FULLY DEFINED SKETCHES 1

[TateJ]

In my class... I try to drive home the point that all shetches should be fully defined. Typically, I deduct a few points from the homework if a student leaves undefined sketches in his model. I'm confident that this is the rule I want to keep for reasonably simple parts. But, I'm curious to hear the opinions of you guys that model a lot of swoopy stuff... aka surfaces. I know that it's practically impossible to fully define a sketch that contains a spline (but I won't bet my paycheck on that statement). And I suspect a lot of your surface built models might contain a few splines.

So tell me please, should I ease up on underdefined sketches?

Windows XP / Microsoft Wireless Optical Mouse 5000
SolidWorks 2007 SP1.0 / SpaceBall 5000
Lava Lamp
www.Tate3d.com

FAQ376​

 

[J-C]

We lock it down when the design is frozen....you don't want something moving unexpectantly later when the design is tweaked.

Jason

UG NX2.02.2 on Win2000 SP3
SolidWorks 2006 SP5.1 on WinXP SP2
SolidWorks 2007 SP2.1 on WinXP SP2

 

[CorBlimeyLimey]

It is usually safer to have the sketches fully defined, but an under-defined sketch is sometimes useful during a concept design stage.

 

[ctopher]

I always make everything defined. If not defined, there is a chance some geometry could move and can take a long time trying to figure out which one moved.

Chris
SolidWorks 06 5.1/PDMWorks 06
AutoCAD 06
ctopher's home (updated 02-10-07)

 

[TheTick]

If a user can't tell his sketch is properly constrained, then he shouldn't be modelling. As an instructor, I would hope you could divine whether a sketch was properly constrained, even if it wasn't technically fully constrained. Do you bother to check the sketches, or do you just automatically knock off the points?

Sometimes fully defined is more trouble than it is worth. Especially true when using construction geometry (other than "infinite length" lines). A construction arc or spline may be adequately defined but with "loose" endpoints. Constraining endpoints of construction geometry often just adds unneeded clutter.

I could be the world's greatest underachiever, if I could just learn to apply myself.

http://www.EsoxRepublic.com

-SolidWorks API VB programming help

 

[toycept]

I'll give another point of view.... I do a lot of concept stuff with mechanisms....which I eventually build either in the shop or on a RP machine I have. With what I do ...knocking something quick out is more important than any final engineering work. Being that I rarely get involved in final engineering and I'm more interested in putting together a quck concept prototype, sometimes I don't bother with fully defining sketches, unless I get the feeling it will be of benefit further down the line. If I'm doing more of a form study and working out the styling, I will endevour to fully define. I have to admit, since using 2007 and the "fully define sketch" feature was introduced, I have made more of an effort to fully define, being that the fully define feature has made it fairly easy.

 

[TateJ]

As an instructor, I would hope you could divine whether a sketch was properly constrained, even if it wasn't technically fully constrained. Do you bother to check the sketches, or do you just automatically knock off the points?

I expand the feature tree & look fot + or -... then I investigate. If it's just the end point of a center line that's causing the sketch to be undefine... that's OK.

I don't grade very harshly either. The students that fail are the ones that take ZEROs on assignments & never do them. You'd have to make an effort to fail my class - if you know what I mean.

Windows XP / Microsoft Wireless Optical Mouse 5000
SolidWorks 2007 SP1.0 / SpaceBall 5000
Lava Lamp
www.Tate3d.com

FAQ376​

 

[TheTick]

I hope I didn't sound harsh. I didn't mean to.

 

[TateJ]

...not harsh at all.

BTW... that flaming bag on your front porch might have poo in it...

Windows XP / Microsoft Wireless Optical Mouse 5000
SolidWorks 2007 SP1.0 / SpaceBall 5000
Lava Lamp
www.Tate3d.com

FAQ376​

 

[fcsuper]

Well, two benefits from fully defined sketches. One: more stable model. Two: Model rebuild times are supposedly reduced everything is constrained because SolidWorks is doing less work figuring out what/where everything is during the rebuild. So, technically, SolidWorks runs faster with everything being fully defined. Has anyone ever tested this?

Matt
CAD Engineer/ECN Analyst
Silicon Valley, CA

http://sw.fcsuper.com/index.php

 

[Theophilus]

Fully-defined stuff can make sense once you freeze your design. In fact, when I release a design to manufacturing, I will also disassociate all parts such that I have no dangling references to the assemblies. (I also keep a non-disassociated directory intact for possible future changes--makes things handy and hard drive space is almost "free" anyway.)

However, while designing, I use a different rule. I define only as much as is needed for its given purpose. Your example of using swoopy stuff is a good one. I have little need to define all the points on a spline, so I don't. In fact, I normally don't even define these things when I "freeze" a design. No point in making things difficult to edit, and spline points don't magically move themselves when I'm not looking. In fact, the same goes for lots of geometry--if it's not anchored to something else outside the sketch (like an assembly reference), it won't move itself spontaneously either. If I move something upstream in the tree, that could affect my undefined stuff, but then it's other geometry or sketches that change, not my undefined stuff (quite easy to fix the rebuild errors if I do this).

I also leave alone specific geometry-driven things that I want to refresh with an assembly's changes.

Since I'm an ID guy, things have to be flexible enough to work with. Normally, I start a junk assembly with all sorts of criss-crossed relationships among parts. I establish working movement and other relationships in that assembly. When things take on sufficient form, I begin severing those relationships and build up a clean, new assembly. Tweak the parts within the new assembly and I'm finished with the design. Clean method, saving myself (and my clients) time, which for me is expensive.



Jeff Mowry

http://www.industrialdesignhaus.com

Reason trumps all. And awe transcends reason.

 

[Eltron]

Sometimes fully defining sketches just isn't necessary. For example, if somebody comes up to me with a doodle on the back of a bar napkin and asks me to make it into a pretty picture for a report, then there is no need to define all of my sketches. Some of the sketched within the model must be defined in order to accurately place a feature, but fully defining items like sketch patterns, splines, ect. will add more time than I have to give sometimes.

Having said that, when I am actually designing a prodict that will eventually need to be constructed or tested virtually, I will always fully define my sketches. I seem to have less problems with the models, and I always gain a little more insight into my design when I actually have to put some thought into how each sketch relates to everything else in the design.

That's my story, and I'm stickin' to it.

Dan

http://www.eltronresearch.com

 

[markruble]

With the new feature in 2007 defining a sketch is now very simple. I've used this feature several times when trying to figure out why a sketch isn't defined. It's now a simple mouse click.

 

[handleman]

Although it may not be imporant in practice for some instances, it is important that a student understand what it takes to make a sketch fully defined. The only way to show that understanding (in my opinion - worth what it costs) is to show a fully, properly defined sketch - properly meaning that the design intent is followed. Once you really understand how to properly define sketches you can decide when to relax and leave a sketch underdefined, especially when swoopiness comes into play. BTW, SolidWorks doesn't think that free endpoints of construction geometry are underdefined, even though they can be grabbed and moved.