Rablcube

Wow, end of week 6 already! I really need to update more often, so much to say. We are nearing the middle of the semester and no matter how much I do, I still feel I am behind on something. I have been and am still pretty swamped with all the work I have from everything University related this year. So much to do, so little time.

I am progressing well on the design but I need to really press on and get it finalised so I can start the build and coding.  Only recently have I got internet at my flat and this has really set me back over the past few weeks. I am currently focusing on my other modules to catch up as I have important tasks in them soon. I recently had to give a 5 minute presentation on my project then answer questions and that seemed to go well so I think I know what I am talking about so far. We still have the hardware side and the software side to this project and both are presenting me with some pretty unique tasks and problems.

Many of the ideas of this project are either directly from, or inspired by, various similar projects involving LEDs in 2D or 3D displays from other people. Many of them are also University projects, but many are also examples in data sheets and professionally created devices. There is a plethora of information on the topic of LED displays. It has taken me 5 weeks to just skim the surface of the bits I need.

Check the photos in the right hand menu column for some pretty amazing pictures of all sorts of led cubes and their circuitry. I will also put my own designs and notes up there soon after this post. The designs have so far gone through at least 5 or 6 major different layouts and 10 or 20 hardware changes. The current design is one in which I have 3 major sections (4 kind of). There is the PIC controller board (1), that connects to the LED controllers and power supply (2) and then there is the LED matrix itself (3). The PC connects through a USB header to the PIC so I suppose that’s number 4. Currently I am designing and creating for a cheap design so I am not using any nice PCBs or surface-mount soldering. I want to be able to put this together by my own hands and remove components and fix them if needs be. So that leaves me with copper strip-board (see picture). The design of my circuit is very large so getting this on to a standard piece of strip-board is turning out to be quite hard. There just isn’t enough copper tracks, I am slowly getting there though.

The LEDs specification, that I really would like, need to be very high brightness, long leaded and diffused lens. A major problem with my design is it is going to use blue LEDs which are just not at the same production level as the other colours due to being much newer on the market. The level of brightness is so hard to achieve as the material the blue LEDs are made from is much more expensive and harder to produce at high quality. I need that brightness level because of the way my cube works by turning on each LED for a 6th of the refresh rate (i.e. each layer on for 1/6  of 75Hz = each layer on for 2.22mS). This results in only getting roughly 1/6th of the final brightness.  However, I have come up with LEDs that match everything but long leads so I have designed a new matrix cube construction technique (see picture) that will allow me to use these LEDs. It uses galvanised steel wire to support the cube and create the wired connections I need. Some might call me crazy to try to solder galvanised steel, but it can be done. High lead (Pb) content solder and a decent soldering iron. Matrix construction is done using a wooden jig I have created which holds the LEDs in all the right places while I solder.

And then there is the software, I can program fine in C but this program contains rather complicated vector mathematics along with various other equations which are not fun let me tell you. (Maybe some mathematicians like this sort of thing? If so, contact me.) I have never had such a hard time trying to imagine what happens to 3D objects when I perform certain alterations to their equations. I am quite glad the majority of this can just be simple code to take in binary LED voxel data and then output it. However to improve the design and to make it a bit more of a hard 4th year project I have to try.

Hardware List:

• PIC Microcontroller: PIC18F4685,
• LED Control/Drivers: TLC5940/41, (TLC5924/44 are better but are surface mount so I cannot easily solder them, too small)
• LED Control/Drivers: Medium/High Power Transistors,
• LEDs: Blue, 5mm, Diffused, Ultra Bright ~5000mcd (millicandela) (or better?), 3.3v, 20mA,
• USB: FTDI USB Header, USB Port B,
• Capacitors,
• Variable Potentiometers,
• Resistors,
• Diodes,
• Power LED,
• Power Control ICs,
• 40 Pin Ribbon cable connectors,
• 40 wire ribbon cable,
• Straightened Galvanised Steel Wire,
• Single Pin Sockets,
• Strip Board.

I have also sourced a lot of my own components through free samples and Ebay such that I now have the ability to create my own cube at home as well. It will probably be a 4×4x4 full colour cube. Should be a challenge and I look forward to it.

That’s about all I have to say for now. I have missed out loads but I will get back to it later. Signing off for the weekend.