I have been trying to build a triple scale version of Dave Kerzel's 2-cylinder CO2 engine. Here's an in-progress pic from a few weeks ago:

I think I may have run into some intractable laws of physics. The picture shows my initial fitting with a 3/8" ball as the valve. For reference, the base of the engine is ~2.5" wide, and the bore is 3/4".
With this size ball, the air pressure at 40 psi puts such force that I can't move the ball by turning the flywheel by hand. So it became quickly evident that I would need a much smaller ball. So after making 4 or 5 different valve seats I'm using a 3/32" ball. At 25 psi I can lift the valve fairly easily by hand turning the flywheel, but the engine still doesn't run.
So I need to experiment with how much lift is needed. If there is too much lift then the ball rises too early in the cycle and the pressure kicks the piston back in the opposite direction. Too little and there is not enough gas introduced to force the piston back around again. I am also concerned whether there is enough space around the lift pin for enough air flow into the cylinder. I might also need a very heavy flywheel to maintain the angular momentum.
If anyone has had experience with larger CO2 engines I'd be happy for their feedback.
I've enjoyed the machining part of building the pieces, but it gets frustrating solving the physics of scaling up. Perhaps I've bitten off more than is chewable.
FWIW, materials are: crankcase, cylinders and heads, crank are 6061 aluminum. Pistons, valve seats, lift pins are 360 brass. Crankshaft, crank pin, connecting rod, wrist pins are silver steel, flywheel is 3x1" 12L14 steel.