What major problems do you see in the electronics?
As you said:-
from switching 15A of power on/off quickly
(Separate 5V PSU with lots of capacitors on the output? Thick wires for all GND connections?)
If it were only that simple.
15A!? Please double check your math. Each layer is 10x10, with 3 pins per LED, for a total of 300 pins. Each pin will drive a max of 20mA, 300x20mA = 6A. I'm planning to use this power supply to drive the LED's: http://adafruit.com/products/658
. With large enough decoupling caps between Vcc and Gnd on each chip, this shouldn't be an issue. Not really sure what your issue is here. I also bought high power mosfets rated for 100V and 40A to handle the switching. http://www.irf.com/product-info/datasheets/data/irf5210pbf.pdf
I probably won't be running the LED's at the full 20mA anyways, since it might be too bright to look at.
That is what worries me. You need to make a realistic prototype of a tiny corner of what you plan, run it at the rate the real thing will run before you know if your design is feasible.
I plan on building a 4x4x4 with a scaled version of the circuit and to practice different techniques of assembling the cube. However, it's hard to tell from a smaller version if something this large will work because I won't be hitting the same limitations.
Good so you know what is involved with 500 LEDs, now you want to scale it up to 3000. Scaling up is not easy, you don't just make more. There are lots of things to consider with switching so much current so quickly. I would have thought the TLC5940 was not the best choice as you have to synchronise the multiplexing to the refresh rate.
I thought they would be a better choice than shift registers because I have to chain together a large amount of IC's. I used shift registers for my 8x8x8. If I use shift registers I would need to drive a much faster clock through the board to manually handle the PWM. What do you think is the best choice and why?
Best of luck with your project. Please post in the exhibition section when you have it working.
Will do, but getting there will not be trivial, and I appreciate the guidance from those who are more experienced in this area than I am. I'm considering making a build thread to document my progress for those who are interested.
Having built a 5x5x5 cube, while finding designing, building, debugging the hardware challenging it was at least something I enjoyed and progress was mostly trouble free. The software was another kettle of fish however. The basic scanning of the cube was pretty simple but I found that the 'creative part' of coming up with interesting pattern shows was so very very tedious and very time consuming and I soon lost interest after defining about six basic patterns including a simple random pattern using a RNG. What I really longed for was a PC based GUI pattern editor that I could use to 'visually' build patterns, scan by scan, and then be able to upload the pattern into the arduino driven cube. But of course my PC software skills would never allow me to write such a application. So now when I watch cube being demonstrated on U-tube or wherever, I've really come to appreciate the artistic effort that some have mastered. So in effect the software is the biggest overall factor is the quality of the finished project.
I'm a verification engineer, so my job involves both hardware and software, so I'm not really worried about the software side. I will make a GUI for this, and I have a few other ideas planned as well. That being said, what I am concerned about is the low-level code on the MCU to handle the multiplexing of all the layers. There are libraries available that I can adapt to my needs, but I don't know if they will be fast enough to handle what is essentially a 300x10 matrix.