I just thought I would share a couple of things picked up in the last
year or so that overcame some aggravations with the feature tree in
1) When I would try to reorder a feature to just under an expanded
feature (an extrude, sweep, loft, or other sketch/curve based feature
that has been expanded to show the absorbed curves) I would get one of
those 'no smoking' signs indicating I couldn't drop it there.
I used to curse, abort the drag, collapse the expanded feature (so the
absorbed sketches/curves were no longer visible), so I could then drag
and drop the feature I wanted to reorder.
Well, I was a little chagrined when I accidentally learned that I was
dropping at the wrong point. If I would try to drop below the absorbed
sketches, I would get the no-smoking sign. If I went further and
dropped the moving feature directly on the expanded FEATURE (not below
its visible sketches/curves), the feature I was moving would reorder
2) When in rollback, you can see all your features under the rollback
bar, some of them with absorbed sketches/curves that you ought to be
able to access. Those sketches/curves would be above the rollback bar
if the feature tree were 'flat', allowing relations, sharing, or
other uses if the feature tree were flat. But unfortunately they were
sucked below the rollback bar by the feature that used them.
The presence of the plus next to the rolled back features was a tease -
clicking the '+' once to expand like in the regular tree doesn't
show the absorbed item. So I would roll down, expand to show the
absorbed sketches/curves, then roll back up to where I wanted to work
so I could make the reference.
I recently learned that you CAN expand features that are in rollback by
DOUBLE clicking the + next to the feature. Depending on history, you
can now easily access the absorbed items for references/sharing
3) Similar to tip 2, when going into parent-child on a rolled back
part, we only get to see the kids of the rolled back features, not the
grandkids or the great grandkids below the rollback even though there
are plusses next to the kids in the box. Well, yup, if you double
click the plus next to a kid you can see what the grandkids, etc are.
Downside - it doesn't show the same as if the part were not rolled back
4) Lofts and sweeps absorb multiple sketches/curves. If you expand the
feature and rollback somewhere in the midst of the absorbed
sketches/curves you can get a "flat" representation of where those
sketches/curves are in history.
Unfortunately, you often can't drag an unused sketch or curve from
higher in the tree into the middle of this flat history of previously
absorbed curves to make references - you get another of those
However, you CAN add a new feature when the rollback bar is in the
middle of this flat representation (I use planes). You can then drag
anything down to below the plane, reordering where you were not allowed
to reorder just a second ago. When done, I delete the workaround
5) Folders often mess with flat history of the tree mentioned in 4.
Just suck it up and delete that folder, and you can get a true flat
history. When done, remake the folder (Ctrl+C the name before
deleting, then Ctrl+V into the new folder name to save typing)
I'm sure a lot of the above is widely known, but I figured I would
post it just in case it helps someone
Speaking of saving typing - I thought I would save Wayne a little
"That's just too funny - WHERE HAVE YOU BEEN??? :-))) I don't think
ever admit to something quite that bad - too funny! (I know, paybacks
hell.......) Hey, at least you know now, though. :-))"
(just kidding, Wayne)
I'm working on a project that entails a lot of assembly-level features.
I have had trouble with features that can not be reordered, in spite
of no problems with parent-child relationships. Some of these are
features have no reason to be difficult, i.e. datum axis defined by
origin and "Front" plane.
Back in Orlando I was lobbying for a new feature tree that actually
showed the true relationships of features by branching out
horizontally as well as vertically. Several comments on the NG have
seemed to point in that direction recently. The parent/child explorer
gives a little better view of the interrealtedness of features and
there is some functionality there. The current feature tree really
doesn't capture the true connections between features nor does it point
out which features are required parents and which are not.
Ahh, I love it when light bulbs come on, whether it be my room or someone
else's. I knew about #1, but not the others. I just tried #2 and although
I can double-click to make the sketch visible in the tree, I still could not
convert that sketch into a new sketch that I was trying while rolled back.
If I roll below that sketch, then I can select it to convert it, but the
actual sketch circle will not show up for me in the graphics area if it's
below the rollback bar. I'll have to watch for that one.
Oh, and no, I don't think those were as bad as the Muggs light bulb... :)))
(Thanks for the laugh this morning.)
In regards to the assembly feature tree and not being able to re-order
things you should be able to, I discovered a couple days ago that you
can get those arnry files that don't want to re-order or tell you they
aren't contiguous, to go wherever you want by using the rollback
state... say you've created 2 planes and were able to get them
re-ordered above any parts/sub-assemblies you may have in your
assembly, and you want to go place an axis in relationship to those 2
planes... that axis then shows up below your "mates" folder and refuses
to move above that folder giving you the not smoking sign... what you
can do is, select one of the planes you created and right click it in
the tree and select rollback (also can be done by dragging the rollback
bar)... now re-create the stubborn axis and it will stay above any
parts/assemblies you may have in your assembly, unless you move it...
hope that makes sense, if not, i will try to explain again...
As long as we're on the topic, ever notice how you can get the "no
smoking" sign when you should be able to reoder a feature? And then it
reorders the feature anyway. Or you want to put a number of features
in a folder, and you get an error message stating that one or more
features cannot be put in the folder. Click "Ok" and all the features
are moved anyway? Nice interface. Or not being able to move a feature
into a folder for no apparent reason. Have to delete the folder and
recreate. Usually happens with patterned features.
Another gripe is not being able to move features en masse. You have to
move each one idividually. What a pain in the a**.
Ok, I'm done whining now. Been a bad Friday. SW has crashed
I think I get it. This amounts to recreating the axis prior to adding
parts or mates. Seems to me that an axis can be dragged to the top if
there is nothing depending on it. The most important reason to have
reference geometry, layout sketches and envelope parts come first is
that anything after the mates folder are resolved after the mates are
resolved which can lead to the "double rebuild".
Another trick along this line that it is absolutely essential is to
pull the rollback bar up to the mates when mating in an assembly with
patterned components. If you don't and accidentally mate to a component
in a pattern don't bother calling tech support when your assembly
starts behaving in a bizarre manner.
P.S. arnry should be ornery
I have attempted in SWks 2006 to delete 1-2 dozen parts out of the
Feature Tree of an assembly (after SAVE AS w/new name) so I could send
a simplified smaller assembly to someone.
Deleting multiple items one after the other resulted in a consistent
crash. Had to save after each delete. One more glitch in the Feature
Wayne, I must not have made myself clear.
You can access sketches if they *should* be available if they were not
absorbed. It does not apply to any sketch whose proper place in
history is below the rollback bar.
The way to test - expand the feature (since its you, most likely an
extrude :)) then rollback to between the sketch and the feature. This
will show that sketches proper place in history in the tree.
You can use #2 to access the sketch as long as the rollback bar is
below that sketches proper place in the tree (from the above test). If
your workflow is 1)make sketch 20make feature, you have no need for any
tricks. If your workflow is 1)Make layout sketch 2)Make all sorts of
features 3)use layout sketch later to make a feature, then it will
Besides, don't all of your models have only 2-4 features (make plate,
drill some holes, go get some barbecue for lunch?)
thanks for the spell check TOP :) they need one those thingies on
here... whats weird about those ornery axis and planes, that refuse to
move up the tree, is that their parents are above the
parts/sub-assemblies in the assembly, so it should move no questions
Aargh about sums it up. I'm glad I almost never need assembly features.
other odd things -
layout sketches in the assembly appear below the mate groups, yet I can
mate to lines in the layout sketch (which is good to me). Your post got
me thinking, and yup, even if the assembly is rolled back before the
sketch (so I can't access the sketch) the mate is hunky dory - I just
can't edit it. Not a big deal - just an odd curiosity. Why rollback
One sore point for me is that when doing a parent/child on a sketch at
the assembly level, it always reports no kids. Isn't a mate to that
sketch a kid of that sketch feature? Aren't any in-context references
by components children of that sketch feature? I hate it when I see a
lot of sketches in an assembly because I have no way (that i know of)
to tell if they are critical layouts or if they are throwaway sketches
used to interrogate the asm or jot down an idea.
"TOP" wrote If you don't and accidentally mate to a
So are you saying that if, for example, you have a stack of shims which you
have created witha component pattern, that you can't (reliably) make a
coincident mate to the end shims?
The seed shim, yes - ought to be bombproof. If not, yell a lot.
Mating to the last shim is risky - what would you expect to happen if
you decide that instead of three shims, you need four? Should the
mate be to the last one, or still be to the third one?
In those cases, where things are still in flux, rebuilding is more
predictable (in my experience) when you mate to a sketch describing the
boundary of the shim stack, and use an equation to divide the length of
the sketch boundary by the number of shims to drive the pattern