ATTENTION Inventor Developers! This part has repeatable crash

[Anonymous]

Open this part, unsuppress fillet12 (just above EOP). This action causes features 25 steps previous to fail. How does a fillet reach back in time to corrupt other features?

I couldn't find the email address for anyone to just send this to. If you are a developer who deals with this kind of problem, please post your email or an email I can just send parts to rather than posting. This time I will write the email down on my monitor with sharpie.

Okay- I cannot post the part here!

"The content type of the file 'Corrupt Part.ipt' is not allowed"

So if someone wants to reply with their email contact, I will be happy to send the corrupt part.
Thanks
Phil

[JDMather]

>The content type of the file 'Corrupt Part.ipt' is not allowed

In Windows Explorer right click on the file name and select Send to Compressed (zipped) folder. Attach the *.zip file here.

[Anonymous]

For some reason I didn't see that instruction when attaching..... Here you go.

Only now you have to move EOP to the bottom on your own.

[Anonymous]

I did not get a crash, I got the attached error message, I'd
start at the top of the list and fix one at a time until the error goes
away.

You have multiple adaptive features in this part.
Why?

This part could have been built with far less
features.

Many of your sketches starting with sketch 2 are
unconstrained. Why? Unconstrained sketches can cause all sorts of
problems.

Software developers cannot help a poor design workflow. Need
to work on that area. What is the Design Intent in this model?

--
Dennis Jeffrey, Autodesk Inventor Certified
Expert
Autodesk Manufacturing Implementation Certified
Expert.
Instructor/Author/Sr. App Engr.
AIP 2008 SP2, AIP 2009-SP1
PcCillin AV
HP zv5000  AMD64 2GB - Geforce Go 440, Driver: .8185
XP
Pro SP2, Windows XP Silver Theme

href="http://teknigroup.com">http://teknigroup.com

[Anonymous]

Thank you for your prompt reply. I will try to answer your questions. I also have a few more for you. If you don't consider the attached image to be a "crash", what is it? When updating a part, and having to remake certain features, a fillet with new inputs causes features 25 steps previous to fail, what do you call that? Even if I suck at modeling, as you have stated, I have a thousand parts on my machine that didn't fail in this fashion. This is new behaviour for IV. Are you telling me this is a predictable outcome and only my awful skills at modeling have caused this? If this is a bonafide function of IV, I'd like to know who thought it would be a good idea to do this to people every now and then, and why. Yes I have adaptive features. Isn't that what IV is supposed to do when you model in assembly? Isn't that the default mode for IV? Relating one part in the assembly with the others is what adaptivity is, or no? Thanks for the review of my model. Please forward the slim and trim version to me ASAP so I can see how a real expert does it. Unconstrained sketches... how do they relate to a fillet? Do they reach forward in time to find out if they are problems 25 steps later? Thousands of parts with unconstrained sketches have never failed this way, nor failed to be made into fine consumer products either, BTW. Sorry to bother you. I guess I should go back to school and get a better work flow. Do you have any suggestions of where I can get instruction in this matter? Edited by: Discussion_Admin on Oct 29, 2008 10:13 AM

[Anonymous]

Did you mean "far fewer" features?

The workaround went just fine. Part is now finished without rebuilding or constraining my sketches or good work flow. Some people just do it different I guess.

[Anonymous]

Hi Phil,

When I unsuppressed the last fillet, I got a string
of errors, but no crash. I noticed that from the iProperties that you have SP1
installed. I haven't loaded the service pack on this machine, so it may be an
issue with the service pack.

I will get this logged into our system so
development can take a look at it.

Loren Jahraus

Autodesk Inventor Product
Design

[Anonymous]

No offense given or taken. I was providing advice. Un
(under)constrained sketches can allow features built on them to sort of float. I
provided sound advice. You have a LOT of projected geometry in this part. And,
yes, this part could be simplified, and in that, more stable. So, what
"workaround" did use use to fix this?

--
Dennis Jeffrey, Autodesk Inventor Certified
Expert
Autodesk Manufacturing Implementation Certified
Expert.
Instructor/Author/Sr. App Engr.
AIP 2008 SP2, AIP 2009-SP1
PcCillin AV
HP zv5000  AMD64 2GB - Geforce Go 440, Driver: .8185
XP
Pro SP2, Windows XP Silver Theme

href="http://teknigroup.com">http://teknigroup.com

[Anonymous]

I simply put the correct fillet onto the part, and deleted the suppressed one.

Autodesk should consider that people like me, who have been using Inventor for years and years (R4), do not necessarily do things "the right way". What do you think a nube would think? Put it back in the box and send it back perhaps.

I have seen a lot over the years. Stuff works one day, doesn't the next release out. Waiting to install until SP1, etc. When something like this happens it is EXTREMELY rare, and noteworthy as such. Often I cannot get crashes to repeat. When I can, I want to share the parts with people who might make IV run a little better next release.

In my business- we test products before we sell them. Worst case scenarios for: user knowledge, install integrity, proper use of, and unforseeable user actions ABSOLUTELY have to be anticipated, planned for, and designed in. Otherwise, people's houses and cars burn down. Bummer.

Can Autodesk, or you, honestly say that IV hasn't cost any business money because of things like this, and the lack of foresight to produce something that can work no matter how poorly the work flow is laid out? Caveat Emptor, I understand. But then, who likes to know their products might cost jobs, create stressful situations, heck, even cost lives for the medical community that uses IV?

GIGO is not an answer anymore. This is not DOS.

Thanks
Phil

[Anonymous]

I'm not a certified professional, nor would I say I'm an expert INV user, but I do know from experience that constraining sketches completely (which doesn't take much extra time at the start), keeping projected items to a minimum and having as few adaptive parts as possible, make the end product so much more stable and makes it possible to go back and change stuff without problems. When I started playing around with INV, I didn't know how to constrain sketches properly and was getting a long way through the design before something would move and I would have to go back and fix it all up. Leaving something unconstrained is like leaving a ball sitting on the edge of a shelf, the slightest movement and it will come crashing to the floor. It has to make sense that constrained sketches and parts can't move without you wanting them to. Constraining from the start saves time in the future. Just my 2c worth.

[Anonymous]

Hi Phil,

As I mentioned in an earlier message, I have sent
this to development. I agree that it is odd that the errors only show up when
you add that last fillet.

Loren Jahraus

Autodesk Inventor Product
Design

[JDMather]

>I don't really need a lesson in how to model with IV

Based only on this file as evidence I would recommend that you get more training. You may very well have 999 other files that better demonstrate your techniques.

>been using Inventor for years and years (R4)

I can't imagine how someone can use the software since R4 designing thousands of parts can start out on a part of this complexity and not even bother to constrain the first sketch. (Some of the arcs in Sketch1 aren't even tangent - Inventor does this for you, no extra work.) Especially considering the mirrored symmetry from left to right.

>Do you have any suggestions of where I can get instruction in this matter?

I offer many industry classes on using Inventor. Where are you located? I could suggest this book as well http://www.amazon.com/gp/p roduct/0470293144?ie=UTF8&tag=mc02c-20&linkCode=as2&camp=1789&creative=93... there are many tips in there about creating robust parts. Seems to me I recall Sean having something to say about constrained sketches. You might also read this document http://home.pc t.edu/~jmather/AU2007/MA105-1L%20Mather.pdf

There is nothing inherently wrong with using adaptive parts, but building a house of cards on top of quicksand is probably not the best way to ensure a robust part. It might sound like the comments here are "kicking sand" but more like an attempt to lend a helping hand out of the quicksand.

>Can Autodesk, or you, honestly say that IV hasn't cost any business money because of things like this, and the lack of foresight to produce something that can work no matter how poorly the work flow is laid out?

I'm quite sure I could sit down with you and demonstrate how less than disciplined modeling techniques are costing your company thousands of dollars. You should be able to easily recoup my fee on the first project.

So Autodesk is demonstrating a, "lack of foresight to produce something that can work no matter how poorly... ...laid out", but you seem to be rejecting the same principle in your work.

>GIGO is not an answer anymore.

GIGO is always relevant. Your part is garbage. Nice looking garbage, but all the same garbage. Yes you should have been given a warning much sooner. (Sketch16 has sick geometry.) Maybe we will see something like SolidWorks has in a future Inventor. In SolidWorks there is a switch that can be set to not allow features to be created from unconstrained sketches. Of course for the most part very few unconstrained sketches should make it out of a first parametric modeling class.

>But then, who likes to know their products might cost jobs, create stressful situations, heck, even cost lives

When professional people ignore GIGO we have bridges collapse into rivers, cranes collapse onto apartment buildings and school buildings collapse onto children. Your designs might not have such concern for GIGO, but it is no less relevant.

I agree that I learn much about the program from unconventional techniques and I would encourage you to submit problems such as this one. But at the same time I would listen to the comments. An open forum like this brings together many different experiences that enlighten us all. Sometimes it is a little messy and argumentative but in the end there should be something of value within the arguments.

[Anonymous]

Full constraint of sketches isn't a unique requirement of Inventor. It is a basic part of good solid modeling technique.

Inventor is a sophisticated tool, with its own unique limitations and quirks at the limits of capabilities. Unpredictable behavior due to broken/poorly constrained geometry is hardly surprising.

[Anonymous]

On Fri, 24 Oct 2008 18:31:26 +0000
philpdx <> wrote:

>I SENT THIS IN BECAUSE THE
> DEVELOPERS NEED TO SEE THINGS LIKE THIS.

I agree with the other side of the fence. JD and DJ have given very
good advice to you and many others here over and over.

I'm wondering what the developers think when they see the
underconstrained sketches, it would be nice if we could get some of
that feedback here, in this topic.

SW was the first solid modeling software I was trained on many yrs ago
(read, I am not an expert) but I remember one thing from that training
that I practice with Inventor - *always fully constrain your sketches*
at least to the point they cannot move in a direction you cannot have
in a design.

Good luck with your "mission" philpdx. I hope *somebody* learns
something from this.

I have, now where is the ole plonk button at...

--
Inventor 10 Series
SP3a + TS1069046 (Inventor Text Broken By Microsoft update)

[Anonymous]

So far, it appears that no one has "fixed" the sketches to see if the fillet failure still occurs. That would be interesting, no ? I can't for the life of me comprehend how a feature that's created later in time (Inventor is a feature/history based modeler, after all) can cause previously created features to fail. That is a bug and Autodesk does need to look at it. I think Loren agrees with that as well as he stayed out of the bickering side of this discussion.

[Anonymous]

> So far, it appears that no one has "fixed" the sketches
> to see if the fillet failure still occurs.
> That would be interesting, no ?

Yes it would. It would happen if more of the people here were more interested in fact and comprehension than appearance and sales.

> I can't for the life of me comprehend how a feature that's
> created later in time (Inventor is a feature/history based
> modeler, after all) can cause previously created features
> to fail. That is a bug and Autodesk does need to look at it.

It would seem so, at least based on experience with other history based modelers. Either a bug or a "unique" set of data processing
and solver algorithms. Memory grows dim but it seems IV does allow certain operations that would result in circular references in
other systems indicating the possibility of unique algorithms and possibly unique requirements such as manually fully constraining
sketches (good practice but if there is necessity it should be understood). One always has to consider what happens behind the
buttons to understand the nature of failures. One could hope for factory explanations / documentation / demonstration but I think
there's still an unresolved question regarding nonmanifold failures with multiple region sketches floating around. That should be a
simple one.
>
Beware those that don the mantle of "expert" and, especially, the fans and true believers. ;^)

[Anonymous]

I dont make any money off of helping people, andIi spend about 10 hours each week helping others on my time. No one in these replies has
suggested that you are "wrong" or "bad at your job " . There are hundreds of ways to skin each cat.

That being said there are some ways that result in a happier cat.

This forum is for constructive results , you take what you want and leave the rest. Its obvious you got your feelings hurt and no one intended to do that at all.

I have not yet looked at your file because it wont download, but I will be glad to send back my naked cat as soon as I can.

I personally have found that I save significant time when keeping slim and trim. and if the stuff I design fails people die.


never stop learning , and definitely never stop calling autodesk to task for things that cause you to lose productivity.

feel free to send it to me at
chenry@engineersupportservice.seeohhmmm

[JDMather]

>So far, it appears that no one has "fixed" the sketches to see if the fillet failure still occurs. That would be interesting, no ?Yes it would. It would happen if more of the people here were more interested in fact and comprehension...

I used to learn a lot from the unconventional techniques you used post here. But in this case I see no real indication that the OP is really interested in the 10 hours or so that it would take me to go through the convoluted mess and fix it up. I routinely do just that for those who are interested and just for my own interest. Obviously the first sketch has nothing to do with the problem reported - that is a separate matter. But I see the first sketch as just the first of many indications I see of a bit of thrashing about.
I estimate I could create the same part with at least half as many features. I would definately start over clean rather than try to use this file. From my experience this much thrashing around is bound to cause problems. Better to take what is learned about the geometry along the way and start over. If anyone wants to try to fix it up I think I saw Sketch16 - well back in the history tree - had sick geometry (pink). Might be a place to start.

[Anonymous]

> the convoluted mess and fix it up. I routinely do just
> that for those who are interested and just for my own interest.

I understand and anyone and everyone should appreciate that effort
be grateful for the leg up on sound modeling practices and procedures.

>

In this instance however I think the goal ~should~ be to
establish whether or not adding or unsuppressing a feature at
or near the end of the history tree normally causes previous
features to regenerate and, in this case, fail. If the full
regen is normal (hard to imagine but ...?) then I guess we
might be back to some ragged edge geometry definitions. Even
then, if they are causing a crash (abnormal program termination)
Autodesk should be happy to receive the data set for examination.
Less than ideal geometry definitions are a fact of life and any
insinuation (not pointing at you) that it's ok for that to take
down a system needs to be examined.

>

And, well... you know; I have a preternatural tendency to question things. ;^)

[Anonymous]

Well, I saved you all "10 hours" or so of untangling things.

Ha, ha. It only took 20 minutes of auto dimensioning and constraining each sketch. I also deleted any extra geometry that was hanging out, for good measure. I also didn't find any adaptive geometry. I did break link with all projected cut geometry.

Results: the part didn't crash. I still had to manually rebuild the last fillet since the root edge for it was now different.

Lesson learned: sure, I could be more disciplined in my modeling.

But then, this odd behaviour would not have raised it's ugly head, I wouldn't have had a file to send in, and I would have been done with this project HOURS AGO.

You see- funny thing here- by constraining all my sketches, and doing my best to have a clean part, I SAVED NO TIME. Funny how that works- once I am done with a part I rarely edit it, and therefore don't waste any time because of this sloppy work on my part. Even with the sketches un-constrained, the editing time is the same.

So what I learned is it just doesn't matter that much to me. (the part is the same no matter how many constraints it has in it's build)

And I don't curse Inventor (I think that's what JD and DJ were on about to begin with- thinking I was blaming their sweetheart IV when I am a mere sloppy constrainer), I have had very good luck, productivity, creativity, and learned a lot from IV over the years. So next time I post a corrupt part, don't lecture me, enjoy the chance to see a real pro in action- breaking the program. (another thing about my business- we are very happy to find out just how things break- poor use is no reason to blame the user)

Breaking the program is a good thing, in other words.

Now- am I going to make parts this complex in the future? You bet. Am I going to constrain the heck out of everything? Only with fingers crossed that the CAD gods smile on my efforts.

And this time I mean it: thanks for pointing, uh, shoving me in the right direction.