update on the project Slowcreen
In the last article on Slowscreen I wrote about Pauline and I working on a much larger screen in order to display more complex shapes on the thermochromic cloth. We have spent a while sewing and soldering our various components and most of what we anticipated worked as expected. The schematics turned out to be accurate, except for a small detail on working with the MOSFETs (which I’ll mention in the corresponding caption).
The schematics does not include the schematics of the fabric but is fairly accurate for the rest of the electronics. Click/tap for full resolution.
On the left, the fabric with its 750 modules. It’s all flexible, so it can be used in a variety of contexts. Ribbon cables lead the current in the fabric, where it heats the modules, and out to the ground.
Each module is composed of 4 resistive threads, and has its own coordinate system from A to Y and 1 to 30 for easier coding and bug-checking .
On the other side of the fabric, one diode and three crimp beads for each module. Copper threads connect everything.
The 750 diodes are all in the same direction. They are necessary because they stop the current from reflowing “up” the fabric.
The connection between each row and column is done on the fabric, on the left and bottom. The current flows in horizontally and is allowed to escape at the bottom. Colored ribbon cable is really necessary here to keep track of where each thread goes and leaves.
Ribbon cables use connectors in order to separate the fabric from the command module. Much easier to transport this way, and each can also be repurposed.
The ribbon cable is then separated and each cable is soldered onto a 5mm copper strip. These copper strips are actually tape, which is pasted on a plexiglas 30×30 cm square. This is an alternative to regular PCB, especially useful when the width of the copper strips needs to be controlled. In this case, the large copper strips allow for the current to flow easily, with very few resistance.
A battery of MOSFETs (of the N-Channel type) act as electric switches. MOSFETs immediately before the ribbon cable that leads to the rows of the fabric allow for the current (5V) to flow to the fabric where needed. MOSFETs placed after the ribbon cable that comes from the columns controls the flow of the incoming current to the ground (GND).
The great thing with using plexiglas is that it is easy to drill tiny holes that lead to the other side. You can see diagonal strips of copper pasted on the bottom side of the plexiglas, that transport the 5V directly from the source (red crocodile clip at the bottom of the photo).
What is called the “gate” of the MOSFETs is what controls the flow of the current between the other 2 pins (the source and the drain). Here the gate pin is folded toward the back of the MOSFET and goes to another cable, connected to a shift register.
Shift registers also act as electronic switches. The difference is that they can control a bunch of outputs, when a specific signal is sent (i.e. 00001000 means “let the current flow only on the fifth pin”). They theoretically could be connected to the ribbon cables (5V 1A) directly, except that they would not be able to dissipate the heat that they produce fast enough, and could burn out in the process. Instead, they just use the lower power from the Arduino to control very effective MOSFETs.
On this picture you can also see the resistors. One 10 kOhm for each MOSFET, which is something I discovered in the process: the current sent to the MOSFET to tell it to open needs to be removed when it is supposed to be closed, but has nowhere to go without the resistors. These resistors go straight to the ground for this reason.
Shift registers can be daisy-chained in order to minimize the number of outputs needed to control them. They share the same “pipeline” so the information travels between them. See the blue, yellow and green cables that enable the connection. The red and grey cables bring the 5V and GND from the Arduino to the breadboard.
The second circuit, on top, is there to control the columns. On this one, the MOSFETs connect the ribbon cable to the GND. The rest of the circuit is the same. The leftmost shift register only controls one MOSFET because there are only 25 columns (letter A to Y) and each register is able to control 8 outputs.
The Arduino that controls the whole thing. The program is embedded directly in there when playing back an animation, so it just need to be plugged in. Meanwhile, I am still exploring the option to use a Rapsberry Pi as an intermediate between a web browser and the Arduino, to get the electronics out of the way and just control everything from the browser.
Most of the work is now done, and it’s all moving along pretty nicely. Some optimizations still need to be made to get everything running smoothly on the electronics side, and the Pi hasn’t been connected yet, but I tend to think that the most time consuming part is over. Will document on the next few steps.