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.

Work-plane definition flaw?

Please login with a confirmed email address before reporting spam

Hi - it seems that using "Face Parallel" to define a new Work-Plane, say parallel to a face in x-y plane, changes the origin of the newly created work plane to a value different than the (x,y) coordinates of the origin in the 3D geometry. Here is an example:

In 3D define a new Worke Plane in (x,y) at z=0 which becomes Geom2. Then draw two concentric rectangles, which do not contain the origin at (0,0) for demonstration purposes. Next, Extract the bigger rectangle into Geom1, define a new Work Plane at the top face (4) using "Face Parallel" as Geom2, where the two rectangles drawn earlier appear together with a projection of Geom1.

Unfortunately, the projection does not coincide with the bigger rectangle, so it seems that the origin has shifted during the Work-Plane definition. Defining the new Work-Plane using "Quick" with a defined z-offset circumvents the shifted origin.

Is that a desired behavior or a flaw. If it's desired, what would be the proper way to use "Face Parallel" to define a new Work-Plane. For me it's important since I imported two DXF files and would like to maintain the alignment between the two.

Thanks and best regards,
K.P.

6 Replies Last Post 22.12.2009, 16:41 GMT-5

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 19.12.2009, 11:07 GMT-5
EDIT: Erm, it seems that you're right, looks like a flaw or bug. I'm testing this and it seems there's strange shift there. What I wrote is approximately true, but the coordinates aren't exact. I tested it with a square at (1.4,.4), and then did a face parallel on the top face. On the new work plane, the drawing was at (1.4,.4) (which is to be expected), but the projected face is not at (0,0) (where I thought it would be),but at (8e-4,4e-3). I'm officially baffled.

Well, one more reason why vertices are my friends.
------------------------------------------------------------
I'm guessing it's expected behavior. When you make a face parallel plane, it will set the zero at a subjective origin with respect to the face. This is usually quite useful.

It's something you'll have to live with, as far as I know. I think that if you wish to use face-parallel, the easiest way is to move your whole drawing into position if you coordinates shift.

One thing, though: I'm not a big fan of face parallel. I prefer to use vertices, since they allow me better control of the coordinates (so you don't end up with drawing 2 being 90° from drawing 1), and you end up with the same results for two clicks more.

EDIT: Erm, it seems that you're right, looks like a flaw or bug. I'm testing this and it seems there's strange shift there. What I wrote is approximately true, but the coordinates aren't exact. I tested it with a square at (1.4,.4), and then did a face parallel on the top face. On the new work plane, the drawing was at (1.4,.4) (which is to be expected), but the projected face is not at (0,0) (where I thought it would be),but at (8e-4,4e-3). I'm officially baffled. Well, one more reason why vertices are my friends. ------------------------------------------------------------ I'm guessing it's expected behavior. When you make a face parallel plane, it will set the zero at a subjective origin with respect to the face. This is usually quite useful. It's something you'll have to live with, as far as I know. I think that if you wish to use face-parallel, the easiest way is to move your whole drawing into position if you coordinates shift. One thing, though: I'm not a big fan of face parallel. I prefer to use vertices, since they allow me better control of the coordinates (so you don't end up with drawing 2 being 90° from drawing 1), and you end up with the same results for two clicks more.

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 21.12.2009, 09:44 GMT-5
Hi

I havent fully managed to duplicate your case, but I know that one of the "features" of the workplanes is to set the origine for the new workplane in the lower left corner of it's natural x,y,(z) coordinate system, as I have been explained this is by design and no bug, but not naturally understood by us users as such.

I suspect it has to do with the treatment of the signs in 3D to stay in the first "positive" quadrant

Personally it complicates my way of doing things, as I do not necesarily want the origin to remain on the original 3D one, nor in the "lower left corner" but often prefer a more symmetric position, this means a two step WP creation and some empty WP to deal with, the other way is clearly 3 vertices or as I use the "advanced" as I have seldom vertices at the right place.

I expect this to be changed in V4, perhaps not immediatly, we will see but at least some improvement on WP and coordinate system identification in the graphics the "apply" is already an improvement shadow planes would be even better, no ?

Could you pls verify if your observations follows my explanations and report back ?

Good luck
Ivar
Hi I havent fully managed to duplicate your case, but I know that one of the "features" of the workplanes is to set the origine for the new workplane in the lower left corner of it's natural x,y,(z) coordinate system, as I have been explained this is by design and no bug, but not naturally understood by us users as such. I suspect it has to do with the treatment of the signs in 3D to stay in the first "positive" quadrant Personally it complicates my way of doing things, as I do not necesarily want the origin to remain on the original 3D one, nor in the "lower left corner" but often prefer a more symmetric position, this means a two step WP creation and some empty WP to deal with, the other way is clearly 3 vertices or as I use the "advanced" as I have seldom vertices at the right place. I expect this to be changed in V4, perhaps not immediatly, we will see but at least some improvement on WP and coordinate system identification in the graphics the "apply" is already an improvement shadow planes would be even better, no ? Could you pls verify if your observations follows my explanations and report back ? Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 21.12.2009, 12:25 GMT-5

I suspect it has to do with the treatment of the signs in 3D to stay in the first "positive" quadrant

Well, sort of. I just tried again and upon creating the WP using Face Parallel, the geometry is not entirely in the first quadrant but the lower left corner is at (-1.3e-5, 2e-5) - so there seems to be a "randomness" in choosing the position of the origin (see comment of Mario).



Personally it complicates my way of doing things, as I do not necesarily want the origin to remain on the original 3D one, nor in the "lower left corner" but often prefer a more symmetric position, this means a two step WP creation and some empty WP to deal with, the other way is clearly 3 vertices or as I use the "advanced" as I have seldom vertices at the right place.

A quick test with my geometry where I used 3 vertices to define the WP resulted in a geometry tilted by 90deg which is not helpful in my case as I need to maintain the orientation and alignment between multiple DXF files I imported. For the moment it is okay for me to define the WP by using 'Quick' and specifying a z-offset. For other geometries it seems that people have found different work arounds, but reconsidering the way of defining a WP seems advisable.



I expect this to be changed in V4, perhaps not immediatly, we will see but at least some improvement on WP and coordinate system identification in the graphics the "apply" is already an improvement shadow planes would be even better, no ?


I did not use 'Apply' before, but doing so now shows me where the origin will be, which does not help to set it to the desired value. So a possibility to define the origin in a second step seems reasonable. If I understood you right, you meant that after defining a WP as it is now, the new WP is shown (in 2D) and then a second window appears which lets you choose the origin by either entering the coordinates manually, choose the coordinates of the global origin, or let you select a point with the mouse.

I am not familiar with 'shadow planes'.

Thank you very much for your responses.
Best regards,
Kurt
[QUOTE] I suspect it has to do with the treatment of the signs in 3D to stay in the first "positive" quadrant [/QUOTE] Well, sort of. I just tried again and upon creating the WP using Face Parallel, the geometry is not entirely in the first quadrant but the lower left corner is at (-1.3e-5, 2e-5) - so there seems to be a "randomness" in choosing the position of the origin (see comment of Mario). [QUOTE] Personally it complicates my way of doing things, as I do not necesarily want the origin to remain on the original 3D one, nor in the "lower left corner" but often prefer a more symmetric position, this means a two step WP creation and some empty WP to deal with, the other way is clearly 3 vertices or as I use the "advanced" as I have seldom vertices at the right place. [/QUOTE] A quick test with my geometry where I used 3 vertices to define the WP resulted in a geometry tilted by 90deg which is not helpful in my case as I need to maintain the orientation and alignment between multiple DXF files I imported. For the moment it is okay for me to define the WP by using 'Quick' and specifying a z-offset. For other geometries it seems that people have found different work arounds, but reconsidering the way of defining a WP seems advisable. [QUOTE] I expect this to be changed in V4, perhaps not immediatly, we will see but at least some improvement on WP and coordinate system identification in the graphics the "apply" is already an improvement shadow planes would be even better, no ? [/QUOTE] I did not use 'Apply' before, but doing so now shows me where the origin will be, which does not help to set it to the desired value. So a possibility to define the origin in a second step seems reasonable. If I understood you right, you meant that after defining a WP as it is now, the new WP is shown (in 2D) and then a second window appears which lets you choose the origin by either entering the coordinates manually, choose the coordinates of the global origin, or let you select a point with the mouse. I am not familiar with 'shadow planes'. Thank you very much for your responses. Best regards, Kurt

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 21.12.2009, 12:28 GMT-5

It's something you'll have to live with, as far as I know. I think that if you wish to use face-parallel, the easiest way is to move your whole drawing into position if you coordinates shift.

This is exactly what I try to avoid, especially since there seems to be some "randomness" in assigning the new origin. The alignment between multiple imported DXF files is essential.
[QUOTE] It's something you'll have to live with, as far as I know. I think that if you wish to use face-parallel, the easiest way is to move your whole drawing into position if you coordinates shift. [/QUOTE] This is exactly what I try to avoid, especially since there seems to be some "randomness" in assigning the new origin. The alignment between multiple imported DXF files is essential.

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 22.12.2009, 07:50 GMT-5
Hi

Have you noticed that if you select "face parallel", select a face, with or without offset and up/down changes, then do an "apply", and thereafter (before OK) hit the "advaced tab":
you see then the natural system selected, try to correct it in here in the advanced tab and then "apply" for a check and finally "OK" you might get a better result closer to your expectations.

Furthermore, if you play with the up/down normal (+"apply"), you see how COMSOL changes the orientation of the WP, it's coherent in saying X-Y in the WP then looking down the "Z", but you can chage this in the advanced tab (once "applied")

You have also the coordinate frames

Hope this helps
Ivar
Hi Have you noticed that if you select "face parallel", select a face, with or without offset and up/down changes, then do an "apply", and thereafter (before OK) hit the "advaced tab": you see then the natural system selected, try to correct it in here in the advanced tab and then "apply" for a check and finally "OK" you might get a better result closer to your expectations. Furthermore, if you play with the up/down normal (+"apply"), you see how COMSOL changes the orientation of the WP, it's coherent in saying X-Y in the WP then looking down the "Z", but you can chage this in the advanced tab (once "applied") You have also the coordinate frames Hope this helps Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 22.12.2009, 16:41 GMT-5

Have you noticed that if you select "face parallel", select a face, with or without offset and up/down changes, then do an "apply", and thereafter (before OK) hit the "advaced tab":
you see then the natural system selected, try to correct it in here in the advanced tab and then "apply" for a check and finally "OK" you might get a better result closer to your expectations.

Actually I did not notice, but that's a good one - thanks a lot. With this little trick I can set (x,y) of the first point to (0, 0) and I get the WP with the same coordinates than the global origin (I did not extensively test that but at a first glance it appears to be the case).

Problem solved, thanks a lot.
Kurt

[QUOTE] Have you noticed that if you select "face parallel", select a face, with or without offset and up/down changes, then do an "apply", and thereafter (before OK) hit the "advaced tab": you see then the natural system selected, try to correct it in here in the advanced tab and then "apply" for a check and finally "OK" you might get a better result closer to your expectations. [/QUOTE] Actually I did not notice, but that's a good one - thanks a lot. With this little trick I can set (x,y) of the first point to (0, 0) and I get the WP with the same coordinates than the global origin (I did not extensively test that but at a first glance it appears to be the case). Problem solved, thanks a lot. Kurt

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.