So the Lindbergh can communicate directly with the MIDI board through a baudrate signal converter, but there appears to be one factor missing or handled differently and that's holding force.
In my MIDI experiments, and indeed in the way I am driving the MIDI board with the NAMCO translator, I have 83 commands as holding force. In my MIDI logs I can see that holding force is established during the initialization sequence. In the LINDY log above, I'm not see that happen, so the realworld result is that the wheel goes through it's automatic init sequence where it turns a quarter circle to the left (I assume under some self-directed holding force) and then stays there in spite of commands to move due to holding force being too weak.
Perhaps the RS422 board has a higher default holding force, making it so that holding force does not have to be established during initialization.
As it is, without establishing a holding force, the MIDI wheel will fail initialization because it does not re-center. I can manually re-center the wheel at the correct time and pass, and even get some FFB effects in game. ID4 and 5 seem mostly to keep the wheel centered. I wasn't noticing much jerking back and forth when hitting walls or the rival car. Hummer seems to mostly center and rumble ALL-THE-TIME, so not sure what to make of that one. Rtuned passed init, but I never felt any effects, not even self-centering. Sega Race TV mostly self-centered. OR2SP in standard mode would not boot, giving an error that some board is not plugged up. I saw a few FFB commands issued over COM1 before the error, but not sure what to make of that.
At any rate, those were all just quick tests with the Arduino passing along the CMDs and ACKs without any intervention. I will probably start by focusing on one game and look at logs closer to see what's happening.
For sure the MIDI board seems to be giving ACKs for every command coming from the LINDY, so it doesn't appear that there's an updated protocol. It may just be operating under some different assumptions that the where the MIDI board has different defaults than the RS422 board.
Also interesting is that one of the games, not sure off hand, sent a command that resulted in 4-byte ACKs (normally they're just one byte) that seemed to be changing based on wheel position... which is pretty interesting. I hadn't been able to figure out how to get positional data outside of monitoring the wheel pot.
Anyway, I'm thinking I can inject a holding force after init and see if it behaves better. I know when self-init is happening because the ACKs are always 0x44 during and then go to 0x00 after. So if I see 0x44 I know it's initializing and when I start seeing 0x00 I can send a command for holding force.