My version of axle support

Marwayne, I understand you're not an engineer. I am, and all I'm trying to do is help.  

Your point about Toyanvil's design is valid, it does create a moment (torque) on the riser. That's why I suggested he bring the diagonal load up as high as possible to keep the moment load on the riser down to minimum. He did that and I think he'll be OK, but I didn't run the numbers. In my design, all the forces are vertical so its not a problem.

What GlueGuy and I are pointing out is that your design doesn't adress the underlying problem which is a cantilever which is too long. The bending stress on the axle tube starts at zero out at the wheels and reaches a maximum at the axle supports. The longer the distance between the wheel and the support point (the cantilever length) the higher the bending stress is. 

If the bump loads on both wheels are the same (say you go over a speed bump too fast) then the stress on the axle tube is the same between the supports as it is at the supports. Changing the distance between the supports  (which is what your design does by adding an additional support point in the center) doesn't change the bending stress on the tube. So, you will still reach the yield stress point of the steel at the same bump load as if you didn't add your support. Once you hit yield the steel tube will not return to its original curvature and you'll have negative camber. 

I hope that all makes sense. 

Re testing to see if any one of these solutions solves the problem is not really practical or necessary. You'd have to set up a controlled experiment where you overloaded your trailer axle with the reinforcement by an known amount, and show that everything is OK. Then remove the reinforcement and load it to the same point and show that the axle fails. Oops, now you have a bent axle  .  

No one is going to actually do that and its not necessary. The engineering of cantilever beams is very well understood. Buildings are designed using these same formulae (no testing needed)  and occupied by thousands of people safely. 

i think there is a misunderstanding here of the forces involved. As GlueGuy says, the issue is that the cantilever created by the inboard axle mounting points to the frame is too long. Toyanvil's and my approaches adress this by supporting the axle tube outboard of those mounting points, reducing that cantilever load on the axle tube. 

Unfortunately, Marwayne's approach does not provide this support. 

Here is the load diagram and the stress equations (just think of it as flipped upside down).