You are mostly right.
I currently have 4 main designs on the drawing board right now:
- Ice40 up 5K boards : Alloy (feather) and BlackIce V tile/stack,
- ECP5 : Amalgam_stack (tile based) and Amalgam (SOM based)
I need to manufacture the ECP5 boards first as I already have the parts in storage and need to produce them. Also there seems to be more interest in ECP5 than Ice40. However do I go the SOM or stack route? If I go the SOM route all the ECP 5 boards would be more expensive if I wanted to go Stack based, it doubles the number of boards. For the higher end loaded version the cost would be less of an issue, however at the more entry ECP5 level cost is king. I also believe I can get the entry level ECP 5 board closer to what the BlackIce board is as a stack board but that would not be the case for a SOM board as it takes 2 boards..
In addition the Tile/Stack design is currently untested, we will only know if this does what it is designed for when we make some test units and try things out in the real world (i.e. not just in CAD). The lowest cost and quickest way to do this is via the BlackIce V Tile style board which I think I can build in the next few weeks along with at least one fully functional Tile (the 4 axis motion control Tile). This also fits in with my planned nMigen development steps I want to cover in the stream.
To add to this I have gotten really frustrated with the Feather mechanical format for my current development of the Alloy stuff which is software heavy, I think the BlackIce V tile would have made this much easier to develop with, but this is a personal thing and I am just frustrated with my current Alloy rig I prob should have done a BlackIce V stack first. maybe.
BTW the ECP5 amalgam Stack does have the potential for a Mixmod or quad Pmod on the main board which may be good for backward compatibility.
I hope that helps a little, I will do more specific threads for the tile spec detail etc.. shortly..