Kent,
> Hope you understand that it was a honest question, and not trying to say
it is done
> wrong currently.
Of course, I understood that. Sorry if my reply sounded argumentative
(although I didn't mean to!). IMO, written communication can get one wrong
more than oral communication. I've seen Sean D. experience it in this NG
;-) I even had the same experience with my wife when we exchanged our
e-mails. Since then, I don't use e-mails to communicate with my wife any
more 🙂 Kent, I truly respect you as a gentleman as I've seen you in both
Inventor and MDT NG's.
> For me anything that makes the message show up less often is the best. I
don't think
> using the part origins for the sketch origins would be a issue, but I can
see where the
> axis would be for many (probably not for me generally because 95% of my
stuff is square to
> the origins.)
>
> Thinking back at the times I have seen the invalid axis/origin message, it
seems like it
> is generally the origin it looses. Seldom is the edge that the axis is
based on removed or
> lost.
I agree with you. The sketch origin is more vulnerable than the sketch
axis. Right now, I'm looking at a bug that our QA reported. In that model,
a downstream sketch orgin is a vertex which we track as an intersection of
three adjacent faces (say F1, F2, and F3). When an upstream sketch
dimension is changed, the topology gets changed and the downstream sketch
loses its origin because we couldn't track the vertex, which now intersects
F1, F2, and F4. It's a legitimate failure, so I will return this bug to QA
as designed.
> Not sure if you are aware of the problems I have had in migrating to 6,
but one of them
> is/was that I would consistently loose the axis/origin on a part when I
did a update all
> of the top assembly. This is one of the few part sketches I have that has
no square
> corners and so the UCS attached to the edge of a fillet. I would think in
that case
> attaching to the origin might have been more stable?
The migration failure could be from some other cause. If there are no model
changes (either topology or geometry), then rebuild all should be
successful. Otherwise, it's a bug. I've seen your model once (it was an R6
migration bug), but the problem seemed to be something to do with project
file if I remember correctly. I'm not sure if we're talking about the same
model.
> At any rate, thanks for the response, and keep up the good work.
Thank you always.
Glenn
Inventor Part Modeling