Commit 5f9035d5 by Gilles Decroly

(almost) finishing week 18

parent 7480cd4d
Pipeline #97618 passed with stages
in 28 seconds
......@@ -25,13 +25,7 @@ Some scientific publications:
## Making single thread artificial muscles
As reported in the scientific papers, there are multiple ways to include constrains in the thread and make those artificial muscles. We will focus on two ways: even by twisting it on themselves, even by coiling them around an axis.
### Solution 1: twisting the thread
......@@ -39,7 +33,7 @@ Some scientific publications:
To twist the tread, we imagined a small setup controlling the number of turns:
> image
![](../images/week18/principle.jpeg)
The key parameters are the following:
......@@ -53,7 +47,7 @@ The key parameters are the following:
Basically, we want to know the number of turns to apply and the weight we need to twist the thread. Based on the paper [Artificial muscle from fishing line and sewing thread](https://science.sciencemag.org/content/343/6173/868)), here is a little script that allows to estimate those values as a function of the thread diameter and its length.
Note that this estimation is only valid for small thread diameter (between 0.1mm and 0.3 mm).
** overestimate for small diam and underestimate for large diam **
**After testing it, it overestimates the number of turns for small diameters and overestimates it for large diameter. It can however be used to estimate a number of magnitude.**
<form id="flexForm">
<table>
......@@ -109,7 +103,6 @@ Note that this estimation is only valid for small thread diameter (between 0.1mm
#### Twisting bench
> picture
The twisting bench is composed of the following:
......@@ -118,55 +111,48 @@ The twisting bench is composed of the following:
* A belt and a bearing for the transmission of the rotation
* An optical turn counter
* A small PCB to control the setup
* A LCD and a button as interface
* Some 3D printed pieces (supports and axis)
The test bench is made modular and through to be placed on an optical breadboard.
> stl file
> code of microcontroller
The test bench is made modular and designed to be placed on an optical breadboard. <a href="../files/week18/stl_files.zip" download="stl_files.zip"> The stl files can be downloaded here </a>
The thread can directly be fixed on the axis, or through a paperclip, idem for the weight. The setup must be placed on a high enough support (i.e. a table), and the weight must be place to slide along a vertical wall, to limit its rotation.
![](../images/week18/setup_twist.jpeg)
![](../images/week18/setup_twist2.jpeg)
![](../images/week18/setup_twist3.jpeg)
>Image
70 cm
test1= 1550 tunrs (program): too musch: broke (cf image)
> code of microcontroller
(but weight ok)
The thread can directly be fixed on the axis, or through a paperclip, idem for the weight. The setup must be placed on a high enough support (i.e. a table), and the weight must be place to slide along a vertical wall, to limit its rotation. It is possible to set a number of turns using the small interface.
test 2: 1222 turns and more slowly
After one or two tests for the calibration, we were able to make easily twisted thread. The thread is subject to break if the weight is too heavy (but the program above is good at this level), or if the number of count is too high!
A good solution is to set first a lower number of turns and increase it step by step. After Twisting, the length reduces almost by a factor 4!
ok! (even too much?)
#### Testing the thread
* test 1: 0.2 mm, 200g for twisting, 4oog Weight
--> once heted: elongates, but can aslo observe small contractions
* Same test and 550g
We made some tests to with the fabricated threads. The results were not as impressive as in the papers, but we were able to obtain a contraction of the thread. We suppose that it could be improve by controlling better other parameter (weight, actuation temperature, coil ratio, type of nylon, ...).
first heting: elongates, then contracts (just a little bit)
--> problmes: number of coils!, heating!
To obtain contraction, an heavier mass than used for the twisting must be lifted, to elongate a little bit the twisted thread, and allow it to contract. **The thread must be first heated once, during which the thread will elongate**. The contraction will be obtained when the thread is heated again. We used a hair drier to heat the artificial muscle.
Here is an example of results with a 0.2 mm diameter thread:
* thread with our bench
![](../images/week18/test_02.gif)
85g + 240: nylon broke!
With a 0.3 mm diameter thread and a weight of 240g (two times the weight used for twisting):
<video width="400" controls>
<source src="../images/week18/test_03.mp4" type="video/mp4">
Your browser does not support HTML5 video.
</video>
* thread with our bench
And with a 0.3 mm diameter thread and a weight of 240g (two times the weight used for twisting). It was this time heated using Joule effect (the thread broke when the heat generation was concentrated). Note that the lower actuation time is due to the low increase of temperature.
85g + 120: nylon broke when heated!
* Threas 0.325
adapt the count number to be less dependant of the thread diam
works! firste heating then see it appear
* Idem with heating wire:
works also but break
<video width="400" controls>
<source src="../images/week18/test_03_wire.mp4" type="video/mp4">
Your browser does not support HTML5 video.
</video>
### Solution 2: coiling around an axis
......@@ -174,28 +160,52 @@ works also but break
The other solution to coil to make the artificial muscle is to coil it around an axis. This solution is more suitable for larger thread diameters. We also tested it, and it gave good coiling results!
> setup image
![](../images/week18/coil_setup.jpeg)
![](../images/week18/coil1.jpeg)
![](../images/week18/coil2.jpeg)
#### Testing the thread
1 mm and weight of 25 g
However, when testing the thread, we were not able to see contraction. The problem was that no constrained had be included in the coil.
--> no muscle effect.. bit plastic deformation
2 mm and 28g: Idem
![](../images/week18/coil_test.jpeg)
without wight: idem: it untwist and elongates..
We tried to make coils with constrains, but it broke quite rapidly...
## Making a small loom to make fabric
With the working twisted thread, we want to make a small "smart fabric". To do this, we first made a [mini loom frame weaving, found and nicely documented on instructables](https://www.instructables.com/id/Laser-Cut-Mini-Frame-Loom-Weaving/).
Making the frame with laser cutting:
![](../images/week18/frame.jpeg)
And making the fabric:
![](../images/week18/weaving1.jpeg)
![](../images/week18/weaving2.jpeg)
![](../images/week18/weaving3.jpeg)
![](../images/week18/weaving4.jpeg)
![](../images/week18/weaving5.jpeg)
![](../images/week18/weaving6.jpeg)
![](../images/week18/weaving7.jpeg)
![](../images/week18/weaving8.jpeg)
![](../images/week18/weaving9.jpeg)
![](../images/week18/weaving10.jpeg)
https://www.instructables.com/id/Laser-Cut-Mini-Frame-Loom-Weaving/
When unloading the fabric from the frame, we have to ensure that the twist in the threads is not released, so we made small laser-cut fixations. If the fabric is not maintained, it will twist macroscopically, due to the constrains of each thread. <a href="../files/week18/stl_files.zip" download="stl_files.zip"> The laser-cutted files can be downloaded here </a>, or on the [instructables page](https://www.instructables.com/id/Laser-Cut-Mini-Frame-Loom-Weaving/) for the frame.
photo
## Testing the artificial muscle
## Testing the artificial muscle fabric
We could finally test the fabric. After a first heating causing an elongation, we were able to observe contraction!
## Future improvement
<video width="400" controls>
<source src="../images/week18/test_fabric.mp4" type="video/mp4">
Your browser does not support HTML5 video.
</video>
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment