Using the DeWalt. I'm going to keep posting updates as I work through this in case it sparks any suggestions or in case this thread is useful to someone else.
I was REALLY hopeful that @WillAdams suggestion that it could just be tool deflection would be correct, so I threw the suspect bit back in, with barely more out of the router than I needed for the 3/4" deep cut.
Okay so that's a no. That made me think it was the bit, I ran it too hard and bent it or something. GWizard during my trial was like "OMG MAX THAT SPEED OUT!" but I'd prefer to be conservative. All these tests were at 20ipm with a 1/8" square bit and a .1" cutting depth. So, to be thorough I threw in the new bit, which in my last post produced clean results.....
Suspect bit is bottom left, new bit, this being it's 2nd cut ever, produced the top right, damn near identical results.
I did one final cut, and it kind of looks like one side is deflected 'out' and one side is deflected 'in', which lends itself to the idea that it's not square. I used a normal square and visually it looks fine but of course that is only so precise. I guess I'm spending $100 on a tram testing jig, and hoping that resolves it.