Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.
Inconsistencies in geometry when converting to solid
Posted Jul 2, 2015, 5:58 p.m. EDT 1 Reply
Please login with a confirmed email address before reporting spam
Hello,
I'm currently working on a tip like geometry for a simulation, and I'm struggling with something that seems like it would have an obvious fix.
What I am doing is defining the profile of the tip using a function. I'm taking sections of the tip with vertical lines in increments to be used for the meshing later as I go down. It all goes well until I get to the end of the tip, where I now change the function to a parabola to finish the tip. The actual creation and plotting of this function on the geometry is fine and well, however when I go to select my boundaries and "Convert to Solid", all of a sudden a new line as been created that is some sort of lower resolution imitation of the old one, and is being used to define the boundary of this new solid. Even though I have selected the same line as a boundary, the created solid is different. This is unacceptable because now it's going to be impossible to rotate my geometry with these strange false shapes.
My guess is that it is related to these repair tolerances that we're trying to understand, but it seems that no matter how we change any of the repair tolerances on anything, it doesn't really change how the geometry looks. Am I possibly just too small? The 'radius of curvature' of the end of the tip is one the order of 250 nm, but I don't see how this is the issue. Is there somewhere I can increase the resolution in order to prevent it from making these guesses/truncations that it appears to be doing?
Thank you,
Graham
I'm currently working on a tip like geometry for a simulation, and I'm struggling with something that seems like it would have an obvious fix.
What I am doing is defining the profile of the tip using a function. I'm taking sections of the tip with vertical lines in increments to be used for the meshing later as I go down. It all goes well until I get to the end of the tip, where I now change the function to a parabola to finish the tip. The actual creation and plotting of this function on the geometry is fine and well, however when I go to select my boundaries and "Convert to Solid", all of a sudden a new line as been created that is some sort of lower resolution imitation of the old one, and is being used to define the boundary of this new solid. Even though I have selected the same line as a boundary, the created solid is different. This is unacceptable because now it's going to be impossible to rotate my geometry with these strange false shapes.
My guess is that it is related to these repair tolerances that we're trying to understand, but it seems that no matter how we change any of the repair tolerances on anything, it doesn't really change how the geometry looks. Am I possibly just too small? The 'radius of curvature' of the end of the tip is one the order of 250 nm, but I don't see how this is the issue. Is there somewhere I can increase the resolution in order to prevent it from making these guesses/truncations that it appears to be doing?
Thank you,
Graham
Attachments:
1 Reply Last Post Jul 3, 2015, 4:53 p.m. EDT
Please login with a confirmed email address before reporting spam
Posted:
9 years ago
I didn't find a fix but instead a workaround for anyone with a similar problem:
I went and generated my own ~1000 point interpolation of the function (in Matlab) and then had COMSOL import this CSV and plot it as an open polygon. Worked way better than allowing COMSOL to try to calculate the curves itself. I will be doing this for every curve I generate from now on and not relying on COMSOL (which is disappointing.)
I went and generated my own ~1000 point interpolation of the function (in Matlab) and then had COMSOL import this CSV and plot it as an open polygon. Worked way better than allowing COMSOL to try to calculate the curves itself. I will be doing this for every curve I generate from now on and not relying on COMSOL (which is disappointing.)