Best approach please for end InterpCrve tangent to a circle


What is best method to draw a curve using InterpCrv and end it tangent onto a circle ?
task indicated in jpg.
Curve starts elsewhere, often with S start tangent, and cntrl osnap OnCrv to select start line, though that has no bearing on how to end it !



This is not very different from this

as far as I can see. I’d review that thread.


i think there are 5 of these threads here :smile:

here’s another:

Aware of those, but…
I wished to keep this simple as those threads became a mix of methods not all actually grasping what my problem and aim was. Whilst showing various methods which I now have stored should I need them, I was unable to apply them in this case ,maybe I missed something.

In this post I am starting away from the circle and ending at it.

The answers were about starting from the circle which was my task back then, now in some cases that’s the same, doesnt matter, for what I now have to do I need to end at circle, I cant start at it.

Jeff had said Any point would work, maybe so technically as all four are tangent, but any point chosen on my circle would give such twists and not allow my curve to sweep and kiss, a moon orbit rather than a moon crash ! Even the original task still required an in orbit type apprach of the curve.

e.g. menno said

Why not put the circle in later using Circle Tangent to 2 or 3 curves?

My circle is one that is not decided by three or even two curves, and has to be where it is and of the diameter it is, it dictates the curve, it does not get dictated by it.

There is only one spot on this circle that lets my curve line flow nicely (tangentially) into the circle.
I wish to find that spot where as I move my mouse about and the curve moves up and down , it ‘kisses’ the circle or curve. Thats the point I am after finding, at that point it will be tangential and give me the natural sweep it was wanting to do to the circle. see the two others where i picked a tangent point as per some other suggestions, curvature graph tells all even if the mk1 eyeball doesnt.

Unless I am god I cant pick that point with line tangent to circle tool.

Maybe this in the first thread linked to here, an entry by Jim, is the best shot at this,

In the enclosed file shows how you might get the curve you want. The two
red lines are drawn as helpers. The red lines are tangent to the circle
with the other end about midway from the two start and end points in
your original interpolated curve. The cyan curve is what you get using
this method.

see that thread for his file.

He starts at the circle and using line tangent from curve draws a line out to halfway to the first point I will be drawing through, a visual aim at that point, (as extending the line misses the point).
He would have had the end of that line at the circle moving about a little as he did this. If I was to have my curve going through that point on circle, would the resulting circle when trimmed and joined to my curve be ok given this ‘subtle variation’ by eye junction for use as a profile for surfacing with sweep ? being aware of what subtle inaccuracies can do to surfaces in sweeps, I am after technical correct here.

Was that method of Jim better than my current method of ending curve …
Here is storyboard of my ‘visual’ method, is it as good as or better than Jim for technical acceptance in Rhino , how good would my result be v Jim for being used in a sweep ?

What is best way of establishing the kiss point on the circle ? I then trim to that and draw my curve ENDING at that point, and do I need to use n for end tangent when doing so ? and my curve enters moon orbit tangential to the moon !


I’d probably eyeball it as close as possible, and then adjustable curve blend the curve to the circle.

If you overshoot the circle, it will make it easier to see if you’re eyeballing it close. Keep your curve outside of the circle, and you should be fine.




googled adjustable curve blend bit found nothing. Just which command is that ?

Cheers Pascal, videos invaluable and I would like to see such in a location on the website one and all users can access. I searched on loft earlier and found no results found. (see tailplane tip post)
it needs some use of the pause button and slow drag of timeline to spot some entries, I was i

One could do with a yellowy glow at cursor, like Camtasia does. not always easy to see where cursor has just ticked etc, unless you hover and rotate/squiggle about the area you are altering choice wise. Certainly keyb’d entries dont appear, camtasia has keyb’d entry pop up. Mouse clicks dont appear as a ripple or red spot as camtasia, right click being blue. Cant remember what Jing did as havent installed it yet. Glad to see anything video wise :slight_smile:

I guess sound bangs up the file size ?

I see that you typed a command (match) in during the command, I never think to do that, picked the circle with tan osnap, the options panel preserve tangent radio buttoned not proving a snap function but a constraint, and continuity keeping the curvature through to the first point , am I right ?

rotated that point around centre (as tangent point rotates about centre)

What was the thinking behind the decision as to where to stop rotation, looking good or a known stop point ?

I also have non circles to deal with, rotation of tan not of use then, what would be done there ?


yeah, in that thread, i said any point would work as far as giving a tangent endpoint… and that is true but in one of the other similar threads, i refined my stance and said i do think there’s an exact point to be found…

i don’t understand the mathematics involved since the start of the interpolated curve is still moving around when looking for the end tangent but i’m pretty convinced there’s an exact point which could (or should?) snap into place…

if you’re sliding the endTangent point around the circle, it would be the point that’s exactly on the threshold of the interpolated curve crossing the circle…

but again, how to find that point -exactly- is beyond me.

out of curiosity… where are these inner points coming from? they seemingly need to be exact and the circle on the end needs to be exact etc… so why don’t the plans include the start/stop point?
or, how were the original start points determined? with a french curve or something?(which would be akin to ‘looking good’)

For the task of touching upon, a glancing blow, the circle, one example… the aerofoil profile comes with a blunt end, upper and lower curves just stop without being related to the frame… The circle of given diameter comes from the frame drawing and placing it where it should sit regards distance from the spar sees it sit with a gap inside of the section. so drawing the latter portion to be as close to the original angle and tangent onto onto circle is needed. In my current task it was travelling to a circle.

As for a curve instead of the circle , various needs arise in other parts of the aircraft, or other tasks.


This thread maybe has some value in itself to learn a few curve manipulation tricks. But knowing the context into which these curves are needed for, I think it’s not the right way to approach this.
Instead of extending the section curves with arcs, it would be better to create the main surfaces and then use fillets to create round ends, instead of trying to create everything in one go.

Hi, Filleting the blunt end though would not create say a 1.5mm radiused tip where the blunt end was 2mm above chord axis ?

If you use a value of 1.5mm in the command parameters, you’ll get a surface with a 1.5mm radius.