It's the bottom of the plane. When it does it I bail to inverted flight.
Plane is a Vector 40. PA 40 UL on a pipe on 60ft lines. Running a 5.5 lap time. I did notice a little lightness on the lines in the top of the out sides around the 2 o'clock position. And now that you mention it I think it did it in the first turn of the inside squares. It made me thing to take it easy on the triangle and I flew a real odd egg shape so did not see it there.
I really banged the first turn of the hour glass and that is where it really showed it.
Off-hand is seems a bit slow, but that will likely just make the problem easier to see, not cause it.
Obviously I think you should seal the hinge lines, but that may not solve the problem. What you seem to be describing is line whip, and you are seeing the second motion it takes after the corner. I think it is probably yawing nose-out due to the leadouts being set too far aft, and rolling inboard in the corner from inadequate tipweight, then after you straighten out, it yaws nose in and rolls out as the lines whip back in reaction to the corner. Carefully check what the yaw angle does during the corner, I suspect it will yaw out dramatically first, then back in at the time you noted. It may or may not roll in, because the yaw reaction will tend to obscure the lack of tip weight.
I suggest this. Observe the roll angle through the intersections of the round 8s (horizontal, vertical, and overhead). See which way it rolls through when you switch from inside to outside and outside to inside. Set the tipweight to minimize the roll change through the intersection. I would start adding it until was obviously rolling away from you in both loops (see the bottom on the insides and top on the outsides), then back off, so you know for sure that you have too much, then you can take it out in small bits (1/8 ounce at a time) until you get a minimum roll change through the intersection.
Then check the yaw angle in the corners. I expect you will see drastic outboard yaw on insides and almost nothing on outsides, or at least a lot less outboard yaw. Start moving the leadouts forward 1/8" at a time until the yaw reaction in the corners is minimized. As you do this, keep checking the roll angle change through the round intersections. I bet you will need to keep adding tip weight at you move the leadouts forward. At some point you will get a minimum reaction in roll AND minimum reaction in yaw. Assuming you have the rudder straight ahead, this will be the best you can do. Then you can adjust slightly around there and see if there is any improvement.
If you have rudder offset or rudder airfoil, your minimum yaw reaction will still be pretty large. At which point you should go put it dead straight ahead and adjustable (cut it off and refinish of necessary), put the leadouts where LINEII tells you to, and then do the same thing except changing the leadouts, you adjust the rudder in *tiny* increments to do the same thing.
I am not entirely sure that I understand the issue correctly from your description, but you are likely to get pretty close to right following the above procedure even if it's some other issue. It's more-or-less consistent with the Paul Walker flowchart and the "Functions of trim adjustments" thing I wrote. In any case, it does require careful observation of the roll and yaw angle while you are flying, which seems to be an ability that varies wildly from individual to individual. IF you need more experienced help, then it is always worth seeking it out.
Brett