Yunjia Textile · Technical guide
Textile Sensors & E-textile Guide
How fabrics and threads sense pressure or stretch, how to connect them, and which test data to ask for. Includes resistance curves measured on our own piezoresistive fabric.
Key takeaways
- Pressing or stretching a textile changes its resistance or capacitance. The structure decides which.
- The sensing region should change; leads and joints should not. Test them separately, then together.
- Fabric weight doesn't predict response. In our tests, the 210 and 220 GSM fabrics differ 18-fold at 0.1 kgf, so ask for a curve for each fabric.
- A number means little without its definition, loaded area and readout.
How do textiles sense?
Pressure or stretch rearranges the fibers, loops or layers in a textile. A resistive design reads that as a change in resistance; a capacitive design reads it as a change in capacitance.
Press → resistance drops
R ↓
More fiber contacts, easier current path
Press → capacitance rises
C ↑
Electrodes move closer; insulation keeps them apart
Stretch → resistance changes
R ↕
Loops and contacts in the knit rearrange
One input, different ways to read it
You apply pressure or stretch; you measure resistance or capacitance. In a resistive sensor, load changes how easily current flows through touching fibers. In a capacitive sensor, load squeezes an insulating layer (the dielectric), so two electrodes move closer.
Know which part should change
The sensing region should change in a useful, repeatable way. The leads should change as little as possible, and elastic yarn alone doesn't guarantee that.8 Test the sensing region and its connections separately, then together.
This guide covers resistive and capacitive sensing; piezoelectric and triboelectric sensors need their own guide.
What makes up a textile sensor?
A textile sensor is a sensing region, electrodes, leads and a readout. Decide which contacts should move and which must stay fixed, and remember that a reading includes everything between its terminals.
Start with the roles, not the layer count
Layouts follow the mechanism. A resistive fabric can carry both electrodes on one side, and a knit can sense stretch with no layered stack at all.1,7
Decide which contacts move and which stay fixed
Some contacts must move to sense; a permanent joint must not. Decide which is which before you sew, bond or encapsulate. In one study, hot-melt bonding placed directly on the sensing interface sharply reduced the response.2
A sensing point is not yet a pressure map
An array has to tell where as well as how much. Shared wiring can make one pressed point show up on another channel (crosstalk), and a cover can spread the force. Check the map with the multiple contacts you expect in use.
Know what the readout includes
A two-wire reading includes everything between its terminals, leads and contacts too. A four-wire setup removes most of that.10 A voltage-divider output depends on the circuit, so don't compare it directly with a resistance reading.
What roles do materials and methods play?
Give each material one job: sense, make contact, carry the signal or hold the joint. The weak spot is usually where conductive thread meets a rigid module.
Fabric: a sensing region, electrode or support.
Yarn & stitches: a placed sensing path, lead or contact pad.
Flexible connection: a signal route that has to survive movement.
Read a pattern as a possible circuit
Before choosing a layout, map which stripes or areas conduct, which are isolated and where they end. Check again after cutting and attaching. Appearance alone doesn't prove an electrode pair or a sensing matrix.
For yarn and fabric basics, see the Conductive Textile Structure Guide.
Placement can matter as much as material
Where the conductive yarn sits in a knit changes which contacts move when it stretches, and so changes the signal.7
A sewn route needs a reliable end connection
Sewing and embroidery put conductive thread exactly where it's needed, as a sensing path, electrode, lead or contact pad. The weak spot is usually where the thread meets a rigid module.
Compare the construction, not just the metal: silver-plated polymer thread, silver-coated copper hybrid yarn and stainless-steel thread are built differently. Give the joint as much care as the thread: in one embroidered connection, simply crossing the yarns did not hold, while a larger pad with polymer fixation did.10
Closely spaced stitch lines need a clean thread
When conductive thread is sewn or embroidered in closely spaced lines, such as parallel leads or neighboring electrodes, uniformity matters. One stray fiber can bridge the gap and cause a short that comes and goes as the fabric moves.
Stitch a test piece at the real spacing and machine settings. Check continuity along each line and insulation between lines, then check again after bending.
Our 260D silver conductive sewing thread on production spools is specially treated to reduce hairiness. For first trials, a 50-yard (45 m) spool of 650D thread is also available.
Can it stretch, recover and keep carrying the signal?
Stretchy doesn't mean stable: repeated stretching can shift a connection's resistance.8 Ask three questions:
- During stretch: does the signal change or drop out?
- After release: do the length and resistance return?
- After many cycles: are the conductor and both ends still sound?
How should test results be read?
Ask for a resistance curve measured under stated conditions, and keep the absolute values. Hysteresis, drift and repeatability show whether a reading returns and holds.
Ask for a load–resistance curve, not just a datasheet
Width, weight (GSM) and surface resistance describe a flat sheet, not how it responds when pressed. For pressure sensing, ask for a resistance curve for each fabric, measured with a stated load on a stated electrode.
Weight alone predicts little. In our tests, the 210 and 220 GSM fabrics differ 18-fold at 0.1 kgf, while the 130 and 240 GSM fabrics started far apart and read within about 2% of each other from 0.8 kgf up.
Resistance vs. load, five fabric weights (Yunjia test data)
A curve describes the material in one setup, not a finished sensor. Electrode size, preload, covers and readout all shift the numbers, so compare curves only under matching conditions. The fabric shown is our piezoresistive pressure-sensing fabric, available in all five weights.
Hysteresis
Does it come back the same way?
Drift
Does the reading hold steady?
Repeatability
Does another run agree?
Keep the baseline and working range
Normalized curves such as (R − R₀)/R₀ show relative change but hide the starting resistance. Keep the absolute reading too, and note whether the baseline was reset each cycle, because resets can hide drift.
Read the definition behind the number
Two “hysteresis” percentages can use different formulas, and a steep slope over a small range is not the sensitivity of the whole range. Response time needs a threshold and sampling rate; resolution needs a noise level.
Match the load, area and repeat being compared
Pressure is force over the loaded area (P = F/A), not the whole cloth, which is why our curve is given as load on a fixed 1 cm-diameter electrode. Also say what was repeated: another cycle, location, specimen or batch. A clean trace on one sample doesn't prove consistency across production.
Cycle count isn't durability: measure what changed
A cycle count says how much stress was applied; the change afterward says what it did. A broken conductor can regain contact without truly recovering, so watch for interruptions during use, not just after.9 One laundering test can't set care rules for every assembly.
For test engineers
Background for setting up or comparing tests. Items marked “From the literature” are results reported in the cited papers, not specifications for Yunjia products.
Why similar materials sense in different ways
From the literature
The arrangement gives a material its role. A carbon-treated nonwoven can be the pressure-sensitive path,1 while carbonized cotton around a compressible dielectric becomes the two faces of a capacitor.6 In resistive stacks, contact with the electrodes is part of the signal; in knitted strain sensors, the contacts between loops are. These are models of particular structures, not measurements of every contact.3,7
Attachment and joints
From the literature
In the hot-melt test, heat damage and adhesive soaking into the fabric were both possible causes; they were not separated.2 Bonding away from the sensing interface is the alternative: conductive strips on a non-conductive backing, or conductive tape with its non-adhesive face toward the sensing cloth.3,4 These examples use different materials, so they don't point to a best adhesive. The capacitive design in the layer diagram uses TPU-reinforced carbonized cotton electrodes, a porous Ecoflex dielectric and stitching to hold the three layers.6
Connections under movement
From the literature
Baseline shift: a 20 cm silver-coated PBO/rubber-core yarn rose 137.87% in resistance after 2,150 cycles to 25% strain, an absolute increase of only 0.0781 Ω. It was read after cycling, so brief interruptions during motion would be missed.8 Rigid ends: woven interconnects with a soft transition at an encapsulated board reached the 100,000-cycle test limit in 180° bending at 1 Hz.9 Joints: ultrasonic welding can fix the microwires of a hybrid yarn by softening its polyester, but broken microwires appeared among failed joints, so check resistance and physical damage together.10
Sensitivity and hysteresis definitions
From the literature
A large slope may cover a small interval. One resistive fabric sensor reports normalized-current slopes of 0.43, 0.18 and 0.10 kPa⁻¹ over 1–4, 4–40 and 40–150 kPa;3 pairing the largest slope with the full range would overstate sensitivity.
The same change can give different percentages: at constant voltage, halving resistance doubles current, so −50% in resistance is +100% in current (arithmetic, not data).
| Research result | How it was calculated |
|---|---|
| Resistive fabric · about 11.7%3 | Largest loading–unloading gap ÷ full-scale response; cycle to 150 kPa, largest gap at 70 kPa (Suppl. Fig. S1). |
| Capacitive stack · 8.8%6 | Area between the two curves ÷ area under the loading curve; 0–100 kPa (Eq. 3, Fig. 7). |
Different devices, ranges and formulas mean these numbers can't rank the materials.
Readout settings
From the literature
Readout settings answer different questions: a 100 kHz AC excitation drove one capacitance measurement;5 a glove patch was sampled every 50 ms with a ten-point moving average.6
What makes two tests comparable
- Specimen and assembly: material, size, direction, electrodes, attachment, support and number of samples.
- Loading: contact area, preload, range, rate, dwell, recovery, cycle history and environment.
- Readout: measured quantity, contact positions, excitation, sampling, filtering, baseline and formula.
- Outcome: absolute and normalized readings, variation, failures and pass criterion.
From the literature
The loading tool sets the nominal area. With 20 × 20 mm plates (400 mm²), 6 N is an average 15 kPa and a 0.5 N preload is 1.25 kPa; local stress in the textile can differ.2
Keep resistance units distinct: Ω is terminal to terminal; Ω/m needs a defined length and tension; Ω/□ needs a sheet geometry; Ω·m needs a bulk geometry. Don't convert between them with assumed dimensions.
Durability and other reported results
From the literature
One pressure sensor lost about 4.3% of its response after 2,000 cycles at 30 kPa.3 Woven and knitted electrodes responded differently in one capacitive stack, though their fabric weights also differed.6 One textile reads strain through coated-yarn resistance and pressure through an ion-gel capacitance channel; each was shown with the inputs applied separately, so for real use apply each input alone, then together.5
How are textile sensors designed for use?
Decide what the sensor must tell you, then test it mounted, with its real cover, motion and readout.
Define the task before choosing a material
Detecting touch, following movement, estimating force and mapping pressure are different jobs. Set the working range, spatial detail and acceptable variation before you compare materials.
Mounting becomes part of the sensor
Follow bending
A knitted sensing region on a mechanical joint gave a clear bending signal.7 Reading a joint angle on a person still needs calibration for fit and slippage.
Sense a grasp
On a glove, grip and contact area change with every hold, so a reading describes the grasp, not a calibrated finger pressure. Test with the real hand movements you need to detect.
A soft surface adds its own movement
On a curved or inflatable body, the surface under the sensor moves too: one fabric array gave noisier readings on a soft robotic link than on a flat rig.4 For seats or insoles, also test sustained load and repeated positioning.
Sensor fabrics, conductive threads and related electrode materials are in our textile sensor collection.
References
Sources for the research examples above. The examples explain design choices; they are not specifications for our products. Diagrams are our own simplified drawings; no figures or curves from these papers are reproduced.
- Xu, H. et al. (2020). Flexible Waterproof Piezoresistive Pressure Sensors with Wide Linear Working Range Based on Conductive Fabrics. Nano-Micro Letters, 12, 159. DOI: 10.1007/s40820-020-00498-y.
- Okss, A., Valisevskis, A., Briedis, U. & Baltina, I. (2023). Functional fabrics for multilayered textile pressure sensors: comparison of structures and sensitivity characteristics. Engineering for Rural Development, 419–426. DOI: 10.22616/ERDev.2023.22.TF089.
- Zheng, X. et al. (2025). Daily carbon ink functionalized conductive fabric for simple, low-cost, and wide range flexible pressure sensor with multifunctional applications. Diamond & Related Materials, 158, 112661. DOI: 10.1016/j.diamond.2025.112661. The hysteresis comparison uses supplementary Fig. S1.
- Day, N. et al. (2018). Scalable fabric tactile sensor arrays for soft bodies. Journal of Micromechanics and Microengineering, 28, 064004. DOI: 10.1088/1361-6439/aab221.
- Keum, K. et al. (2022). Mechanically robust textile-based strain and pressure multimodal sensors using metal nanowire/polymer conducting fibers. iScience, 25, 104032. DOI: 10.1016/j.isci.2022.104032.
- Ko, Y., Vu, C. C. & Kim, J. (2021). Carbonized Cotton Fabric-Based Flexible Capacitive Pressure Sensor Using a Porous Dielectric Layer with Tilted Air Gaps. Sensors, 21, 3895. DOI: 10.3390/s21113895.
- Kim, Y.-H. et al. (2024). Assessing the Role of Yarn Placement in Plated Knit Strain Sensors: A Detailed Study of Their Electromechanical Properties and Applicability in Bending Cycle Monitoring. Sensors, 24, 1690. DOI: 10.3390/s24051690.
- Schwarz, A. et al. (2011). Electro-conductive and elastic hybrid yarns – The effects of stretching, cyclic straining and washing on their electro-conductive properties. Materials & Design, 32, 4247–4256. DOI: 10.1016/j.matdes.2011.04.021.
- Veske, P., Bossuyt, F., Thielemans, F. & Vanfleteren, J. (2023). Measuring the Flex Life of Conductive Yarns in Narrow Fabric. Micromachines, 14, 781. DOI: 10.3390/mi14040781.
- Dils, C. et al. (2022). Interconnecting embroidered hybrid conductive yarns by ultrasonic plastic welding for e-textiles. Textile Research Journal, 92(21–22), 4501–4520. DOI: 10.1177/00405175221101015. A 2023 corrigendum updates the funding statement.