Tangent constraint ambiguity?

[Anonymous]

I'm developing a sketch that is heavily driven with parameters. This fact will become important later.

When constraining two circles with the tangent constraint there is no distinction to which side the constraint should act on. When placing the constraint the sides that are closest are the ones constrained.

To repeat this experiment create two circles with one much larger than the other. Fully constrain the large circle. The smaller circle only constrain on one axis. Now you can apply the tangent constraint. Notice how it is ambiguous to which side of the line it is constrained too.

This becomes a mess when the sketch is parameter driven because often the constraint will flip.

So far I've been able to get out of this mess by finding alternative ways to constrain that act in the same fashion as a tangent.

I believe my problem might be a misuse of the tangent constraint but I thought I might ask on here since I've searched on quite a few occasions of this problem. Let me know if my question is confusing and I'll post a screencast to give example.

[johnsonshiue]

Hi! Could you attach the part you are working on? It will be easier to tell where the problem is.

Many thanks!

[Anonymous]

I'd hope that my textual explanation would be enough as the sketch I'm working on is complicated. Although the sketch is complicated I'm confident that the issue is as exact as what I've written above.

I did make an attempt now to recreate a parts file with a sketch that simulates the ambiguity problem of tangents but the part file I could come up with doesn't really help support the problem. Actually getting inventor to flip the tangent apparently harder in a simplified example sketch.

I would like to assert that I feel like you didn't actually read my post and your response was an automatic response. Which I understand is fully justified to ask for a parts file but until I can actually recreate a basic version of the sketch I would hope my textual explanation is enough.

[johnsonshiue]

Hi! I am a human. I am not a machine. I don't do any automatic reply. I think I know what you are talking about and I have seen a case like this. However, it is geometry specific and constraint set specific. This is why you are not able to reproduce the behavior using a simple example. As a result, I would not be able to reproduce it on my side without your file.

I suspect the solution shown in Inventor sketch is mathematically valid but it does not obey your intent. If possible, please send the file to me directly (johnson.shiue@autodesk.com) so I can take a look to find out where the problem is. At least, I would like to find an explanation of such behavior.

Many thanks!

[rhasell]

Following.

When working with circles, I have always had issues with tangent, I will go to great lengths to find alternate ways to constrain sketches and parts (Assembly tangent included)

I don't think its a software issue, its just the nature of the beast. It's like constraining to a point, IMO the last resort constraint.

Everything works as predicted until the geometry changes, then the constraints do not know where they belong on the arc any more (left/right, top/bottom). It also get tricky not to induce an overconstraint condition by adding helper dimensions and constraints.

So going back to your post, I suspect what you have been doing is probably the best option.

Just my opinion of course, that's why I am following, to see if there is something to learn here.

[Anonymous]

I've actually figured out how to control the tangent constraint. At least for now

My project is to build a fully parameter driven skateboard miniramp/halfpipe design. In about a week about 15 hours worth of work and about 4 throw away part files later I've actually come up with something very reliable.

The concerning point for the tangent was where the deck extended to the coping (circle hollow section). The design concerns get a little bit complicated due to the way the coping can be adjusted but alas for this post all you need to know is that I need a tangent to calculate it.

One thing I observed was how the deck would flip over to the far side of the coping when I adjusted the copings outer diameter more than twice its previous amount. I also observed that it never flipped back.

So with my design I just accepted the flipped version and worked backwards from it. You can see in my screenshots how I strangle the copings OD with a rectangle. So far no matter what I do it'll always calculate correctly.

I intend to put this concept into practice on another place I need a tangent constraint.

[Anonymous]

I was just about to post a few more screenshots as examples to show how the coping is adjusted in extreme increments but then I noticed "the tangent" was actually flip flopping right to left and back again.

Note: I'm referring to the sketch line and tangent constraint relationship as just "the tangent".

Surprisingly I managed to get "the tangent" to flip back and forth but because the design accepts a flipped tangent it didn't break my design.

In this screenshot you can see "the tangent" gets flipped. But thats okay because it just means it sits on top of the intended edge.

I hope this helps people who read this to understand. Please feel free to ask questions.