Discussing Master sketch approaches

[user1234]

That's my point.
For testing, I want to recreate that kind of assembly in FreeCAD and having a setting for to change thickness helps a lot visually (as stated by the original poster of this thread) specially when there are tons of lines on screen.

Just for completeness, your screenshot would be way too complicated for a skeleton model, even for CATIA (or other commercial CADs). A skeleton model also should be simple as possible, as described a plain skeleton (like the pdf you linked). Not that it does not work, but the change to interlace and get stuck is higher.

Greetings
user1234

[Lonfor]

Just for completeness, your screenshot would be way too complicated for a skeleton model, even for CATIA (or other commercial CADs). A skeleton model also should be simple as possible, as described a plain skeleton (like the pdf you linked). Not that it does not work, but the change to interlace and get stuck is higher.

Then you saw nothing yet, that skeleton is both, a sub-assembly and top-level assembly from which a few more skeleton sub-assemblies are derived.
These stacked sketches are necessary to avoid circular references altogether, and at the same time, the need of avoiding assembly constrains as much as possible in order to eliminate once and for all, any possible point of failure when changing parameters. If assembly constrains cannot be avoided, they must refer to datum planes or axis first, linked parameters second, never ever to geometry.
Until now this system has worked like a charm.
See pic 1 and 2, these assemblies rely only on the master sketch plus "mirror" and "circular pattern" operations to position all the parts. Zero (0) assembly constrains have been used.
See pic 3, 4 and 5 are the same case as above. Also notice how the assembly changes shape when a parameter propagates across the master sketch.

Sorry for the off-topic and kinda hijacking the thread.

[chrisb]

I have split this topic from https://forum.freecadweb.org/viewtopic. ... 19#p591019, where it went completely off topic.

[chrisb]

Then you saw nothing yet,

That's very obviously not true if you look at user1234's other posts.

that skeleton is both, a sub-assembly and top-level assembly from which a few more skeleton sub-assemblies are derived.

In a company there is often a certain culture of modeling; and of naming things. Of course it is possible to describe in a master sketch many details of the final model, and of course you can name this a skeleton or a master sketch, because you derive things from it.

However, that is not really sensible in FreeCAD, nor does it seem so in general as the pdf linked by yourself shows. I will not discuss here the general case, but in FreeCAD it is extremely sensible to derive things from simple sketches, which are very unlikely to be subject of serious changes, especially those adding or deleting elements. This is due to the topological naming problem

[adrianinsaval]

I've learned not to use Sketch for Mastersketches

how so? What do you use then?

[Palmstroem]

I just was told, that sketches are not made for that. I was not told, what to use instead. Actually trying to get into the Draft-workbench which crashed 5 times since and doesn't seem to me to be useful for mastersketches. Drawing is outdated and TechDraw is also not meant for master sketches. So at the moment I'm confused.

[adrianinsaval]

You may have been told that sketches are not meant to be overly complex like that example but AFAIK sketches are the best method for master sketches, I'm not sure what the actual intended purpose of Draft is but I don't think is meant for parametric mechanical design, I think the main goal is the architecture stuff. Techdraw and Drawing workbench definitively aren't for master sketches.

[Lonfor]

I just was told, that sketches are not made for that. I was not told, what to use instead. Actually trying to get into the Draft-workbench which crashed 5 times since and doesn't seem to me to be useful for mastersketches. Drawing is outdated and TechDraw is also not meant for master sketches. So at the moment I'm confused.

Hang on, I'm doing inroads to reproduce the NX skeleton method in FreeCAD. Give me a couple of days.

[FBXL5]

I'm quite convinced that you can build a skeleton of master sketch and dependent sketches, but haven't tried myself yet.

For clarity you can colour the sketch in the 3D view. A master sketch may display more geometry than needed/allowed to create 3D objects and then you can use further sketches to derive (filter) the contour geometry in one and auxiliary geometry in another.

You are right if you find it complicated to load external geometry and then rebuild it manually to get a usable contour. It would be helpful if we could use the external geometry directly using a toggle like that for construction geometry.

[user1234]

Sorry for the off-topic and kinda hijacking the thread.

That happens, and i think i was the first hijacker.

Extreme complex or interlace skeleton can work, but only on a (few) circumstances. For example when only one person work on a complex skeleton, then of course, then for this one person it can work. But i you have exchange people, then it will be hard to figure out, what is linked in the skeleton and what not (even something is sketched in the skeleton, it must not be linked or can generally deviate). I know that in some companies, this way of work is forbidden.

Also it can work (with more people), when it is always 100% the same part/s / assembly/ies with 100% same topology and only different dimensions. But this is very seldom and maybe build up in an other way. Or when all parts always extreme homogeneous and a quasi internal norms.

Also it can work, when the build up is pretty obvious. For example your laminated springs, i would absolute agree, that is a good example to do this in a skeleton way. But also the pretty opposite is your example the frame. It makes changing things unnecessary hard and when elements linked to other attached parts, which maybe other people work on, the result can sometimes intersect. Also when skeleton attached to other parts, for example on sections, it could inveigle to people to think, it changes all correct automatic with it. Maybe it is true for the section, but maybe not to other things like: wall thickness, mounting access/possibility, fabrication/machining access/possibility, ...... . So when you know, it does not change automatic, at least you are nudged, to look at the other parts exactly. And when change elements in a complex skeleton way and it have impact on this points and you have to change things on it, it can happen, that you have to change also the skeleton itself (topological), which can have impact on also other parts which is also linked to the skeleton, which maybe do itself not change per se, but you have to relink/correct some elements.

So often, not always, a simple skeleton for assembly, movement, or (simple) section definition is more preferred, as a complex skeleton model, at least in the mechanical engineering. In the vehicle manufacturing (which i worked many years ago), this is maybe a little bit different, but at that time i worked on it, it was strictly forbidden (reason for dismissal).

Greetings
user1234