I guess H. Ford had the same experience?
A lot of empirical "tuning" - just to get to a condition you think will be correct - just to find that things are not as expected and iot doesn't work, so then lots of trial and error to get to a working condition.
I would have guessed that a drawing or a few calculations should have given you a closer "starting point" (Radius of 0.100" instead of 0.375" ?)? - But you got there anyway.
Well done Sir!
I am learning a lot about this engine. - Fascinating!
K2