Its along a couple of the lengthwise welds (horizontal), has actually cracked the relatively thin rocker sheetmetal rather than the overly thick SSM's (SouthSide Machine subframe connector). I only noticed it because of the telltale hairline rust mark it left so I looked closer and sure enough... its cracked. So I have 2 cracks, about 1.5" long each.
Here's a link, I am not going to copy it all. There's a section on torsion for those not afraid of books or worms.
http://en.wikibooks.org/wiki/Solid_Mechanics
In order for torsion to occur in the connector itself, the tube itself needs to twist around its centerline (or thereabouts), one end relative to the other. If you were to assume that its connected at front and rear, worst case scenario is if the front twists clockwise, and the rear (like where it attaches, LCA bolt likely it looks like, like my SSM's do) twists counterclockwise. In the twisting case, there's no deflections horizontal, vertical, or any other direction, it only twists. In order for it to twist in that manner in the clockwise direction, it has to do the following, based on how my SSM's attach and likely how the Spohn's attach and probably the UMI:
Twist the rocker (since its physically attached to the SFC and is parallel)
Twist the frame rear rail (since its physically attached to the SFC and parallel)
Pull/stretch the floorplan
Push/compress the rear 1/4 and inner 1/4
It'd probably compress most if not all of the roof
Pull/stretch the opposite side rear 1/4
At some point the two push/pull would converge, likely the opposite side bend from the roof and 1/4, as a wild guess based on where these cars like to stress crack.
I'm ignoring a few things like the other frame rail, etc, but the principle is the same. In any case, either there is some major deflection in the rocker and rear frame rail caused by torsion, or the whole car is twisting about the SFC in order for significant torsion to occur IN the SFC. Two things would happen that would be obvious if this were true:
1. Damage at the rocker and/or front and rear frame rail, or crossmember, from trying to tear the rail or rocker from its welds or compression of the sheetmetal causing bends. Ever seen that happen? I've only heard of some fast drag cars pulling the LCA so hard it damages the frame, much like the brand F guys do much easier because brand F cant build a frame.
2. Visual deflection in the body of the car, if the whole body twisted. We're talking like 9 second drag car twist here, to induce any significant torsional force in the SFC worth talking about, and that would be twist around the SFC itself, and not the center of the car.
I've attached a picture of what torsion looks like. It requires a relative angular deflection along the part in question. Take an exaggerated example of a car, make it 5000 feet wide. Put a small round tube at the edge of that car. If you twist the car around its center 45 degrees, how much twist is in that tube? Not much. Move the tube to the middle of the car, repeat the process. Now how much twist is in that tube? A whole lot.
Look at it another way, if both right and left frame rails are in such heavy torsion, the car is trying to tear itself apart down the middle. If torsion is the issue at the rockers, you should be attaching the right to the left to prevent them from twisting.
Personally, I would be much more concerned about bending. Like the kind of bending that allows the rear frame rail to move away from the front when the rear goes over a bump. Same kind of movement that transfers load through the roof and cracks every hard driven or high mileage F-body hardtop right at the rear top of the side windows and the base of the windshield.
