On Jun 14, 2006, at 8:33 PM, Josh Babcock wrote: > Jon S. Berndt wrote: >> Snap roll: >> >> This is indeed the recipe for a snap roll: starting from a speed >> slightly >> above the stall, apply a sudden yaw with the rudder, apply >> opposite aileron, >> and pull back on the yoke. SNAP! --- One wing stalls and the plane >> rolls >> over. >> >> [I liked the clever use of the word, "recipe" with the phrase >> "snap *roll*"] >> >> This would be hard to model using lookup tables, but it might be >> possible >> using JSBSim functions and a table or tables, together. Could be >> fun. I need >> to think about this one. The first idea that comes to mind is that >> if the >> aircraft speed minus the yaw rate times some characteristic >> lateral length >> (span/2?) falls below the stall speed, then a rolling moment would be >> generated - maybe a yawing moment, too. >> >> Jon >> >> >> >> _______________________________________________ >> Flightgear-devel mailing list >> Flightgear-devel@lists.sourceforge.net >> https://lists.sourceforge.net/lists/listinfo/flightgear-devel >> > > My understanding is that a snap roll is initiated by yaw-roll > coupling. > The lower wing is put into the turbulent flow behind the fuselage > by the > hard yaw. This imparts a strong roll moment, and the result is that > the > AOA of the upper wing goes down, while the AOA of the lower wing goes > up into the stall region. At that point the partial loss of lift on > the > down wing becomes almost complete, while the upper wing only loses a > small amount of lift. > > If it were done at a low AOA you would only get roll damping as the > low > wing would go into a high AOA high lift regime, while the upper wing > would go into a low AOA low lift regime. You need to be close > enough to > stall that the lower wing goes past the high lift regime and into the > stall regime. > > I may be wrong about that. If the roll were initiated on the back side > of the lift curve, the upper wing would actually gain lift in the > roll, > and the lower one would lose it as it goes into stall. I'm not sure > which is right, but I'm pretty sure that to get a roll going fast > enough > to get only one wing into a stall you have to have the yaw-roll > coupling. Otherwise roll damping would limit you to a mere barrel > roll. > > So for JSBSim, you would need to add another dimension to your lookup > tables that indicates the loss of lift as an airfoil goes through the > turbulent wake of other elements like the fuselage. Not a bad idea > really, but it's a lot of data and probably pretty hard to find. You > would also need separate R/L wing elements. > > Josh >
I don't believe so. Remember that a snap roll is a type of spin. So probably this discussion should focus on simulating spins more generally. It is easy to enter a spin. For a simple spin the aircraft is just put in slow flight, near stall, and then some yaw is added, but a lot is generally not necessary (depends on the aircraft). I don't believe that it is enough to blanket the lower wing behind the fuselage. I believe that the main effect is just that the yaw rate is enough that the two wings are flying at different speeds. Essentially, like a helicopter, there is a retreating wing and an advancing wing. If both wings are near stall, then that retreating wing , going a little slower (or effectively a higher AOA), will stall, and you will end up in a spin. A snap roll is only different in that the aircraft is generally at a higher airspeed, so a high pitch rate is necessary to get the aircraft to that stall AOA, instead of just causing a climb, and a slow airplane. But the dynamics are still just high AOA, then yaw rate causing a wing to stall. It might be possible to simulate spins (and snap rolls) by just adding a new term to JSBSim that is essentially "loss of lift due to yaw rate" (or something similar...maybe normalize to Cl difference?). But it would have to be a table that would look like a Cl_Alpha curve, so that it would capture the stall on the wing. I believe, as Jon mentioned, that it could be done with a table, showing wing lift difference, from average alpha and yaw rate. And, just to complicate things, inverted spins are fun to do, and there could be some sign issues to keep in mind for the above table to also work that way as well. --Adam _______________________________________________ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
