USER FORUM
(you are viewing a thread; or go back to list of threads)
Defining Workplace with New Normal (by Matthew Lugg)
Hi,
I've been learning SolveSpace through experimentation for a while (I got a 3D printer and have been occasionally modelling bits), and so just wanted to confirm that I'm doing something the "correct" way.
Say I want to make some geometry at a weird angle to the standard reference planes. (Most recently, I wanted something which was at 45 degrees to the horizontal and then angled 45 down.) The thing that makes this difficult is that I have no pre-existing normal in the direction I need. Here's the strategy I've figured out for doing that.
* Sketch in some 3D group (I guess it's best to define a new group for this?)
* Define a workplane in that group about some point that should lie in it
* Constrain the normal vector, probably by defining its angle to two other lines or similar
* At this point, the only remaining DOF rotates the workplane about its normal. Activate this workplane, and define its orientation by marking a line horizontal or vertical. (This might be a line from an earlier group, or maybe a new construction line if necessary.)
* Sketch in New Workplane off of the workplane I just defined, and work from there.
This works okay, and I have no real problems with it,.but I just wanted to confirm that it's the "best" way to achieve what I want here. In particular, the trick of marking a line as horizontal/vertical within the created workplane to constrain it's orientation took me a hot minute to figure out, so I wanted to check that it is what I should be doing here.
Small additional questions which feel silly to make a separate thread for:
* I'm currently running stable 3.1. Are there any changes on master which it might be worth updating for?
* In particular, are there any improvements to NURBS solving? False positive naked edge issues happen in almost every model I make, and I always end up having to force triangle mesh. I know I can generally work around it by adding a tiny amount of overlap, but that tends to be quite inconvenient to do once I've already finished my 2D sketch, especially since I can't do a two-sided extrusion biased towards one direction.
Thanks!
I've been learning SolveSpace through experimentation for a while (I got a 3D printer and have been occasionally modelling bits), and so just wanted to confirm that I'm doing something the "correct" way.
Say I want to make some geometry at a weird angle to the standard reference planes. (Most recently, I wanted something which was at 45 degrees to the horizontal and then angled 45 down.) The thing that makes this difficult is that I have no pre-existing normal in the direction I need. Here's the strategy I've figured out for doing that.
* Sketch in some 3D group (I guess it's best to define a new group for this?)
* Define a workplane in that group about some point that should lie in it
* Constrain the normal vector, probably by defining its angle to two other lines or similar
* At this point, the only remaining DOF rotates the workplane about its normal. Activate this workplane, and define its orientation by marking a line horizontal or vertical. (This might be a line from an earlier group, or maybe a new construction line if necessary.)
* Sketch in New Workplane off of the workplane I just defined, and work from there.
This works okay, and I have no real problems with it,.but I just wanted to confirm that it's the "best" way to achieve what I want here. In particular, the trick of marking a line as horizontal/vertical within the created workplane to constrain it's orientation took me a hot minute to figure out, so I wanted to check that it is what I should be doing here.
Small additional questions which feel silly to make a separate thread for:
* I'm currently running stable 3.1. Are there any changes on master which it might be worth updating for?
* In particular, are there any improvements to NURBS solving? False positive naked edge issues happen in almost every model I make, and I always end up having to force triangle mesh. I know I can generally work around it by adding a tiny amount of overlap, but that tends to be quite inconvenient to do once I've already finished my 2D sketch, especially since I can't do a two-sided extrusion biased towards one direction.
Thanks!
(no subject) (by Andrew)
You can create on new sketch on a workplane defined by a point and two lines that are on the desired plane. Select them and start a new sketch. If the points and line already exists, such as slanted face created by extruding a shape, use those. Further, you can add construction lines to any group, and you can switch in and out of sketch on plane in a sketch group.
(no subject) (by Daniel Engineering Solutions)
To answer your first question, take a look at the example I've attached. I've found that sketching and then extruding planes give the most "traditional" feel to solvespace. It also allows the sketch plane to be hidden when required. I think your method is perfectly acceptable as well though.
(no subject) (by ruevs)
There are a fair number of changes on master compared to 3.1. Some of them are rather interesting. The changelog https://github.com/solvespace/solvespace/blob/master/CHANGELOG.md has most of them listed.
No NURBS fixes thoug - I've had very little time and they are hard.
The edge build is up to date so if you are on Windows or macOS you can just download it an use it. https://github.com/solvespace/solvespace#via-automated-edge-builds
No NURBS fixes thoug - I've had very little time and they are hard.
The edge build is up to date so if you are on Windows or macOS you can just download it an use it. https://github.com/solvespace/solvespace#via-automated-edge-builds
Defining Workplace with New Normal (by Jaroslav Kopal Czech republic)
My example embedded here, you can beautifully use for experiments. I also printed it on a 3D printer! For clarity, the individual parts are colored.
Post a reply to this comment: