Awesome, another very legit question! So far, there´s none I’m afraid. I’m only gathering some random thoughts for now. The feature creep madness phase seems to be wearing off, as I haven’t barely added anything new to the requirements for few days. I’m very much aware of the need for a plan. A complex project like this is bound to fail without one. So, let’s draft one!

The first step is to call the feature creep madness complete. That’s something I’m ready to do right now. Sure, some small new features will pop up at some point, but I’m very much satisfied with my current list.

Except for one small thing. As DRAM may not be fast enough to respond withing a single cycle to the CPU, I will implement some sort of cache. A smaller SRAM IC, handled again by the MMU will acomplish this.

There’s a whole new set of skills I need to learn.

  • 65816
  • Verilog / VHDL
  • DRAM
  • Memory Cache
  • Video generation
  • Screen Controller
  • Audio generation
  • SD card / Filesystem management
  • USB (Storage and HID)
  • RTC
  • Battery Management

Not bad! In parallel, as I progress with the above list, work on a first prototype will proceed as follows:

  • Core block diagram. Extend this random thoughts into a proper block diagram for the system core
  • Prototype schematic. Will have to include CPU, FPGA, SRAM and DRAM footprints. Video will also have a place. SPI bus from the MMU should be exposed, to use it in the peripherals phase
  • Preliminary boot sequence (No SD support yet)
  • First run. FPGA will forward all writes to SRAM. DRAM doesn’t even need to be present
  • Test memory accesses, clock stretching and bus disconnect
  • Add DRAM. No cache, all writes goes to DRAM
  • Implement cache
  • Add video by mirroring upper memory bank
  • Implement video DMA

Then, time to work on the peripherals. It’s very likely all of them can use a SPI bus, of which the MMU should be the master.

  • Simple serial
  • Keyboard support
  • Mouse support
  • SD and filesystem
  • USB Storage
  • RTC
  • Network
  • Battery

SO Development will be done parallel to all the above. As I move on, more and more features will be made available. However, by the time the hardware is complete, I will need to fully focus on the software. This is still too far into the future, so I won’t think to much about it. I will probably start with a simple, text only console, running some kind of Basic. A fully graphical OS is right now beyond my knowledge. We’ll talk about in a few years.

It seems I now have a plan.

⇠  Some random thoughts on the noVa architecture