still have something binding in track on CAT 225. Best way to move w gimp track

I simply cannot pinpoint the problem with the track on the 225 and I cannot put more time into it right now. We were thinking the track brake was causing problems so I rebuilt both of them. Upon tear down of the problem side track brake, I realize the input shaft bearing was toast and the input shaft was cracked. I attributed the input shaft and bearing failure to someone leaving the drive motor loose,which basically was hanging on the input shaft, and clutch material and trash built up in the bearing. There was Also water in the system because the drive motor was tight and allowed water to enter the brake and the final. I rebuilt all that really thinking that I had found the problem and all would be good. None of this fixed the problem and even with several eyes on the tracks when running, no one can figure out what is causing this. Basically if I creep along very slowly, the track will run pretty smooth, all links seems to be free, and I have more than the recommended slack in the chains. We have combed this issue before on this site, suspecting chain stretch was beyond limits and the bushings were striking the tip of sprocket causing it to bind. I determined the pitch was stretched but not to max limit and the bushings were hitting a reasonable zone on the sprocket. Basically when the machine tracks, at slow speed, both tracks move together as they should. There is a point in the travel that is speed dependent that one track starts to momentarily stop and bind. The engine loads down indicating there is load. I feel this issue is purely mechanical, not hydraulic. When the track binds up, it stops the track which jolts the machine a bit. This happens with the track off or on the ground. My father suspects someone lodged something between the sprocket and case and bent the output shaft but I am having a hard time with that one. I am also thinking if the output shaft bearings are beat up this bad, I should have found metal when draining the final and there was none. The gear faces that I could see looked brand new. I also wondered if bearing failure this bad would be seen as a leak at the sprocket and it is all dry. Maybe wishful thinking.... I have run the hydro drive motor uncoupled and it runs perfectly, no play in the shaft, no leaks, and simply not convinced it could be that motor. When this binding event happens, I cannot trace it to anything rotationally consistent with a part of the track. I have tried to stop tracking the second I feel it try to bind and inspect everything and cannot see anything at the sprocket/chain interface that would lead to the cause. Simply put, I stumped on this one. And needing the machine right now, I need to let it do what it can, even if that is crawl at snail pace. I suspect trying to turn with the machine will be a concern as this loads the tracks up even more. I am curious if you guys think lifting the front of the machine and pivoting by swinging the house to turn the machine is much safer than trying to turn it? Or is that just as hard on things? Obviously I want to find the cause of the problem and correct it but I can't come up with any better test than popping the track off to run the final drive separately. But, due to my limited time and resources, I am trying to make do.