Max, is that you?
Indeed 
This FPGA is fine to host the CPU too. As a benchmark example, popular Z80 model (Daniel Wallner's T80), takes about 1/4 of this FPGA. STA reports, that the T80 core can run faster than 50 MHz. OCM fits too.
Choosing bigger FPGA will push the layout to 4 layers, automatic/semiautomatic BGA assembly and testing (expensive). BGA assembly still can be done at home, but requires the skill and experience.
There is another design based on the XC6SLX25 in the 484FBGA package with the same 16 bit SDRAM but also in BGA package. Can be adapted to the MSX cartridge outline too if necessary. (PCB cost penalty + LC stencil + oven reflow required).
I could propose, however, to start with Procyon, and if it becomes too limiting, then jump to another system (100% compatible but bigger).
The goal here was to manage the signal integrity on 2 layers (very low cost) with the LVDS (HDMI) lines running beyond 1GBit/s and no BGA components. From the start it wasn't clear whether the signal integrity is feasible (2 layer layout violates the HDMI spec). Now, the answer is - yes it is feasible, it works.
About small series:
Batch of 5+ will require a bit of logistics: manufacturing, component procurement, assembly, test & packaging.
Also I've got to improve the PCB design at first (there are a few little errors).
The timeline I see the following:
1. Card debug - make sure, that it works like planned (in progress).
2. PCB improvements, DFM must be included for the automated assembly.
3. Documentation for the automated assembly.
4. Then, either we have to find a low cost manufacturer, which can make the PCBs, assemble, test and distribute them. Otherwise the participants of the PCB POOL will get the bare PCBs, docs, order themselves the component kits @ Digikey or Mouser and voila - soldering @ home for a few days.
What do you think?
I guess we can cut PCB costs starting from batch of about 20 - let's call it public beta-testing. From response I guess we can have this amount of orders quickly. I shall take a board for sure (don't mind soldering).
Yes, the price would drop drastically from 25+ units. +1 bare.
This sounds an awesome project. I for one would be up for either a self assembly or fully assembled board when finished. Can't wait to turn my HX-10 into a MSX2+.
Excellent! +1 bare.
Hi guys,
I'd buy one of your boards if it would be pre-assembled (not interested in assembling myself :-))
+1 preassembled.
IMHO, there are two types of wishlists:
- Bare board - self-assembly
- Soldered and maybe boxed board
Too early for kickstarter, IMHO. Using Egor's term 
"public beta testing" should be done before.
I haven't tested yet the AUDIO & SDRAM though (in a week or two).
If bare & pre-assembled shares will be equal - better to go pre-assembled for all. Let's see. Also, this board is generic enough to emulate any 8 bit computer of 80s. It needs a slot adapter though to support atari, ZX, c64, SMS, NES, ....
Too early for kickstarter, IMHO. Using Egor's term "public beta testing" should be done before.
I haven't tested yet the AUDIO & SDRAM though (in a week or two).
Fair enough, just launching some ideas :-)
If bare & pre-assembled shares will be equal - better to go pre-assembled for all. Let's see. Also, this board is generic enough to emulate any 8 bit computer of 80s. It needs a slot adapter though to support atari, ZX, c64, SMS, NES, ....
You may want to check http://enterpriseforever.com - an ELAN enterprise forum (mostly Hungarian, but English area) with some hardware geeks around. You could post your board there as well with the specific notes you mentioned. Who knows people will jump on this as well. (of course, this is probably for more z80 based systems, but MSX and Enterprise are the only ones I am really following :))