USER FORUM
(you are viewing a thread; or go back to list of threads)
Unexpected "Solve Failure" (by Walter)
In the enclosed Test-file I designed a box with rounded edges.
To keep the thickness of the wall even in roundings the same,
I wanted to constrain the center points of the circles "OnPoint".
This ends in a "Solve Failure".
My unexpected WorkAround is to make them at a distance of 0.01
or even less, but not 0.
Please verify it in the enclosed Test.slvs
To keep the thickness of the wall even in roundings the same,
I wanted to constrain the center points of the circles "OnPoint".
This ends in a "Solve Failure".
My unexpected WorkAround is to make them at a distance of 0.01
or even less, but not 0.
Please verify it in the enclosed Test.slvs
(no subject) (by Andrew)
Your problem is redundant constraints, meaning two (or more constraints have the same effect). In the case of your example, removing the eq-pt-line-distance constraint eliminates the error. the tangent constraints make the line offsets the same as the tangent points are the same distance from the center of the radius, and the centers are the same distance from the edges.
Unexpected "Solve Failure" (by Walter)
Thank you for answering.
For me it is not clear why a distance of 0.01 is OK,
but a distance of 0.0 is NOT OK.
There are no other constraints, just a value is different.
The respective value constrains the radius of the inner circle, as one can see by varying the value.
So there should not be a difference in constraints, or is there anything I dont understand?
I will experiment a further more, perhaps I can find whats wrong in my design, the value of 0.01 is a practical solution, but its not "perfect".
For me it is not clear why a distance of 0.01 is OK,
but a distance of 0.0 is NOT OK.
There are no other constraints, just a value is different.
The respective value constrains the radius of the inner circle, as one can see by varying the value.
So there should not be a difference in constraints, or is there anything I dont understand?
I will experiment a further more, perhaps I can find whats wrong in my design, the value of 0.01 is a practical solution, but its not "perfect".
(no subject) (by Tom)
Distance of 0.0 doesn't usually work right, I guess because of some kind of degeneracy as the constraint goes from something like a circle or other curve (nonzero distance) to something like a point (zero). It would be nice if that gave some kind of warning, or automatically converted to a point-coincident constraint.
(no subject) (by Paul)
@Walter
A points-coincident constraint in a workplane creates 2 equations equivalent to X1=X2 and Y1=Y2 to constrain points to the same 2d location. In 3D it covers 3 degrees of freedom by including Z1=Z2 as a 3d equation. This is definitely not the same as setting a distance to a constant value, which is one degree of freedom in both 2d and 3d.
I often have trouble with arcs because they seem to be defined by 3 points in a workplane which would suggest they have 6 degrees of freedom, but they really only have 5, as there is an implicit constraint that the 2 endpoints are equidistant from the center which reduces it by one DoF.
This is really the first time I looked closely at the text window after over constraining. It does indicate 4 options of constraints that can be removed to resolve the issue, and hovering over each of them highlights the constraints on the sketch. That's a really nice feature. I usually just UNDO when it happens and then figure out why.
A points-coincident constraint in a workplane creates 2 equations equivalent to X1=X2 and Y1=Y2 to constrain points to the same 2d location. In 3D it covers 3 degrees of freedom by including Z1=Z2 as a 3d equation. This is definitely not the same as setting a distance to a constant value, which is one degree of freedom in both 2d and 3d.
I often have trouble with arcs because they seem to be defined by 3 points in a workplane which would suggest they have 6 degrees of freedom, but they really only have 5, as there is an implicit constraint that the 2 endpoints are equidistant from the center which reduces it by one DoF.
This is really the first time I looked closely at the text window after over constraining. It does indicate 4 options of constraints that can be removed to resolve the issue, and hovering over each of them highlights the constraints on the sketch. That's a really nice feature. I usually just UNDO when it happens and then figure out why.
(no subject) (by Andrew)
>For me it is not clear why a distance of 0.01 is OK,
Without the distance constraint, the arc center is free to move along a 45 degree line from the corner, that it it has one degree of freedom. A non zero distance removes one degree of freedom, while coincident points removes two, and therefore over constrains the model. Removing either the equal distance constraint, or one tangent constraint give it two degrees of freedom, and so points coincident no longer over constrains.
Without the distance constraint, the arc center is free to move along a 45 degree line from the corner, that it it has one degree of freedom. A non zero distance removes one degree of freedom, while coincident points removes two, and therefore over constrains the model. Removing either the equal distance constraint, or one tangent constraint give it two degrees of freedom, and so points coincident no longer over constrains.
(no subject) (by ruevs)
> This is really the first time I looked closely at the text window after over constraining.
😯
> It does indicate 4 options of constraints that can be removed to resolve the issue, and hovering over each of them highlights the constraints on the sketch. That's a really nice feature.
Indeed. I'm really surprised you never noticed it.
😯
> It does indicate 4 options of constraints that can be removed to resolve the issue, and hovering over each of them highlights the constraints on the sketch. That's a really nice feature.
Indeed. I'm really surprised you never noticed it.
(no subject) (by ruevs)
Andrew - as usual gives a perfect explanation.
Post a reply to this comment: