A light weight EE Canberra

If I just do an ellipse it prints a bottom layer as usual & the walls continuously in a spiral.
Putting a circle in the centre (like an eye) with the perimeter of the circle at the same diameter of the minor diameter of the ellipse so they are coincident:-
With the Surface Mode parameter set to "Normal" Cura just generated the ellipse with a bottom layer.
With the Surface Mode parameter set to "Surface it prints the circle as well but no bottom layer & not spiral walls.

To do the windscreen for the Dragonfly I used two different size part-ellipses and additive & subtractive lofting which gave a 1.0mm or 1.5mm thickness which could be different at the bottom & the top. As I understand it the nozzle size determines the thickness of Vase walls which are uniform.

Nice work trying various bracing options.  Interested in the choice to use the CURA slicer to do the lofting rather than do the loft in the CAD tool.  Can you say a bit more about that.

Sorry jon my mistake. The loft is done in FreeCad not in CURA.
A FreeCad lofted stl although acceptable to CURA does have some limitations. For some reason it can only be printed in the direction of the original loft, i.e vertically from the bed. If you turn the print on its side using CURA it throws a wobbler and breaks up the print.
In my case the lofted D box structure would be much stiffer if it could be printed along the line of the span rather than as horizontal layers.

Still not sure why a conventional stl file can be freely rotated as desired but a FreeCad lofted stl is obviously missing a parameter that CURA requires. 

Magpie
It is possible to print a circle within an ellipse and with each layer as a continuous print.
 
By arranging that the complete print has a single path for each layer and ensuring the path is at exactly the right distance apart where contact is required however it is not a true Vase print. Both the ellipse and circle have to have a unique start and end point and each start and end have to be joined so the complete shape is created in one go.
In the video you can see each layer has a start and the complete shape is printed without interruption until returning to the start point and the next layer started
www.youtube.com/shorts/J8yg8VcvMug
There is of course a "seam" left on the outer surface at the layer start point.
I hope this helps.   

Vase print
Your video is the same as how my eyeball printed. The nozzle changed direction when it moved from the ellipse to the circle & vice-versa. A vase should spiral in the same direction on both as if it were a train track with the points on one junction set to change track & the other set to straight on.
A vase also has a bottom layer. Like mine yours hasn't.

Loft print.
I'll attach a screenshot of my Windscreen, double loaded into Cura & sliced.
The back one is how Cura loaded it & is vertical in the direction of the lofts. Obviously I used supports to actually print it.
The front one is rotated so that it it lies on the straight edges. I didn't printed it like that because of all the overhang.

Both sliced OK
 

 

One of the outer wing panels.
 

  
The D box and the ribs are printed all the rest including the rib cap strips are 2mm Depron.
Now to do the other side!
 

A bit of progress. Both outer wing panels done waiting for the aileron servos.
 
And the tail feathers.
 
Still got to work out how to move the elevators with its dihedral tail plane.
Note its bank and yank!
 

I did use a divided push rod on the B2 with simple pin joints. It used a 2g servo! 
 
But for landing protection it was all above the tail plane so one link had to pass through the fin!
It worked OK but it was always a bit "flaky". 

Things have moved on quite a bit.
The outer wing panels are now tissue covered, doped and glued on to the nacelles.
 
The aileron servo wires have been run through the nacelles (but over the duct inside) across the inner wing panels and into the fuselage. Very fiddley and time consuming. 
In this picture the nose is not fixed but just positioned to get a feel for the overall size.
The next task is to cut an opening in the nose for the long tear drop canopy to create a battery hatch but first the ESCs will have to be connected to the EDFs and built in to the inner wing leading edges as they are constructed.

A slight change of plan.
I was concerned that if the ESC were in the wing leading edges the ESC to battery leads would be rather long so I cut of the rear section of the nose and fixed it directly to the fuselage to allow the ESCs to be installed in the fuselage and the PR9's "trade mark" extended inner leading edges to be completed.
 
It also allowed both the EDFs to be checked for correct rotation and the aileron servos for equal travel.
I really must get on an install an elevator servo! 
       .  

I have certainly experienced ESC failure on my early set ups on my Concord when I place the ESCs right next to the motors at the back and had to have the battery well forward in its long nose over a meter away.  Moving the ESCs forward and the battery back a bit cleared the issue completely.

The problem is caused by wire inductance generating short duration peak voltages, greater than the battery voltage, from the high frequency pulsed DC current taken from the battery by the ESC. Made worse by running the +& - wires together in close proximity. 
All ESCs have a capacitor across the +/- battery input specifically to protect the "chips" in the ESC. The capacitor is sized and rated to protect the ESC providing the battery and ESC wire lengths are not greatly exceeded. It has nothing to do with voltage drop from wire resistance. Long wires over work the ESC's capacitor causing it (and thus the ESC) to fail sooner or later and of course it will happen most likely during a period of maximum load.

The ESC to motor wires do also suffer from inductance voltage spikes but as the motor has no ICs it is not spike sensitive at least to 1000+Volts! The wires to the motor can thus be as long as you like but of course the motor will suffer a power loss from wire resistance.