Yes, I think that’s right too. That was my original thought, that the motors would rotate relative to the rings (and the rest of the elevator, going from 11 degrees off perpendicular to perfectly perpendicular at top, then 11 degrees off at the bottom of the opposite leg.
I didn’t double up on the rings on purpose, but that may be a mistake on my part when I twisted the motors thinking they were separate from the rings.
Then we are on the same page - just a matter of how to draw it.
Well, at least it seems to be done correctly in at least some of the many iterations so far 
We are close, I think.
My dinner is soon, but I will check back later.
Arch 8c.skp (2.7 MB)
I placed an elevator at the top of the arch and a few other places. In no case did I have to rotate the elevator in any way other than accounted for in the current design, though I didn’t turn the tracks in these quick samples.
But note that my pushpulled section does not leave a flat top the way the original cut out did, so as you say, clearly it is a matter of how it is first drawn.
OK, will check back later. Thanks.
I’m really struggling here.
Helix, stretched, just doesn’t fit - my intuition isn’t doing well on that one. When I think about it, the edge of a helix can’t lie in a plane, which the sloping face does.
Have tried half a dozen times now with FredoScale, and get it wrong every time. Will have another go, but if that doesn’t work, will give up for tonight.
Sorry, it’s taking a long time and is trickier than I thought it would be.
Thanks. I think that whatever you did in archb (which I made into archc) was mostly there except for other problems with the arch itself. The “roof” of the channel was flat so that worked, and the angle of the lower channel was managable with the tilting motors and tracks. Maybe go back to that method?
I wonder if a more useful shape would be to just tilt the outside face, then push-pull a 90° channel from that face? To get the elevator horizontal, it would have to hang off a frame with approx 11° angle.
I’ll try that too, after I get the original idea of v10 to work with the correct channel dimensions.
I’m reluctant to make the elevator do more bending and twisting. It’s already doing two directions. I think flattening the arch roof is the better way to go…and one you’ve already done apparently. Maybe it’s possible to create the channel separately, and then move it into place and remove extra geometry.
But I don’t want to make more work for you and 8b was pretty close, channel wise.
Hmm, if you scale it though, the channel will get stretched. Maybe there is another way?
I thought I was there with the FredoScale solution, but the inner arch is around 2ft too low at the top. Tried to fix it with a bit of scaling, but can’t. I can’t really understand it. Maybe I have to scale the inner and outer arches separately after they are drawn as half circles.
Will have one more go tonight, then stop.
Finally got it, exact.
Here’s the file:
Arch 11.skp (752.1 KB)
Looks like this:
Channel dimensions:
At top
12’6 measured difference in radius on blue axis
12’9 on the slope (to nearest inch)
And at the bottom:
But what a palaver! Interesting, though. First time I’ve used FredoScale in a long time, in earnest.
Great! But there is one thing off still. It should taper at the top vertically too. 875’ from the bottom of the arch to the ground, in a non-concentric arch to the outer one. But, at least the channel is right and now I can move onto the elevator cowl and mirroring the window on the other side (might need your help with that last bit).
So it should - drat. I missed a step to make the inner arches elliptical before the FredoScale. Sorry. Will try again tomorrow though probably busy on other things till evening at earliest.
I’ve been playing a bit with a simplified model elevator and track. Approximated to round number dimensions from your elevator drawing. As you can see in the model, I’ve separated the concrete block and track from the elevator model - they don’t belong together. And you will need unique copies of the elevator and motor blocks, since each different position around the arch needs to have the motor blocks at a different angle - so you need more clearance to allow them to rotate - which will be fun for someone to engineer!
Arch 11a.skp (464.5 KB)
Had enough for tonight, but here’s where I’ve got to - one elevator at the bottom, and one at 60° up. As you can see from the upper one, the track sections joing at a slighter lesser angle as you go up - which I haven’t done, to prove the point.
Here’s the bottom elevator:
And here’s the upper one:
For you own experiments along these lines, as I suggested very early on in this series of threads, you should usesimplified models as you are working out the details of how things will work at this large scale, and replace simple components with more complex ones only for presentational purposes.
External Elevator 2.skp (3.0 MB)
I inserted two swivels between the motors and the rings. These will work if the angle is moderate - under 15 degrees or so, but technically, the track layers will have to run the tracks slightly closer together when there is an angle to the motors relative to the length of the elevator in order to handle the skew. The greater the angle, the more the tracks would have to be closer to one another. This can all be calculated by computer, I’m sure, and for illustrative purposes it doesn’t matter since we are dealing with angles of 11.31 or less.
Yes, lesser angle at the top. In my presentation, I’ll present the elevator in 3 - 4 positions and people will understand there will be adjustments along the way up and down.
P.S. I’ll need help mirroring the window to the other side.
Well. I think I have finally got it as you wanted - elliptical inner arch profile, with constant width channel.
Here it is:
Arch 12.skp (306.4 KB)
This time:
1 I remembered to scale each inner arc separately, to the correct bulge height after scaling.
2 I made half of the whole structure as a component, then mirrored it, hiding the edges on the midline so they don’t show.
3. I am getting better at using FredoScale, so I got it right first time this time.
Here’s the result, dimensioned:
Hope that is now as you intended.
Be aware that if you measure the 12’6 dimension of the channel using the tape measure at the top, it most easily measures the slant direction (~12’9)., not the radial dimension, which is 12’6 - as in the previous version.
Perfect! This is great. I’ll have to ask you about the mirroring soon too, for the elevator, but I am also laying the columns for the first floor, which will continue all the way up, so will need to mirror those too.
But for now, thank you very much! 
Converting the arc to poly line lets you set the arc length to a poly line, and then some of the SU tools including the poly line divider to the lengths you what , almost. You cannot not have a fraction of number so the number of the segments must be an interger. Some of the SU tools set length between the vertecies and can give you short sections at ends and one divides them evenly for the full length.
Just some thoughts