RE: Grumman XF5F
OK, time to eat crow. I did a quick scale up in CAD of your wing and cowl from the 3view you posted earlier. I don't remember what form of size I was playing with but actually this is quite doable.
At 42 inch span the wing area is 316 sq inches and the cowls are 4.6 round with a 3.2 inch opening. The fuselage is 3.4 inches wide at the trailing edge.
At 44 inch span the wing area is 350 sq inches and the cowls are 4.8 inches round with a 3.4 opening. The fuselage is 3.5 inches wide at the wing trailing edge.
At 46 inch span the wing area is 385 sq inches and the cowls are 5.0 round with a 3.5 inch opening. The fuselage is 3.7 inches wide at the trailing edge.
If you choose to go with the 44 inch or 42 inch option you'll need to keep the weight down. At these sizes the scale effect is hitting hard and the weight and wing loading can't just be linearly reduced. I would suggest that you will not be happly if the 46 incher has a loading of more than 15 oz/sq ft. That equates to 40 oz. To fly the same I would say that the 44 incher would need to be 14 oz/sqft and the 42 incher no more than 12 to 12.5 oz/sqft. Those loadings equate to 34 oz for the 44 incher and 26 to 27.5 for the 42 incher. This is all based on the idea that the Reynolds numbers (that old "scaling effect" thing) are really stacking the deck against us small model flyers.
My own feeling is that if you can build it light enough that the extra wing area of the 44 or 46 inch version is easily worth the extra cowl size. At that size and given what I know about the Norvel power output if you can keep the weight down to about 2.5 lbs (40 oz) you'll have a great flying model that will offer good climb, reasonalble speed and yet still slow down nicely but it won't be a floater by any means.
At any of these sizes I would also suggest that you choose to use the Selig 8036 airfoil. It's a thicker but low camber shape that is picking up a good reputation for having a gentle stall and slightly lower landing speed on heavier wing loading scale typle models.
But if you truly want a fast and zoomy model I'm afraid that while this is a fantastic looking project that you're doomed to failure in the speed department due to the two built in parachutes on the wings of this model. Something with inline engines would be far better.
On a positive note the size of the cowlings and the style of the retract gear would make doing your own design gear quite an easy option with the legs being made from carbon fiber legs and the side brackets to hold the pivots being from simple plywood. There's easily enough room in there to let you do this in a larger but still light method that supports the leg pivots without point loading the structure highly.
The big issue here is the lack of prop diameter that the Norvels can swing. They produce lots of power but won't put it into a larger diameter. Something that occurs to me is that the 46 inch option with two OS 15LA's in it would swing the prop sizes needed while still only putting out the same power as the two Norvels.