3.3 mm spacer fabric 4.4 mm spacer fabric 6 mm spacer fabric
-0.90 -0.80 -0.70 -0.60 -0.50 -0.40 -0.30 -0.20 -0.10 0.00 0.10
603.34 787.17 314.24 94.27
∆F/∆P
∆P
2.5 x 2.5 cm
Single jersey 1:1 rib
Spacer 3.3 mm Spacer 4.4 mm Spacer 6 mm
