Note: the blocks in the picture are an attempt to prevent the brackets on the bottom rail from causing deformation of the frame. One would think that supporting the frame at the corners would be conducive to a square frame....  nope.

Gap closed when frame inverted; gap opens when frame is right side up on blocks. Conclusion: Frame materials (specifically the corner fittings and corner to side joinery) are flexing badly. In this situation it is virtually impossible to get the screen stack flat using frame tension (compression gaskets installed on all four sides).

Not bad...   Much improved!! If I was using a bright, indirect flash rather than the point flash in the camera, you'd see virtually no distortion at all no matter what light intensity was used. There is still a bit, but I'm confident that with a little time, I can eliminate the last 5% or so.

After removing all but the top rail gaskets. I re-installed the side and bottom clamps. I then inverted the frame and re-seated the screen stack in the upper clamp/gasket assembly (as in Part I), paying particular attention that the stack was centered side-to-side and that no gaps existed between the stack and the top frame rail.

Then I carefully (to avoid racking the assembly) turned the assembly right side up and re-installed the Front Mask Assembly onto the chassis.

Essentially, the screen stack is suspended from the top rail and is "floating" with respect to the side and bottom rails. I plan to leave it like this for a couple of weeks to observe any changes due to environmental conditions, frame/clamp flexing, etc.

Read on to see what my thoughts are currently for addressing the last 5% of the lenticular distortion. I say lenticular distortion because frankly it doesn't really meet the criteria for "sag" anymore. 

Discussion: I think the main cause of the unstable frame is simply that it is made out of a plastic in such dimensions that some flexing is inevitable. This inherent material instability is exacerbated by the fact that the rails and corner fittings are held together with only 2 screws each, rather than six screws as is evident from the design of the corner fittings. 

Additionally, the clamps also tend to flex excessively due to 1) construction in plastic, and 2) excessive spans between mounting screws. Clamp flexing results in the screen stack actually only being retained by smaller intermittent static contact areas, rather than a uniform static contact "strip" along the entire clamp/frame bearing surface.

Possible Courses of Action:

1) Employ necessary measures to increase the rigidity of the frame such that the screen stack can be securely clamped in on all sides under tension to maintain a uni-planar screen surface.

2) Do nothing. Accept the current results as "acceptable", and continue to use a gravity/free-floating screen assembly.

3) Refine the gravity/float method to achieve near zero lenticular distortion by utilizing the following approach:

a) Increase frame rigidity with the installation of 16 (4x4) additional screws in the corner fitting to frame rail joints. Frame will be secured in a truing fixture while additional hardware is installed to insure the frame is square.

b) Add at least two additional mounting points to the top clamp/frame rail to spread the suspension loading of the screen stack across the entire width of the stack.

c) Analyze the gasketting requirements for the sides and bottom of the frame to 1) maintain a free float, 2) limit the fore/aft travel of the screen stack to a minimum, and 3) encourage maximum contact between the fresnel and lenticular screen elements.

Chosen course of action: Number 3....  obviously. Number 1 would require a completely new build of the frame to achieve the necessary rigidity; the plastic frame as designed probably could not be reinforced enough (at least economically) to accomplish a tension mount. While I am happy with the improvement achieved thusfar, I am NOT satisfied. Not when it appears relatively certain that near perfect results can be achieved with a little more time and effort.

I will continue to refine this procedure, and will re-write it in final form once the entire procedure is complete and proven. The re-write will exclude the "misadventures" and simply present a step-by-step procedure to get from Point A to Point B.