Bruce L. Bergman wrote:

If this area is cast iron that gets shock-loaded in tension during
use, it won't take much to make it crack again.
Yes, that's the main problem.
Since you tried a
fishplate on the bottom side, and it just cracked where the loads were
moved to one end of the fishplate, I'm thinking there may be an
inherent design weakness in the platen. (Bet you when they changed
over from "Peck Stowe and Wilcox" to "Pexto" they made it heavier...)

Maybe, this is a 30" shear, I can't find a gauge rating on it, and it
only weighs 150 Lbs. It has 18 (nameplate) 90 cast into the upper blade
support, so it must be a very early design.
How about putting a tension rod through or under the platen, right
below the shear throat edge, thread the ends and use some fine thread
nuts to put some external compression on the platen? Maybe even a
V-block (or two on the webs) at the center like the tensioning rod on
your garage door, to counter any downward forces in the unsupported
center of the platen? THEN you remove the old fishplate, drill the
ends of the cracks, preheat and weld them up again.
It seems to me that compressing the piece first and then welding is
going to cause even worse problems (unless you never remove the
preload again). If you do ever remove the preload, it might just
crack right then. I arranged to leave the first crack slightly spread
(maybe .020") before doing all the welding. After I was done, the crack
had closed up quite nicely.

You can check the finished side of the platen for bowing, and stop
tightening the tensioning nuts when it gets a very slight positive
arch from the preload.
I'm not sure this is going to be stiff enough. I think having a thick
piece of steel sheet that supports the platen along the entire length
would be a lot stiffer than a rod. It might not even need to be welded
to the platen at any point, just attached to the legs.

Jon