ORIGINAL: bwick
It's going to take a little bit of messing around, but I would try much more aileron, and maybe even a bit more rudder. If it's going into a barrel roll there's probably not enough aileron to get the wings around fast enough. If you're having trouble getting a clean break, a cg adjustment is probably in order.
Again, it's hard to diagnose over the internet. A lot of planes react differently and need different set ups depending on how they're trimmed. It just takes a lot of practice and tinkering.
Kjell,
I don't own an Integral, but I can't say what my deflections are compared to normal -- aside from that I've got more elevator, less rudder, and more aileron. However, the inputs and amount of input I give depend on the snap that's being performed, so it's not accurate to say that I use all of that deflection and in those proportions for every snap.
Rendegade,
Take a look at the wing's LE radius. If the radius is large, on order of 1/4 inch or so, the wing would benefit from stall strips. These don't need to be elaborate. Strips of 1/8" balsa, 1/2" wide and 12" long. Pin them to the wing's LE on edge, starting at the tip and working to the root. If you are happy with the result, then finish them out properly and make the installation permanent.
It also looks like you could be carrying more stab area than desirable for Pattern models. If it's more than about 25% of the wing, snaps will tend to wallow and be somewhat upredictable. Large stabs will tend to overstabilize an aerobatic model....but that's an oversimplification. There are several other factors that influence stability and stab % plays a major role. Suffice it to say that large stabs work great on trainers
MattK