Advice sought on this network intersection as fixing it wrecks correct profile

Hi
V5

C and D show intersections that dont ‘X’…i e meet up. Also A and B need a slight tweak.
Those at X are ok.

adjusting (red line) to create an intersection ruins the profile (lower part of jpg). The arcs at leading edge of each profile are radiused based but my red line will break that trend/rule. It certainly doesnt follow the steep sides.

Drawn to a 70yr old dimensioned plan I am stuck on quite what to do, Rhino needs intersections if I am going to use NetworkSrf.
and anyway these profiles should intersect, but faced with profiles created from dimensioned plans, and looking good, like onion skins, I am not sure what to do here.

SECONDLY …given those at A and B, how can one tweak one part of a curve without altering (losing ) its correct intersections at points labeled X. Turning on control points and adjusting sees all but start and end of curve move from the X points. I need something that protects the X’s and also snaps the edit to the desired point at A and B.

THIRDLY …I need to surface this then trim the front tip off at indicated line. I have no profile for that last 1mm or so, but this should skin up I would think after I have sorted out red line.

File attached.Aerofoil_Network_Dilemma.3dm (67.8 KB)

Steve

Dunno- it seems clear that one or both of the intersecting profiles need to be changed, I don’t see any way around that… You can start part of the curve at the desired intersection point and work back from that location in either direction- Match, possibly Merge, when done, to make sure the two parts are continuous.

-Pascal

Hi, I cant see how I can alter curves without messing up those at X X X which are good.

I have spent days trying to make lines cross, every time I edit a curve it moves another intersection out of cross roads mode and into flyover !

what is the edit technique to alter part of a curve and ensure other parts dont lose their crossroad status and become flyover, this can develop into a killer of a tailchase,

The fore aft curve in my example is drawn to dimensioned plans. somebody hasnt got this part right. but as you can see, the profiles are onioned ringed in both directions.

anyone care for an attempted tweak ? I am exhausted on it. 15 hrs :frowning:

Steve

Try SoftEditCrv to move only part of a curve.

Unfortunately SoftEditCrv has a problem which limits it’s usefulness. I started another thread about the bug: SoftEditCrv bug?

FIRST: Never trust any hand drawn, 70yr old plan. Trust the dimensions, use the drawing as a hint only. I have yet to find any hand drawn construction plan where the floor plan matches the elevation, matches the section drawing… If they were good to begin with, they may have been distorted while reproduction or during scanning.

SECONDLY: In general: Loft the sections for one direction. Create sections for the other direction and compare them with the ones from the drawing. do it the other way around. Now decide which direction to trust and redraw the curves to better match the opposite direction.
Specifically: Wings are build from sections starting from the wing base to the tip. Those ribs are covered with sheet metal or plywood. So this is like a loft from wing base to tip along given ribs. The other sections are the result of cutting through that loft. They are secondary information and do not necessarily represent the wing exactly.

THIRDLY: The way wings are built, normally requires the ribs to be similar. You can extrapolate the complete wing by moving the last complete section to the wing base and scale it appropriately. Then cut away unnecessary stuff.

Hi,
As you say, dont trust drawn shapes. I never do, they are drawn deliberately ‘off’ so as not to be traceable, so as to ensure users back then didnt use copies of copies. The eastings and northings dimensions that are with them should be enough back then to plot the form, with splines.

This item was a small tailplane tip so very thin cane splines indeed,

also as you say the profiles in one direction and their spacing should be enough to form the shape in 3D. Chances of the secondary set matching are slim.

I did eventually manage to intersect all these, before realising what I just wrote, I also somehow managed to use NetworkSrf on them all, and it worked, yet today I cant get it to do so at all.

Wings are often straight sweeps from inboard to outboard profiles, sometimes and intermediate rib. This was not a wing as such but an elevator tip and I had a feeling seeing the real thing it was shaped with a need to round it off towards the tip.

On the topic of what to use so as not to upset curve through points, I ended up using curve edit points and curvature graph, adding a point and snapping point to target point and moving those nearby a tad, all with nudges, until I achieved intersection.
SoftEditCrv with this bug is a loss, just my luck to find V5 has this.
I sem to find V5 has knobbled me a bit with the things I took for granted or need here. layer bug, SoftEdit bug, networkSrf beware advice, single surface for filleting limitation…all happening on this one task. :frowning:

Steve

SoftEditCrv is apparently working as intended, but is not as useful as if it could be. SoftEditCrv bug?