Loft With Guide Curves - Can More Guides Be Added "Later"?

Yea, you can. You should not have to delete the loft or start over just to add or delete a gc of profile. You should be able to add/delete a GC in the list at anytime. If there is a problem, it's usually because the GC is not intersection the profile(s) (although I've seen a loft work even though the GC's are not intersecting exactly).

Or you can also add more loft profiles by placing your cursor above or below in the list or moving the arrows on the side of the list bumping up or down the order of the selected profile.

Generally, adding or deleting GC's is not a problem. And you can reorder them as well but I do not believe it matters or has it ever mattered what order they have?

BUT, the loft interface is not without problems... I have often seen a problem where adding profile will mess up the order of the loft so bad that you have to start over or remove some of the profiles and change the start/end condition and exit, they reinstate the profiles. Also, sometimes when removing profiles or if a profiles start/end direction (edge of a surface or curve is reversed) the loft interface gets confused and will not allow for any other profiles accept the ones uses earlier. In that case, if you run into it, you may have to delete the loft and start over? (you'll know what this happens, you can not select anything or it rejects all selections,... one of those consistency things or the loft feature becomes corrupt?) Yep, a very strange interface sometimes and I believe it is due to all the different algorithms which the loft interface has used over the years. (redefining lofts between releases and sp's is not always pretty.)

Hmm,.. you're right, I can not find anything in the help with specifies this? But you should be able to add/delete at anytime.

SNIP

Paul / Sean

It does matter

Ed gives an surface loft example in Curvy Stuff 201 , slide 36, (see thread above) showing the difference reordering guide curves can make. The first guide curve, as I understand it, drives the "flow" direction of the isoparm lines. If the guide curves are radically different in character, this can make quite a difference to the outcome.

Wouldn't it be nice if SolidWorks would tell us this stuff in the documentation, rather than us having to rely on the brainpower of guys like you and Ed to reverse engineer this information, not to mention the altruism to then take the trouble and time to share it around? Anyone who's "been there" knows how much longer it takes to pass knowledge on than it does to acquire and apply it.

I doubt that SldWks were unaware of the special significance of the #1 guide curve. I shudder at the waste of effort that is inevitable from their decision that we, the users, do not "have to worry our pretty little heads" about stuff like this. I'm not just thinking of the effort of Ed to infer and document this, but the waste of effort for guys like us, (well, me anyway) who get perplexed because of the -at best- inscrutable, and at worst capricious behaviour software exhibits when it is obeying rules which are kept secret from us.

Interesting question: does this pearl of wisdom appear in the new 1000 page pdf?

Andrew,

For a surface loft by itself, gc order should not change the way the surface looks? (Otherwise, I'm not sure what gc order or issues Ed is referring to in the 201 and 102 ppt?)

But, for downstream referencing, order can be an issue... for instance,..

(this is a touchy subject for me, see note comments)

..

Andrew Troup wrote:

Ah, doing a few more test... I think I understand the picking order Ed is talking about... it's related to cases where the initial "profile" picking order or how/where you pick the profile makes the gc order dependent?

But, yeah, I've seen this dependency, and imho, another screw up by SW Corp, imho.

So, if/when you do see or have a problem with the loft failing, reordering of the gc's such that they are in the proper "flow" sequence (like the profiles are) is the remedy.

..

Paul Salvador wrote:

Also, to solve the problem, you can "re pick" (clear the list, and pick from the opposite side, i.e., the end profile first.) your initial "profiles" so the gc's are not dependent.

Basically, when selecting profiles for your loft, you pick on a end or location where the initial (or imaginary if a gc is not used) gc would pass/flow through the sections. If they are open profiles (surfaces), then you pick near the open ends where the gc would pass/flow.

And similarly, when selecting gc's, you should pick them in order (but that is because of old habits, which seem to help due to limitations with SW).

..

Paul Salvador wrote:

Paul - Thanks for your advice - I knew that somehow it shoudl be able to be done.

Here was my original catch 22 - I make two loft sketches and a guide and form my loft surface. Now I need to add another guide, so I add another sketch to touch the loft in the right points . . .etc and shoot . . . I need to edit the feature, which makes the new guide curve suppressed as it comes later in the tree, so maybe I need to add something before, but . . . nope . . . it has to relate the loft sketches, so it can't come before . . . "so he asks"

After reading your message I knew I was not fully insane and went back with victory in site. What I found is that if I rolled back to the last guide curve, in the feature, the feature would temporarily "unabsorb" and let me add another guide (which I had to add manually to the loft later, but it worked).

It told me: "You have chosen to rollback to Sketch3 which is absorbed in feature Surface-Loft2. The following features will be temporarily unabsorbed for editing purposes: Sketch1 Sketch2 Sketch3"

This unabsorbtion was something I had not seen too much, but now it's crystal clear once I went thru it.

Thanks again-

SMA

Ed's example definitely shows a surface loft, with multiple profiles. but only 2 guide curves. This loft yields quite a different shape, using the exact same sketches, when you reorder (not repick) the 2 guide curves, using the "Up/Down" arrows provided alongside the guide curve list for that very purpose. There is no sample file provided (Ed's speaker's notes say that it is included in Curvy Stuff 101 - 2 - Four types of loft, but it's not in the download I made last year) I have seen this on my own models, and I'll investigate further when I get a moment.

Andrew,

Well, I had to download Ed's file to understand what you are talking about,.. the "slide 25" file. There are actually 3 examples using similar profiles and gc's.

The file first shows the issue of the initial profile chosen and how it makes the gc order dependent. (in this case the 3 gc's reside between the ends of two open profiles) He then shows the same profiles chosen from the other end of the profiles and the gc's are now not dependent on order.

Only these two cases use the same profiles and gc's "but" the profiles start/initiate in different order/direction, which changes the way the face looks. And, for some reason the "Start tangency, All faces of

1.00" is not checked?" Anyhow, they are not exactly the same, but very similar with face tangency applied.

He also shows that adding a gc at the end(s) of the two profiles corrects the gc order. (the extra end gc makes it a different loft)

So... the examples are not the same, and it does not show how gc order changes the look of the face.

see image...

This whole exercise (although appreciated) reinforces the problems SW Corp has with lofts.

.. (so much fun)

Andrew Troup wrote:

Paul

On further investigation, I found (in the same presentation- (Curvy Stuff

101-2-Four kinds of Loft) a downloadable model with a surface loft in it which is near enough to the one used for slide 26

Slide22_B - C2 tangent guides better.SLDPRT is the model

Surface-Loft4 is the feature

If you reorder the guide curves and change nothing else, the look of the face changes in a similar way to that shown on slide 26

Andrew,

Well then, sadly, the two files of the "seat back" are showing problems with lofts in a different way. In this case, the gc order is changing the way the loft solves and the surface is not passing through the sections properly.

Files: slide22_A - C1 tangent guides freaks the model out slide22_B - C2 tangent guides better

If you open up the file in SW2004 sp2.1 and ctrl-q the loft changes. If you delete the loft and remodel it, the loft is different looking.

BTW, the illustration on the PPT page 26 has 4 more profiles to correct some of the problems.

I feel like I'm somehow defending gc order? But, I'm not surprised, I just have not had many if any problems with gc order. Maybe it is the way I model or choose my curves... and I thought I experimented enough with profiles and gc's??? (this is a bad dream)

What this all reaffirms and is telling me are lofts in SW are totally inconsistent. And that does not help you or me in explaining or solving some of these issues.

SW Corp needs to stop wasting our time and solve these inconsistencies.

..

Andrew Troup wrote:

The guide curve order thing is subtle and easy to miss - the seat back just shows a really, really dramatic example anyone can understand. I first ran into the problem when a part didn't match up to what it was supposed to be, because of a change of a few thousandths due to guide curve reordering. I have no idea how many other parts I have done where I didn't notice, or where the subtle change didn't matter at all. I have been trained by my experiences to never delete a loft and remake it (there are always tricks that can be ustilised to save it so the legay algorithm is used), and to number my guides so I can always tell which is the first one. It prevents many issues that pop up with service packs. As a matter of fact, I think SP2.1 is the first that I remember where lofts that had not been redefined actually changed substantially (instead of the usual rebuild error where it palin couldn't complete). Of course my experience is clouded by usually not changing service packs in the middle of a job.