I'll try; Harold and Eric can chime in with their thoughts to if they wish.
Careful layout is important (somewhat obvious I suppose). First, the exact shape of the desired flange is drawn on the sheet. In this case, it is traced from the one piece firebox, as the actual firebox shape is a little different than the print. This final flange shape is drawn as the inside surface of the flange. Once that's done, the centers of rotation are determined for the curved areas. We then pick a standard offset from the flange line. For straight areas, the offset line is parallel. For curved areas, the standard offset is swung as an arc off the center of rotation. Points of tangency, centerljnes, and sometimes another bisecting set of lines, are drawn on both the flange edge and the offset line.
The pointer on the flanger anvil is set at the offset distance from the face of the die being used.
The toggle is brought down to 90 degrees, the anvil brought up to one sheet thickness away from the toggle shoe face, and the backside of the anvil blocked to that location off the anvil stop. That's 0 location for that operation. The anvil is then blocked either in by 3", and the plate bent about 20 degrees at each index mark for that operation. The anvil is then backed off to 0 + 2" and bent to about 45 degrees. On the easier curves, we then back off to 0 + 1" and bend to about 75 degrees. On harder bends, adjacent areas must be bent first.
As expected, some buckling develops on the sharp corners. Here's the cool part: the flanger's available 200 tons exerted on the flange simply smushed those bumps flat. Pretty much like they never existed.
The flanger never maxed out, but does exert more effort when bending more plate on a curve, against adjacent areas which aren't flanged as much. Again, expected.
We're achieving exact knuckle radii and flange locations within 1/32".
Our indexing method of lining up index prick punches to the pointer, under the plate, is a little tedious, but is a very direct method. We're considering going to a template and follower system for flanged plates requiring tighter tolerances.
It's been a lot of fun finally seeing this machine in use. I've had it in mind since 2006, when we hot formed No 9's flanged boiler plates.
See ya
Jason
