Bearing path projection/sweep/deboss geometry issue

[Michael]

Please, take the Gene Kranz approach:
"I don't care what it was designed to do, I care about what it can do."

Please don't just reply with 'it can't be done'. 😉👍

Hello, and thank you for taking a moment to review my issue.

I am attempting to create a varying slope track for a pair of bearings or chrome pin within a cylindrical tube.
I have attempted many methods over the past week and ran into issues with each one. The best option I have found thus far is creating a debossed inlay from a 2D projection of the track. The issue is there is no way to specify the focal distance (or infinite geometry) and so the algorithm assumes the radial is the objective. This leaves the path skewed and no apparent way to fix it. I have tried cutting circular paths and sweeping paths after the fact hoping some combination of tasks would lead to a solution, but so far I've just been shaving yaks 😞. The second closest solution with significantly lesser success is creating 3D geometry with the emboss tool, spit the generated geometry along the cylinder and extracting the 3d sketch, narrowing the 3d sketch to a single path, then using the spline to create a sweep path. as you may observe from the screencast, each has moderate to severe issues. 
The end-all solution would be if a could trim with a solid body along a rail (or rails) that mimics the bearing's exact geometry and path.
I am almost to the point of using my MacroExpress software and creating a mathematic function to perform the move and subtract for every millidegree of rotation until the desired layer height is obtained. This option is beyond kludgy, would leave far too much geometry in its wake, and would take about 262.5 hours at 7 seconds/div for each attempt. All that is IFF it doesn't (and most likely will) crash my poor i3 slim client, so it's an extreme last-ditch effort.
This isn't an optional thing that if it doesn't work then I'll have to do without, this has to happen someway.
The helical complex is a critical component and several versions will need to be made.
One would have thought this was a simple task, but oh, man.
Thank you so much for reading, and special thanks in advance to anyone who even attempts this ridiculous challenge.
Help me Obi-wan Kenobi! you're our only hope.

EDIT!: The post stripped my screencast and .f3z file so I've re-added it. here:
https://autode.sk/3cQgsm1

error:
Your post has been changed because invalid HTML was found in the message body. The invalid HTML has been removed. Please review the message and submit the message when you are satisfied.

[TheCADWhisperer · Accepted solution]

https://www.youtube.com/watch?v=Wekmai5MqFM

[TheCADWhisperer]

I would not mark as solution until you have gone through and understand the process.

[Michael]

Wow! I have searched for and worked at this for a week now, that was exactly what I was after! I knew there had to be a way, but didn't know how to find it. Thank you so much for making this video. Before your solution, I had tried sheet metal, but it always had the same issues as the current methods with more complexity and weird incongruity in the metrics so I prematurly abandoned that method.
Have a blessed day 😀

[Michael]

Ok 👍
I will follow and complete a couple of the parts to make sure it works for what I need it to do. Thank you for your post.