On a somewhat related note, back around '96 I was working for Boeing doing structures work on the 757-300 stretch. The 757 was originally a paper design but the modified areas of the structure were designed using CATIA V4. One of the largest tasks for engineering was building a digital model set of the existing structures using the paper drawings and mylar flat pattern masters for formed sheet metal parts. In order to expedite the work Boeing developed a process that scanned the mylar flat patterns and created a 2D CAD wireframe of the outline plus the bend lines, mold lines, tooling holes, etc. The 2D wireframe was imported into a custom CATIA V4 application where you would identify particular wireframe elements as part outline, bend line, etc. and input values like bend radius, bend direction, bend angle, material thickness, etc. Then the CATIA V4 application would create a solid model of the formed part.
It was amazing how well the process actually worked and how accurate the 3D models produced from the scanned mylar masters were. There were tolerance standards that each solid model had to be checked against, and except for very simple parts each model did require a bit of manual clean-up work. But given the state of CAD technology in 1996 the process did significantly reduce the amount of labor required to build the baseline 757 digital model database. The process would never have produced satisfactory results unless those 15 year-old mylar masters were drawn with high accuracy and were well maintained.
Attached is part of a supplier quality standard for sheet metal parts including tolerances for those produced by photo transfer from (mylar) master templates. There is a brief description of validating the dimensional accuracy of the templates.
Best regards,
Terry
