Out of shear boredom while deployed, I decided I'd try my hand at making a LED cube.
I didn't do a whole lot of research before I bought the LEDs and some parts. Over the course of 2 days, I made the cube, not really paying much attention to how I was going to control it.
I quickly learned that I couldn't power the cube via my Arduino Mega, and that's where I got stuck.
I did a bit of research, and found what I thought to be a good design. I put it together with a friend's copy of Eagle, but I'm not 100% sure it will actually work, regardless of how I program my Arduino.
Attached is my schematic.
I'm thinking of using the following components for the design:
It looks like you are using the Max7219 which isn't bad, although to make an efficient design, you could only use them to achieve a 4x4x4 cube. Using 4 of the Max7219's isn't bad but if you want to go any bigger, you are going to get some parts of the Max7219 that won't be used.
On top of that, disclaimer: I haven't done much with transistors (crazy for an aspiring computer engineer, I know), I don't think that you'd need the transistor arrays. As you can send commands to the Max7219's individually and be able to explicitly say, I want this one to turn off and this one to go on. If you are thinking like me, each 7219 will control one layer of the 4x4x4 cube (which i'm going to do now since its a great idea) and those 7219's will be connected to your mega, which I believe would have enough PWM ouputs if you can't daisychain the 7219, i can't remember if you can.
