Bug blocking progress on model

Jekle

Hi all,

I'm currently working towards a model that will be 3D printed in spiral mode/vase mode. As a result I need to make cuts to connect the outer surface of a model. Also, my model is symmetric. If you're not familiar with the process, don't worry about it, this is just a bit of context, but that explains why I'm completely blocked from making progress on this one.

Hence: I've modeled one side of my part, pushed back some surfaces on the mirror plane, and ran a mirror feature on it.

However, no matter what I do or which approach I take, if I somehow cut/cross the point in the middle of the circle here, the big red DOT, it will not 'Add' the mirror feature no matter what. The side is overlapping, that's not the problem, even enlarging parts or using a boolean instead, it will not join the halves whatsoever.

What is the issue here and how to fix it? Is this kind of thing a known bug?

Rhett_Robinson

Can you share the document? Have you also tried mirroring a new part rather than adding, and then create a boolean operation? Without seeing more of the model I am wondering if you are running into some sort of manifold body issue.

Jekle

Rhett_Robinson said:

Can you share the document? Have you also tried mirroring a new part rather than adding, and then create a boolean operation? Without seeing more of the model I am wondering if you are running into some sort of manifold body issue.

With all due respect, my model has so many features in it that I doubt you'd find the features that create the issue despite my labelling, so unless you have a couple of days to spare and are utterly to bored death, I'll withhold on the model for now .

I've tried different ways of making said cut: thicken, move face, extruding. Same with variations of trying to get it to merge: I tried side-stepping the issue with a boolean of a new part. I've tried a new mirror feature instead of re-trying the same old one, and then some.

Interesting observation: I was just messing around with a chamfer, on that edge/through that point, initially it all converged to that one point, but when I changed the chamfer size, all of a sudden it all worked

Now I haven't been able to recreate the initial weird convergence, but I do get this consistently:

0,1mm chamfer: 

0,2mm chamfer:


*EDIT*:
Just recreated the convergence when I further reduce the chamfer. It's hard to get a good capture, but here you go:
Turns out that any chamfer smaller than 0,05599 mm will do this

dirk_van_der_vaart

Have you tried to do the mirror first and then use a sweep to remove the chamfershape.
And another thing, what is wrong with sketch 50, every sketch that,s turned red has some fault,s.

Jekle

dirk_van_der_vaart said:

Have you tried to do the mirror first and then use a sweep to remove the chamfershape.
And another thing, what is wrong with sketch 50, every sketch that,s turned red has some fault,s.

I'm not able to reproduce that now, but I did try it that way around as well, but that also didn't work.
Sketch 50 is just a remnant of all the things I've tried to get the cut/mirror to work, and from the top of my head actually the sketch I used to make the cut post-mirror. Hence also all the supressed features in there, I try to go back to clean these kinds of things up as I'm already drowning in features. All-in-all, basically, sketch 50 is nothing to worry about.

dirk_van_der_vaart

Try this workaround, move the face half a millimeter before mirror, there is a gap and don,t use the chamfer

Jekle

dirk_van_der_vaart said:

Try this workaround, move the face half a millimeter before mirror, there is a gap and don,t use the chamfer

That is was my original approach, but it was 0,05mm instead of half a mm 😉 which didn’t work.
Dat de versie met chamfer wel werkte kwam pas na wat spelen met functies 

dirk_van_der_vaart

The where you have problems is the result of a Loft I presume, with 2 sketches and some guidelines, there is something wrong with the connection of the guidelines. That,s creating the gap.

Jekle

dirk_van_der_vaart said:

The where you have problems is the result of a Loft I presume, with 2 sketches and some guidelines, there is something wrong with the connection of the guidelines. That,s creating the gap.

Apart from this bit, the model is solid without any errors or anything like that.
So, what you’re saying then comes down to the software using a point on an underlying sketch of a loft(-guideline) as some kind of reference when generating the solid body for the mirror/extrude/chamfer/etc, causing this bug to trigger?
Because making a cut in my model specifically when that point and that point only is involved causes this to upset the Boolean/mirror.

rick_randall

As a last resort , you may try creating a sketch on the parting surface that's giving you trouble. in this sketch draw tiny circle,( or any appropriate closed loop - this can be tiny). Do a symmetric extrude, include both sides in the merge scope. The value can also be very small. May not be best practice, but sometimes this is all that's needed to trick the merge.

I also find that chasing errors are best done with first occurring fault ,first (top of feature tree). Fix that one, and sometimes all the others will heal by themselves

hope this helps

Jekle

rick_randall said:

As a last resort , you may try creating a sketch on the parting surface that's giving you trouble. in this sketch draw tiny circle,( or any appropriate closed loop - this can be tiny). Do a symmetric extrude, include both sides in the merge scope. The value can also be very small. May not be best practice, but sometimes this is all that's needed to trick the merge.

I also find that chasing errors are best done with first occurring fault ,first (top of feature tree). Fix that one, and sometimes all the others will heal by themselves

hope this helps

This is the first occurring fault in a model with over 300 features , only at the home stretch (at least for now) did errors start occurring. See images below!

I could give that approach a try though, see what that does, but I do believe I've already tried something very very similar so I don't have much hope.

dirk_van_der_vaart

Sorry for your troubles with your model, since you cannot share the document with the Forum, my first commend should have been to report a bug to Onshape, if it,s a bug they would want to know.
I,m using Onshape since the beginning and also reported a bug or something weird.
After you report the bug Onshape will come back to you with a solution and explanation of the thing,s that wrong and your document stay,s private.

Jekle

dirk_van_der_vaart said:

Sorry for your troubles with your model, since you cannot share the document with the Forum, my first commend should have been to report a bug to Onshape, if it,s a bug they would want to know.
I,m using Onshape since the beginning and also reported a bug or something weird.
After you report the bug Onshape will come back to you with a solution and explanation of the thing,s that wrong and your document stay,s private.

I would just prefer to not just throw this model out on a public forum due to it's highly complex nature (well, it's a big reason among several reasons). But I've been to that rodeo before, we'd go off on a thousand tangents and lose track of the actual issue.

I could send you the model in DM if you want, just be aware that it's not one you'll walk through in a few minutes. Just let me know.

I did make a bug report, of which I wasn't aware of the button until now XD.

dirk_van_der_vaart

Try the bug report that,s simple and your work will not be public