OK, Bernie. I put some CA on my fingertips but I didn't wait long enough for it to cure.......... So after I got the keyboard off my finger I thought I'd write this up so we could close this off........ I hope

It seems to me that in a truly coordinated turn the airplane should not experience any return to normal as all forces should be balanced and the airplane should just go round and round until controls are then used to disturb this balance. Dihedral or no a coordinated turn should be stable.

But as you suggest with the need for the outside aileron to counter the wing dropping into the turn on the ASK there is often some other factor that disturbs a truly balanced turn. It might be the speed and lift difference from the inside to the outside or the aircraft may be spirally too stable or unstable causing the pilot to require rudder inputs that are not part of the coordinated turn just to maintain the proper turn bank and radius or any other number of design related factors.

BTW, you, as a full size pilot, should also remember that it's possible to be in a balanced spiral dive and have the T&B indicator show nothing other than straight flight. This was the cause of many early crashes when pilots tried to fly in clouds and got disoriented.

But all this has little to do with how dihedral works. You say it's the projected area difference and I say it's sideslipping air acting to change the angle of attack on the wings differently.

I'd like to close my part with another example. I've flown quite a few Goldberg Eagles with novice students. This model has a fairly strong tendency to return to level flight when the controls are neutrallized. I dont' have the actual numbers with me but from memory this model has about 3 inches of dihedral under each wingtip and the span is 50 inches. My trigonometry is a little rusty so I used my CAD to lift a 25 inch panel 3 inches and I get a projected span of 24.8 inches. That's only a 0.2 inch difference per panel and a 6.8 degree angle. So if we bank our model by 6.8 degrees then the low panel is 25 inch span and the raised panel is reduced by a futher 6.8 degree tilt to 24.3 inches projected. So we would have just 0.7 inch span worth of area to do all the restorative work. For our wing at 50 inch span and 10 inch chord this would be a differnce of 0.7 x 10 = 7 sq inches. Each panel has 250 sq inches so we have a total differnce at this bank angle of 7/250= 2.8 % total difference in wing area to do the work.

On the other hand lets look at the same wing if we allow a side slip of 5 degrees. This isn't much and it's probably safe to say that you'd need that much at least to see a noticable side slip from the ground. From having flown the Goldberg Eagle and neutrallizing the controls while banking I know it can be seen to sideslip noticeably. Using a side slip angle of 5 degrees then the change in AoA for each wing is 0.48 degrees or a 0.96 degree total difference between the two panels. Using the FoilSim applet HERE I "made up a wing and airfoil close to our model. Starting at 2 degree AoA I have a Cl of .675. Adding 0.96 degrees gives me an increase to 0.79 for a Cl difference between the wing panels of 0.115 or a 15.7% difference between the panels based on the average Cl of 0.733.

And just to see it from another perspective I changed the Foilsim units to lift vs angle and at 30 mph flight speed the 2.48 average degree AoA gives us just under 4 1/2 lbs of lift so we must have built a heavy model Factoring in the AoA differences for each wing shows a lift difference between the wings of a whopping 0.845 lbs of lift trying to force our model back to level.

As I said before I have every reason to agree that the projected span difference is certainly a factor but from the numbers above you can see that the side slip induced change to lift is a much more powerful force.

I think I'll go use that CA on the models now instead of my fingers. Cheers.