Main-line trials today went pretty good, but there was one spot on a curve where it consistently derailed. It looks like one wheel climbs up the rail then the other wheel falls into the gauge. So I took the trucks off again to check the back-to back. I have a digital micrometer and I am getting crazy readings with it. Three different measurements on the same wheel set. My conclusion was that I was still to wide, so with a little persuasion from a hammer I was able to tighten up the back to back a bit more, but still can’t get a consistent 1.575" reading.
Getting frustrated again thinking the semi-scale flanges were the problem I grabbed a spare set of USA wheels and shortened the axle on the grinder. The shoulder on the axle would not go into the side frames without drilling out the brass. I’m not ready to bugger them up quite yet, so I gave up on that idea.
I have a ready-to-go set of Delton trucks with metal wheels, but I couldn’t come up with any hardware to mount them on the AMS car. Another idea foiled.
Ready to give up again, I put the AMS truck back together and did one last test run indoors. Now it’s was back to dropping into the gauge on one switch. Just for the heck of it, I spun the offending truck around 180 degrees and tried again. It worked!!! Took the car back out on the main-line. Ran it through switches at speed and around the curve that it derailed on earlier. The car tracked smoothly with no derailments.
It looks like the problem is solved. I just wish I knew what I did to fix it.