Project Logs

Collapse

• Hardware data

  mclien • 06/10/2015 at 08:13 • 0 comments

  I loaded up the hardware docu for 2 Versions (sorry this took ages):

  V1 is the one is the exact one of the logs (might miss some bits and pieces)

  V2 is improved to make the build a bit es tedious mainly by using WS2812B already soldered on a lttle PCB breakout board, which also holds a capacitor 8current stabilization) and a resistor (over volatage protection). The V2 is not testet by me in the Türtris configuration, but the Breakouboard are used in sisams social gaming cube and apear to be less prone to damage while soldering and less prone for failure in use (in the door some LEDs miss one color).

  If you note something missing in the docu, make a version with other measurements or have any other questions/ suggestions, feelfree to ask.

  Note: V3 is planned to use lines of read soldered LED stripes, but will require quite some rework, because the grid has to fit the stripe pattern)
  

• Software

  mclien • 06/03/2015 at 13:48 • 0 comments

  I finally managed find teh arduino sketch back and uploaded it here.

  http://mclien.de/files/tuertris/tuertris_sw.zip

  Files of the design are to follow
  

• Soldering

  mclien • 01/07/2015 at 15:13 • 0 comments

  Having glued all wires in place, I first solderd the main power "rails" in place, which is easy: just lay it crossed to the "V"shaped parts that will supply the LEDs and solder it in place. I didn't interconnect them to each other at this time, to be able to test each column separatly. First step is to drop a LED in each "box" from the rail, then I pressed them into the holes with a tweesers and flipped the whole thing over to solder them. Starting with the data connection:

  Here we see the data connection already soldered and the preperationfor the trickier one: SOldering the power connection with a capacitor of 3.2x1.6mm. While these little parts tend to stick to the liquid solder with more force tha the gravity will hold them down, one will need the ever missing 3rd hand (that is the one holding the 1mm wire pressing onn the cap from top). My solution is this: Just some piece of metal with a wire taped to it, so it hase some force downwars under which I place the cap.

  After some practise I found the right position of the positioning of LED, cap and wire (which I was able to bend in position, because gluing was done a bit away). So when nicely done I ended up like in the right picture. After each column I did a testrun with a little sketch (written by sisam) which just blinks through each LED with color white, which makes it instantly visable, if one color is missing. Non coloumn worked at the first try (I tend mostly to miss one solderspot or overheaded the LED and had to exchange it.

  When all columns worked fine I closed the gaps in the data and powerlines and tested the whole thing with a new sketch which led 10 LEDs on run over the panel in each color (RGB).
  

• get the wire in shape

  mclien • 01/06/2015 at 13:30 • 0 comments

  I used silver coated copperwire for the electrification. Large amounts of that are delivered on coils, which means it tends to keep a curved shape. To straighten it you have to pull them a bit in lenght just to the point they start strech plastic.

  From the pics on the last log you could see/read, that there are "V" shaped pieces of that. Being lazy AND kinda perfectionistic bending these parts manually without any way tho reproduce a similar shape to all of it, I built a little apparatus:

  Mostly some spare/ leftover metal/ wood parts lying around and 2 quickreleases (normally used for seapost clamps on bicycles). The edges of the metal parts are rounded a bit (1-2mm radius) to prevent the wire from breaking throught notch effect.

  After a bit trial and error I had the quick releases adjusted to the right force (you have to find the balance between the wire just slipping underneth the quick release and being stretched for straightnes) and could start:

  After some tests and repeats (roll up wire, close release, press down bar, open release, cut wire, repeate) I was good to do most of the parts blind while watching a movie (or two or three..). Afterwards I had to go through it again to get the correct and similar lenght for all. And soon enough I had more than the 315 I needed for the door:

• wiring

  mclien • 01/05/2015 at 13:45 • 0 comments

  This is the pinout of the ws2812b LEDs:

  quite nicly here is that at a 45° angle the power supply can be from both sides, while data goes "trough" it. Which led me to the following design:

  This

  This is the view from the backside. on each LED we have power from the sides, which are supplied by the +5V/ground "rails". Data "flow" starts on bottom turns on top and goes down the next line of LEDs (blue dashed arrows), so the whole matrix is one single LED strip. As you can see I left a notch in each part for the data line (green circle), but not to the power line (was afraid to weaken the parts. red circle). So I went for the upper design, but will change the CAD data with new notches for the power wire aswell, to avoid the V-bended wire parts (see next log entry). Firstly I glued all wire parts from the packside to have all tips in place:

  I glued the wire in place just by a spot of epoxy with syringe and blunt cannula. You can also see the gluing of the wooden parts, which were appied in the same way and only from behind.

  So with alittle it of bending I had all wires in place. I then set the LEDs in place from top prevented from falling trought by the wire ends and holding in place for soldering by frictioin (holes are 5mm rectangele, solder pads stick out a little from that).
  

• Lasercutting and assemble

  mclien • 01/03/2015 at 13:21 • 0 comments

  For lasercutting I prepared 300x600mm pieces of 3mm MDF (medium density fiberboard), which were leftover from another project. I think about 15 of them (I'll load up the set of dxf files once I sort it out). After a session of cutting I ended up with this:

  6 parts for the packplate and a pile for the horizontal and vertical parts. Next step is putting them together all parts with the open end facing to the floor:

  I noted down the positions of the different length of the vertical parts to know which length to start with in which position. (on the paper you see in the bottom left corner). As soon as the grid advances to a point one could apply the next backplate piece:

  Last step is to putt all the outer edge pieces in place, which have a slighly different design, because they don't intersect with the backplate (which isn't such a nice idea for they tend to fall of). Most of it sits tightly in place by friction and tolerace, but I applied some epoxy glue in the corners of the intersections anyway. And first check:

  
  

• CAD design

  mclien • 01/02/2015 at 16:22 • 0 comments

  As soon as it was decited we need a LED matrix, I started to design that thing on my CAD program. I use VariCAD, because some years ago this was the only affordable 3D program running on linux machines. So knew I would have a nice lasercutter at hand in our local fablab/open garage thingy (Die Werke). This way you can easily use rectangle shapes to make system, where the parts nice link together (holes for the LEDs missing):

  The size of the cubic was tested with one of the LEDs some white cardboard aon a breadboard.
  

  After designing that for the whole thing, it had to be cut into pieces that would fit into the lasercutter and still interlink nicely for stability reasons:

  red markings show the link sction, black one how it's "captured" by the horizontal parts. Backplane is splitted in 5 parts, which interscet also. I made several links with different parts in lenght, so I could spread this weak points evenly in the construction. Here is the whole picture:

  horizontal parts needed noe splitting

  vertical parts: The way I designed them, there are two parts, depending whech end they end up (because the design of my splitting line dosen't allow flipping). So I made them even (blue) and uneven (green) to end up with one type on each end of the row. The backplanes have cutting lines which prevent them crossing the complete hight/width.

  I will upload all CAD data (dxf parts already sorted for cutting on a 300x600mm bed and the uncutted parts as well as the 3D step data)
  

• How could this happen?

  mclien • 01/02/2015 at 15:18 • 0 comments

  While rebuilding our house, it happened that the door of my sons room ended up at the other hinge side than planned. To my own supprise the only spare door was one with a glas window :

  while this is nice in living rooms and hallways to spread some light, it's not as private and/or dark as you moght want it for privacy and sleeping. Buying a new door or appling a piece of wood instead of the glass was out of question, so we ended up with a quick duct tape and cardboard solution temporarily.

  So we had some time to ponder a bit about which might a cool, nerdish or funny idea. That was about the time I stumbled over some cool coffetable LED build in ideas und alike.
  

• First log

  mclien • 12/29/2014 at 10:11 • 0 comments

  Just a quick setup of the page, to have something to direct people to. It is because there are a lot of quiestions right now at the 31c3, which I cannot answer (too much dutx elsewhere), so this is the starting point.

  I will update the information the days after the 31c3 with all the files for the lasercutting, photos of the building, code for the arduino and such, just stay tuned or conntact me.

  I will be back after 31c3 for updates.

  thanks for the interest
  

View all 9 project logs