Hackerspace textile, 11-15 rue de la Réunion, Paris

Outils pour utilisateurs

Outils du site


Etextile matrix sensor V1.1


  • Size : 35 cm square
  • Physical Resolution : 256 sensors (16 rows and 16 columns)

Bill of materials

ID Product Ref Size Quantity Price
0  Fabric substrate NA 37 x 37cm 2 NA
1 Conductive fabric TODO 37 x 37cm 2 TODO
2 Eeonyx piezoresistive fabric Eeonyx @ Hitek 20K ohms per square 30 x 30cm 1 TODO
4 Statex conductive thread 117/17 2-PLY HC + B 4 meter 1 TODO
5 Teensy board Teensy @ Lextronic 3.2 NA 1 18.89€
6 Etextile matrix sensor PCB V2 Sources ? 1 20€

Sewing process #0

  • Each conductive strip measures 18mm wide
  • The space between strips is 1/2 mm

Etching process #1

This method will be used for small production quantity. The process details are presented in the DataPaulette materials library.

Physical Resolution : 256 sensor

The e-Textile bus

This component is the black ribbon on the side of the sensor. The conductive lines are insulated into the textile ribbon. The white lines are indications to help during the assembling process.


  • Conductive lines : 16
  • Conductive lines material : SHIELDEX conductive thread 235f34 2 X 2 X 2 PA Ag / nm 8 (10 meter = 1,25g)
  • Pitch : 1,54 mm

The eTextile matrix

The eTextile matrix is composed of two fabric layers, one for the X axis one for the Z. In the first prototype we were using fusible conductive fabric but it will be easier to use a zebra fabric. This zebra fabric will contain 3mm conductive traces spaced from 1.5 to 2mm with non conductive lines. The picture below shows the first knitted prototype.

The piezoresistive layer

This layer is a fabric that allows pressure sensing. It is made out of fabric composed by polyester and carbon dyeing process. We plan to make our own DIY formula to give access to everyone to this useful material.


Technical Evaluation

On the X and Y axis, the current sensor uses 16×16 conductive stripes and on the Z axis we have analog values on 13 bits (but we don't really need more than 8 bits). For the visualization and finger blob detection (with openCV), the analog Z allows interpolating X and Y from 16×16 to 64×64.

a) Are the 16 strips independent?
Yes, if we touch strip number 3, strip number 2 and strip number 4 won't bet triggered.

b) Is the interpolation useful?
It only helps the blob detection, for now we use open frameworks and open CV but we plan to port the algorithm used by openMV.


We performed a series of tests for the technical evaluation:

a) placing object of different weights on the cloth, to check if the observed pressure is both constant and what the smallest measurable difference is (report variability in measure or minimum required difference)

b) placing weights of different size but identical weight on the cloth to determine when cross-talk starts (report the largest size/weight combination that does not cause cross-talk)

c) placing the same weight at different locations (that we measure with a caliper) and comparing its actual position (report the difference in observed and measured position as a percentage of the spacing of the strips, report and discuss if there are repeating distortions due to your interpolation)

The results are available as:

atelier/projets/matrice_textile.txt · Dernière modification: 2018/03/13 14:12 par maurin